ALOSI: Aspect-Level Opinion Spam Identification

Opinion mining, an upcoming area of computational linguistics, deals with analysis of public sentiment polarity of a product giving rise to a new platform on social media and e-commerce for users to express/explore opinions on products and/or product manufacturers for the betterment of the product development. A tentative product rating can be derived from the product opinions mined. ALOSI takes opinion mining to a deeper level performing opinion mining at aspect level giving well-analyzed results at product component level. But the data received is not always totally relevant leading to false rating of the product or aspect(s). This irrelevant data can be labeled as spam and is important to identify as it may restrict the model. The presented work models a product-based aspect-level sentiment analysis on public reviews with spam identification and filtering, hence, highlighting the difference in the rating of product or aspect(s) with the presence and absence of spam reviews.