A Model to Analyse Social Media Data to Gain a Competitive Edge

Data mining, text mining and opinion mining in one form or another has occurred since people started maintaining records in the modern era. Online posting on products from different popular sites like twitter, flipkart, amazon etc. has been increased, so research on sentimental analysis is going for long time. Due to increase in number of users for online shopping, accessing social media websites that generates huge amount of data in form of user reviews, messages, blogs and tweets results in. So there is need for sentimental analyses, which classify and categorizes these reviews in order to gain insights from this user generated data. This research aims to automate the process of gathering online, end user reviews for any given product or service and analysing those reviews in terms of the sentiments expressed about specific features. This includes pre-processing of inapplicable and unusual data, classifying that data into different classes and finally providing end user with the summarized data about the expressed sentiments. In this research author have designed an application, which analyses the data (mobile reviews of different users) collected from amazon and classify this data into positive, negative and neutral using two machine learning algorithms like naï ve base and K-nearest neighbour. The results show that the accuracy of KNN is better as compared to naï ve base.