Dissertation > Excellent graduate degree dissertation topics show
Improve Decision Tree Algorithm Based on Association Rules and Its Application
Author: XuLiMei
Tutor: LinJianLiang
School: South China University of Technology
Course: Probability Theory and Mathematical Statistics
Keywords: Association rules C4. 5 algorithm Information gain ratio Best demarcation points
CLC: TP311.13
Type: Master's thesis
Year: 2011
Downloads: 77
Quote: 0
Read: Download Dissertation
Abstract
|
Making it the most active research directions in data mining classification mining widely used in business and other fields . Which decision tree for its excellent efficiency of data analysis , intuitive features , the acclaimed . Existing decision tree is mainly focused on the use of various heuristic information to measure the properties of the extent , or take advantage of a variety of strategies to prune the decision tree . But in many cases , the data contained in the original property or the presence of redundancy, or the amount of information covered , which will undoubtedly affect the decision tree construction . Raw data covers less than the amount of information as a starting point , using classic Apriori association rules algorithm to generate the new property first , and then draw on the idea of information gain , and the support and confidence of association rule mining parameters proposed new property evaluation criteria , identify high confidence approximate precise rules . The new property is added to the original data attributes , C4.5 decision tree algorithm to classify forecasts , thereby improving the accuracy of prediction algorithm . Apriori algorithm exist to verify the candidate frequent the k_ items set need to scan the entire database is very time - consuming defects ; C4.5 decision tree algorithm in the continuous -valued attribute discretization process , the need for all the division test also take up more time . In order to avoid the expense of the efficiency of the algorithm at the same time improve the algorithm accuracy . Fayyad and Irani proof : regardless of the data set for learning the number of categories , the category of how the distribution of continuous value attribute always the best division point at boundary points ; continuous C4.5 decision tree algorithm the value attribute discretization process improvements . Continuous value attribute more UCI database contains 15 test experiments of the data in the database from the complexity of the algorithm , the efficiency of the algorithm , the algorithm accuracy of C4.5 algorithm and improved C4.5 algorithm confirmed the improved C4.5 algorithm reduces the complexity of the algorithm to improve the accuracy of the algorithm , and without sacrificing the efficiency of the algorithm . Finally, the improved C4.5 algorithm is applied in the aviation industry churn prediction , by comparison with C4.5 results , further improved C4.5 algorithm has high practical value .
|
Related Dissertations
- Based on data mining research tax audit case selection,F812.42
- The Application of Association Rules Algorithm in Higher Vocational Colleges’ Endorsement of Impoverished Students,G717
- Data Mining of Application in the School Management and Training Students,TP311.13
- Researeh on Data Mining Technology for Teahing Assessment in Technique College,TP311.13
- Research Offingerprint of Communication Behavior,TP311.13
- Research of Dynamic Association Rules,TP311.13
- An Intrusion Detection System for High-Speed Networks,TP393.08
- Supercomputer based on log analysis method of forecasting errors,TP338
- Data mining in the student evaluation system application,TP311.13
- Privacy -oriented association rule mining research,TP311.13
- Users trading behavior analysis and demonstrate - in modern barter industry in the application,TP311.13
- Figure based on association rules and emotional tendencies sort of sentence,TP391.1
- Intrusion detection based on data mining technology research,TP393.08
- Association rules algorithm and its pharmacy system in the intelligent application of,TP311.13
- Based on Data Warehouse Management System NCMS,TP311.13
- Research on Risk Monitoring of Subway Foundation Pit Construction Based on Association Rules,U231.3
- The Research on Medication Laws of Arthralgia Syndrome in Medicalcase Prescription Based on Clustering Analysis and Association Rules,R255.6
- Aplastic anemia of TCM Syndrome typing literature data mining research,R259
- The Website Optimization Research Base on Web Usage Mining,TP393.092
- Design and implementation of intrusion detection system based on association rules,TP393.08
- Research on the Stream Data Mining in Network Traffic Analysis,TP393.06
CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer software > Program design,software engineering > Programming > Database theory and systems
© 2012 www.DissertationTopic.Net Mobile
|