Mining online review for predicting sales performance

Abstract

Information technology field getting raised day-by-day and due to the growth of internet associated services the online shopping and social media is become more popular as well as attain a great development. The involvement of users in social media as well as e-commerce sites are increasing a lot and all are willing to share their opinions as post in the social medium as well as the increasing volume of reviews raising several ideologies over both commercial and non-commercial firms. The commercial vendors think like for obtaining the reviews from customers to identify the sales level of the product as well as the vendors think like based on these reviews identify the best product as well as provide the best service to customers accordingly. So, these reviews are most important to deal with as well as through this review mining logic sales and its associated performance will be improved in drastic level. These kinds of reviews estimation factor is called as Sentiment Analysis, in this paper a new sentimental analysis logic is introduced called as Sales and Performance of Products Analysis based Sentimental PLSA (S-PLSA). By using this formulation, the review of the product is considered and the sentimental vectors are measured by using complex sentimental estimation logic based on the reviews generated by the user. The logic is estimated with the help of training association and model creation principles, these models are cross-validated with the help of real-time live data review reports. The experimental results prove the proposed approach of SPPA provides good response in results as well as the Quality-of-Service to customers to prove the effectiveness of the proposed algorithm called SPPA.