Dissertation > Excellent graduate degree dissertation topics show

A Study on an Optimal Feature Selection Algorithm Using Minimum Joint Mutual Information Loss Criterion

Author: ZhangYiShi
Tutor: ChenChuanBo
School: Huazhong University of Science and Technology
Course: Software Engineering
Keywords: Classification Feature selection Invariant feature subset Minimum joint mu-tual information loss Fast conditional mutual information estimation
CLC: TP311.13
Type: Master's thesis
Year: 2011
Downloads: 48
Quote: 1
Read: Download Dissertation

Abstract


With the rapid development of social economy and computer technology, studying onFilter feature selection algorithms nowadays is of great importance not only in theory but alsoin practice, for a good feature selection algorithm has broad application prospects.In this thesis, an optimal feature selection algorithm using minimal joint mutual infor-mation loss criterion is proposed based on the concept of invariant feature subset. By ap-plying joint mutual information as the criterion, this algorithm decomposes the feature se-lection problem into two sub-problems. It first discovers an invariant feature subset by amaximal conditional mutual information principle, and then eliminate potential redundantfeatures based on the minimal joint mutual information loss criterion. From the reliabilitypoint of view, this criterion can also abate the disturbance caused by sample insufficiencyin conditional mutual information estimation. Finally the algorithm will discover a featuresubset which best compresses the discriminative information contained in the original set offeatures.A fast implementation of conditional mutual information estimation is proposed andused to tackle the computationally intractable problem in conditional mutual information es-timation. The definition of local mutual information is introduced, and the relation betweenconditional mutual information and local mutual information is then discovered, which trans-form the estimation task to several local mutual information solving tasks.Ten benchmark datasets and challenge datasets are selected in order to verify the perfor-mance of the proposed algorithm and the efficiency of the fast conditional mutual informationestimation. Those selected datasets are differ greatly in the sample size (2000～72626) aswell as the feature dimension (22～139351), which can provide a comprehensive testingunder different conditions. Empirical results show that our algorithm outperforms severalrepresentative feature selection algorithms. In addition, experimental results for the executiontask show that the proposed implementation for conditional independence test achievesa better performance on both time efficiency and RAM utilization, especially on highdimensional dataset Thrombin containing 139351 features.

Related Dissertations

  1. K Company’s Improving Planning and Forecasting for the Reasonable Allocation of Inventory,F224
  2. Research on Target Tracking and System in Wireless Sensor Networks Based on Character of Movement and Terrain Restriction,TP212.9
  3. Study on Growth Monitoring Technique Based on Pixel Un-Mixing Method and HJ Remote Sensing Images in Paddy Rice,S511
  4. The Impact of Tourism on Typical Vegetation in Luya Mountain Nature Reserve, Shanxi Province,S759.9
  5. The Perineal Sonography in Diagnosis and Analysis of 30 Cases of Stress Urinary Incontinence,R445.1
  6. Early in Northeast China , Middle Triassic stratigraphic division and lithofacies palaeogeography,P531
  7. The Application and Research of the Theory of Metadiscourse,H136
  8. Design and implementation of mobile communication network system based on business analysis,TP311.52
  9. The Cumulative logistic Regression Classification of Students’ Poverty Data,O212.1
  10. The Simulation and Analysis of Spectrum Sensing Based on Higher Order Cumulants,TN911.23
  11. Empirical Research of RMB Exchange Rate and Sino-Germany Bilateral Trade,F752.7;F224
  12. Text classification based on Bayesian filtering technology research and implementation,TP393.098
  13. Clustering algorithm based bank card customer segmentation analysis system design and implementation,TP311.52
  14. Application Research of Model Construction in Biology Teaching in Senior High School,G633.91
  15. Study and Conservation of the Germplasm Resources of the XieBei Tea in Longyan City, Fujian Province,S571.1
  16. The Application Research of Food Production Enterprise Quality Credit Supervision System in Guangzhou City,F426.82
  17. Application Reseach in L Supermarket Chain Based on the ABC Classification,F721
  18. Research on Data Rectification Methods for Process Industry,TP274
  19. High-performed Kernel Classification Methods Based on Multi-kernel Learning,TP391.41
  20. Research on Feature Extraction, Selection and Classification Algorithms for Pulmonary CAD,TP391.41
  21. Design of Environmental Protection Game Based Flash,TP317.4

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