The study aims to use Facebook data to create a user profile to be able to recommend personalized gifts and help users to choose the right gift for a certain occasion. The Facebook data includes posts, comments, liked pages and user’s biography. The data gathered are then preprocessed to create a user profile and item profile. The preprocessing stage includes data cleaning, and POS tagging. These profiles can be classified as book lover, fashion fiend, outdoor enthusiast, foodie, music lover and sports fan. These profiles are then mapped through content-based and multicriteria collaborative filtering. In content-based filtering three criteria are used, namely, receiver’s personality, cosine similarity and user’s chosen event. The events include birthday, valentines, wedding, anniversary, father’s day, mother’s day and graduation. Multicriteria collaborative filtering uses Pearson Similarity to distinguish similar users who would likely like the same product. Combining these results, a hybrid system is produced and a desirable list of items is recommended.
CITATION STYLE
Buctuanon, M. M., Alegado, J. C., Daculan, J., & Ponce, L. C. (2019). Utilizing social media analytics to recommend personalized gifts using content-based and multicriteria collaborative filtering. In Advances in Intelligent Systems and Computing (Vol. 886, pp. 423–437). Springer Verlag. https://doi.org/10.1007/978-3-030-03402-3_29
Mendeley helps you to discover research relevant for your work.