Proposed Projects for B.Tech

PROJECT-1

Proposed Projects for B.Tech 4th Year (2015-16)

Proposed Project Title: Duplicate Bug Detection by using Natural Language Processing and Bug Triaging in Bugzilla by Evolving Developer Social Networks

Project Abstract: We have to introduce Developer Social Network (DSN) in bugzilla by using the database of Eclipse and/or NetBeans (Open Source Software). The idea was taken from the paper “Evolution of Developer Social Network and Its Impact on Bug Fixing Process” by Gupta A. et al. This paper shows the evolution of DSN by comparing the Social Network properties like Average path Length, Clustering Coefficient, Modularity, Average Distance etc.

DSN in Eclipse bugs database and its assignment for solving the bug introduce the detection of duplicate bugs; therefore we will device a mechanism to detect duplicate bugs reported. Detection of duplicate bugs can be done by using natural Language Processing, information Retrieval and execution information.

Expected Features & Novelness: Use of NLP and execution information, extracting information are the features of project. Evolution of DSN in bug fixing process is another feature of this project.

Objectives: In every organization where we have to communicate, apply social network features to that organization and take advantages of social network in that organization. We can divide our findings into the following objectives:

1. Evolving Developer Social Network in Bugzilla (Eclipse/NetBeans/chrome etc,)
2. Duplicate Bug Detection and Bug Triaging in Bugzilla by Evolving Developer Social Networks

Expected Results and its Significance/Usefulness: In this project, we can present a novel approach to assist triagers in detecting duplicate bug reports. Unlike existing approaches, this approach further considers execution information. Furthermore, this approach employs two heuristics to combine the two kinds of information. The expected results can be the comparison with the best performance of approaches using only natural language information, this calibrated approach (with the classified-based heuristic and using only the summary) leads to approx. 15% to 20% increase in recall rates on the two experimental bug-report sets respectively.

To be Proposed Technology/Platform JAVA/DOT NET, SQL

Any Additional Resources/Support Required: The database of Bugzilla of approx. 10GB.

PROJECT-2

Proposed Projects for B.Tech 4th Year (2015-16)

Proposed Project Title: Identification and Detection of classified and non-classified Objects through Outlier Analysis

Project Abstract: In real life there are many observations, we have observed that some elements may have similar characteristics and few of them behave differently. Outlier Analysis is an important concept, it is an observation (or measurement) that is different with respect to the other values contained in a given dataset. In data mining literature different definitions of outlier exist: such “An outlier is an observation that deviates so much from other observations as to arouse suspicions that it was generated by a different mechanism “(Hawkins, 1980). “An outlier is an observation (or subset of observations) which appear to be inconsistent with the remainder of the dataset” (Barnet & Lewis, 1994).