Work in Progress
I am trying to keep this work-in-progress page updated:
- Automated Bug Triaging. We have currently finished the collaboration with a Czech-based company to explore the introduction of an automated bug triaging system. The interest is on looking into aspects that differ in proposing such systems withi the context of industry, compared to usual open source software projects evaluations. Tools used: mostly Python scripting with NLTK libraries;
- Cost Sensitive Classification for Severity of Bug Tracking. Following SEAA'14 paper, we are looking into the evaluation of several cost sensitive classification schemes to deal with the problem of data imbalance - quite relevant in the context of issue trackers considering bugs severity. Tools used: mostly Python with NLTK libraries;
- Anomaly detection on Smart Grids data. The aim of the research is to support the decision making process within a Smart Grids Infrastructure by an evaluation of possible anomalies from Smart Meters. The proposed approach is based on frequent itemset mining and unsupervised clustering and is under current improvement. An initial paper has been submitted at the SMC conference. Improvements will go in the direction of both scaling-up the approach to allow online learning, as well as to improve the results based on experts' opinion (more info on (private) BitBucket page). Tools used: R-project scripting;
- Software Design Pattern Implementation Violations: Looking into violations of design pattern implementation that due to code aging as a form of technical debt: provision of a software tool that can mine software repositories and providing scoring for design patterns integrity. Tools used: mostly Java implementation;
- Software Requirements Prioritization and Machine Learning. The interest is into looking about the application of machine learning and data mining for the requirements prioritization process, that is what we can still use from the research on issue trackers (severity, triaging), and what are specific challenges in the way requirements are elicited. We are mostly interested in the different open source software projects communities and the differences in requirements management for the prioritization process;