Optimizing Decision Tree Through Attributes Generation Using Genetic Programming for Clinical Data

摘要: Objective: To intend towards increasing classification efficiency of J48 classifier by introducing attribute set on the basis applied genetic programming. The constructed not only enhances data capabilities but also increased space for algorithm giving more accurate results. Methods/Analysis: datasets related to heart and liver disease were selected from UCI machine learning repositories. experiment has been conducted with help WEKA tool, which is an open source tool mining. Finding: After experimentation it found that better accuracy reduced error rate when after inclusion newly generated attributes adding induced programming, significant boost can be seen in 74% 83% 68% 72% respectively. Improvement: We obtained results compared existing literature chosen clinical datasets.