Sistem Identifikasi Famili Secara Otomatis Berbasis Teks Menggunakan Dokumen Etnofitomedika

Abstract

This research represents a text-based system that can be used for automatic identification of plant families according to the taxonomy of plants. The identification was done by utilizing information from etnofitomedika documents on plant characteristics that can represent the family of each plant including morphology, habitats, habitus, and biochemical compounds. The method used in this research is Chi-Square method to select important words in each document and Naïve Bayes method to classify the words. The experimental results showed that the critical value using significance level of 0,001 had better accuracy than the critical value using significance level of 0,01. The accuracy of classification systems based on family category using K-fold cross validation with significance value of 0,001 was 85,5%. It was found this system is useful in helping users especially researchers and taxonomists for plant family identification. Furthermore, it can help strengthening the knowledge on biodiversity and the use of medical plants.