IMPLEMENTASI METODE NAÏVE BAYES CLASSIFIER UNTUK IDENTIFIKASI JENIS JAMUR

Abstract

Being in a country that has abundant natural resources and biodiversity, of course, something that should be grateful for, such as its diversity of flora. The climate is very suitable for growing various types of plants. One of them is the food sector, especially horticulture, a horticultural commodity that is often found and widely used consumption materials by the community are mushrooms. The types of fungi of the genus Agaricus and Lepiota are often found around the community. There are types of mushrooms that are edible and some are poisonous. The type of fungi genus agaricus and lepiota which have poisons can cause pain in people who consume them, can even have an impact on death. To distinguish between mushrooms that are safe for consumption and those that are poisonous is very difficult. Therefore the need for problem solving to contribute to identifying the fungi is safe to eat or toxic. In this study the researcher tried to contribute in solving the problem in order to identify fungi using the naive bayes classifier method. Naive bayes classifier will classify the fungi dataset of genus agaricus and lepiota using probability calculations. In the scheme the testing data will be inputted with certain attributes, with subsequent training data with the NBC method testing data will be classified with training data so that the results of the testing data are included in the edible or poisonous category. The data used in this study is a collection of mushroom data (genus agaricus and genus lepiota), this dataset is taken from the UCI Machine learning repository. The mushroom data has 23 attributes based on morphological characteristics, with a total of 8124 mushroom records. The mushroom data has 23 attributes based on morphological characteristics, with a total of 8124 mushroom records. After using the 10-fold cross validation scheme, the average method of naive bayes method in the history of agaricus and lepiota mushrooms was found to be an accuracy value of 86.64 using the method of good classification.