Professional Documents
Culture Documents
COLLEGE OF COMPUTING
DEPARTMENT OF INFORMATION SYSTEMS
December 2019
1 12/26/2020
Topics to be Covered
Introduction
Architecture
Application
Problem solved
Approach followed
Result achieved
Conclusion
Recommendation
References
2 12/26/2020
Introduction
Information filtering: We must go through an enormous amount of
The information filtering tasks are generally more complex and more
difficult to automate than information retrieval tasks.
Information retrieval (IR) involves typically queries from relatively static
3 12/26/2020
Cont..
A more interesting however approach is followed by the
4 12/26/2020
cont…
Information filtering is a particularly important function in KM
5 12/26/2020
Information Filtering Agents
Information Filtering agents find the content of interest to a
user.
Information Filtering agents could gather information from
different sources
They could filter information based on user’s personal interest
Information filtering agents may use Information Retrieval
techniques
Vector space models, where a document is represented as a
vector of attributes.
Tree structure, which represents a hierarchical view of a
document
6 12/26/2020
Cont…
Metrics: The main metrics used to test the accuracy of the
retrieval/ filtering algorithm are precision and recall.
These are defined as:
7 12/26/2020
Example of information-filtering agents
An advanced bot that finds and categorizes relevant information based on the
Search Pad user’s preferences, also learning from them.
An agent that carries out Net searches by simultaneously consulting the most
Copernic important search engines
KOS (Knowledge A new class of intelligent information retrieval Tools built by modeling how
Object Suite) we learn. Cognitive science, collaborative knowledge sharing, and
knowledge modeling that continues where search drops you by “reading” the
knowledge in search results.
8 12/26/2020
Architecture IF
9 12/26/2020
Advantage and disadvantage
Advantage
To make sense of data
Reduced information overload
Information profile
Disadvantage
number of website rename
The recommendation based on existing interest of user
10 12/26/2020
Building Software Agents for
Information Filtering on the Internet:
A Genetic Programming Approach
11 12/26/2020
Introduction
Now a days, The Internet is one of the largest public available
"databases" of documents.
Number of services have arisen on the Internet to help users
search and retrieve documents from servers around the world.
Software robots are autonomous agents that interact with real-
world software environments such as operating systems and the
World Wide Web. Number of software agents have been built to
help users.
Agents for information gathering on the Internet has to solve at
least two problems in common. The one concerns where and how
the relevant documents are retrieved.
The other is to determine the preference of documents according
to the current specific interests of the user.
12 12/26/2020
Statements of the problem
A human user to access a wide range of information resources by
14 12/26/2020
Methodology
This study focuses on the six step procedure that involves two
agents cooperate each other i.e search agent and a filtering agent.
1. Initialize the knowledge bases of the search agent and the
filtering agent.
2. The search agent uses his knowledge base to search for
documents on the Internet.
3. The filtering agent uses his own knowledge base to give
preferences to the articles suggested by the search agent.
4. The user reads the articles in a decreasing order of preference
and gives relevance feedback for each document.
5. The filtering agent learns from the relevance feedback and adapt
his knowledge base.
6. Go to step 2.
15 12/26/2020
Methodology …
In simulations, 100 documents were used for evolving agents.
16 12/26/2020
Experimental Results
The test domain was the evolutionary computing area. Each document was
given a relevance feedback.
The relevance feedback is either positive (+1) or negative (-1) and determined
by the user.
A half of the documents were positive examples and another half were negative
examples.
17 12/26/2020
Experimental Results…
The function set consisted of three logical operations:
AND, OR, and NOT. Crossover exchanged the sub trees
of two parents, selected at random. Mutation changed the
terminals and functions at a rate of 0.03 for each node.
Elitist strategy was used in combination with
proportionate selection so that the best individual always
survives the selection.
These agents achieved a 97% performance for the
sample data.
18 12/26/2020
Conclusion
An information filtering system is a system that
removes redundant or unwanted information from an
information stream.
Strength
To gave valuable information at the limited time.
19 12/26/2020
Weakness
20 12/26/2020
Recommendation
The recommendation is if it is possible the information
21 12/26/2020
References
1. 0. Etzioni, N. Lesh, and R. Segal, Building softbots for UNIX, Technical Report, Dept of
CSE, University of Washington, 1992.
2. 0. Etzioni and D. Weld, A softbot-based interface to the Internet, Communications of the ACM,
37(7):72-76, 1994.
3. J. Horng, B. Liu, and C. Kao, A genetic algorithm for database query optimization, In Proc. of
IEEE Conf. on Evolutionary Computation, IEEE Press, 1994, pp. 350-355.
4. J. R. Koza, Genetic Programming: On the Programming of Computers by Means of Natural
Selection, Cambridge, MA: MIT Press, 1992.
5. D. Kraft, F. Petry, B. Buckles, and T. Sadasivan, The use of genetic programming to build
queries for .information retrieval, In Proc. of IEEE Conf. on Evolutionary Computation, IEEE
Press, 1994, pp. 468-473.
6. Y. Lashkari, M. Metral, and P. Maes, Collaborative interface
22 12/26/2020
THANK YOU!!!
23 12/26/2020