Sentimantics: Conceptual Spaces for Lexical Sentiment Polarity Representation with Contextuality

Current sentiment analysis systems rely on static (context independent) sentiment lexica with proximity based fixed-point prior polarities. However, sentimentorientation changes with context and these lexical resources give no indication of which value to pick at what context. The general trend is to pick the highest one, but which that is may vary at context. To overcome the problems of the present proximity-based static sentiment lexicon techniques, the paper proposes a new way to represent sentiment knowledge in a Vector Space Model. This model can store dynamic prior polarity with varying contextual information. The representation of the sentiment knowledge in the Conceptual Spaces of distributional Semantics is termed Sentimantics.