Ανάλυση συμπεριφοράς καταναλωτή
Καραγιασσώτη, Παναγιώτα Ειρήνη
KeywordsΑνάλυση ; Συμπεριφορά καταναλωτή ; Marketing ; Collaborative filtering ; Recommendation system ; e-Shop
This paper aims to present the functionality of a system of recommendations in the context of e-commerce. The system suggests a particular group of products (top N products) among many candidate types of products that the user will potentially buy for the first times . Using the collaborative filtering method, the probability of choosing the item by the current user is calculated, and the most suitable products are suggested on the page are displayed. The main advantage of the method is that it works both in cases with limited content, as well as in others where the analysis of objects becomes difficult. The development of the e-shop platform was made using the MVC Model / View / Controller, which is the most popular method, as it manages to separate the dependencies between the different parts of the code. Finally, using the Weighted Slope One method on the platform, it is concluded that in order for the recommendations system to work better, the user must rate at least one product in the class, while other products must also be rated by other users.