Application of Grayscale Co-occurrence Matrix (GLCM) Method for Classification of Quality Type of Guava Leaves as Traditional Medicine Using Neural Network Algorithm

Currently, there are still many people who use traditional medicine such as the use of guava leaves as anti-diarrhea medicine. But the types of guava leaves have different qualities for traditional medicine, some types of guava leaves have different leaf shape characteristics, and it will be difficult to distinguish the quality of the leaves. The purpose of this study was to classify the quality of guava leaf species with the classification of "good" quality and "bad" quality. The method used in this study starts with data collection, then the next process is Pre-Processing. After doing the Pre-Processing, the GLCM method will be applied with the Matlab application, the results of the application of the GLCM method will produce a data matrix which will later be used for the process of implementing the neural network algorithm on RapidMiner for the classification process and will produce an accuracy value. The results of this study produce several attributes, namely Angular Second Moment (ASM) as data attribute 1, contrast as data attribute 2, Inverse Different Moment (IDM) as data attribute 3, entropy as data attribute 4, and correlation as data attribute 5 in the sample type. Guava was tested, then the final result of the application of the neural network algorithm in this study resulted in an accuracy of 95%.