<![CDATA[In choosing skin care products, many consumers often make mistakes due to a lack of understanding of skin type and a lack of knowledge about skin care products available on the market. This often makes it difficult for them to find suitable products. This research aims to design an application that is able to provide skincare recommendations based on previous product preferences. The method used is Content-based filtering. The recommendation process is carried out by comparing product content to produce the highest to lowest ranking, as well as calculating the minimum support and confidence values to determine association rules for itemset combinations. To calculate the similarity between words using the cosine similarity algorithm, product descriptions will be given a value using TF-IDF (Term Frequency-Inverse Document Frequency) calculations. After that, the similarity weight will be calculated using the cosine similarity algorithm, the similarity weights from highest to lowest. In this research, the product with the highest similarity value was obtained with a value of 0.722.]]>
Copyrights © 2024
