News sentiments grading using multiple dictionaries based lexical approach

Abstract

In sentiment analysis, traditionally the results are thought of on scale of three grades, positive, negative and neutral. Going beyond the traditional emphasis, an effective and result oriented methodology for step by step news sentiment grading into three, five and seven different categories based on scale and intensity of sentiment using multiple dictionaries based lexical approach is going to be detailed in this paper, so they can be well understood easily, with the domain not restricted to just one section or subjective part. Potential differences in addition to unveiling of the tricky and challenging parts of news sentiment processing and grading process are also part of the details. Implementation and experimentation using MPQA dataset at the end reveals the effectiveness of methodology in addition to basis and formalization.