Evaluating The Effectiveness Of Various Similarity Measures On Malay Textual Documents

Abstract. This paper analyzes the effect of various similarity measures namely inner product for un-weighted query terms, inner product for weighted query terms, cosine of the angle between query and document vectors and Euclidean distance. The study was motivated by the fact that many researchers especially in Malay document retrievals tend to use simple method of calculating similarity measure that is based on inner product for un-weighted query terms. This paper shows that Euclidean distance outperforms other similarity measures significantly. The results suggest that Euclidean distance should be used to improve performance of document retrieval systems