Extraction of Characteristics from SNS Guidelines -Analysis using Guidelines Issued by National and Public Universities-

Abstract

This study tries to classify the social networking service (SNS) guidelines created by university administrators. These guidelines describe the rules for using SNS. Although SNS is a powerful and useful tool for sharing information, some students share unethical content, which is a social issue. Therefore, universities create SNS guidelines to prevent troubles. In this study, we try to extract differences between the guidelines using a neural computational method called ‘potential learning’. In our previous study (FSS2017), we tried to extract characteristics from SNS guidelines. However, this study only uses SNS guidelines issued by national and public universities. Moreover, to track the changes in characteristics, classification was conducted several times. From the experimental result, it was found that guidelines were classified by ‘penalty related words’, followed by ‘natural products related words’, ‘animal related words’, and ‘animal-part related words’.