I Wanna Lend With Somebody, but I Still Haven’t Found What I’m Looking For
(2024) NEKP01 20241Department of Economics
- Abstract
- This thesis explores the application of Recommender Systems (RSs) to match borrowers with participant lenders in the syndicated loan market. It aims to determine whether historical interactions in the syndicated loan market contain sufficient information to reveal predictable patterns without relying on large sets of variables related to the lenders, borrowers, and the loans themselves. Two types of Collaborative Filtering (CF) algorithms — user-based and item-based — were employed, with additional clustering of borrowers to address data sparsity issues. Both CF models exhibited average F1 scores higher than random chance, suggesting the presence of predictable patterns. Nonetheless, the absolute average F1 scores were modest, with a... (More)
- This thesis explores the application of Recommender Systems (RSs) to match borrowers with participant lenders in the syndicated loan market. It aims to determine whether historical interactions in the syndicated loan market contain sufficient information to reveal predictable patterns without relying on large sets of variables related to the lenders, borrowers, and the loans themselves. Two types of Collaborative Filtering (CF) algorithms — user-based and item-based — were employed, with additional clustering of borrowers to address data sparsity issues. Both CF models exhibited average F1 scores higher than random chance, suggesting the presence of predictable patterns. Nonetheless, the absolute average F1 scores were modest, with a maximum of 0.06. Clustering borrowers improved the scores, reaching up to 0.21, yet the CF algorithms did not consistently outperform
the most-popular model baseline. Future research can focus on integrating more advanced RSs to enhance effectiveness. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9159666
- author
- Molin, Elisabeth LU
- supervisor
- organization
- course
- NEKP01 20241
- year
- 2024
- type
- H2 - Master's Degree (Two Years)
- subject
- language
- English
- id
- 9159666
- date added to LUP
- 2024-10-01 13:19:35
- date last changed
- 2024-10-01 13:19:35
@misc{9159666, abstract = {{This thesis explores the application of Recommender Systems (RSs) to match borrowers with participant lenders in the syndicated loan market. It aims to determine whether historical interactions in the syndicated loan market contain sufficient information to reveal predictable patterns without relying on large sets of variables related to the lenders, borrowers, and the loans themselves. Two types of Collaborative Filtering (CF) algorithms — user-based and item-based — were employed, with additional clustering of borrowers to address data sparsity issues. Both CF models exhibited average F1 scores higher than random chance, suggesting the presence of predictable patterns. Nonetheless, the absolute average F1 scores were modest, with a maximum of 0.06. Clustering borrowers improved the scores, reaching up to 0.21, yet the CF algorithms did not consistently outperform the most-popular model baseline. Future research can focus on integrating more advanced RSs to enhance effectiveness.}}, author = {{Molin, Elisabeth}}, language = {{eng}}, note = {{Student Paper}}, title = {{I Wanna Lend With Somebody, but I Still Haven’t Found What I’m Looking For}}, year = {{2024}}, }