BLACK TEA GRADE CLASSIFICATION USING PROBABILISTIC NEURAL NETWORK (PNN)

Abstract

North Sumatera tea is known as black tea, one which is produced by PT Perkebunan Nusantara IV Unit Bah Butong which produce 16 types of black tea. This research aims to classify black tea grades using Probabilistic Neural Network method and determine the accuracy value of black tea classification using Probabilistic Neural Network method. Data used are data of characteristics of 16 black tea types with the attributes id, types of tea, colour density, particel weight, particel size and particel shape. To get  the best accuracy result, training data and testing data are divided  using K-Means Clustering. Futhermore, classificy the testing data using Probabilistik the result obtained a grade 1 classification class totaling 7 records with a brownish  black appearance  and granutes particel, grade 2 totalling 7 record with a beownish appearance and grade 3 totaling 2 records  with reddish appearance and not determine its shape and obtain an accuracy  value of 80,00%.