Enhanced Pattern Matching Algorithms for Searching Arabic Text Based on Multithreading Technology

Abstract

Many exact pattern matching algorithms have been developed and optimized for better efficiency and accuracy. In this regard, English is a well-served language compared to Arabic, which suffers some difficulties when searching its text. A few to mention include: Arabic writing direction is right-to-left, letters often change their shapes depending on their position within a word, Arabic language uses different encoding techniques for computing and digital storage and Arabic is a vocalized language, where diacritics might show above and beneath its alphabets. Hence, efficient search algorithms for Arabic text must be sought. This paper has two fold aims. The first one is to enhance three well-known exact pattern matching algorithms, namely: Brute-force (BF), Boyer-Moore (BM) and Knuth-Morris-Pratt (KMP) to search Arabic text. While the second aim is to speed up these search algorithms using multithreading technology. Experimental results showed that multithreading technology has positive influence on the performance of the original search algorithms when the size of the text is relatively large.