Naive Bayes Classification Framework Model for Optimizing Prediction of Agrotourism Products Orange Gerga
Yogi Isro Mukti, Vike Itteridi, Iskandar Sulaini, Yuni Widiastiwi · 2022 International Conference on Informatics, Multimedia, Cyber and Information System (ICIMCIS) · 2022
One of the mainstay products of Agrotourism is Orange Gerga, one type of tangerine that is widely enj oyed by tourists and developed in various products. The production activity of Gerga Orange produces reports in the form of data which is a source of information, so it is necessary to carry out in-depth analysis by the director as a basis for making decisions. Of course, this task is not easy, mainly if it is carried out by a director with a lot of work. Therefore analytical activities can be carried out by business intelligence which is a remarkable development that can produce knowledge classifications that directors can use in decision making. Naive Bayes is one of the classification algorithms that have accuracy in making predictions and is good in reputation in classification, especially in speed in learning compared to other machine learning classification algorithms. The results showed that experiments using the nave Bayes classification framework model for optimizing the production of Gerga harvests using x-fold validation with linear sampling showed that the accuracy was 100%, and the AUC was 80%. Compared with shuffled sampling, the accuracy was 97.32% and an AUC of 90%, while stratified sampling produces an accuracy of 97.14% and an AUC of 100%. The experimental results show that testing the climate change data set using the Naive Bayes algorithm shows more accurate results in the excellent category.