Professional Documents
Culture Documents
3/2014 1
Editors-in-Chief
Prof. Adela Bara, PhD, University of Economic Studies, Bucharest, Romania
Prof. Marinela Mircea, PhD, University of Economic Studies, Bucharest, Romania
Secretaries
Lect. Iuliana Botha, PhD, University of Economic Studies, Bucharest, Romania
Lect. Anda Velicanu, PhD, University of Economic Studies, Bucharest, Romania
Editorial Board
Prof. Ioan Andone, PhD, A.I.Cuza University, Iasi, Romania
Prof. Anca Andreescu, PhD, University of Economic Studies, Bucharest, Romania
Prof. Emil Burtescu, PhD, University of Pitesti, Pitesti, Romania
Joshua Cooper, PhD, Hildebrand Technology Ltd., UK
Prof. Marian Dardala, PhD, University of Economic Studies, Bucharest, Romania
Prof. Dorel Dusmanescu, PhD, Petrol and Gas University, Ploiesti, Romania
Prof. Marin Fotache, PhD, A.I.Cuza University, Iasi, Romania
Dan Garlasu, PhD, Oracle Romania
Prof. Marius Guran, PhD, University Politehnica of Bucharest, Bucharest, Romania
Lect. Ticiano Costa Jordão, PhD-C, University of Pardubice, Pardubice, Czech Republic
Prof. Brijender Kahanwal, PhD, Galaxy Global Imperial Technical Campus, Ambala, India
Prof. Dimitri Konstantas, PhD, University of Geneva, Geneva, Switzerland
Prof. Hitesh Kumar Sharma, PhD, University of Petroleum and Energy Studies, India
Prof. Mihaela I.Muntean, PhD, West University, Timisoara, Romania
Prof. Stefan Nithchi, PhD, Babes-Bolyai University, Cluj-Napoca, Romania
Prof. Corina Paraschiv, PhD, University of Paris Descartes, Paris, France
Davian Popescu, PhD, Milan, Italy
Prof. Gheorghe Sabau, PhD, University of Economic Studies, Bucharest, Romania
Prof. Nazaraf Shah, PhD, Coventry University, Coventry, UK
Prof. Ion Smeureanu, PhD, University of Economic Studies, Bucharest, Romania
Prof. Traian Surcel, PhD, University of Economic Studies, Bucharest, Romania
Prof. Ilie Tamas, PhD, University of Economic Studies, Bucharest, Romania
Silviu Teodoru, PhD, Oracle Romania
Prof. Dumitru Todoroi, PhD, Academy of Economic Studies, Chisinau, Republic of Moldova
Prof. Manole Velicanu, PhD, University of Economic Studies, Bucharest, Romania
Prof. Robert Wrembel, PhD, University of Technology, Poznan, Poland
Contact
Calea Dorobanţilor, no. 15-17, room 2017, Bucharest, Romania
Web: http://dbjournal.ro
E-mail: editordbjournal@gmail.com; editor@dbjournal.ro
2 Database Systems Journal vol. V, no. 3/2014
CONTENTS
The paper deals with the problem of automatic labeling output variables in Mamdani-type
fuzzy rules generated by using heuristic algorithms of possibilistic clustering. The labeling
problem in fuzzy clustering and basic concepts the heuristic approach to possibilistic
clustering are considered in brief. Labeling consequents procedure is proposed. Experimental
results are presented shortly and some preliminary conclusions are made.
Keywords: Data Mining, Clustering, Fuzzy Rule, Consequent, Labeling
1 Introduction
Fuzzy classifiers play an important
role in different data mining approaches.
If xˆ 1 is Bl1 and and xˆ m1 is Blm1
, (1)
Thus, the problem of generation of fuzzy then y1 is C1l and and y c is C cl
rules is one of more than important where input variables xˆ t1 , t1 1, , m1 are
problems in the development of fuzzy antecedents and output variables yl ,
classifiers. l 1, , c are consequents of fuzzy rules,
There are a number of approaches to
learning fuzzy rules from data based on Blt1 , t1 {1, , m1 } and C ll , l {1, , c} are
techniques of evolutionary or neural fuzzy sets that define an input and output
computation, mostly aiming at optimizing space partitioning. A fuzzy classifier which
parameters of fuzzy rules. From other is described by a set of fuzzy classification
hand, fuzzy or possibilistic clustering rules with the form (1) is the multiple inputs,
seems to be a very appealing method for multiple outputs system.
learning fuzzy rules since there is a close The principal idea of extracting fuzzy
and canonical connection between fuzzy classification rules based on fuzzy clustering
clusters and fuzzy rules. The fact was was outlined in [1] and the idea is the
shown in [1]. following. Each fuzzy cluster is assumed to
Let us consider in brief some basic be assigned to one class for classification
concepts. We assume that the training set and the membership grades of the data to the
contains n data pairs. Each pair is made clusters determine the degree to which they
of a m1 -dimensional input-vector and a can be classified as a member of the
с -dimensional output-vector. We assume corresponding class. So, with a fuzzy cluster
that the number of rules in the fuzzy that is assigned to the some class we can
inference system rule base is с . So, associate a linguistic rule. The fuzzy cluster
Mamdani and Assilian’s [2] fuzzy rule l is projected into each single dimension
within the fuzzy inference system is leading to a fuzzy set on the real numbers.
written as follows: An approximation of the fuzzy set by
projecting only the data set and computing
the convex hull of this projected fuzzy set or
approximating it by a trapezoidal or
triangular membership function is used for
4 Labeling Consequents of Fuzzy Rules Constructed by Using
Heuristic Algorithms of Possibilistic Clustering
described with the aid of a partition So, the family of fuzzy sets
l
matrix Pcn [u li ] , l 1, , c , i 1, , n . ( X ) { A | l 1, c, c n} is the possibilistic
The set of all fuzzy c -partitions will be partition of the initial set of objects
denoted by . So, the fuzzy problem X {x1 ,..., x n } if condition (5) is met.
formulation in cluster analysis can be Obviously that the conditions of the
defined as the optimization task possibilistic partition (5) are more flexible
Q extr under the constraints (3), than the conditions of the fuzzy c -partition
P ( X )
(3).
where Q is a fuzzy objective function.
In order to be applying the found cluster
The best known optimization approach to prototype as classifiers, they need to be
fuzzy clustering is the method of fuzzy given reasonable names. One can then use
c -means [6]. The FCM-algorithm is these names as column titles of the
based on an iterative optimization of the membership matrix when using the recall
fuzzy objective function, which takes the function of the FCM-algorithm. This helps
form: in the interpretation of the results.
c n 2
Q FCM ( P, ) u li x i l , The process of assigning class names to
l 1 i 1 cluster prototypes is called labeling. A
(4) labeling method for the fuzzy c -means
where u li , l 1, , c , i 1, , n is the method is to inspect the cluster prototypes
membership degree, xi , i {1, , n} is the and their respective membership values of
data point, { 1 , , c } is the set the various attributes, and to assign a label
manually.
fuzzy clusters prototypes, and 1 is the
However, usually, it is already known when
weighting exponent. training a classifier which objects belong to
The purpose of the classification task is which classes. This information can be taken
to obtain the solutions P ( X ) and into account to use so as automatic the fuzzy
1 , , c which minimize equation (4). c -means labeling process. The
Some other similar objective function- corresponding labeling procedure is
based fuzzy clustering algorithms are described in [7] in detail. The principal idea
considered in [1], [5] and [6] in detail. of the procedure is to present a sample of
However, the condition of fuzzy c - objects to the FCM-classifier whose class
partition is very difficult from essential membership are known in the hard form of 0
positions. So, a possibilistic approach to or 1 values and have also been calculated by
clustering was proposed by the procedure. By means of the given cluster
Krishnapuram and Keller in [3] and membership values for each fuzzy cluster
developed by other researchers. Major prototype, the fuzzy cluster prototypes can
algorithms of possibilistic clustering are be associated with their respective classes.
objective function-based procedures. On the other hand, all objective function-
A concept of possibilistic partition is a based fuzzy clustering algorithms are
basis of possibilistic clustering methods iterative procedures and the initial fuzzy c -
and membership values li , l 1,, c , partition P ( X ) is initialized randomly. So,
i 1,, n can be interpreted as the values coordinates of fuzzy clusters prototypes and
of typicality degree. For each object xi , values of membership functions will be
i 1, , n the grades of membership different in each experiment for the same
should satisfy the conditions of a data set, because the result of classification
possibilistic partition: is sensitive to initialization. Moreover,
major objective function-based fuzzy
6 Labeling Consequents of Fuzzy Rules Constructed by Using
Heuristic Algorithms of Possibilistic Clustering
(7)
3. Basic concepts of the heuristic is called a typical point of the fuzzy α -
approach to possibilistic clustering
cluster A(lα ) , α (0,1] , l [1, n] . A set
Let us remind the basic concepts of the
heuristic method of possibilistic K ( A(lα ) ) {τ1l , , τ ll } of typical points of the
clustering [4]. Let X {x1 ,..., x n } be the
fuzzy cluster A(lα ) is a kernel of the fuzzy
initial set of elements and
T : X X [0,1] some fuzzy tolerance on cluster and card K ( A(lα ) ) l is a cardinality
X with μT ( x i , x j ) [0,1] , xi , x j X of the kernel. If the fuzzy cluster have an
being its membership function. Let α be unique typical point, then l 1 .
the α -level value of the fuzzy tolerance Let R zα ( X ) { A(lα ) | l 1, c, 2 c n} be a
T , α (0,1] . Columns or rows of the family of fuzzy α -clusters for some value of
fuzzy tolerance matrix are fuzzy sets tolerance threshold α , which are generated
{ A1 ,..., A n } on the universal set X . Let by a fuzzy tolerance T on the initial set of
Al , l {1, , n} be a fuzzy set on X with elements X {x1 ,..., xn } . If condition
c
μ Al ( xi ) [0,1] , xi X being its μli 0 , xi X ,
l 1
membership function. The α -level fuzzy
(8)
set A(lα ) ( xi , μ Al ( xi )) | μ Al ( xi ) α, xi X
is met for all A(lα ) , l 1, c , c n , then the
is fuzzy α -cluster. So, A(lα ) Al , family is the allotment of elements of the set
α (0,1] , Al { A1 , , An } and μ Al ( xi ) is X {x1 ,..., xn } among fuzzy α -clusters
the membership degree of the element { A(lα ) , l 1, c, 2 c n} for some value of the
xi X for some fuzzy α -cluster A(lα ) , tolerance threshold α . It should be noted
α (0,1] , l {1, , n} . The membership that several allotments Rz (X ) can exist for
degree will be denoted μli in further some tolerance threshold α . That is why
symbol z is the index of an allotment.
considerations. The membership degree
Obviously, the definition of the allotment
of the element xi X for some fuzzy α -
among fuzzy clusters (8) is similar to the
cluster A(l ) , α (0,1] , l {1, , n} can be definition of the possibilistic partition (5).
defined as a So, the allotment among fuzzy clusters can
be considered as the possibilistic partition
Database Systems Journal vol. V, no. 3/2014 7
and fuzzy clusters in the sense of (6) are Thus, the conditions (9) and (10) can be
elements of the possibilistic partition. generalized for a case of particularly
Allotment separate fuzzy clusters. So, a condition
c
RI ( X ) { A(l ) | l 1, n, (0,1]} of the set
card ( Al ) card ( X ),
l 1
of objects among n fuzzy clusters for
some tolerance threshold (0,1] is the A(l ) Rс( z ) ( X ), , (11)
initial allotment of the set X {x1 ,..., xn } . (0,1], card ( Rc( z ) ( X )) c
In other words, if initial data are and a condition
represented by a matrix of some fuzzy T card ( Al Am ) w, A(l ) , A(m ) ,
then lines or columns of the matrix are , (12)
l m, (0,1]
fuzzy sets Al X , l 1, n and -level
are generalizations of conditions (9) and
fuzzy sets A(l ) , l 1, c , (0,1] are
(10). Obviously, if w 0 in conditions (11)
fuzzy clusters. These fuzzy clusters and (12) then conditions (9) and (10) are
constitute an initial allotment for some met. The adequate allotment Rс( z ) ( X ) for
tolerance threshold and they can be
considered as clustering components. If some value of tolerance threshold (0,1] is
some allotment a family of fuzzy clusters which are
l
Rс ( z ) ( X ) { A( ) | l 1, c, c n} corresponds elements of the initial allotment RI (X ) for
to the formulation of a concrete problem, the value of and the family of fuzzy
then this allotment is an adequate clusters should satisfy the conditions (11)
allotment. In particular, if a condition and (12). So, the construction of adequate
c allotments Rс( z ) ( X ) { A(l ) | l 1, c, c n} for
Al X ,
l 1 every is a trivial problem of
(9) combinatorics.
and a condition Allotment RP ( X ) { A(l ) | l 1, c} of the set
card ( Al Am ) 0, A(l ) , A(m ) , of objects among the minimal number c ,
,
l m, (0,1] 2 c n of fully separate fuzzy clusters for
(10) some tolerance threshold (0,1] is the
are met for all fuzzy clusters A(l ) , l 1, c principal allotment of the set X {x1 ,..., xn } .
of some allotment Several adequate allotments can exist. Thus,
l
Rс ( z ) ( X ) { A( ) | l 1, c, c n} for a value the problem consists in the selection of the
unique adequate allotment Rc (X ) from the
(0,1] , then the allotment is the
set B of adequate allotments,
allotment among fully separate fuzzy
clusters. B {Rc ( z ) ( X )} , which is the class of possible
Fuzzy clusters in the sense of definition solutions of the concrete classification
(6) can have an intersection area. If the problem. The selection of the unique
intersection area of any pair of different adequate allotment Rc (X ) from the set
fuzzy clusters is an empty set, then
B {Rc( z ) ( X )} of adequate allotments must
conditions (9) and (10) are met and fuzzy
clusters are called fully separate fuzzy be made on the basis of evaluation of
clusters. Otherwise, fuzzy clusters are allotments. In particular, the criterion
called particularly separate fuzzy clusters c 1 nl
F ( Rc( z ) ( X ), ) li c ,
and w {0, , n} is the maximum number l 1 nl i 1
go to step 1.2;
4. A labeling procedure for fuzzy rules 1.3 Check the following condition:
consequents if the typical point l is labeled
Fuzzy classifier can be generated directly then l : l 1 and go to step 1.2
by some heuristic algorithm of else go to step 1.4;
possibilistic clustering [4]. A fuzzy rule is 1.4 Check the following condition:
associated to each fuzzy α -cluster of the if all typical points l , l 1, c
obtained allotment, Rc ( X ) . So, a number are labeled then go to step 2.
of fuzzy rules is equal to a number of 2. Perform the following operations for
fuzzy α -clusters and equal to a number each output variable yl , l 1, c and
of output variables y l , l 1, , c . each typical point l , l 1, c :
The results obtained from heuristic 2.1 Let l : 1 ;
algorithms of possibilistic clustering are 2.2 A label of typical point l
stable. The set of kernels should be assigned to output
1 c
{K ( A(α ) ), , K ( A( α ) )} and the set of labels variable yl ;
{label 1, , label c}are inputs for a 2.3 Check the following condition:
labeling procedure [8]. We assume that a if a condition l c is met then
condition card K ( A(lα ) ) 1 is met for l : l 1 and go to step 2.2 else
stop.
each kernel K ( A(lα ) ) , l 1, c . In other
words, the set of typical points { 1 , , c } That is why the proposed labeling procedure
is given. for consequents of fuzzy rules can be
Each output variable { y1 , , yc } considered as an extended version of the
corresponds to a fuzzy α -cluster of the procedure for labeling fuzzy α -clusters [8].
obtained allotment, Rc ( X ) . There is a
5. An illustrative example
two-step procedure which can be The Anderson’s Iris database [10] is the
described as follows: most known database to be found in the
pattern recognition literature. The data set
1. Perform the following operations represents different categories of Iris plants
for each typical point l , l 1, c and having four attribute values. The four
each label label m , m 1, c : attribute values represent the sepal length,
1.1 Let l : 1 and m : 1 ; sepal width, petal length and petal width
1.2 Check the following condition: measured for 150 irises. It has three classes
if l corresponds to label m Setosa, Versicolor and Virginica, with 50
then the label label m is the samples per class. Examples of records in
the database are presented in Table 1.
label for the typical point l and
go to step 1.3 else m : m 1 and
The Anderson’s Iris data form the matrix The matrix of fuzzy tolerance
of attributes Xˆ 1504 [ xˆit1 ] , i 1,,150 , T [ T ( xi , x j )] was obtained after
t1 1, ,4 , where the sepal length is application of complement operation
denoted by x̂1 , sepal width – by x̂ 2 , petal T ( xi , x j ) 1 I ( xi , x j ) , (17)
length – by x̂ 3 and petal width – by x̂ 4 . to the matrix of fuzzy intolerance
The data was normalized as follows: I [ I ( xi , x j )] , i, j 1, ,150 obtained from
xˆit1 previous operations. The labeled training
xit1 .
max xˆit1 data is shown in Fig. 1.
i
By executing the D-AFC(c)-algorithm [4]
(15) for c 3 using the normalized Euclidean
So, each object can be considered as a distance (16), we obtain that the typical
fuzzy set xi , i 1, ,150 and
point of the first class τ 1 is the object x23 ,
xit1 xi ( x t1 ) [0,1] , i 1, ,150 ,
the typical point of the second class τ 2 is the
t1 1, ,4 , are their membership object x95 , and the typical point of the third
functions. The matrix of coefficients of class τ 3 is the object x98 . The clustering
pair wise dissimilarity between objects
result is presented in Fig. 2. The set of labels
I [ I ( xi , x j )] , i, j 1, ,150 can be
is {SETOSA, VERSICOLOR,
obtained after application of some VIRGINICA} and these labels were
distance to the matrix of normalized data assigned to corresponding output variables.
X 1504 [ xi ( x t1 )] , i 1, ,150 , t1 1, ,4 . The performance of the generated fuzzy
In particular, the normalized Euclidean classifier is shown in Fig. 3 in which
distance [11] l 1, ,3 is the number rule.
1 m1
e( xi , x j ) 2
xi ( x t1 ) x j ( x t1 ) .
m1 t11
(16)
was applied to the normalized data.
Fig. 3. Performance of the fuzzy classifier which was generated from Anderson’s Iris data
So, the result of application the proposed database. These perspectives for
labeling procedure seems to be investigations are of great interest both from
satisfactory. the theoretical point of view and from the
practical one as well.
6 Conclusions
The fast procedure for labeling Acknowledgment
consequents of fuzzy rules generated by This investigation was done in framework of
using heuristic possibilistic clustering a project of the scientific program
results is proposed in the paper. Stability “Monitoring-SG” of the Union State of
of results of heuristic possibilistic Russia and Belarus.
clustering is a basis of the developed
procedure. The proposed procedure can References
be very useful in process of adding new [1] F. Höppner, F. Klawonn, R. Kruse and
records to database. For the purpose, all T. Runkler, Fuzzy Cluster Analysis:
records in a database can be classified by Methods for Classification, Data
using heuristic possibilistic clustering and Analysis and Image Recognition,
fuzzy classifier can be generated on a Chichester, Wiley Intersciences, 1999.
basis of clustering results. That is why [2] E.H. Mamdani and S. Assilian, “An
the corresponding label can be assigned Experiment in Linguistic Synthesis with
to a new record which added to the a Fuzzy Logic Controller”, International
12 Labeling Consequents of Fuzzy Rules Constructed by Using
Heuristic Algorithms of Possibilistic Clustering
Business Intelligence has evolved in the last decade increasingly relying on real-time data.
Business analysis has become essential. This involves actions in response to results analysis
and instant change of business processes parameters making BI beneficial for several
reasons.
Keywords: Business Intelligence, Performance Management, Cognos
1 Introduction
Business Intelligence (BI) has two
different basic meanings related to use of
large number of business information in
concrete performance indicators and finally
the organizational knowledge. BI
the term intelligence [1]. The first implementations are efforts involving
meaning, less frequent, is referring to the multiple issues, from organizational strategy
capacity of human intelligence in to organizational processes and
business or in activities. Business management, from application management
Intelligence is a new and powerful area of to information infrastructure changes. BI
investigation of the applicability of projects do not aim to teach managers how
human cognitive abilities and artificial to make the right decisions; instead they
intelligence technologies in management help them make decisions based on facts and
and decision support in various business figures, not on assumptions. Companies
issues. The second meaning displayed in collect vast amounts of data through
Fig.1 refers to intelligence as information transactional systems (e.g. ERP, CRM, and
valued for its use and relevance. SCM) that have implemented over the years
and which they use daily to perform a
variety of corporate functions. Before the
notion of BI was launched, there was no
concept that allowed the use of large
amounts of data by integrating and
transforming them into information [3].
Development of BI concepts and
technologies creates a management
environment where current and new data can
be used to improve the quality in the process
of decision making. In addition, the
Fig.1 Information and business existence of large volumes of transactional
intelligence [2] data, especially transactional data with a
high degree of specificity and particularity,
Application. BI applications include creates opportunities for management to
decision support systems, query and improve forecast accuracy.
reporting tools, online analytical
processing and also forecasting and data BI Tools:
mining systems. Ultimately, the final Spreadsheets
results of BI implementations are depth Query software and OLAP (Online
analysis, refining and concentrating a analytical processing)
Database Systems Journal vol. V, no. 3/2014 15
IBM Cognos Business Intelligence is an and monitor outcomes and metrics so that
integrated business intelligence suite that they can prepare efficient business decisions
offers a broad range of functionality to [8].
help understanding an organization’s IBM Cognos 8 integrates the following
information. Everyone in the system can business intelligence activities, shown in
use IBM Cognos 8 to view or create table 1, in one Web-based solution.
business reports, examine information,
Table 1 Components and activities [8]
IBM Cognos Connection Publishing, managing, and viewing
content
IBM Cognos Insight Managed workspaces
IBM Cognos Workspace Interactive workspaces
IBM Cognos Workspace Advanced Ad hoc querying and data exploration
IBM Cognos Report Studio Managed reporting
IBM Cognos Event Studio Event management and alerting
IBM Cognos Metric Studio Scorecarding and metrics
IBM Cognos for Microsoft(tm) Office Works with IBM Cognos BI content in
Microsoft Office
IBM Cognos Query Studio Ad-hoc querying
IBM Cognos Analysis Studio Data exploration
on Cognos BI servers. This software allows
multiple users to download and install IBM
“IBM Cognos Connection is the Web Cognos Insight on their computers from the
portal for IBM Cognos Business Cognos Connection interface.”
Intelligence.” It is the beginning spot to In IBM Cognos Workspace, the user can
access BI information and the create sophisticated interactive workspaces
functionality of IBM Cognos BI. The using IBM Cognos content, as well as
Web portal can be used to publish, find, external information sources such as TM1
manage, organize, and view Websheets and CubeViews, according to
organization's business intelligence specific information needs. The user can
content, such as reports, scorecards, and view and open favourite workspaces and
agents. Possessing the necessary reports, manage the content in the
permissions, various studios can be workspaces, and e-mail the workspaces.
accessed from the portal and this can be Also, can be used comments, activities and
used for content administration, including social software like IBM Connections for
programming and distributing reports, collaborative decision making.
and creating businesses. In IBM Cognos With IBM Cognos Workspace Advanced,
Insight, the user can analyze information, the user can perform advanced data
explore scenarios, and influence exploration and create simple reports. When
decisions by creating personal or the user is in a workspace in IBM Cognos
managed workspaces. The interactive Workspace and wants to perform deeper
workspaces can be used to pass the analysis and report authoring, he can
results to managers. Because Cognos seamlessly graduate to Cognos Workspace
Insight supports write-back, the user can Advanced, where he can perform more
also use these workspaces to collect and advanced data exploration, such as adding
consolidate management targets, additional measures, conditional formatting,
commitments, and forecasts. “IBM and advanced computations [7]. The user
Cognos Insight is provided with IBM can also launch Cognos Workspace
Cognos BI. IBM Cognos Connection Advanced from the IBM Cognos
Installer for Cognos Insight can be used Connection portal. With Cognos Workspace
in order to install provisioning software Advanced, the user can create reports with
Database Systems Journal vol. V, no. 3/2014 17
relational or dimensional data sources, access to all IBM Cognos report content,
and that shows the data in lists, crosstabs, including data, metadata, headers, footers,
and charts. He can also use his external and charts. The user can use predefined
data source. With this feature, a report reports, or create new content using IBM
authored in IBM Cognos Report Studio Cognos Query Studio, IBM Cognos
can be accessed; the objects that can be Analysis Studio, or IBM Cognos Report
inserted only in Report Studio can be Studio. By importing content into Microsoft
seen (such as map). However, he cannot Excel spreadsheet software, the user can
modify these objects. work with the data and leverage Microsoft
Using Report Studio, report authors Excel's formatting, calculation, and
create, edit, and distribute a wide series presentation capabilities. The user can also
of professional reports. They can also use the formatting and charting features of
define corporation standard report Microsoft Excel. By importing content into
templates for use in Query Studio, and Microsoft PowerPoint and Microsoft Word,
edit/modify reports created in Query various reports and charts can be included to
Studio or in Analysis Studio. enhance the presentations and documents.
Report Studio can be used for reports Using Query Studio, users with little or no
which are intended for a large audience, training can quickly design, create and save
exist long enough to require maintenance reports to meet reporting needs not covered
for updateing requirements and data or by the standard, professional reports created
require a thorough check on the in Report Studio [8].
appearance. Report Studio provides In Analysis Studio, users can explore,
powerful features, such as transferring in analyze, and compare dimensional data.
burst, prompts, maps and create advanced Analysis Studio provides access to
graphics and offers many ways to dimensional, OLAP (online analytical
customize reports. processing), and dimensionally modelled
Event Studio is an action oriented agent relational data sources. Analyses created in
which notifies the users when an Analysis Studio can be opened in Report
bussiness event occurs. Agents can Studio and used to build professional reports
publish details to the portal can deliver [12].
alerts by e-mail, can run and distribute IBM Cognos Transformer is a multi-
reports based on events and can monitor dimensional data modelling component
the status of events. designed for use with IBM Cognos Business
In Metric Studio, the user can create and Intelligence. This component is useful to
deliver a customized scorecarding create a multi-dimensional model: a
environment for monitoring and business presentation of the information in
analyzing metrics throughout one or more different data sources that share
organization. Users can monitor, analyze, common data [8]. After the user add the
and report on time-critical information by needed metadata from IBM Cognos
using scorecards based on cross- Business Intelligence packages, reports, and
functional metrics. other various data sources, model the
In IBM Cognos for Microsoft Office, dimensions, customize the measures, and
the user can work with secure IBM apply IBM Cognos BI secured views with
Cognos Business Intelligence content in dimensional filtering, he can create IBM
his familiar Microsoft Office Cognos PowerCubes based on this model.
environment. He can extract report These cubes can be deployed in order to
content from a variety of IBM Cognos support OLAP reporting and analysis around
applications, including IBM Cognos BI the globe.
and IBM Cognos PowerPlay. IBM IBM Cognos Business Intelligence
Cognos for Microsoft Office provides Administrators ensure that IBM Cognos BI
18 BI solutions for modern management of organizations - Cognos
runs without interruption and at optimal downloading - can view the downloaded
performance. They can: pages and interact with them without the
Define connections to data sources of need to wait for complete download of
the organization the report.
Define security permissions for users 2. Dashboards
and groups in the organization With the ability to create dashboards,
Specify distribution lists, contacts and Cognos Business Intelligence allows
printers users to create and customize their ways
Manage servers and dispatchers, and of viewing dashboards that they can
can finely adjust performance for shape the needs.
IBM Cognos BI Historical data with variable data and
Predefine links to an entire package what-happens-if scenarios provides
that authors can easily add it to their extended insight of performance.
reports Portable dashboards can be created that
Customize the appearance and can support decisions on when and
functionality of IBM Cognos BI where to create
IBM Cognos BI is secured by setting 3. Analysis
permissions and by enabling user With analysis capabilities of Cognos
authentication. When anonymous access Business Intelligence, users can use
is enabled, IBM Cognos BI can be used analytical reporting, trend analysis,
as a specific user without authentication. statistical analysis and some other
In IBM Cognos BI, administrators define intuitive tools that can easily analyze
permissions so that users can access information.
functions. For example, to edit a report Financial and business analysis can
using IBM Cognos Report Studio, estimate the short-term business through
security permissions and the appropriate advanced analytics and predictive
license are required. In addition, each statements what-happens-if.
entry in IBM Cognos Connection is Support for operational and strategic
secured by defining those who can read decision cycles may engage the right
it, edit and run [8]. people at the right time for analysis.
IBM Cognos BI comes with the Cogons Business Intelligence comes
following capabilities: with options for mobile devices, analysis
1. Reports and offline interactivity along with
Cognos Business versions for Microsoft Office to the
Intelligence features professional appropriate user needs.
report generation capabilities that are 4. Collaboration
easy to use and help minimize the The ability to form communities, to
effort to build them. capture annotations, views and share
Users can create their opinions help streamline and improve
own reports (queries) or edit existing decision-making groups.
reports. Workflow and ability to handle tasks
Collaborative are designed to connect users and to
reporting features built into Cognos enhance activities coordination.
Business Intelligence helps users to Using IBM Connections and its
communicate between them to make social capabilities, users can share their
decisions and to gain additional opinions and request ideas.
insights. 5. Scorecarding
Users can access Using measurement capabilities inside
reports on mobile devices and can (scorecarding) of Cognos Business
already interact with the report while Intelligence, organization can compare
Database Systems Journal vol. V, no. 3/2014 19
questions that we might put, as we try to distribution, we click on Asia Pacific and
solve a problem or understand the then tab Metrics; we click the measuring
information. index Asia Pacific - Returns by Reason %
7. In the right pane, the History tab - wrong product ordered and on the
(figure 7), we click on the list. The superior toolbar choose Add the watch list
information in the History tab, answer the (figure 8), and press OK. In the left panel,
question "When?". We see the actual we click on My Folders on the Watch List.
values on the target metric data for Index Asia Pacific - Returns by Reason %
previous periods. Also see half-yearly - wrong product ordered appears on our
and annual summaries. list of supervisors
agiu.corneliudragos@yahoo.com, vlad.mateescu30@gmail.com,
muntean.iulia12@gmail.com
The need for information has always been present, but what defines the present time is the
big amount of data available everywhere and the need to have all the answers we need in
a very short time. Because of the agglomeration of markets and the dynamics of economic
environment, the ability to collect data and turn it into useful information for decisional
process may be the element which makes the difference. Thus, the term of “business
intelligence” was first introduced by Gartner Group in the mid of 90’s, but it still exists
since the 70’s, when the reporting systems were static, two dimensional, with no
analytical skills.
Keywords: Business Intelligence, management, Data Mining, CRM, ERP
1.Analysis 2.Design
5.Evolution 3.Development
4.Implementation
extensions for the instruments are technical support department for a better
frequently necessary.The real challenge for guidance of the end user.
the BI applications comes from the BI that 4.5. The evolution stage
is hidden in the data of the organization, This stage can be called, contrary to the
that can only be discovered with data title, the consumption stage because the
mining instruments. Developping metadata user uses the information received to
deposits becomes a subproject of the main change the business and make decisions.
BI project. The main goals of this stage are:
4.4. The implementation stage Measuring the success of a project
After all the BI components were Extending the application to the
thoroughly tested, the application is enterprise level
implemented at user level. Regardless of
the technology used, the success of a Increasing the exchange of
project will depend on the training done at information between business and
the user level and the support of a functional, both internally and
dedicated team, especially in the first stage externally.
of implementation. This phase requires an Business-oriented aspects are often
iterative approach with extra training ignored at this stage - since the BI solution
sessions so that customer needs are met. is technically capable, the user is free to
This step also requires a preliminary make any decision. Many projects stalled
development of specific reports and in this phase because responsibility of
analysis for the business users type. This passing to the user is not officially sent to
will create a foundation for future a business manager. Therefore a good
advanced analysis. study is dividing this stage in several sub-
All support and guidance phases to show more clearly how this
operations will be provided by the IT team methodology can lead to the end of a life
and the users will serve as consultants cycle of BI.
during the execution of this stage. Of Shown below it is presented the schema of
course, the IT team will have to train the the evolution stage subdivisions:
a.Descovery b.Access
e.Change c.Decision
Evolution
d.Division
Often the organization does not understand something did not come out as expected
how the solution can lend itself to the organizational changes can occur.
exact needs of the customers. End users e. The changing stage
along with the first version of the solution Changes made can trigger a fundamental
should create the main environment for process of reengineering. On this stage all
using an effective solution, but for there technical resources together with COE
has to be a mutual agreement over a basic shall review the issues and help resolve
system where the solution can run at first. them.
b. The access stage Any BI cycle that ends should start again
Having identified the indicators and the from the first stage, but the methodology
valuable information during the discovery approach will have a new level of focus:
stage, the end users begin to follow, - The analysis stage
understand and manage the information - The reevaluation stage
leading to deeper perspectives in the - The modifying stage
chosen business solution. The users can - The optimization stage
ask for assistance or contact the - The adaptability stage
organization for further clarification. This determines that the benefits of the
c. The decision stage experiences should be added again to a
The end users make definite decisions process and maintain the relevant BI cycle.
based on the newly obtained information. Using a BI methodology helps the
The Center of Excellence (COE) may be organization to understand and produce a
involved in verifying the solution so that sequence of steps to develop and
the user will make a better decision. implement a successful Business
d. The division stage Intelgence site. Some methodologies may
The decisions and analyzes made are serve as a guide for consuming the
distributed in the company to better resources effectively and/or funding from
analyze whether the BI solution given to other successful companies.
the user was indeed a good one, and if
(Online Analytical Processing) as well as
the data mining concept.
5. BI Components: OLAP, Data Mining
the complex data of a company. In the dimensions as are needed for the business
OLAP concept a cube is a model model. For dimensions with large number
specifically designed to work with data of records, hierarchical dimensions are
stored on multiple dimensions. The made. They use a field named "parent"
concept of a dimension used with the which groups all fields of the same type.
OLAP technology refers to a characteristic For example, if a store sells shirts with
of the data, not an actual dimension in three colors, red, green and purple, instead
space. of having one field named "Color", there
For a data model to be defined, it is will be a parent-field called “Shirt” and the
necessary to define the model structure child-fields will be "Red", "Green" and
(the objects of the model and the relations "Violet".
between them), the operators (those acting There are two main models for
on the structure), and the integrity implementing cubes, MOLAP
constraints (rules and constraints imposed (Multidimensional OLAP) and ROLAP
to ensure the correctness of the model). (Relational OLAP). MOLAP assumes that
The structure of the model consists of a cube is multidimensional and has the
dimensions (structures spanning different data stored in a multidimensional way.
hierarchical levels that the data are Thus, the data is copied from the data-
grouped by) and tables of facts (containing store to the data-store of the cube, and
measures and foreign keys to dimension aggregates of different size combinations
tables). A time dimension, for example, are pre-calculated and stored in the cube at
may include days of the week, months, or a given vector. This means that the
even years. response time will be very short, but there
Generally in Business Intelligence there is a possibility that explosion of data could
are used three types of schemes: star occur when all combinations of size
schema, snowflake schema and aggregations are stored in the cube.
multidimensional cube. In a star schema ROLAP, as the name suggests, uses the
type (Fig. 5), there is a table of facts, of relational model and considers that it is for
which the dimension tables are connected the best to keep the data stored in the data
to. Dimension tables are not connected warehouse. As in the MOLAP, the data
among them, otherwise the scheme would can be aggregated and pre-calculated using
be called snowflake. A star schema does materialized views. [6]
not correspond to the third normal form, b. Data-Mining
since the dimension tables are made up of Data Mining is the process of finding
several adjacent tables. But this is hidden patterns and associations,
preferable, due to the loss of performance constructing analytical models, achieving
when the schema is in third normal form classification and prediction and
and running operations on large data sets. presenting the obtained results. In other
If it is wanted a variation in the third words, the process of data mining can be
normal form, a snowflake schema can be viewed as examining past data to find
used. [6] useful information that can be used as a
Given the fact that the users who use the guide for the future. Thus, data mining
OLAP technology want quick answers to gets the answers to questions like "What
questions like "How many bikes have sold happened in the past?", "Why it
to customers in Amsterdam in the last four happened?" and "What is likely to happen
months?" several queries are performed on again in the future?”.
large databases and united, in order to To be effective, the process of data mining
provide an answer. For this reason, it is needs "clean" and organized data,
preferred to use a multidimensional model, therefore preferring them to be stored in
the Cube. A cube is consists of as many data warehouses, because they have
32 Business Intelligence overview
already been cleaned, processed and that trends and patterns found,
organized by ETL processes. answer the questions defined in the
According to [7], data mining process is first stage .
divided into the following steps: 6. Launching and updating the model
1. Defining the problem - as in any – in the last step, the model is
Business Intelligence process, first launched in execution so that the
up, is the fully understanding of the users can use it. While in use, it is
problem and the questions need possible that a new set of questions
answers. It is recommended to which need to be answered could
focus on what we want to ask, not occur and the model must be
how we will respond. modified.
2. Preparing data - once the necessary The functions that a system based on
information is well defined, the data mining process performs, are:
data is being prepared. Data mining - Classification - maps data into
algorithms are very intelligent in predefined groups and classes
solving the problem, but will fail if - Clustering - groups similar data
they receive the data in a foreign into clusters
format. As mentioned, it is - Analysis of linkages - discovers
recommended that the data should relationships linking certain
be retrieved from a data data
warehouse. - Finds data that misbehaves
3. Exploring Data - to understand the from the general behavior of
result of the data mining process, the data.
the data must first be covered and A natural question is "why data
studied. This is usually done using mining?". Data mining can be
one of the two methods : successfully used in data analysis
a) Means and extremes: on processes and support of decision
different sets of media making. For example, in a system of
numbers, the means are marketing and management analysis,
calculated and the standard the first elements that are studied, are
deviation from the mean the modes of the origin of data
too. Also, the minimum and (transactions, payments, gift vouchers
maximum values are and so on). The second step is to find
determined. This process clusters of customers who have
helps to understand the data common features (ways of spending
range used for working. the money, income, etc.). Further,
b) Basic statistics: different operations marketing analysis can be
data sets are taken and performed, establishing associations
simple statistical operations and correlations between product sales
are performed. so that predictions based on these
4. Creating the data mining model - assumptions can be made. Also, by
queries are created using DMX clustering and classification, the types
language (Data Mining Extensions) of customers that buy certain products
or using a software wizard. can be determined and may make
5. Exploration and validation of the assumptions about what factors will
model – once the model is build, it attract new customers. [8]
should be explored to see the
trends and patterns that the model A data mining system is generally the
produces and validate it, to ensure architecture shown in Figure 7.
Database Systems Journal vol. V, no. 3/2014 33
User interface
Results evaluation
Iulia MUNTEAN - The passion I have for IT has started 2 years ago,
during a project I made for school. That’s the moment when I decided I
would love to work in this domain. Since then, I started to learn everything
about programming. I am currently working as a web programmer and I
love this job. Every day is a new day to apply what I learned theoretical in
the past and see it’s working.In my spare time, I like reading historical
books, travel and don’t miss any chance to learn something new.
Database Systems Journal vol. V, no. 3/2014 37
This paper focuses on the current state of knowledge in the field of software development
methodologies. It aims to set the stage for the formalization of a software development
methodology dedicated to innovation orientated IT projects. The paper starts by depicting
specific characteristics in software development project management. Managing software
development projects involves techniques and skills that are proprietary to the IT industry.
Also the software development project manager handles challenges and risks that are
predominantly encountered in business and research areas that involve state of the art
technology. Conventional software development stages are defined and briefly described.
Development stages are the building blocks of any software development methodology so it is
important to properly research this aspect. Current software development methodologies are
presented. Development stages are defined for every showcased methodology. For each
methodology a graphic representation is illustrated in order to better individualize its
structure. Software development methodologies are compared by highlighting strengths and
weaknesses from the stakeholder’s point of view. Conclusions are formulated and a research
direction aimed at formalizing a software development methodology dedicated to innovation
orientated IT projects is enunciated.
Keywords: software development, project management, development methodology
1 Introduction
The management process of a
software development projects follows
development projects are notorious for
frequently changing initial planning and
specifications. What do you believe are the
the basic rules of project management but main reasons for changing specifications
also includes particular features. A during the development stage? Project
software development project manager managers responded on 3 out of the 11
has to deal with challenges and setbacks groups. The 3 groups are: PMO – project
that are proprietary to the IT industry. management office, Agile Project
Also in software development there are Management Group and International
benefits and strong points that help ease Society for Professional Innovation
the burden of management. Management. Based on the information
collected from the LinkedIn discussions and
Software development projects are on the author’s own experience as software
notorious for frequently changing initial development project manager specifications
planning and specifications. In order to change due to the fact that:
identify the main reasons for changing the project owner identifies new
specification during the development business opportunities and decides to
stage of a software product debates were integrate them into the software
started on LinkedIn project management being developed;
groups. Debates were initiated on 11 due to the technical nature of
LinkedIn groups starting the discussion software development projects there
with the following introduction: Software is a lack of shared understanding of
38 Comparative Study on Software Development Methodologies
Table 1 summarizes the impact that behaviour patterns. The project team is
software development characteristics have responsible for evaluating the requirements
on the project management and implicitly from a technical perspective. The project
on the project team. The only team will have to research the frameworks,
characteristic that does not have a positive API’s, libraries, versioning tools and
impact is frequently changing of hosting infrastructure that will be required
specifications. in order to build the software product.
research, planning, programing, testing and their core characteristics. The analysis
setup. includes specifying the scale of the project
the methodology is suited for, the stage
The above mentioned stages are generally project owner feedback is delivered and a
agreed by the software development graphic representation of the methodology.
community as being the cornerstones of For coherence reasons stages defined in the
every software development project. second section of the article are also used
Depending of software development in depicting the graphic representations of
methodology they may be found under the methodologies.
different naming conventions, they may be
overlapping changing order or missing Waterfall is the first methodology
altogether. generally acknowledged as being dedicated
to software development. Its principals are
3. Current software development for the first time described by Winston W.
methodologies Royce even though the actual term
A software development methodology is a waterfall is not used in the article [24]. It
set of rules and guidelines that are used in emphasizes meticulous planning and it
the process of researching, planning, outputs comprehensive documentation.
designing, developing, testing, setup and The Waterfall methodology is linear-
maintaining a software product. The sequential process where every stage starts
methodology also includes core values that only after the previous has been completed.
are upheld by the project team and tools Each stage has its own deliverables. The
used in the planning, development and Waterfall methodology is predictable and
implementation process. This paper values rigorous software planning and
reviews 20 of the most popular software architecture.
development methodologies and highlights
The project owner’s feedback is received software application one step at the time in
after the prototype is completed. The the form of an expanding model [3]. Based
Prototyping methodology is suitable for on initial specification a basic model of the
large scale projects where is almost application is built. Unlike the prototype,
impossible to properly define exhaustive the model is not going to be discarded, but
requirements before any actual codding is is instead meant to be extended. After the
performed. Prototyping methodology is model is tested and feedback is received
also suitable for unique or innovative from the project owner specifications are
projects where no previous examples exist. adjusted and the model is extended. The
process is repeated until the model
Iterative and incremental is a becomes a fully functional application that
methodology that relies on building the meets all the project owner’s requirements.
The project owner’s feedback is received methodology has 4 major phase: planning,
after each iteration is completed. The risk analysis, development and evaluation.
Iterative and incremental methodology Project will fallow each phase multiple
emphasizes design over documentation and times in the above mentioned order until
is suitable for medium and large projects. the software application is ready to be
setup on the live environment. The Spiral
Spiral is a methodology that focuses on methodology emphasizes risk analysis and
identifying objectives and analysing viable always evaluates multiple alternatives
alternatives in the context well documented before proceeding to implementing one.
project constrains [4]. The Spiral
Database Systems Journal vol. V, no. 3/2014 43
The project owner’s feedback is received much faster development and higher-
after the first iteration of the spiral is quality results than those achieved with the
completed. The Spiral methodology is traditional methodologies. It is designed to
suitable for medium and large scale take the maximum advantage of powerful
projects. It has also proven more effective development software [15]. Rapid
in implementing internal projects as application development imposes less
identifying risks proprietary to your own emphasis on planning tasks and more
organization is easier. emphasis on development. Development
cycles are time boxed and multiple cycles
Rapid application development is a can be developed at the same time.
development lifecycle designed to give
The project owner’s feedback is received projects with the constraint that projects
after each module is completed. The Rapid have to be broken down into modules.
application development methodology is
suitable for small, medium and large scale Extreme programming breaks the
conventional software development
44 Comparative Study on Software Development Methodologies
process into smaller more manageable and the other is supervising. They change
chunks. Rather than planning, analysing, roles at regular intervals. For reducing the
and designing for the entire project at once, number of errors it relies heavily on unit
extreme programming exploits the testing and developers are required to write
reduction in the cost of changing software tests before writing the actual code. There
to do all of these activities a little at a time, is a collective ownership code policy
throughout the entire software where any developer can change any code
development process [5]. It enforces pair sequence even if it was not written by him.
programming where two developers use The project owner is the one that decides
the same computer. One is writing code the priority of the tasks.
The project owner’s feedback is received suitable for small and medium scale
in the form of acceptance testing after the projects.
entire application is completed. The V-
Model development methodology is Scrum is a methodology for incrementally
building software in complex
Database Systems Journal vol. V, no. 3/2014 45
needs of the business [17]. Dynamic standards at the beginning of the project
systems development methodology is an and it sets non-negotiable deadlines.
iterative development model that uses a Testing is done early and continually
timebox approach and MoSCoW task throughout the entire development cycle.
prioritization. It defines strict quality
Project team and project owner share a the-box roadmaps for different types of
workplace, physical or virtual, in order to software projects and it provides guidance
facilitate efficient feedback at every stage for all aspects of a software project. It does
of the project. The Dynamic systems not require the project team to engage in
development methodology is suitable for any specific activity or produce any
medium and large scale projects. specific artefact. It provides guidelines that
help the project manager tailor the process
Rational unified process methodology if none of the out-of-the-box roadmaps
provides a disciplined approach to software suits the project or organization [8].
development. It comes with several out-of-
The project owner’s feedback is received is suitable for small, medium and large
as agreed at the start of the project as the scale projects.
methodology does not enforce a rule on
project team – project owner collaboration. Lean software development is a product
The Rational unified process methodology development paradigm with an end-to-end
Database Systems Journal vol. V, no. 3/2014 47
focus on creating value for the project also insures that team members are
owner and for the end user, eliminating motivated by empowering them to make
waste, optimizing value streams, significant decisions regarding the
empowering people and continuously application. The Lean software
improving [9]. Value is defined as development methodology does not
something that the project owner would enforce a certain process in terms of
pay for. Anything that is not adding value conducting the project. Project manager
is considered waste and has to be and team members are free to use any
discarded. Lean software development process they see fit as long as it stays true
delivers early iterations of working code. It to core lean development principals.
The project owner’s feedback is central to new functionality. If the tests work then
the Lean software development there is no need to write any code as the
methodology. The Lean software functionality already exists, it was just not
development methodology is suitable for know to the developer. This scenario is
small, medium and large scale projects. often encountered when dealing with
legacy code. If tests do not compile then
Test-driven development is a the developer will write the code and run
methodology developed around unit the tests again. The process is repeated
testing. Before writing any actual code the until all requirements are met [10].
developers write automated test cases for
The project owner’s feedback is received ways of handling requirements [12]. Based
after the application is already setup but on project owner’s requirements a
constant interaction between development metamodel is define. The metamodel is
team and project owner takes place actually a platform independent model that
throughout the entire length of the project. can be migrated onto any environment.
The Feature-driven development UML is usually used for building the
methodology is suitable for small, medium metamodel. The metamodel is then
and large scale projects. converted into a model that is specific to a
certain platform. Based on the model the
Model-driven engineering is a complex actual code is then generated.
methodology that uses domain models as
In open source software development there The methodology is suitable for small,
is no project owner to provide feedback. medium and large scale projects.
Database Systems Journal vol. V, no. 3/2014 51
The project owner’s feedback is received They were not included in the current
after deployment. The Microsoft Solutions comparative study as they are variation of
Framework is suitable for small, medium other methodologies and their
and large scale projects. characteristics were already submitted to
analysis.
Apart from the above mentioned
methodologies a further research of the 4. Strengths and weaknesses
topic might consider analysing the Software development methodologies
following methodologies: follow one of two paths: heavyweight or
Dual Vee model – a variation of the lightweight. Heavyweight methodologies
V model; are derived from the waterfall model and
Evolutionary project management – emphasizes detailed planning, exhaustive
forerunner of the current agile specifications and detailed application
methodologies [23]; design. Lightweight methodologies are
Agile Unified Process – a derived from the Agile model and promote
simplified version of the Rational working software, individuals and
Unified Process; interactions, acceptance of changing
Essential Unified Process – a requirements and user feedback. The
variation of the Rational Unified Microsoft Solutions Framework includes
Process; variations for both heavyweight and
Open Unified Process - a variation lightweight philosophies.
of the Rational Unified Process
[22].
This paper tries to show that the need to implement BI solutions that integrate social networks
data is increasingly more acute. Companies need to understand that efficient decision making
has to take into account the importance that social medial has nowadays. In addition, the
software developing companies have to keep up to date with new trends generated within
these communication channels and adapt their solutions to this dynamic environment.
Keywords: Business Intelligence, Social Network, Social BI, Market Trends
1 Introductory notes
Social BI is a management technique
that integrates the sharing of social
resources in order to improve
Currently, through the contribution of
social networks to the scope of BI, the
focus is increasing on their ability to bring
a lot of customers for small companies that
existing company projects, products or are producing for niche markets. In
business processes. This type of BI is addition to this, they provide analysis of
usually handled by software products that trends and statistics with greater relevance
offer the possibility to analyze social to the company and the end customer that
media indicators while, at the same time, is involved in creating the products for
managing projects like a traditional which he is the target of. The next figure
solution. Also, the main characteristic of allows us to see the features outlined by
this type of BI is that it welcomes the social media networks into the BI domain
involvement of the client from the starting [2].
point of business processes and long time As networks and internet applications are
before the product development end with becoming increasingly complex, given
its market release. increasingly more detailed data can be
Social BI is seen as an ideal way to take collected about users/consumers and sent
full advantage of human capital, both to companies operating in narrow markets,
inside and outside the company. From the which they can use for commercial
employee’s point of view, this kind of purposes (e.g. loyalty campaigns,
application offers transparent information promotions, etc.).
distribution across project assignments, no At the same time, user interaction can lead
matter if they are involved or would like to to favorable information dissemination
be involved in these projects. Also, it about the company which can become
creates a widely accepted environment for major waves, this being among the most
merit acknowledgement. On the other effective ways for market advertisement
hand, outside stakeholders (mainly clients) today. Also, this information can turn into
receive the means to actively give reliable sources for consumers who want to
feedback on product/services development, document about a specific product that
ensuring themselves that they want to purchase. Propagation speed
they will have a higher level of satisfaction of this medium is constantly increasing,
of their needs [5. actively using friend networks and
acquaintances of users.
58 Social BI. Trends and development needs
Companies need to recognize the need for to develop into a significant source of
virtual existence in the social environment personal data at the individual, opinion,
because the potential for loyalty and sales relationship and interests level. As a result,
increase is considerable. Even if the the lasting parallel existence of these two
purpose is just notifying the end user that concepts, led inevitably to the compund
the company exists is huge and at the same concept of Social BI.
time, the effects are measurable and the Social BI, conceptually wants to extract
measurements reliable. relevant information that would lead to
The fact that both companies that are effective decisions, all based on data from
trying to sell products or services and also social media. Also, it seeks the
social networks have as a common goal the participation in the design of BI solutions
meeting of the needs of the same that contain tools for analysis and
customer/end user, will lead over time to measurement of social metrics. In the
the strengthening of relations of present context, in which more companies
interdependence. This will actively be succeed in implementing BI systems, the
reflected in new products for BI techniques addition of this new component leads to
and processes. These tools will provide the accentuated competition and a major
link that will connect the customer to the benefit for the end customer [2].
companies that serve him and bring
increased efficiency and competitive 2 BI Interaction Management
advantage for businesses that will invest in
these solutions. Management Systems of social networking
BI areas of influence and social media as an integral part of BI, is a group of
were, since their appearance, the subject of applications or methods used to manage
continuous development and research. and follow business processes in a
While BI aims to help decision-making at distributed environment. These systems
company level by providing reports can be manually or automatically
containing relevant data, social media aims maintained and they allows managers to
track, aggregate, publish and take part in
Database Systems Journal vol. V, no. 3/2014 59
many social media channels using a single able to adjust or program the
software product.[4] occurrence of one or more messages;
The operating processes have three main Ability to manage social data. The
features: system allows managers to view
Connection to social media aggregated data and have various
channels; analytical reports with different levels
Ability to quickly publish from of complexity and detail.
anywhere inside these channels being
Fig. 2. BI interactions
Each of these basic functions can be reach the end user. Interacting with social
extended to a very high level of detail. media, CRM uses information provided by
Social media platforms are proving to be it in order to publish messages that should
an important component of any marketing reach both current customers as well as
strategy. At the same time, marketing has prospective customers. By integrating
come to use many of the BI tools to reach these two components, potential customers
for the target audience, often extracting will know what to expect if they choose a
data directly from the internal business product provided by a certain company.
processes. Consequently, we see the need Inside the social media component, gaining
for marketing to track customer response credibility is a matter not only of content,
in real time. but also of a good social interaction. A
CRM (customer relationship management) common mistake is to approach social
focuses on providing services, loyalty media as platforms to send some mass e-
techniques and sales analysis that address mails and not as some communication
actual client needs in a more direct way processes. They should be considered more
that before on what they consider to be the like publishing tools through participation.
main issues of company-customer As mentioned previously, CRM integration
relationship (prices, warranties, shipping, with social media result in a long term
etc.). marketing strategy. The problem here is
CRM is fundamentally different from the huge amount of data coming through
social media components in that it is much these two channels. This is the time when
less immediate. This is the mechanism BI techniques must intervene in order to
through which contracts and products interpret the data into relevant reports
60 Social BI. Trends and development needs
available for the decision-making factor. Oracle and IBM which provide huge
What BI cannot do, however, is improve processing capabilities and complex
the quality of input data. External analysis.[3]
consultants are needed to standardize the
data stream so that it becomes valuable for 3 Social media dynamics and associated
the company. The existence of these metrics
expensive consultants does not preclude
Regarding the most important changes that
the use of free, but very simple services
(e.g. the number of brand occurrence over social media has been through since its
a given period of time). Top market share appearance, we can summarize the
service providers in this area are SAP, following in Table 1.[2]
Regarding the most important social media this can be achieved by the same methods
metrics that companies have to be of filtering online content.
responsive to, consultants identify the
following:
4 Development possibilities
1. Monitoring views of a target Characteristics of social media together
demographic in a certain geographical area with design elements of BI currently
by filtering used keywords in natural contribute as a foundation for any social BI
language to identify opinions of strategy. As areas of development for the
satisfaction or negative feelings associated above mentioned tools, we mention the
with a product or company. This process following:
can prevent a viral explosion of negative
opinions. If negative opinions are needed Users and customers: the inclusion of
direct intervention is needed to solve valid social media related set of data available
issues of the most influential individuals. for analysis can attract new BI users within
2. Identifying cases where negative an organization. The importance attached
opinions could turn into a PR crisis is to social interaction between users is an
achieved through effective monitoring of example of the new design requirements of
social media to identify key individuals BI. Also, if the BI system allows and
before the problem gets proportions. encourages interaction with customers, this
3. Identify individuals, blogs and web will generate new technical needs.
applications that are most influential to a Products and services: along with the
brand should lead to discussions on topics availability of data from social media, new
with which they have shown interest. products and services within the BI can be
4. Real time tracking of the evolution offered for use. Inside this context,
of views and opinions, especially about development directions include the
new products and services leads to discovery of methods to integrate
shortening the time between the signaling traditional BI results and analysis related to
of the problem and its solution. the structure of social networks or opinions
5. Measuring the link between trends. In addition, the limitations
marketing efforts and customer response regarding the quality and security of data,
at certain times of day and divided into will constantly update the need for
geographical areas leads to the appropriate tools.
customizing level that marketing Processes: some BI processes should be
campaigns need to be more efficient. adapted for the integration of social media
6. Determining the type of media and functionality. However, the research and
platforms that have the most success development requirements are relatively
allows the company to focus its efforts. For low in this case, contrary to the situation in
example, a company may divert resources which the BI system is used directly in the
to video campaigns because this means of interaction with the clients’ calls via live
communication has proved most effective channels.
in attracting clients. Data: nearly all of the social media
7. Monitoring competitors must be impacts data management architecture.
done using all available means to redeploy There will be a need for a broad spectrum
their tested business formula into the approach to BI tools if they are to process
company’s own processes. both traditional transactional data and data
8. Documenting the industry sector in from social media. Social data type is
which the company operates leads to rapid extremely dynamic and therefore cannot
identification of growth opportunities and guarantee an accurate mapping over a long
62 Social BI. Trends and development needs