A Fast Hybrid Algorithm for the Exact String Matching Problem

Al-Mazroi Al-Mazroi

doi:10.3844/ajeassp.2011.102.107

Abstract

Problem statement: Due to huge amount and complicated nature of data being generated recently, the usage of one algorithm for string searching was not sufficient to ensure faster search and matching of patterns. So there is the urgent need to integrate two or more algorithms to form a hybrid algorithm (called BRSS) to ensure speedy results. Approach: This study proposes the combination of two algorithms namely Berry-Ravindran and Skip Search Algorithms to form a hybrid algorithm in order to boost search performance. Results: The proposed hybrid algorithm contributes to better results by reducing the number of attempts, number of character comparisons and searching time. The performance of the hybrid was tested using different types of data-DNA, Protein and English text. The percentage of the improvements of the hybrid algorithm compared to Berry-Ravindran in DNA, Protein and English text are 50%, 43% and 44% respectively. The percentage of the improvements over Skip Search algorithm in DNA, Protein and English text are 20%, 30% and 18% respectively. The criteria applied for evaluation are number of attempts, number of character comparisons and searching time. Conclusion: The study shows how the integration of two algorithms gives better results than the original algorithms even the same data size and pattern lengths are applied as test evaluation on each of the algorithms.

Highlights

String matching algorithm is an essential segment in computer science presently because of its usefulness in searching and matching pattern and text from vast databases containing huge of complicated data (Hassan, 2005)
Analysis: The hybrid algorithm contains the preprocessing phase of Skip Search and Berry-Ravindran algorithms
The results for the DNA, Protein and English text of the hybrid algorithm displayed enhanced performance than the original algorithms; it is largely right algorithms are of prime of importance when deciding which algorithm(s) to select to form hybrid as some of them cannot perform creditably even when combine together

Summary

Introduction

String matching algorithm is an essential segment in computer science presently because of its usefulness in searching and matching pattern and text from vast databases containing huge of complicated data (Hassan, 2005). Due to increasing rate and complex nature of biological sciences and scientific data nowadays the usage of one algorithm alone for string searching is not efficient the urgent need to combine two or more algorithms to form a hybrid in order to ensure efficient performance (Chen, 2007). This study proposes a new hybrid algorithm (called BRSS), by merging the best properties of two algorithms BerryRanvindran and Skip Search algorithms in order to ensure a better performance during string searching These algorithms are chosen because Skip Search algorithm is more efficient for small alphabets and long patterns, while Berry-Ravindran algorithm is more effective for providing a better shift value from the two successive characters immediately to the rightmost of the window (Tathoo et al, 2006).

Methods

Results

Discussion

Conclusion