An Investigation on Social Network Recommender Systems and Collaborative Filtering Techniques

Nowadays, with the remarkable expansion of the information through the internet, users prefer to receive the exact information that they need through some suggestions from their friends or profiles to save their time and money. Recommend systems based on different algorithms as one of the basic ways to reach this goal through the internet have been proposed but each of them has their own advantages and disadvantages. In this study, we have selected and implemented two approaches which are Collaborative Filtering (CF) and Social Network Recommendations System (SNRS). Based on some limitations to finding a dataset which covers friendship, rating and item categories we generated it for 10 categories, 10 items, and 100 users and compared two approaches. We used Mean Absolute Error (MAE) and accuracy to compare the result of two mentioned approaches and found that the SNRS method as it is claimed to be improved version of CF works more efficiency.