Analisis Komparatif Inferensi Fuzzy Tsukamoto, Mamdani dan Sugeno Terhadap Produktivitas Padi di Indonesia
Abstract
The productivity of rice plants is a measure of the yield produced per hectare of agricultural land. High productivity will increase rice production and the ability of farmers to earn higher incomes. Fuzzy logic models uncertain situations by approaching human thought. Rice productivity is influenced by land area. The wider the land the harvested area increases. Fuzzy Tsukamoto maps input values into fuzzy sets and then operates with fuzzy rules. Fuzzy Mamdani consists of 3 stages: fuzzification, inference, and defuzzification. Sugeno's method is one of the methods in AI systems to create fuzzy inference systems. There are three variables for rice productivity data in Indonesia for 2020-2022, Harvest Area(ha), Productivity(ku/ha), and Production(tons) with a total of 170 data. The prediction results of the 3 methods from the 9 reviewed journals have different results. The Tsukamoto method has nearly the same accuracy as the Sugeno fuzzy method and then the Mamdani fuzzy method. Of the 9 journals, 4 journals produced the Tsukamoto fuzzy method with the best accuracy and 4 journals produced the Sugeno method with the best accuracy and 1 journal produced the Mamdani fuzzy method with the best accuracy. Results with Tsukamoto fuzzy, Mamdani fuzzy, and Sugeno fuzzy: 73.67;292.5; and 274.55 tonnes. The results between fuzzy mamdani and fuzzy Sugeno are not that far apart. This is also the same as the results in the 4th variation where the results are 64.45;320; and 300.11 tons. The results of this study were the results of the Mamdani fuzzy and Sugeno fuzzy methods, the results obtained for each variation: the first variation produced 274.55 tons, the second 86.23, the third 135, the fourth 300.11 tons. The results of the comparison of the 3 methods produce Tsukamoto fuzzy and Sugeno fuzzy have the best accuracy and the results obtained in the production research using the Mamdani fuzzy and Sugeno fuzzy methods have a slight difference compared to the Tsukamoto fuzzy.
