Document Type |
: |
Thesis |
Document Title |
: |
TIME-AWARE RECOMMENDER SYSTEM FOR E-COMMERCE APPLICATIONS نظام التوصية الواعي بالوقت في تطبيقات التجارة الإلكترونية |
Subject |
: |
Faculty of Computing and Information Technology |
Document Language |
: |
Arabic |
Abstract |
: |
As e-commerce websites began to develop, users found it difficult to find the most appropriate choice from the immense variety of items. Recommender systems have been applied to several domains such as online streaming and e-commerce to assist in decision making. A recommender system is a subclass of information filtering. It seeks to predict a rating that a user would give to an item. Recently, contextual information has been recognized as a useful factor in improving the quality of recommendations in different fields. However, it is under investigation in the area of online shopping that uses the purchasing data instead of the user’s rate. Among all contextual information, time is considered as one of the most important dimensions. This work integrates time dynamics with implicit feedback (add to cart, and purchase or transaction) in an online shopping recommender system using three algorithms: Matrix Factorization (MF), Nearest Neighborhood (KNN), and Sparse Linear Method (SLIM). The integration is done using two approaches: The first approach is Bias in which the time is used as the third column in the user rating matrix. The second approach is the Decay function which produces new ratings by aggregating implicit feedback with time dynamics and gives a higher weight to the new items over older ones. Using the “Retailrocket” online shopping dataset, the experimental results demonstrate the effectiveness of decay function over the traditional context-aware Matrix Factorization (MF) and Sparse Liners Method (SLIM) in terms of precision, recall, and Mean Average Precision (MAP). However, the Nearest Neighbor Algorithm (KNN) results in a decrease in effectiveness for all recommendations list lengths with 0.04%, 0.37%, 0.10% of precision, recall, and MAP respectively. |
Supervisor |
: |
Dr.Etimad Fadel |
Thesis Type |
: |
Master Thesis |
Publishing Year |
: |
1441 AH
2020 AD |
Added Date |
: |
Wednesday, June 17, 2020 |
|
Researchers
آيات يحيى محمود | Mahmoud, Ayat Yehia | Researcher | Master | |
|