Dissertation > Excellent graduate degree dissertation topics show
Research and Implement of Structure Learning Algorithm for Hybrid Bayesian Networks
Author: YanPengFei
Tutor: DongLiYan
School: Jilin University
Course: Computer Software and Theory
Keywords: Data Mining Bayesian networks Node sequence Pseudo - Bayesian network Minimum d- separation set
CLC: TP183
Type: Master's thesis
Year: 2010
Downloads: 156
Quote: 0
Read: Download Dissertation
Abstract
|
The goal of data mining is a useful and easy - to - understand information discovery rules, implicit in the massive data mode or correlation , integrated multidisciplinary field knowledge , roughly five categories of tasks to be completed , that prediction , classification, clustering association rules and time series . This article will focus on the structure learning of Bayesian network classifiers . Bayesian network structure learning process often requires experts in the field to provide some a priori information to reduce the network search space , making the classification accuracy of the resulting network model is heavily dependent on the knowledge of experts in the field . To address these deficiencies , we propose a hybrid structure learning algorithm HBN , the algorithm is a three-stage algorithm and does not require any a priori completely from the data set to learn a Bayesian network classification model . The HBN algorithm was first introduced in the pseudo - Bayesian network structure thinking , in the structure of the learning process , we have designed a new scoring function , is used to measure the network model and the data model differences in the expression of marginal independence and the first-order conditional independence level. The process of closed networks , we also consider the node probability categorical attributes and data attributes directly dependent , which makes HBN model combines the advantages of the NBC model and the BN model , while avoiding their disadvantages . In addition, we also find the approximate minimum set of d- separation algorithm is presented from the perspective of the implementation of efficiency improvement . A large number of experiments show that , the HBN model compared to the the NBC model and TAN model in classification accuracy has been greatly improved . Inadequacies HBN algorithm uses a greedy search strategies to find a sub-optimal network structure , the algorithm is more suited to construct a network of small and medium-sized structure .
|
Related Dissertations
- Yellow phosphorus storage tank area of Safety Evaluation,TQ126.317
- A Study on Healthcare Product Marketing Based on Data Mining Technology,F426.72
- Gao Zhong-ying academic thought and experience and use of Bufei Decoction treatment of common diseases of the respiratory system drug law,R249.2
- Bing- thick academic thought and clinical experience and empirical studies apply to turtle soups treatment of chronic kidney disease,R249.2
- The Design and Implementation of Bicluster Data Analyzing Software,TP311.52
- Research on Clustering Algorithm Based on Mutation Particle Swarm Optimization,TP18
- Research on Fuzzy C-Mean Clustering Algorithm Based on Particle Swarm Optimization and Shuffled Frog Leaping Algorithm,TP18
- Research on Clustering Algorithm Based on Genetic Algorithm and Rough Set Theory,TP18
- Based on data mining research tax audit case selection,F812.42
- Community-oriented education, personalized learning system and its implementation,TP391.6
- Association rule mining based Intrusion Detection System Research and Implementation,TP393.08
- Data warehouse technology in the banking customer management systems research and implementation,TP315
- Design to E-learning System in Senior Vocational School Base on Moodle,TP311.52
- Design and Development of Teaching Quality Assessment System Based on Data Mining,TP311.13
- The Application of Association Rules Algorithm in Higher Vocational Colleges’ Endorsement of Impoverished Students,G717
- Based on Data Mining Technologies in Urban Water Supply Analysis and Decision,F299.24;F224
- Research on Application of Data Mining Technology in Degree of Satisfaction Analysis of Television Customers,TP311.13
- Web Usage Mining and the Research of Personalized Recommendation,TP311.13
- Data Mining of Application in the School Management and Training Students,TP311.13
- Research on Employment Monitoring System of University Graduate,G647.38
- Design and Implementation for Decision Support System of Drug Administration Based on Data Warehouse,TP311.13
CLC: > Industrial Technology > Automation technology,computer technology > Automated basic theory > Artificial intelligence theory > Artificial Neural Networks and Computing
© 2012 www.DissertationTopic.Net Mobile
|