Abstract:Emotion analysis of posts is still challenging because of the limitedcontextual information that they normally contain. . This issue is resolved byconstructing a separate space for emoticon as emotional space which isrepresented as feature matrix and projections of emotions and words is donehere. this is based on semantic composition. Performance of emotion analysis isimproved as proposed method is capable to capture more emotion sematic thenother models which is done by projecting emoticons and words into emoticonspace. This helps identify subjectivity, polarity and emotion in microblogenvironment. This paper in course gives insights for design of ECNN. Facialexpressions give important clues about emotions. Sentimental analysis is othernatural language processes for more task. Thus, many approaches have been putforth to classify human affective states.
InFacial expression analysis, features used are local spatial position or missingpoints or regions of face. In audio analysis, global statistics of acoustics isused as feature.
Keywords:EmotionRecognition, Amazon Product Review, Text Mining, Natural Language Processing(NLP), Sentiment Analysis, Convolution Neural Network