Use this URL to cite or link to this record in EThOS:
Title: The use of stratified sampling to improve the performance of classification trees
Author: Gill, Abdul Aziz.
Awarding Body: University of East Anglia
Current Institution: University of East Anglia
Date of Award: 2004
Availability of Full Text:
Access from EThOS:
In this thesis we study two primary areas for improvement in classification trees, namely accuracy and consistency of prediction. Classification trees are generally easier for humans to comprehend and have been used in a wide range of disciplines including medical sciences and statistics. This thesis reviews existing improvement techniques in detail and addresses those areas where the existing techniques show poor performance. Additionally, their strengths and weaknesses/ limitations in the light of statistical background have been investigated. Some important contributing factors towards improving accuracy have been explored. Four innovative sampling techniques are introduced. They are thoroughly tested on various real world data sets from the UCI repository. Results of experiments show that these techniques work well and produce significant improvements for classification in terms of predicting accuracy and consistency of performance.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available