Font Size: a A A

Research On Software Functional Searching Technology

Posted on:2017-06-22Degree:MasterType:Thesis
Country:ChinaCandidate:Y Z WangFull Text:PDF
GTID:2348330518470803Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
As there is more and more open source code on the Internet, it is much more difficult to understand how the project do work or which codes make up of functions with searching query.The source code searching engine nowadays can only return code fragments with query, but the structure of query which means a function or module. So this essay will research how to find the classes or module information with query.The method this thesis proposed is called function search technology, which combines topic analysis and software clustering. There are three aspects of work this paper done. First,this thesis proposes an approach of how to get the features of object-oriented language like Java and how to compute the characteristic vector. And as the input of hierarchical clustering, the result of clustering is also good. Second, it's different with traditional source code searching technique that this paper proposes a method of substituting file with classes based on topic model. The result of searching will not be fragments of source code while it will be functions of projects. At last, with the aim of more accurate of searching result, this thesis combines software clustering and topic model. And return more accurate results finally.The demo system this essay designed is success. The results of software clustering is much better than Bunch. And it also gets more accurate results of searching then query by words like GitHub because it can get rid of bad results automat. The approach which based on software clustering and searching technique, can make a good role in reading open source project, and legacy system. And it can also improve efficiency of developers.
Keywords/Search Tags:Software Clustering, Searching Technology, Topic Model, Function Location
PDF Full Text Request
Related items