Effective similarity measures in electronic testing at programming languages

Abstract

The purpose of this study is to explore the grammatical proper ties and features of generalized n-gram matching technique in electronic test at programming languages. N-gram matching technique has been success fully employed in information handling and decision support system dealing with texts but its side effect is size n which tends to be rather large. Two new methods of odd gram and sumsquare gram have been proposed for the improvement of generalized n-gram matching together with the modification of existing methods. While generalized n-grams matching is easy to generate and manage, they do require quadratic time and space complexity and are therefore ill-suited to the proposed and modified methods which work in quadratic in nature. Experiments have been conducted with the two new methods and modified ones using real life programming code assignments as pattern and text matches and the derived results were compared with the existing methods which are among the best in practice. The results obtained experimentally are very positive and suggested that the proposed methods can be successfully applied in electronic test at programming languages.