Enhanced sentimental analysis using visual geometry group network-based deep learning approach

In this digital era, there are many reasons to investigate the positive and negative opinion: a lot of irrelevant or error-related information exists, and major changes can strengthen or undermine past knowledge due to the newly discovered information. Speculation and negation principles are used as a significant influence in determining the factuality of occurrences or words. Negation reverses the validity of argument to maximize or decrease the doubt regarding the opponent and conjecture. Deep learning recently demonstrated has increased ability to distinguish truthful and non-true knowledge. Moreover, different languages were used in most of the previous approaches. To our understanding, the negative or hypothetical concept in biomedical texts in English is not known in previous work. This work will establish an adaptive visual geometry group network (VGG Net) based on CNN models for the identification and rejection of statements (negated or speculated sentences, etc.). The applied models are tested by considering the English corpus BioScope. Finally, as a result, the statistical study confirms that the performance of the proposed models surpasses our original baseline on F1-Score.