Sistem Pakar Identifikasi Varietas Tanaman Kunyit

Abstract

Along with the increasing development of genetic plant resources, the number of accessions of the turmeric plant is also increasing, making it difficult to differentiate between varieties, due to similar appeareances. This reserach aims to develop on expert system to identify the varieties of turmeric (Turina-1, Turina2, and Turina-3). Knowledge acquisition for building the expert system is performed in conjunction with an expert from the Indonesia Spices and Medicinal Research Institute, Drs. Cheppy Syukur. The knowledge acquisition process resulted in 13 input variables to describe a variety. The input inference process divides these 13 variables into two types: fuzzy and non-fuzzy. The non-fuzzy variables are: color of flowers, the base of leaf, and the meat colors the rihizome which are used to decide wheter the input description represents a Turina-1, Turina-2, Turina-3 variety or a nonvariety. The remaining 10 variables fuzzy variables are: number of flowers per stem, plant height, number of tillers, leaf length, leaf width, rhizome weight per hill, number of parent rhizome, the number of primary roots, secondary roots and the levels of curcuma, which are used to determined whether the input description represents a Turina-1, Turina-2 or Turina-3 variety, using the Mamdani method for Fuzzy Inference System (FIS). In the testing phase, we use 100 data, which are already labeled according to the decision by the expert. The result shows that the system is able to correctly identify 89 of the input data.