Master Seminar:
Applications of Machine Learning in Software Engineering
(WS 14/15)
Professor: Prof. Bernd Brügge, Ph.D.
Instructors: Tobias Roehm, Stefan Nosovic
Objectives
This seminar focuses on techniques from machine learning and data mining and surveys their application in software engineering.
In particular, the seminar investigates how such techniques aid program comprehension, help to analyze and interpret large amounts of user input, and facilitate the extraction of relevant knowledge from unstructured data.
During the seminar each participant will read a scientific paper and present it to the seminar audience. More specifically, each participant presents a particular machine learning technique and a particular application in software engineering.
Organizational Issues
- An information meeting will be held on Wed, 2.4.14, at 14:30 h in room 01.07.058.
- Please indicate your interest in the seminar by sending an e-mail to Stefan Nosovic containing a CV and brief description of previous ML/ Data mining experience. Application deadline is Sunday, 6. July 2014, 23:59.
- The main seminar language will be English but report and presentation can also be done in German.
- Previous experience with Machine Learning and Data Mining is not mandatory but preferred.
- The seminar presentations will be done in 1-2 days at the end of the semester.
Possible Topics
Topics will be assigned in the first seminar meeting.
Paper Title | Authors | Used Technique |
Summarizing the Content of Large Traces to Facilitate the Understanding of the Behaviour of a Software System | Hamou-Lhadj and Lethbridge | Statistics |
An information retrieval approach to concept location in source code | Marcus et al. | Latent Semantic Indexing |
Design pattern mining enhanced by machine learning | Ference et al. | Decision trees, neural networks |
How Long will it Take to Fix This Bug? | Weiß et al. | a-kNN |
Duplicate Bug Reports Considered Harmful... Really? | Bettenburg et al. | SVM |
Paper Title | Authors | Used Technique |
A Recommender System for Requirements Elicitation in Large-Scale Software Projects | Castro-Herrera et al. | Collaborative Filtering |
Towards an Intelligent Code Search Engine | Kim et al. | Summarization based on AST information and K-means |
On-demand Feature Recommendations Derived from Mining Public Product Descriptions | Dumitru et al. | Incremental Diffusive Clustering |
Classification of Software Behaviors for Failure Detection: A Discriminative Pattern Mining Approach | Lo et al. | Frequent iterative pattern mining |
Discriminative Pattern Mining in Software Fault Detection | Di Fatta et al. | Frequent pattern mining: FREQT |
Modalities
Grades will be based on the following criteria:
- Ability to do independent research
- Oral presentation
- Written term paper
- Active participation in all the other presentations (compulsory attendance)