University of Twente Student Theses
Predicting the Success of Crowdfunding Campaigns on Kickstarter
Lysin, Serhii (2024) Predicting the Success of Crowdfunding Campaigns on Kickstarter.
|
PDF
711kB |
| Abstract: | This thesis analyses the ability of machine learning models to predict the success of crowdfunding projects on the biggest reward-based platform - Kickstarter. This paper uses a dataset with more than 150,000 campaigns from 2009 to 2018 to evaluate feature importances and critically review the performance of three tree-based models: Random Forest Classifier, XGBoost and LighGBM. This research identified that duration, subcategory, and fundraising goal are the most significant features that influence the success of projects on Kickstarter. All prediction models were evaluated based on cross-validation accuracy, precision, recall, F1 score, ROC ACU score, and log loss. The findings imply that XGBoost and LightGBM performed at the same high level, while the Random Forest Classifier was chosen as the best model for the given task. Case studies validated the models’ predictive reliability. This study contains certain limitations and weaknesses, including problems with biased dataset. Finally, possible future development was discussed, considering the integration of real-time analysis and expanding the utilised dataset. |
| Item Type: | Essay (Bachelor) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 54 computer science |
| Programme: | Business & IT BSc (56066) |
| Link to this item: | https://purl.utwente.nl/essays/101151 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page