Abstract—The analysis of emotions can be performed on various sources of input, namely: text, speech or voice, images, and videos. This paper discusses a graph-based technique that prepares the ground for flexible and expandable emotion analysis from text documents. We take into account the complexities of human emotion representation in text that is more challenging and susceptible to inconsistent outcomes. The design of the proposed system aims towards a generalized solution for such kinds of text mining applications and emphasizes a simplified (yet flexible) emotion representation and analysis system that can be subjected to fine-tuned analysis, depending on future requirements. The aim of our work is to develop a prototype that can be used for (unsupervised) learning from some given text and can be used for possible extension of the emotion dictionary.
Index Terms—Affective Computing, Natural Language Processing, Opinion Mining, Pattern Clustering, Text Analysis, Text Mining.
Rekha S. Sugandhi is with the MIT College of Engineering, Department of Computer Engineering, Pune, India, (e-mail: rekha.sugandhi@mitcoe.edu.in).
Aneesh S. Mulye was Persistent Systems Ltd., India. He has enrolled for MS (Computer Science) from the Georgia Institute of Technology, Atlanta USA (e-mail: aneesh.mulye@gmail.com).
Dr. Vijay M. Wadhai is with the MIT College of Engineering, Pune, India as Principal, He is also Director, Research and Development, Intelligence Radio Frequency (IRF) Group, Pune. (e-mail: wadhai.vijay@gmail.com).
[PDF]
Cite: Rekha S. Sugandhi, Aneesh S. Mulye and Vijay M. Wadhai, "A Framework for Extensible Emotion Analysis System,"
International Journal of Engineering and Technology vol. 3, no. 5, pp. 540-546, 2011.