Predictive Insights: Machine Learning and FOSS Project Sustainability
Authors
Klockmann, Steffan ; Laursen, Mathias Kudahl ; Tarp, Jonas Højen
Term
4. Term
Publication year
2024
Pages
58
Abstract
This thesis investigates whether machine learning—computer models that learn from data—can predict which Free and Open Source Software (FOSS) projects in the Apache Software Foundation Incubator (ASFI) will graduate or be retired, using sustainability metrics. We trained models on indicators of long-term project health, such as community size, development activity, growth, how often people communicate, and turnover (changes in contributors). The models showed potential to forecast outcomes for ASFI projects, and the results suggest a link between sustainability and whether projects graduate (leave the incubator) or retire (end in the incubator). We recommend combining quantitative data with qualitative assessment to improve understanding and prediction accuracy.
Dette speciale undersøger, om maskinlæring—computermodeller, der lærer af data—kan forudsige, hvilke FOSS-projekter i Apache Software Foundation Incubator (ASFI) vil graduere eller blive pensioneret, ved hjælp af bæredygtighedsmålepunkter. Vi trænede modeller på indikatorer for projekters langsigtede sundhed, såsom fællesskabets størrelse, udviklingsaktivitet, vækst, kommunikationsfrekvens og udskiftning blandt bidragydere. Modellerne viste potentiale til at forudsige udfald for ASFI-projekter, og resultaterne peger på en sammenhæng mellem bæredygtighed og om projekter graduere (forlader inkubatoren) eller pensioneres (slutter i inkubatoren). Vi anbefaler at kombinere kvantitative data med kvalitative vurderinger for at forbedre forståelsen og forudsigelsernes nøjagtighed.
[This apstract has been rewritten with the help of AI based on the project's original abstract]
Keywords
FOSS ; Sustainabiilty ; Apache ; ASFI ; Machine-Learning ; ML ; Prediction