PREDICTING SOCIAL NETWORK ADDICTION USING VARIANT SIGMOID TRANSFER FEED-FORWARD NEURAL NETWORKS (FNN-SNA)

Abstract

Researchers have reflected on personal traits that may predict Social Networking Sites (SNS) addiction. However, most of the researchers involved in the findings of personality traits predictor for social networking addiction either postulate or based their conclusions on analytical tools. Moreso, a review of the literature reveals that the prediction of social networking addiction using classifiers have not been well researched. We examined the prediction of SNS addiction from a well-structured questionnaire consisting of sixteen (16) personality traits. The questionnaire was administered on the google form with a response rate of 95% out of the 102-sample size. Additionally, a three (3) variant sigmoid transfer feed- forward neural networks was developed for the prediction of SNS addiction. Result indicated that pertinence (β = 0.251, p  0.01) was the most powerful predictor of social networking addiction in general and less obscurity addiction (β = 0.244, p  0.01). Experimental results also showed that the developed classifier correctly predict SNS addiction with 98% accuracy compared to similar classifiers.