Contributor
Giannis Christodoulou

AQAvit - Deep AQAtik


Mentors
Shelley Lambert
Organization
JBoss Community
Technologies
python, bash, seaborn, sklearn, gensim
Topics
machine learning, nlp, doc2vec, LDA
The current proposal deals with the task of document similarity utilizing dense text representations. Furthermore, it leverages topic modelling to investigate potential non-informative sets of words which could be removed in the preprocessing phase. The latter step will improve training time and hopefully will increase performance on the document similarity task. The task occurred from the need of matching Github issues with the relevant Jenkins logs in order to correlate ci pipeline errors with prior identified issues.