No abstract

No abstract

No abstract

Faults are extracted from the source code, from the pre-processed metrics and the related fault data. The generated faults are represented as a data set and are categorised into code, design and other features. Selection of features and identifying the importance of the attributes from the given set of test cases is one of the important tasks in the software testing phase. A Fuzzy based evaluation of features using the Entropy measure and Hurwicz criterion has been carried on the code and design metrics for different test cases. The datasets have been further analysed using the Random Forest Approach for identifying the feature that has the higher priority. The results obtained using the Fuzzy Entropy measure and the Random Forest approach exhibits a similarity of 95 % in identifying the feature importance. The results show that the feature NUM_OPERANDS is having the highest impact from the given list of features by applying Fuzzy Entropy measure and the model built using Random Forest approach.

No abstract