Dissertation > Excellent graduate degree dissertation topics show

Based on Multi- relational Decision Tree Algorithm

Author: SongGuangLing
Tutor: HaoZhongXiao
School: Harbin University of Science and Technology
Course: Computer Software and Theory
Keywords: Multi- relational data mining Multi- relational Decision Tree Tuple identity propagation Background Properties
CLC: TP311.13
Type: Master's thesis
Year: 2009
Downloads: 178
Quote: 1
Read: Download Dissertation

Abstract


Multi-relational data mining in recent years, rapid development of important areas of data mining. Efficient and scalable data mining has been an important research topic. Consider multi-relational data mining, this issue is particularly important. Multi-relational data mining task complexity on the performance of the algorithm put forward higher requirements. With traditional data mining algorithms, multi-relational data mining algorithms search space becomes more complex, even more. Learning algorithm for multi-relational data, improve efficiency of the algorithm is the main bottleneck hypothesis space. To solve the above problems, this paper mainly done the following work: First, this paper data mining theory, the theory of relational data mining research, especially in multi-relational data mining classification algorithm - multi-relational decision tree algorithm and multi-relational data mining of the latest technology - tuple communication technologies conducted in-depth research. Secondly, we propose a multi-relational decision tree algorithm. Multi-relational decision tree to improve in two major areas: a multi-relational decision tree algorithm in order to improve the scalability, this article will tuple virtual connection and communication technologies applied to the improved multi-relational Decision Tree algorithm; 2 To reduce the time to explore alone reduce the system searches Useful properties of time and improve user satisfaction, this paper presents the user classification task under the guidance of the background attribute transfer technology, and the technology to improve the relationship between the decision tree multi. Finally, the improved algorithm for multi-relational Decision Tree theoretical proof and experimental validation. The main advantage of this experiment PKDD CUP'99 the Loan, Account, Transaction three relationships, two methods for general multi-relational decision tree algorithm and improved decision tree algorithm for relational comparative experiments. The first method, a fixed number of records in the same three relationships, increased the number of attributes for each relationship experiment, the second method, a fixed number of attributes in three relationships unchanged, the number of records to change between experiments . Through the above experimental results, the study suggests that multiple relationships while improving data items in the search tree does not reach the threshold to pass the background attribute, multi-relational decision tree algorithm to improve the operating efficiency is low; when improved multi-relational Decision Tree in Search data item passed threshold reached background properties, improved the efficiency of multi-relational decision tree algorithm is relatively high and increasing by the number of attributes (or number of records increases) less affected.

Related Dissertations

  1. Research on Multi-relational Association Rule Mining,TP311.13
  2. Research of Port Enterprise Statistic Work Based on MRDM and Intelligent Agent,F550
  3. Multi-relational Classification Based on Decision Tree Algorithm,TP311.13
  4. ILP-based multi-relational Classification Algorithm,TP311.13
  5. Research on Multi-relational Decision Tree Classification Algorithm,TP311.13
  6. The Research of Multi-Realation Assoaction Rules Algorithms in Data Mining,TP311.13
  7. Research on Some Problems of Statistical Relational Learning,TP181
  8. Research on Multi-relational Clustering Analysis Approaches,TP311.13
  9. Support Vector Clustering for Multi-Relational Data: Design and Implementation,TP311.13
  10. The Research of Multi-Relational Data Mining Technology and Its Realization in Tax Assessment,TP311.13
  11. Research on F&B Index Structure Supporting XML Query,TP311.13
  12. Query Processing and Optimization in Massive Multi-Database Integration,TP311.13
  13. Research on Compression, Operation and Query Processing Methods of Massive Datasets,TP311.13
  14. Implementation of Data Compression, Operation and Query Processing System Based on BAP,TP311.13
  15. The Design and Implementation of DICOM Middle Software and Access Control Model in Formation Integration Platform,TP311.13
  16. Research and Improvement on K-Means Clustering Algorithm,TP311.13
  17. Study of Data Reduction Technique Based on Manifold Learning,TP311.13
  18. Clustering Method Research Based on Divided and Conquered Method,TP311.13
  19. Incomplete information on the completeness of the system and its knowledge acquisition,TP311.13
  20. Evolutionary Clustering Algorithm and Its Application,TP311.13
  21. Study on the Key Technologies of Domain Database-Oriented Question and Answering System,TP311.13

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