A RECOMMENDER SYSTEM UTILIZING CROWDSOURCING FOR VILLAGE-OWNED ENTERPRISES' PRODUCT RECOMMENDATION

The COVID-19 pandemic period is a good opportunity for village-owned enterprises (BUMDes) to restructure their operations by shifting online. The main challenge of the village-The COVID-19 pandemic period is a good opportunity for village-owned enterprises to restructure their operations by shifting online. The main challenge of village-owned enterprises is managing potential products and selling them to both the community and investors. This study aims to recommend village-owned enterprises' products to investors. A linear kernel content-based recommender system is used to recommend the products of village-owned enterprises to investors, while an embedding sigmoid collaborative filtering recommender system is utilized to determine the quality of the items. The crowdsourcing method is used to collect some data points, such as ratings, reviews, and investment, for collaborative recommendation. The results of content-based recommendation demonstrate that the linear kernel method outperforms cosine similarity by 81% in terms of accuracy with a slightly longer computation time, whereas the result of an embedding sigmoid yielded 79% of the accuracy with the shortest computation time of 30 milliseconds when compared to other methods. In addition, we developed a mobile application enabling village-owned enterprises to recommend their products.