Adjectives Grouping in a Dimensionality Affective Clustering Model for Fuzzy Perceptual Evaluation

Abstract

More and more products are no longer limited to the satisfaction of the basic needs, but reflect the emotional interaction between people and environment. The characteristics of user emotions and their evaluation scales are relatively simple. This paper proposes a three-dimensional space model valence-arousal-dominance (VAD) based on the theory of psychological dimensional emotions. It studies the clustering and evaluation of emotional phrases, called VAdC (VAD-dimensional clustering), which is a kind of the affective computing technology. Firstly, a Gaussian Mixture Model (GMM) based information presentation system was introduced, including the type of the presentation, such as single point, plain, and sphere. Subsequently, the border of the presentation was defined. To increase the ability of the proposed algorithm to handle a high dimensional affective space, the distance and inference mechanics were addressed to avoid lacking of local measurement by using fuzzy perceptual evaluation. By comparing the performance of the proposed method with fuzzy c-mean (FCM), k-mean, hard -c-mean (HCM), extra fuzzy c-mean (EFCM), the proposed VADdC performs high effectiveness in fitness, inter-distance, intra-distance, and accuracy. The results were based on the dataset created from a questionnaire on products of the Ming style chairs online evaluation system.

Keywords

• clustering
• affective computing
• fuzzy
• valence-arousal-dominance
• product evaluation