Web-based Rice Pest and Diseases Detection Using Hybrid Fuzzy and K-Nearest Neighbor Methods

Rice is a food crop that produces rice which is very useful for human life. And this is one of the main food commodities in Indonesia. Therefore, the Government must pay attention to the agricultural sector as a forum for providing food needs for the Indonesian population, which is increasing every year. East Java Province is one of the regions that produces the highest rice harvest in Indonesia. However, in recent years, rice production in East Java has not been able to meet the population’s food demand properly. This is due to disease and pest attacks on rice plants and a lack of knowledge so that rice farmers experience crop failure. Therefore, rice production at the farm level can be increased through web-based technological innovations in early detection of diseases and pests. This research used the hybrid Fuzzy and K-Nearest Neighbor (K-NN) method to detect diseases and problems in rice plants. The fuzzy logic is used to find the membership value of the disease detection class and the K-NN method is used to find the closest distance between the evaluated data and its k closest neighbors in the training data. The results of trials conducted on 200 case data with 13 diseases and pests and 32 symptoms resulted in an accuracy of 98.74% with 160 training data and 40 test data, and the value of k=4.