Secure User Recommendation using Machine Learning Techniques
Durai Kumar. D, Sherly Puspha Annabel L · 2023
A secure user recommendation is a significant and emerging service, which expands and enhances the recommendation service by actively recommending secure user to other users based on the trustworthiness. One primary reason for the need of a secure user recommendation system in current society is that people have too many opportunities to access other users due to the ubiquity of the internet. The existing recommendation system such as content based, collaborative filtering gives user to user recommendation based on their similar choices, activities and location preferences etc. These exiting methods have narrow scope of recommendation and are less efficient. This study proposes a new secure user recommendation system that overcomes the shortcoming of the existing system. The proposed secure user recommendation system determines the trustworthiness of user found in the dataset based on the various attributes like profile, interaction, intensity of relationship etc. The machine learning techniques are used for giving recommendation on various applications. To arrive at the goal of intensifying the user trustworthiness, here the focus is on building the trustworthy model that can predict the trustworthiness level of users under the multiple trust attribute conditions using machine learning techniques. To help the process of finding the most accurate machine learning technique in providing the secure user recommendations, the performance metrics such as precision, recall, F1 score and accuracy of various machine learning techniques are calculated and it is also observed that a random forest technique gives higher accuracy of 97.5% than the other existing techniques.