Data Mining Techniques for Predicting User Interest in Facebook Pages: A Comparison
Nusrat Jahan Farin, Morium Akter, Puthi Roy, Mohammad Shorif Uddin · 2019 1st International Conference on Advances in Science, Engineering and Robotics Technology (ICASERT) · 2019
Now-a-days the use of social media (e.g; facebook, twitter etc.) is increased rapidly. It is rational to exploit interest of users based on their facebook-page activities. However, it cannot be easily predicted manually as there are number of attributes including social interactions. The aim of this paper is to do analysis of 7 data mining algorithms and their variants for determining the best appropriate prediction of people interest in facebook pages. For this, we will analyze and compare various algorithms such as KStar, LWL, IBK, J48, Naive Bayes, Bayes Net, Decision Table. Their performances have been evaluated by using a dataset with 10172 instances of 6 attributes each. Among all the utilized algorithms Bayes Net demonstrates the best execution in regard to the accuracy. However, in respect to the computational time KStar demonstrates the best execution instead of Bayes Net.