Institutional Repository of Vilnius University
Product recommendation systems with neural networks and enhancement opportunities with conjoint analysis /
Abstract
dc:descriptionNowadays, many customers value personalized product proposals. This study analyzes popular product recommendation approaches such as content-based recommender system and collaborative filtering (specifically Artificial Neural Networks). The goal of this paper is to assess current accuracy of recommender systems and to propose enhancement opportunity for collaborative filtering and content-based recommendation approaches using conjoint analysis. The data analyzed is Amazon giftcard reviews. Results of this analysis imply that the best method to recommend a product to a user is collaborative filtering using neural networks. These models were able to reach mean squared error of around 0.01 on test set and mean squared error 0.2 on validation set and it reached 76% prediction accuracy. The model recommends five giftcards for each user which were predicted to be good rated. At the end of this study it is suggested that the accuracy of predictions can be improved and overall research process simplified by introducing conjoint analysis.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2022
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Maikštėnaitė, Emilija,
Subjects
dc:subject × 1Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- eng
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD146236047&prefLang=en_US