The topic of this thesis was to find out whether the Fedora distribution attracts machine learning developers and also to create a guideline content for newcomers in this new, progressively developed field. In addition, part of the thesis research was also focused on solving the SRPM packaging dependency issue by using recommendation system based on machine learning algorithm. The model has two different variations, one is using the association rule learning algorithms Apriori and FP-Growth, and the other is based on neural network embedding layers approach, which is actually a Word2Vec model implementation from library Gensim. Targeted programming language was Python.
Research of Fedora Status for Machine Learning
Faculty of Informatics
Date of Completion