You are on page 1of 8

Clementine for Public Sector

Discover predictive
power in your data

Combine your data and business knowledge with Clementine

to create and deploy powerful predictive models


Make a difference with the predi
Your organization’s day-to-day activities result in a flood of data.
This information holds — and often hides — the experience
of your organization’s past. By putting this information to work,
you can solve your toughest problems. With Clementine, your
organization learns from its past to improve its future.

“ Clementine sets Improve efficiency with data mining


Discover the patterns and trends that enable your
the standard for government agency to work more efficiently and
open up a whole new world of opportunities. Data
data mining
mining enables you to learn from experience by
workbench tools discovering knowledge in your data. When data
mining with Clementine, you’ll uncover new dimen-
with the level of sions in your data using predictive modeling.

support it Discover and deploy better solutions


with Clementine’s workbench of tools
provides for all Clementine’s comprehensive set of techniques helps
Clementine helps you discover solutions you otherwise
wouldn’t by leveraging two of your organization’s most
stages of the data you find answers to your organization’s toughest valuable assets — data and your business knowledge.
challenges. You begin by combining two of your
mining process.” most valuable assets — your data and your business Clementine empowers you to:
knowledge — with Clementine to create powerful ■ Predict citizen behavior for citizen relationship

Ovum Data Mining Report models. You’ll find a complete set of analytical management (CRM)
techniques: from neural networks and decision ■ Mine Web data by discovering and predicting

trees to logistic regression, visualization and data clickstreams, the paths visitors take through
preparation procedures. With this variety, you are your Web site, and redesign your site to improve
assured of getting the best results and will be ready e-government services
to tackle business problems as they arise. Build ■ Improve military readiness when you maximize

predictive models with Clementine. Then deliver equipment maintenance schedules and ensure
them throughout your organization to achieve real- you attract and keep more recruits
world solutions with concrete benefits — such as ■ Classify your citizens into specific categories to

more targeted service delivery, decreased costs and ensure that the right programs serve the right
improved processes. people
■ Identify likely cases of fraud or non-compliance

Scale the entire interactive data mining process to recoup more money for your organization
Clementine scales to the size of your challenge, even ■ Fight crime by unmasking the characteristics

if opportunities for achieving your agency’s goals are of crime


hidden deep within massive datasets with millions of ■ Forecast service utilization and apply your

rows. Clementine’s approach to scalability leverages resources where they’re most needed
your existing IT investment by pushing processing
back into the database with in-database mining.
Therefore, you don’t need expensive new investments
in hardware and software and complicated configu-
rations to experience scalable data mining.
2
ictive power of data mining
Meet your aggressive goals
for moving services online
How does your agency ensure its e-government initia-
tives respond to your citizens’ needs? Web mining —
data mining using Web data — empowers you to
understand your visitors better. Web mining tells you
how different groups of people use your Web site and
enables you to predict what your visitors expect from
your site. With this information, you can improve your
Web site design and content — and better present
online service delivery. Move beyond counting clicks to predicting visitor behavior.
Understand page affinities, such as the possible relation-
Get faster results from your Web mining projects
ship between the “HomePage” and the “InfoForYou” page,
using the Clementine Application Template (CAT) as shown in this Web graph, and deploy predictive models
for Web Mining. By using pre-built streams of complex to personalize visits.
but common operations, such as accessing and
merging Web log data, you can expedite the time
needed to see results. These templates follow the
industry-standard CRISP-DM (CRoss-Industry
Standard Process for Data Mining) methodology Business
understanding
Data
understanding

for data mining and use real-world Web mining


experience, so you can be confident that your
Data
project will benefit from a proven methodology preparation

and best practices. Deployment Data

Modeling
With Clementine and the Web Mining CAT, you can:
■ Examine visit length

■ Profile key events such as downloads or payments

■ Find and describe different types of visits


Evaluation

■ Identify clickstream sequences

■ Predict the next action in a clickstream using

Capri, an add-on module that identifies sequences The CRISP-DM method, shown above, provides a roadmap for
■ Make recommendations based on visit profiles beginners and a checklist for experienced data miners.

Government body uses Clementine to improve education


Situation Solution
A state’s department of education faces pressure to meet academ- The department selected Clementine to guide its data mining
ic goals and educate its students more effectively. The department efforts. Using Clementine, the department:
needs to improve standardized test scores to demonstrate that it ■ Uncovered patterns in classes to identify the effects of

adequately educates children. curriculum structure on learning


■ Explored the relationship between the sequence of classes

Critical issue and test scores


The department wanted to explore the relationship between cur-
riculum structure and standardized test performance to understand Results
how course sequence affects test scores. ■ Maximized curriculum structure to ensure more effective learning

3
Clementine’s scalability protect
Scalability that supports interactive data mining
Data mining is a discovery-driven process — when
you see a relationship that sparks your curiosity,
you need to interactively explore and analyze the
pattern. For example, if you see a strong relationship
between your Web site’s home page and the search
page, you can drill down to discover if visitors truly
find the information they need. These discoveries
are a necessary step to building higher-performing
models.
“ Without Clementine, Clementine’s approach to scalability supports this
interactivity. Most data mining tools focus on scaling
analyzing this amount narrow steps, which affect only the speed of a part of Interactive graphs invite your firsthand knowledge into the
the process. Instead, Clementine focuses on scaling as data mining process as you work toward a solution.
of data would have many steps as possible to give you faster results overall.
taken months. Using
Perform operations where they can be handled Shorten your project’s life cycle using
Clementine, it took most effectively — via in-database mining time-saving tools
Clementine mines large datasets by intelligently Clementine provides tools for getting started quickly
just weeks.” distributing processing to more efficient database and evaluating your results. Clementine comes with
and analytical server tiers. Clementine delegates application templates to give you a head start for
UK Ministry of Agriculture many operations to the database for processing. tackling common problems. Use the Web mining
Fisheries and Food The operations that a database cannot perform are template to process and enhance raw Web log data
processed on a powerful server tier. In addition to and discover clickstream sequences. Data mapping
shortening model-building time, this approach enables tools facilitate matching your data with these best-
you to visualize, explore and manipulate large datasets practices templates and enable you to share your
more efficiently. templates with colleagues. The ease with which you
can build and combine Clementine’s full complement
of statistical and machine-learning algorithms means
that you find the right combination of models quickly.
You can easily evaluate multiple models with lift,
gains, response and other model evaluation graphs
in a one-step process, shortening your project time.

A government tax agency better detects non-compliance


Situation Solution
Under increasing budgetary pressure, a government tax agency The tax agency focused its search for non-compliant taxpayers
needed to maximize its auditing process and recoup more money. by mining its data with Clementine. By exploiting the associations
between the line entries and the amount of tax that was adjusted,
Critical issue it built a model to predict the likelihood of a new return being
The search for non-compliant taxpayers presents one of its greatest non-compliant – based solely on the contents of the return.
demands for resources, both in terms of money lost and the
amount of time spent on the auditing efforts. Previously, the tax Results
agency used past experience and intuition to establish subjective ■ Maximized its auditing process by detecting the returns most
guidelines to identify the typical non-compliant profile. But these likely to warrant adjustment
unproven theories wasted auditors’ time on compliant taxpayers ■ Recouped millions of dollars of tax revenue that would have
while non-compliant taxpayers were missed. been otherwise lost through non-compliance
4
s your existing IT investment
The details
How the interactive Clementine
knowledge discovery process works model as input into another model. These “meta-
models” consider the initial model’s decisions and
See your solution discovery process clearly can improve results substantially.
The interactive stream approach to data mining is the
key to Clementine’s power. Using icons that represent Understand variations in your business with
steps in the data mining process, you mine your data by visualized data
building a stream — a visual map of the process your Powerful data visualization techniques help you
data flows through. Start by simply dragging a source understand key relationships in your data and guide
icon from the object palette onto the Clementine the way to the best results. Spot characteristics and
desktop to access your data flow. Then, explore your patterns at a glance with Clementine’s interactive
data visually using graphs. Apply several types of graphs. Then “query by mouse” to explore these
algorithms to build your model by simply placing the patterns by selecting subsets of data or deriving new
appropriate icons onto the desktop to form a stream. variables on the fly from discoveries made within
the graph.
Discover information using Clementine’s
interactive streams
Work toward a solution by applying your business
expertise to select the next step in your stream, based
on the discoveries made in the previous step. You can
continually adapt or extend initial streams as you work
through the solution to your business problem.

Easily build and test models


All of Clementine’s advanced techniques work together Clementine’s interactive stream approach to data mining lets
to quickly give you the best answer to your business you quickly test ideas as you work through the solution to
problems. You can build and test numerous models to your business problem.

immediately see which model produces the best results.


Or you can combine models by using the results of one

Clementine Server
helps you discover
better solutions
by leveraging your
distributed computing
architecture to mine
large data tables
faster. Clementine
Server efficiently
distributes your data
mining stream to your
database, application
server and client for
efficient processing.

5
Maximize your ROI when you p
Seize opportunities with data-driven strategic deployment with Clementine Solution
decision making Publisher takes you much further — maximizing
Your employees make decisions every day that affect data mining return-on-investment (ROI) by offering
your service quality. If you don’t share the knowledge deployment options that meet your specific needs.
gained through data mining, you run the risk that Clementine Solution Publisher’s rapid deployment
uninformed employees might make costly decisions approach — which automates data access and pro-
that compromise services. With Clementine, you can cessing steps in addition to models — eases the
put your data mining solutions to work at the point integration of powerful predictive models into your
of decision making — ensuring consistent, data-driven organization. This makes real-time use of models
decisions are made throughout your organization. for decision making possible — and empowers your
organization to provide more efficient service delivery.
Maximize your data mining ROI With a deployed model, front-line managers can
with strategic deployment feed new information into models to get results that
Data mining deployment often comes in the form of boost returns in day-to-day activities. Or the model
“ Deploying a data printed reports given to auditors and investigators. can serve as an analytical engine, working behind
mining solution While this is an important deployment method, the scenes to enable your Web site to offer content
tailored for each visitor.
throughout an
organization is the How organizations deploy Clementine models

big issue . . . a tool Target programs to Web sites offer per- Military personnel can
those who can most sonalized content predict part failure due
like Clementine benefit from them. for each individual to use, manufacture
visitor for more or design failure and
successful maximize maintenance
Solution Publisher e-government schedules to increase
initiatives. military readiness
provides a faster while lowering costs.

means to automate
the power of data Fraud investigators Health care profes- Program managers
prioritize their cases, sionals provide the find better, more cost –
ensuring limited best possible care effective ways to
mining.” resources are spent by identifying trends deliver social services
on getting the largest in treatments and to the right citizens.
return. processes that produce
Wolfgang Martin, the best outcomes.
Vice President, META Group

Armed forces cut maintenance costs


Situation Solution
Modern day armed forces face many challenges, such as improving The defense agency selected Clementine to improve its mainte-
readiness to respond quickly to smaller conflicts and cost-effectively nance process. The agency used exploratory data mining to
maintaining military equipment. understand the relationship between part failure and tank design,
manufacture and usage. The agency built predictive models to
Critical issue streamline the maintenance processes by fixing more parts from
To lower the costs of maintenance and increase readiness, a the same tank at a time, increasing the amount of time the vehicle
defense agency needs to build models to predict tank part can be deployed in the field.
failure, maintenance and overhaul.
Results
■ Lowered maintenance costs by taking steps to avoid part failure
■ Increased the time equipment can remain in the field, thereby
improving readiness

6
ut your solution to work
The details
How Clementine cost-
effectively delivers your results

Publishing solutions to meet your data mining goals


is simple. First, analysts build powerful models using
Clementine’s visual workflow environment. Analysts
add the Clementine Solution Publisher node to the
With Clementine, you deploy a model.
Clementine stream and execute the node to create
all the files needed to build a stand-alone solution.
Then, analysts deploy the solution where it’s needed
most — whether it’s to score a database or to give
real-time responses at the point of decision making.
Then, decision makers “close the loop” by sharing
results with analysts to improve the solution on
a continual basis.
With Clementine Solution Publisher, you deploy
Clementine Solution Publisher enables rapid — and the complete data mining process built with
Clementine — from data access and data preparation
cost-effective — deployment of data mining solu-
through models and results.
tions, empowering non-analysts to leverage work
performed in Clementine. Clementine Solution
Publisher is the only product that enables you to
deploy a complete data mining process — from data Continually improve the data mining
access through results — to maximize data mining process by “closing the loop”
ROI by sharing your knowledge discovery. Clementine Solution Publisher enables you to
“close the loop” in your data mining process.
Clementine Solution Publisher is cost-effective Once a data mining solution is deployed, the
for a number of reasons: analysts can monitor the effectiveness of the
■ Avoid costly programming solution on new data from the field. Then,
Clementine Solution Publisher generates an they can refine the Clementine stream to
image file for the full data mining process for the increase the solution’s effectiveness. Clementine
pre- and post-processing steps in your data mining Solution Publisher is the only data mining
process — instead of just generating the code for product that empowers organizations to con-
a model. You avoid the costly development work tinually update the data mining process and
and time needed to program these steps manually. then redeploy updated solutions to decision
■ Easy maintenance controls the cost of change makers throughout the organization.
You can maintain your data mining application
affordably by changing the stream in Clementine
and re-publishing the process. Clementine Solution
Publisher’s rapid application development process
maximizes your organization’s resources by enabling
you to maintain your data mining applications easily.

7
Clementine features Learn more about SPSS
Clementine Application Templates – Model scoring, including combinations Contact your nearest SPSS office or visit us at www.spss.com.
■ Web mining of models
■ Tools for managing files built using templates – Post-processing SPSS Inc. +1.312.651.3000
■ Runtime environment for executing image Toll-free +1.800.543.2185
Data access file on target platforms SPSS Argentina +5411.4814.5030
■ Data input: ■ Easy update of solutions through small
– Native access to database image file SPSS Asia Pacific +65.245.9110
management systems including
SPSS Australasia +61.2.9954.5660
Oracle, SQL Server, DB2; additional System requirements
access to any ODBC-compliant Clementine client: Toll-free +1.800.024.836
data source ■ Hardware: Pentium-compatible processor or
SPSS Belgium +32.16.317070
– Import delimited and fixed-width text, higher and a monitor with 1024 x 768 resolution
® ®
SPSS and SAS 6, 7, 8 files or higher (support for 65,536 colors is recom- SPSS Benelux +31.183.651.777
■ Data output: mended). A CD-ROM drive for installation is SPSS Brasil +55.11.5505.3644
– Delimited and fixed-width text, ODBC, also required.
®
SPSS, Microsoft Excel, SAS 6, 7, 8 ■ Operating system: Windows 95, Windows 98, SPSS Czech Republic +420.2.24813839
Windows 2000 or Windows NT 4.0 with SPSS Danmark +45.45.46.02.00
Data preparation Service Pack 3 or higher.
■ Generate subsets of data automatically from ■ Minimum free drive space: 80MB SPSS East Africa +254.2.577.262
graphs and tables – Additional option: Clementine Application
■ Choose from various data cleaning options
SPSS Federal Systems (U.S.) +1.703.527.6777
Template for Web Mining: 115MB
■ Manipulate data with complete record and – Additional option: Clementine Application Toll-free +1.800.860.5762
field operations, including: Template for Telecommunications: 60MB SPSS Finland +358.9.4355.920
– Field filtering, naming, derivation and ■ Minimum RAM: 128MB

value replacement SPSS France +01.55.35.27.00


– Record selection, sampling, merging and Clementine Server: SPSS Germany +49.89.4890740
concatenation, sorting, aggregation ■ Hardware: Pentium-compatible

and balancing processor or higher for Windows, SPARC SPSS Hellas +30.1.72.51.925
– Specialized manipulations for showing for Solaris, HP Workstation for HP/UX or SPSS Hispanoportuguesa +34.91.447.37.00
the “history” of values and converting IBM RS/6000 for AIX. A CD-ROM drive for
set variables into flag variables installation is also required. SPSS Hong Kong +852.28119662
■ Operating system: Windows 2000 or
SPSS Ireland +353.1.415.0234
Modeling algorithms Windows NT 4.0 with Service Pack 3
■ Prediction and classification: or higher; Solaris 2.6, 7 or 8; HP/UX SPSS Israel +972.3.6506022
– Neural networks (Multi-Layer Perceptron, 10.20 or 11.0; AIX 4.3.
Radial Basis Function) ■ Minimum free drive space: 25MB for instal- SPSS Italia +800.437300
– Decision trees and rule induction lation; plus at least twice the drive space SPSS Japan +81.3.5466.5511
(C5.0, C&RT) of the amount of data to be processed.
– Linear regression, logistic regression, ■ Minimum RAM: 128MB SPSS Korea +82.2.3446.7651
multinomial logistic regression SPSS Latin America +1.312.651.3539
■ Clustering and segmentation: Clementine Solution Publisher Runtime:
– Kohonen network, Kmeans, TwoStep ■ Hardware: Pentium-compatible SPSS Malaysia +603.7873.6477
■ Association detection: processor or higher for Windows, SPARC SPSS Mexico +52.5.682.87.68
– GRI, Apriori and Web visualization for Solaris, HP Workstation for HP/UX or
■ Data reduction: IBM RS/6000 for AIX. A CD-ROM drive for SPSS Miami +1.305.627.5700
– Factor analysis, principle components installation is also required.
SPSS Norway +47.22.40.20.60
analysis ■ Operating system: Windows 2000 or

■ Combine models for greater accuracy Windows NT 4.0 with Service Pack 3 SPSS Polska +48.12.6369680
■ CEMI interface for custom algorithms or higher; Solaris 2.6, 7 or 8; HP/UX 10.20
or 11.0; AIX 4.3. SPSS Russia +7.095.125.0069
Interactive visualization ■ Minimum free drive space: 4MB for installa-
SPSS San Bruno +1.650.794.2692
■ Query by mouse to explore subsets of data tion; plus at least twice the drive space of
in a graph the amount of data to be processed. SPSS Schweiz +41.1.266.90.30
■ Histograms, distributions and other bar graphs ■ Minimum RAM: 128MB
SPSS Singapore +65.324.5150
■ Line and point plots

■ Web association detection Capri for Clementine: SPSS South Africa +27.11.807.3189
Optional plug-in algorithm for SPSS South Asia +91.80.2088069
Scalability detecting sequences
■ In-database mining and server-tier ■ Hardware: Pentium-compatible processor SPSS Sweden +46.8.506.105.50
processing for scalability or higher for Windows, SPARC for Solaris. A
■ Minimized network traffic via intelligent
SPSS Taiwan +886.2.25771100
CD-ROM drive for installation is also required.
field projection ■ Operating system: Windows 98, Windows
SPSS Thailand +66.2.260.7070
2000 or Windows NT 4.0 with Service Pack
Deployment with Clementine 3 or higher; Solaris 2.6. SPSS UK +44.1483.719200
Solution Publisher ■ Minimum free drive space: 200KB
In addition to these offices, SPSS has a worldwide network of
■ Automated export of all operations, including: ■ Minimum RAM: 128MB
distributors. Contact the SPSS office nearest you for assistance.
– Data access
– Data manipulations
SPSS is a registered trademark and the other SPSS products named
are trademarks of SPSS Inc. All other names are trademarks of their
respective owners. Printed in the U.S.A.
© 2001 Integral Solutions Ltd. CLMPS6BRO-0401W
Data mining makes the difference™
The company delivers solutions at the intersection of customer relationship management and
business intelligence that enable its customers to interact with their customers more profitably.
SPSS’ solutions integrate and analyze marketing, customer and operational data in key vertical
markets worldwide including: banking, consumer packaged goods, finance, health care,
insurance, manufacturing, retail, telecommunications, market research and the public sector.
Headquartered in Chicago, SPSS has more than 40 offices. For more information, visit
www.spss.com.

You might also like