Synthesis Algorithm of Homologous Detection Based on Analytic Hierarchy Process

Abstract

Nowadays, the research of software homologous detection is especially important in the flourishing software market while the existing detection techniques based on text, token, abstract syntax tree always get the real similarity inaccurately. In this paper, a synthesis algorithm based on Analytic Hierarchy Process(AHP) is proposed, which analyzes text, token, syntax tree three comparison algorithms synthetically and reflects software homologous similarity more accurately by artificial simulation plagiarism and experts calculation in the homologous field. This algorithm compares superiorities of homologous detection algorithms based on text, token and abstract syntax tree in performances factors such as miss rate, error rate, location accuracy, It also combines three comparison algorithms linearly according to their contributions to performances factors of synthesis similarity. The result of this synthesis algorithm meets the demand of homologous detection, reflects the reasonable real similarity which is calculated by experts.