An intelligent framework for text-to-emotion analyzer

The expansiveness of the internet encourages people to express their personal feelings via textual medium in terms of virtual communication. People are being influenced to move into this type of communication with the remarkable growth of social sites, messengers, blogs, micro blogs etc. Automatic derivation of the emotion from text is a challenge as it minimizes the misunderstanding by conveying the internal state of the users. Here we propose an intelligent framework to detect the emotion of a text. We divide the framework into two modules, namely Training Module and Emotion Extraction Module. We utilize the concept of Exploratory Data Warehouse (DW) technology to train our system. Therefore, DW relies not only on internal data but also on external (Web) data. The DW is used by the Emotion Extraction Module to detect the emotion of a given text. A comprehensive experimental evaluation, comparing our framework to a solution made with existing systems on a concrete dataset, shows that the proposed framework outperforms the existing approaches in terms of accuracy.