Published Paper Details:

Even after delivering the product by implementing all the features as per customer�s requirement, the improvement requests are frequently proposed by the users. For a product if more numbers of stakeholders are there, then the improvement requests are more in number and also frequency of receiving requests is high. Though the received requests are high in number, all are not suitable to implement. If a developer has to manually go through each and every request to differentiate between implementable and unimplementable requests , it is very time consuming. So to overcome this problem, sentiment based analysis can be used to predict the most likely requests which will be approved for implementation. This approach makes the software applications competetant in the market by helping in implementing the features with in deadlines. In this approach, the first step is to preprocess improvement requests using natural language preprocessing techniques. second is to calculate the sentiment of each improvement request by identifying the words having positive and negative sentiments in the summary attribute. Third is to train the machine learning based classifier to predict whether a given improvement request would be approved. The proposed approach is evaluated with the sample history data from real software applications.