Loading...
Please wait, while we are loading the content...
Similar Documents
A survey on reducing features to improve bug prediction by using cos-triage algorithm.
| Content Provider | CiteSeerX |
|---|---|
| Author | Jadhav, Veena Gaikwad, Ana |
| Abstract | Abstract- Bugs are nothing but Software defects, present a serious challenge for system consistency and dependability. It is very difficult task to predict bugs. To detect the bugs from the software bug prediction is useful way. Machine learning classifiers have emerged newly as a way to envisage the existence of a bug in a change made to a source code. The machine learning classifier is first skilled on software history data and then it is used to predict bugs. Two main drawbacks of existing classifier-based bug prediction are insufficient accuracy for practical use and deliberate prediction time because of a large number of machine learned features. In this paper we have proposed mainly two techniques cos-triage algorithm which tries to utilize both accuracy and cost of bug prediction and feature selection techniques which discard less important features until optimal classification performance is reached. Reducing the feature improve the quality of knowledge extracted and also enhance the speed of computation. |
| File Format | |
| Access Restriction | Open |
| Subject Keyword | Bug Prediction Cos-triage Algorithm Source Code Abstract Bug Practical Use System Consistency Difficult Task Prediction Time Useful Way Serious Challenge Software Defect Insufficient Accuracy Optimal Classification Performance Large Number Software Bug Prediction Classifier-based Bug Prediction Technique Cos-triage Algorithm Main Drawback Software History Data Feature Selection Technique Important Feature |
| Content Type | Text |
| Resource Type | Article |
