Informatics Publishing Limited

A. Adhiselvam ¹, E. Kirubakaran ², R. Sukumar ³

Affiliations
1 Department of MCA, S.T.E.T. Women’s College, Mannargudi, Tiruvarur - 614 016, Tamil Nadu, IN
2 STTP (System), Bharat Heavy Electrical Limited, Trichy - 620014, Tamil Nadu, IN
3 Department of CSE, Sethu Institute of Technology, Kariapatti, Virudhunagar - 626115, Tamil Nadu, IN

Abstract

Background/Objectives: Software bugs are generally the faults or errors that occur in the code that may leads to incorrect results. It is therefore necessary to find software bugs to increase the quality of software and for making the software to meet the user requirements. Methods/Statistical Analysis: The testing enables assessment of software which ensures the system whether it meets the system requirements. Graph mining is an approach to find software bugs and furthermore testing involves more computational complexity and cost therefore tools are developed for testing the software. The challenging task is to locate and fix bugs automatically. Bug localization using high performance map reducing process is very successful in the recent research. Findings: This proposed technique called Map Reduce Technique Based Apriori (MRTBA) Algorithm based on Graph mining in where edge weights are used to takes the program as the input and computes method calls of the program. Here each node represents method and each edge represents a method call. The algorithm aims for detection of faulty nodes. It uses Hash Map for reducing edge weights. JUnit test cases are performed at last for detecting bugs by giving different inputs. The approach mines all nodes not only at same level but also the nodes present at different levels. Conclusion/Improvements: In this work MRTBA, a dynamic control flow centered approach for bug localization has been presented. In future this technique can be extended with the proposal of steps involves in connecting same nodes at different levels. It also further enhances to find out indirect recursions for both the trivial and non-trivial map reduction in software bug localization.

Keywords

Apriori, Call Graph, Graph Mining, Map Reduce, Software Bug Localization, Subgraph

Full Text

   

M. Robinson Joel ¹, M. V. Srinath ², A. Adhiselvam ²

Affiliations
1 Department of Computer Science and Engineering, Shree Motilal Kanhaiyalal Fomra Institute of Technology, IN
2 Department of Computer Applications, Sengamala Thayaar Educational Trust Women’s College, IN

Abstract

Web Usage Mining is the application of data mining technique used to retrieve the web usage from web proxy log file. Web Usage Mining consists of three major stages: preprocessing, clustering and pattern analysis. This paper explains each of these stages in detail. In this proposed approach, the web directories are discovered based on the user's interestingness. The web proxy log file undergoes a preprocessing phase to improve the quality of data. Fuzzy Clustering Algorithm is used to cluster the user and session into disjoint clusters. In this paper, an effective approach is presented for Web personalization based on an Advanced Apriori algorithm. It is used to select the user interested web directories. The proposed method is compared with the existing web personalization methods like Objective Probabilistic Directory Miner (OPDM), Objective Community Directory Miner (OCDM) and Objective Clustering and Probabilistic Directory Miner (OCPDM). The result shows that the proposed approach provides better results than the aforementioned existing approaches. At last, an application is developed with the user interested directories and web usage details.

Keywords

Advanced Apriori Algorithm, Fuzzy Clustering Algorithm, Pattern Analysis, Web Directories and Web Personalization.

Full Text