期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2018
卷号：96
期号：1
页码：248
出版社：Journal of Theoretical and Applied
摘要：The massive recording of the continuous data in the real-time imposes huge burden in analyzing the medical data as it takes huge time for execution and may cause errors due to the data uncertainty and the inability to interpret the highly significant features. As a result, the data is classified according to the rules and the efficiency of the data classification relies on the data feature selection. To address this issue, the paper introduces a new algorithm for the optimal selection of the features that uses a novel parameter for optimally determining the features. The proposed Log Decision Tree (LDT) uses the novel parameter termed as the Log-entropy to retrieve the mutually significant feature from the data such that the future data classification depends on the LDT classification model for determining the disease/condition of the patients, which leads to the easy diagnosis. The Log-entropy function uses the entropy and the weights of the features for computing the importance of the features and based on the selected features, the LDT model is developed. The experimentation proves that the proposed LDT attained a greater value of accuracy, specificity, and sensitivity of 83.7386%, 77.6160%, and 86.4407% respectively.
关键词：Log-entropy; decision tree; medical data classification; Entropy; feature selection.
