CLUSTERING LIRIK LAGU ROHANI MENGGUNAKAN METODE K-MEANS

Abstract

In a spiritual worship, choosing the relevant song with theme is very important to reflection for the congregation in the service. In the fact, the people often having a difficult to choose some relevant songs because, they must see, search and read the song one by one. The problem is a motivation that it needs a searching system for the people. In the system, using a Vector Space Model after that the system does a clustering from the relevant songs using K-Means method. In the searching system, It is to search by unigram and bigram with the maximum keywords are six keywords. The system using NKB (Nyanyian Kidung Baru) as data which nuance of hymne. The Result of precision evaluation from the relevant documents  using MAP (Mean Average Precision), was getting the searching with unigram is the good result than bigram with the percentages is 70.41%. While, the clustering evaluation, using purity with the some trials for k values, was getting k = 4 and unigram given the good result with the percentages is 76.88%.