A Hybrid Algorithm for Pattern Matching: An Integration of Berry-Ravindran and Raita Algorithms

Abstract

AbstractDue to the quick actions of technology, a complicated and huge volume of data deriving from biological sciences are generated which makes string matching patterns a challenging task. This direction has the aim to make the utilization of an individual algorithm for string searching nearly ineffectual as the number of tries and the number of observations continues to increase. So, the solution is in the grouping of more than one algorithm to create a hybrid algorithm for quicker performance. The proposed hybrid algorithm uses the best features of Raita algorithm and Berry-Ravindran algorithm. The reason for choosing these two algorithms is because they achieve better performance in “number of attempts” and “number of character comparisons” tests. New Hybrid Algorithm (BRR) is able to produce better results by decreasing the effort and character comparisons. The data types used to evaluate performance are English text, DNA, and protein. In number of attempts evaluation, for DNA, English, and protein text datasets, the improvement of the hybrid algorithm was 18%, 50%, and 50% in comparison to Berry-Ravindran algorithm and it was 71%, 74%, and 70% in comparison to Raita algorithm. The results show that regardless of the size of the data used, the mix of algorithms yields better results and improved performance than the original algorithm.KeywordsHybrid algorithmPatternString searchingString matching