Classification method of multi-class on C4.5 algorithm for fish diseases

Sucipto, Sucipto and Kusrini, Kusrini and Emha, Luthfi Taufiq (2016) Classification method of multi-class on C4.5 algorithm for fish diseases. In: Proceeding - 2016 2nd International Conference on Science in Information Technology, ICSITech 2016: Information Science for Green Society and Environment, Balikpapan.

Text (Artikel) Classification Method of Multi-class on C4.5 Algorithm for Fish Diseases .pdf - Published Version Available under License Creative Commons Attribution Share Alike. Download (1MB)

Abstract

The background of the research is to analyze data derived from an elucidation of catfish and carp diseases in Kediri, East Java, Indonesia. The research shows that data about fish's disease history have not been used effectively because it is only be collected. Data about fish’s symptom history used by fish trainer only present the number of fish that get disease. Data about fish’s history should be also optimized to discover the relationship among fish’s disease. Thus, anticipation about disease that always attack fish could be prevented earlier. The research is done to understand the relationship history among fish’s disease. Then the accuracy of relationship quality is measured to acquire the quality of data properly so it can be worked to identify fish’s disease. Data relationship quality among fish’s disease symptoms should be understood to know how is the accuracy of datum classification obtained. A proper method is required to extract information from data obtained. There are many data-mining classification algorithms such as CART, CHAID, Rain Forest, and C4.5. But, the C4.5 algorithm is appropriate for this research used to form decision tree for data quality assessed from accurate performance of some multi- class fish diseases. This research uses 1120 data involving six diseases. The data were obtained from Agriculture Board (fishery subdivision) of Kediri Regency. The result shows that C4.5 algorithm is well to do for both a low and high accuracy class at 55.3 and 88.4 percent.

Actions (login required)

View Item