Skincare Recommendation System based on User’s Skin Characteristics Similarity using Collaborative Filtering Approach

In order to obtain the optimal result from skincare products, consumer must use the ones that are suitable for their skin. However, the large amount of skincare products that are available in the market might cause confusion to consumer in choosing the right skincare products for themselves. Therefore, this study proposes a skincare recommendation system that will provide skincare product recommendations to consumer based on their skin characteristics. The proposed model will implement collaborative filtering approach by finding other consumer with similar skin characteristics using cosine similarity method, read-through the review comments to determine the rating of the products and create the relationship between users and products using Singular Vector Decomposition (SVD) algorithm. As a result, the proposed model shows an excellent result with accuracy score of 92.8%, highest among all baseline models. The proposed model also successfully outperformed other system that employed different approach with precision score of 76%. Hence, it is proven that the proposed model can give better skincare product recommendation to consumer.