Performance Improvement of C4.5 Algorithm using Difference Values Nodes in Decision Tree
Handoyo Widi Nugroho, Teguh Bharata Adji, Noor Akhmad Setiawan · 2018 6th International Conference on Cyber and IT Service Management (CITSM) · 2018
The C4.5 algorithm is one of the decision tree algorithms that can be used to generate rules that are easily interpreted and fastest among other algorithms. The algorithm is also able to generate base model subsystem that can be used to support decision support system. So research on improvement of C4.5 algorithm performance is still very interesting to do. There are many features involved in C4.5 Algorithm i.e. data, attribute data, instance, and attribute classes. In some classification cases, this algorithm still produces less than the maximum accuracy. Therefore, this study aims to improve the performance of C4.5 Algorithm by applying the process of modifying equations and adding treatment to improve the accuracy of selection of nodes to be trimmed. Some methods of development of the C4.5 algorithm focus on the pruning phase which still allows for trimming of nodes with high or contributive value information. The fix is to modify the pruning function and will ensure that the pruning process is performed against branches that are completely non-contributive, thus improving the accuracy of the results.