A Recommendation System Implementation For E- commerce Web Sites With Implicit Feedback Data Sets: An Ecuadorian Enterprise Case Study

Nowadays, the internet has become a very important and widely used tool in several human daily activities. E-commerce is one of the sectors being powered by the internet, enabling people to purchase products or services more easily. Due to information overload, enterprise actors are constantly immersed in the search for tools that allow them to offer correctly their products to the great variety of users that visit their e-commerce. This research emphasizes the relevance of the implicit feedback data set when building a recommendation system. Likewise, it explains the state of the art about recommendation systems and mentions the benefits of using this kind of tool in companies. Besides, this degree thesis document implements and evaluates various models of recommendation techniques based on matrix factorization using two data sets. One of these datasets comes from Retail Rocket, a real anonymous e-commerce website which has collected implicit data from customers and has decided to share for research purposes. All the models here- implemented are evaluated and compared regarding two evaluation metrics commonly used in the recommendation systems field. Finally, the models are implemented with the Ecuadorian real data set. This data set was used to reveal how the distinct models might behave under real data provided by an Ecuadorian similar enterprise.

• 主要作者: Román Eras, Osiris Anael (author)
• 格式: bachelorThesis
• 语言: eng
• 出版: 2020
• 主题: Ecommerce; Recommender Systems; Implicit Feedback; Machine Learning; Matrix Factorization; Comercio Electrónico; Sistemas de Recomendación; Feedback Implícito; Factorización Matricial
• 在线阅读: http://repositorio.yachaytech.edu.ec/handle/123456789/193