A comparison of similarity measures for online social media Thai text classification

Social media is widely used as a channel of communication in general purposes, including the comment that are related to retail business. It is a highly effective communication tool for direct interacting with their customers. Growth rate of the users is rapidly increasing, because they use this channel to receive information and share something interesting. In this paper, we present a comparison experimental results on the similarity measurements on Thai social media text. We use the nearest neighbor classifier with the 10 distance functions to classify there data into 4 classes, which the classes were defined by the expert in retail business. The experiments were performed on two feature extractions with the well-known term weightings and compared the classification results with the experts. According to the experimental results, we found the feature extraction and the distance that can be employed with Thai social media text classification.