EngageQuotient aims to enhance the user-based collaborative filtering algorithm and assess its performance against traditional methods. Experimental results show that the improved algorithm achieves lower prediction errors, higher precision, and better recall in sparse data scenarios compared to the conventional approach. By refining collaborative filtering based on user similarity, this enhanced algorithm provides more accurate predictions and reduced processing time, effectively addressing data sparsity and overfitting issues.
- Clone the project
git clone https://github.com/iaayushmaan/EngageQuotient- Go to the project directory
cd ./EngageQuotient- Un-zip the dataset.
- You are good to go!
The proposed collaborative filtering based on user similarity results in minimal prediction error, higher precision, and recall under sparse data compared to traditional methods. This improvement addresses data sparsity and overfitting issues, providing more accurate predictions and better recommendation performance.
The coverage of the proposed algorithm will be tested by computing the ratio of recommended items to the total number of items, illustrating its capability to offer a diverse array of recommendations. Evaluating the recommender system's performance with and without incorporating user ratings will provide insights into its effectiveness in personalized recommendation delivery. This comparative analysis assesses how effectively the system leverages user preferences to enhance recommendation accuracy and user satisfaction, highlighting the impact of personalized versus non-personalized approaches on overall recommendation quality and diversity.