Σύστημα προτάσεων αγορών εξατομικευμένο σε ανάγκες χρηστών
Recommender system for purchases with personalization on users’ needs
View/ Open
Keywords
Σύστημα προτάσεων ; Φιλτράρισμα βάσει περιεχομένου ; Εξατομικευμένες προτάσεις ; Ψηφιακές αγορές ; Ικανοποίηση χρήστηAbstract
In today’s digital age, the rapid growth of online shopping and content streaming platforms has resulted in an overwhelming number of products and content available, making it difficult for users to locate the items they are interested in. Recommender systems have emerged as a key tool for improving user experience, by offering personalized recommendations for user satisfaction. In this thesis, a recommendation system based on an online shopping platform for laptops was developed. The features of the laptops were used to develop a recommender system based on the content-based filtering. In contrast to a typical content-based filtering system that recommends similar products, in this case emphasis is placed on the recommendations of products with a better quality/price ratio than the product selected by the user. The system aims to guide inexperienced users, in purchasing the most suitable laptop based on their needs. It uses filtering techniques and user’s preferences regarding the maximum price and category of the product he wishes to buy, ensuring accuracy in the suggestions made on the end user. The Precision@N metric was used to evaluate the accuracy of the system. A dataset of several different categories of laptops was used, which were pre-processed and adjusted accordingly. The system was implemented using the Python-based web framework Django for the backend and the JavaScript library React for the frontend. Through this thesis, a practical application of recommendation system techniques is made, and their capabilities are investigated in an online shopping environment.