Zeroth Review
This document was uploaded by user and they confirmed that they have the permission to share it. If you are author or own the copyright of this book, please report to us by using this DMCA report form. Report DMCA
Overview
Download & View Zeroth Review as PDF for free.
More details
- Words: 1,035
- Pages: 11
ANALYZING THE PERFORMANCE OF KNOWLEDGE BASED SYSTEMS USING BAYESIAN NETWORKS PROJECT GUIDE : Dr. K. Raja TEAM MEMBERS: S.L. Kokila Bai 20053832 K. Manasa 20053834 N. Manoj 20053835
KNOWLEDGE BASED SYSTEMS
A Computer system that is programmed to imitate human problem-solving by means of artificial intelligence and reference to a database of knowledge on a particular subject. A program for extending and/or querying a knowledge base which is a special kind of database for knowledge management. It provides the means for the computerized collection, organization, and retrieval of knowledge. Knowledge bases are categorized into tw o ma jor ty pe s : Machine-readable knowledge bases Human-readable knowledge bases
BAYESIAN NETWORK
A Bayesian n etwork is a directed acyclic graph (DAG ) consisting of a set of vertices representing var iables of i nterest and a set of direct ed edge s representing depen dency rel ati onships among variables. Each random variable denoting an attribute, feature or hypothesis can hold a finite set of mutually exclusive states and has a con ditional pr ob abili ty ta ble attached to quantify the degrees of belief in the strength of the dependencies. The Bayesian network is a probabilistic graphical model that is based on Bayes R ul e which is given as: For any two events, A and B, p( B|A ) = p( A|B ) x p( B) / p ( A ) , where you read 'p(A)' as "the probability of A", and 'p(A|B)' as "the probability of A given that B has occurred".
PROBLEM DEFINITION The problem solving techniques generally used in knowledge based systems such as rule based reasoning, case based reasoning, etc have certain disa dv an ta ges such as : Rules are “brittle” - cannot handle missing or unexpected information and may have complex interactions. Users might rely on previous experience without validating it in the new situation. Users might allow cases to bias new problem solutions. Large case base adds challenge to storage and search processes When no case matches the requirements, a conclusion is arrived at using "similar" cases and is then added to the case base, this might lead to infinite case additions. Adaptation i.e., modifying solutions of former similar cases to fit for a current problem is one of the major bottlenecks.
On the other hand Bayesian ne twor ks are known to have the following pro pe rtie s : Bayesian networks can readily handle incomplete data sets. Probabilities need not be exact to be useful. Bayes nets are generally quite robust to imperfect knowledge. They allow one to learn about causal relationships. They readily facilitate use of prior knowledge. They are very adaptable. You can start them off small, with limited knowledge about a domain, and grow them as you acquire new knowledge. You can use as much knowledge as is available and the net will do as good a job as is possible with the available knowledge. Given whatever knowledge we have about an instance, based on the best mathematical and statistical knowledge to date, the net will tell us what we can legitimately conclude rather than merely relying on likelihood. Bayesian methods provide an efficient method for preventing the over fitting of data (there is no need for pre-processing).
Hence, based upon the afore mentioned properties of the Bayes Networks which overcome disadvantages of other modeling techniques, we propose to incorporate the Bayesian network to support human decision making via a Knowledge Based System (KBS), and thereby analyze the performance of the Bayesian networks in KBS.
BASE PAPER
Ye Chen and Divakaran Liginlal, “Bayesian Networks for Knowledge-Based Authentication”, IEEE Transactions on Knowledge and Data Engineering, VOL. 19, NO. 5,pp.695-710, May 2007 Knowledge Based Authentication (KBA) is a form of user authentication, that involves verifying a claimed identity by matching one or more pieces of information (factoids) provided by an individual (claimant) against the information sources associated with the claimant. In this paper, the authors present a Bayesian network model of KBA grounded in probabilistic reasoning and information theory and propose a methodology to implement a Bayesian network based KBA system.
Further, they have carried out an empirical evaluation of the relative merits of two Bayesian network structures for KBA, the Naïve Bayes and the Tree Augmented Naïve Bayes and confirmed the hypothesis that the TAN structure is superior in terms of both authentication accuracy and error rates.
OTHE RS :
Rainer Schmidt, Lothar Gierl, “Case-based Reasoning for Medical Knowledge-based Systems”, Institute for Medical Informatics and Biometry, University of Rostock Rembrandtstr, 16 / 17, D-18055 Rostock, Germany which discusses the appropriateness of CBR for medical knowledge-based systems, points out problems, limitations and possibilities how they can partly be overcome.
“The Development of a case-based reasoning system as a tool for residential valuation in Bangkok”, Pacific-Rim Real Estate Society (PRRES) Conference 2000 Sydney, 23-27 January, 2000 which is a study that describes the development of a case-based reasoning system for valuers. Especially, the study examines the usefulness of the system for the valuation of townhouses in Bangkok, Thailand. Rainer Schmidt, Olga Vorobieva, Lothar Gierl, “Case-based Adaptation Problems in Medicine”, Institut für Medizinische Informatik und Biometrie, Universität Rostock Rembrandtstr. 16 / 17, D-18055 Rostock, Germany which summarises experiences with adaptation in medicine and elaborate typical medical adaptation problems and indicate possibilities how to solve them.
REFERENCES
Ye Chen and Divakaran Liginlal, “Bayesian Networks for Knowledge-Based Authentication”, IEEE Transactions on Knowledge and Data Engineering, VOL. 19, NO. 5,pp.695-710, May 2007 http://www.pr-owl.org/basics/bn.php#mdrcpt An Introduction to Bayesian Networks and their Contemporary Applications, Daryle Niedermayer, I.S.P., PMP, B.Sc., B.A., M.Div. http://journals.cambridge.org/action/displayIssue?jid=KER&v http://www.dcs.qmw.ac.uk/%7Enorman/BBNs/BBNs.htm
Rainer Schmidt, Lothar Gierl, “Case-based Reasoning for Medical Knowledge-based Systems”, Institute for Medical Informatics and Biometry, University of Rostock Rembrandtstr, 16 / 17, D-18055 Rostock, Germany which discusses the appropriateness of CBR for medical knowledge-based systems, points out problems, limitations and possibilities how they can partly be overcome. Rainer Schmidt, Olga Vorobieva, Lothar Gierl, “Case-based Adaptation Problems in Medicine”, Institut für Medizinische Informatik und Biometrie, Universität Rostock Rembrandtstr. 16 / 17, D-18055 Rostock, Germany
KNOWLEDGE BASED SYSTEMS
A Computer system that is programmed to imitate human problem-solving by means of artificial intelligence and reference to a database of knowledge on a particular subject. A program for extending and/or querying a knowledge base which is a special kind of database for knowledge management. It provides the means for the computerized collection, organization, and retrieval of knowledge. Knowledge bases are categorized into tw o ma jor ty pe s : Machine-readable knowledge bases Human-readable knowledge bases
BAYESIAN NETWORK
A Bayesian n etwork is a directed acyclic graph (DAG ) consisting of a set of vertices representing var iables of i nterest and a set of direct ed edge s representing depen dency rel ati onships among variables. Each random variable denoting an attribute, feature or hypothesis can hold a finite set of mutually exclusive states and has a con ditional pr ob abili ty ta ble attached to quantify the degrees of belief in the strength of the dependencies. The Bayesian network is a probabilistic graphical model that is based on Bayes R ul e which is given as: For any two events, A and B, p( B|A ) = p( A|B ) x p( B) / p ( A ) , where you read 'p(A)' as "the probability of A", and 'p(A|B)' as "the probability of A given that B has occurred".
PROBLEM DEFINITION The problem solving techniques generally used in knowledge based systems such as rule based reasoning, case based reasoning, etc have certain disa dv an ta ges such as : Rules are “brittle” - cannot handle missing or unexpected information and may have complex interactions. Users might rely on previous experience without validating it in the new situation. Users might allow cases to bias new problem solutions. Large case base adds challenge to storage and search processes When no case matches the requirements, a conclusion is arrived at using "similar" cases and is then added to the case base, this might lead to infinite case additions. Adaptation i.e., modifying solutions of former similar cases to fit for a current problem is one of the major bottlenecks.
On the other hand Bayesian ne twor ks are known to have the following pro pe rtie s : Bayesian networks can readily handle incomplete data sets. Probabilities need not be exact to be useful. Bayes nets are generally quite robust to imperfect knowledge. They allow one to learn about causal relationships. They readily facilitate use of prior knowledge. They are very adaptable. You can start them off small, with limited knowledge about a domain, and grow them as you acquire new knowledge. You can use as much knowledge as is available and the net will do as good a job as is possible with the available knowledge. Given whatever knowledge we have about an instance, based on the best mathematical and statistical knowledge to date, the net will tell us what we can legitimately conclude rather than merely relying on likelihood. Bayesian methods provide an efficient method for preventing the over fitting of data (there is no need for pre-processing).
Hence, based upon the afore mentioned properties of the Bayes Networks which overcome disadvantages of other modeling techniques, we propose to incorporate the Bayesian network to support human decision making via a Knowledge Based System (KBS), and thereby analyze the performance of the Bayesian networks in KBS.
BASE PAPER
Ye Chen and Divakaran Liginlal, “Bayesian Networks for Knowledge-Based Authentication”, IEEE Transactions on Knowledge and Data Engineering, VOL. 19, NO. 5,pp.695-710, May 2007 Knowledge Based Authentication (KBA) is a form of user authentication, that involves verifying a claimed identity by matching one or more pieces of information (factoids) provided by an individual (claimant) against the information sources associated with the claimant. In this paper, the authors present a Bayesian network model of KBA grounded in probabilistic reasoning and information theory and propose a methodology to implement a Bayesian network based KBA system.
Further, they have carried out an empirical evaluation of the relative merits of two Bayesian network structures for KBA, the Naïve Bayes and the Tree Augmented Naïve Bayes and confirmed the hypothesis that the TAN structure is superior in terms of both authentication accuracy and error rates.
OTHE RS :
Rainer Schmidt, Lothar Gierl, “Case-based Reasoning for Medical Knowledge-based Systems”, Institute for Medical Informatics and Biometry, University of Rostock Rembrandtstr, 16 / 17, D-18055 Rostock, Germany which discusses the appropriateness of CBR for medical knowledge-based systems, points out problems, limitations and possibilities how they can partly be overcome.
“The Development of a case-based reasoning system as a tool for residential valuation in Bangkok”, Pacific-Rim Real Estate Society (PRRES) Conference 2000 Sydney, 23-27 January, 2000 which is a study that describes the development of a case-based reasoning system for valuers. Especially, the study examines the usefulness of the system for the valuation of townhouses in Bangkok, Thailand. Rainer Schmidt, Olga Vorobieva, Lothar Gierl, “Case-based Adaptation Problems in Medicine”, Institut für Medizinische Informatik und Biometrie, Universität Rostock Rembrandtstr. 16 / 17, D-18055 Rostock, Germany which summarises experiences with adaptation in medicine and elaborate typical medical adaptation problems and indicate possibilities how to solve them.
REFERENCES
Ye Chen and Divakaran Liginlal, “Bayesian Networks for Knowledge-Based Authentication”, IEEE Transactions on Knowledge and Data Engineering, VOL. 19, NO. 5,pp.695-710, May 2007 http://www.pr-owl.org/basics/bn.php#mdrcpt An Introduction to Bayesian Networks and their Contemporary Applications, Daryle Niedermayer, I.S.P., PMP, B.Sc., B.A., M.Div. http://journals.cambridge.org/action/displayIssue?jid=KER&v http://www.dcs.qmw.ac.uk/%7Enorman/BBNs/BBNs.htm
Rainer Schmidt, Lothar Gierl, “Case-based Reasoning for Medical Knowledge-based Systems”, Institute for Medical Informatics and Biometry, University of Rostock Rembrandtstr, 16 / 17, D-18055 Rostock, Germany which discusses the appropriateness of CBR for medical knowledge-based systems, points out problems, limitations and possibilities how they can partly be overcome. Rainer Schmidt, Olga Vorobieva, Lothar Gierl, “Case-based Adaptation Problems in Medicine”, Institut für Medizinische Informatik und Biometrie, Universität Rostock Rembrandtstr. 16 / 17, D-18055 Rostock, Germany
Related Documents
Zeroth Review
October 2019 8
Zeroth Review
October 2019 4
Zeroth Review
November 2019 4
Zeroth Review Student Instructions
August 2019 18
