R. Janani ¹, S. Vijayarani ¹

Affiliations
1 Department of Computer Science, School of Computer Science and Engineering, Bharathiar University, Coimbatore - 641046, Tamil Nadu, IN

Abstract

Objectives: To retrieve the information after analyzing the contents of the documents which are stored in the desktop by applying string matching algorithms. Methods/Statistical Analysis: To analyze the content of the documents, the various pattern matching algorithms are used to find all the occurrences of a limited set of patterns within an input text or input document. In order to perform this task, this research work used four existing string matching algorithms; they are Brute Force algorithm, Knuth-Morris-Pratt algorithm (KMP), Boyer Moore algorithm and Rabin Karp algorithm. This work also proposes three new string matching algorithms. They are Enhanced Boyer Moore algorithm, Enhanced Rabin Karp algorithm and Enhanced Knuth-Morris-Pratt algorithm. Findings: For experimentation, this work has used two types of documents, i.e. .txt and .docx. Performance measures used are search time, number of iterations and accuracy. From the experimental results, it is realized that the enhanced KMP algorithm gives better accuracy compared to other string matching algorithms. Application/Improvements: Normally, these algorithms are used in the field of text mining, document classification, content analysis and plagiarism detection. In future, these algorithms have to be enhanced to improve their performance and the various types of documents will be used for experimentation.

Keywords

Brute Force, Boyer Moore, Information Retrieval, Knuth-Morris-Pratt, Pattern Matching, Rabin Karp.

Full Text

   

C. Sivamathi ¹, S. Vijayarani ²

Affiliations
1 Department of Computer Science, Bharathiar University, Coimbatore - 641046, Tamil Nadu, IN
2 Department of Computer Science, Bharathiar University, Coimbatore - 641046, Tamil Nadu,, IN

Abstract

Objectives: Utility mining gains more attraction in recent years. Utility mining can be defined as mining item’s utility and revealing high utility items. In this work an algorithm EHUFIM (i.e Enhanced High Utility Frequent Itemset Mining), was proposed to reveal high utility – frequent itemsets even with negative profits. Methods/Statistical Analysis: The proposed algorithm uses utility mining methods to retrieve high profitable items. Then it uses support measure to reveal high occurrence items. The algorithm implements filter procedure of HUINIV algorithm to handle negative profit items. Findings: This algorithm discovers itemset that have more frequency and high utility with negative profit. Discovering such items helps in decision making in super markets, cross product marketing etc. The proposed algorithm was executed and performance of the algorithms was calculated. Application/Improvement: Existing utility frequents mining algorithms does not consider negative profit values. But, this proposed algorithm, takes negative utility values into account.

Keywords

High Utility Itemset, Negative Utility Itemset, Transaction Weighted Utility Property, Utility Frequent Itemset Mining, Utility MiningHigh Utility Itemset, Negative Utility Itemset, Transaction Weighted Utility Property, Utility Frequent Itemset Mining, Utility Mining

Full Text