In this paper we investigate the automatic recognition of emotion in text. We propose a new method for emotion recognition based on the PPM (PPM is short for Prediction by Partial Matching) character-based text compression scheme in order to recognize Ekman’s six basic emotions (Anger, Disgust, Fear, Happiness, Sadness, Surprise). Experimental results with three datasets show that the new method is very effective when compared with traditional word-based text classification methods. We have also found that our method works best if the sizes of text in all classes used for training are similar, and that performance significantly improves with increased data.
This paper consist some new generalizations of some definitions such: j-ω-closure converge to a point, j-ω-closure directed toward a set, almost j-ω-converges to a set, almost j-ω-cluster point, a set j-ω-H-closed relative, j-ω-closure continuous mappings, j-ω-weakly continuous mappings, j-ω-compact mappings, j-ω-rigid a set, almost j-ω-closed mappings and j-ω-perfect mappings. Also, we prove several results concerning it, where j Î{q, δ,a, pre, b, b}.