An Improved Approach based on Balanced Keyword Weight to Traceability Recovery

Software traceability recovery has become more and more important to the software life cycle recent days with the increasing complexity of software system. However, it is a tedious and error-prone task by manual. Many researchers have devoted themselves in building and predicting traceability links. In this paper, we approach a novel method to calculate text similarity and improve our results by a learning to rank method. And a couple of experiments are conducted to verify our method.