Measuring Similarity System and Method of Android Application Using Text Mining

摘要

The present invention relates to a technique which extracts a birthmark from a dalvik executable file of an android application and measures a similarity between applications by utilizing a TF-IDF. A system for measuring a similarity of an android execution file using text mining comprises: a feature extracting part extracting a birthmark of an original application file and a birthmark of a comparison application file, and extracting a contrast term from the original application file; and a similarity calculating part counting the number of the contrast terms included in the birthmark of the original application file and the birthmark of the comparison application file, projecting the number of included times of the birthmark into a vector space, and calculating the similarity of the original application file and the comparison application file.