Leveraging Network Similarity Measures for Recommendation Systems
- Resource Type
- Conference
- Authors
- Anjum, Shagufta; Ali Masood, Mohammed Rayid; Gupta, Yayati; Sukhija, Sanatan
- Source
- 2022 International Conference on Emerging Techniques in Computational Intelligence (ICETCI) Emerging Techniques in Computational Intelligence (ICETCI), 2022 International Conference on. :102-109 Aug, 2022
- Subject
- Aerospace
Bioengineering
Computing and Processing
General Topics for Engineers
Geoscience
Power, Energy and Industry Applications
Robotics and Control Systems
Signal Processing and Analysis
Transportation
Measurement
Collaborative filtering
Electronic commerce
Recommender systems
Engines
Computational intelligence
recommendation systems
bipartite graph
collaborative filtering
matrix factorization
singular value decomposition
similarity metrics
- Language
With the growth of e-commerce websites, efficient recommendation systems are desired to reduce the turnaround time for servicing a customer. This study focuses on understanding the various techniques and algorithms that are used to model real-life recommendation systems. We present a recommendation engine for Amazon products that uses collaborative filtering (CF). Given a list of users and their reviews of Amazon products, our CF-based recommendation engine generates a ranked list of the top k products for individual users. The generated recommendations are based on the preferences of similar users and past purchases. We have created two such systems: a memory-based user-item CF system and a model-based CF system that utilizes matrix factorization techniques. Additionally, we also propose a graph-based recommendation technique that generates a list of the most similar products. This method relies on network-based local similarity metrics to generate product suggestions.