Detail Inovasi Perguruan Tinggi

Tema: Intelligence, Computing and Multimedia (ICM)
Judul: Classification of Hadith into Positive Suggestion, Negative Suggestion, and Information
Perguruan Tinggi: Universitas Telkom
Jenis/sdm: dosen/0420038901#1103130144#1103130141

Tahun: 2017

This paper presents experimental results on new classification labels of hadith, namely positive suggestion, negative suggestion, and information. The preprocessing involves stemming and stopword removal, and the feature extraction methods include bag-of-word and tf-idf of unigram, bigram, or trigram. We created a baseline classifier using simple rule-based string matching technique which generated F1-Score of 0.69. We experimented with Support Vector Machine and Neural Network as the classifiers algorithm, and the best performance was resulted from SVM model with polynomial kernel with F1-Score of 0.89.