Prediction of customer’s perception in social networks by integrating sentiment analysis and machine learning

Abstract Understanding the customer behavior and perception are important issues for motivating satisfaction in marketing analysis. Customer conversation with support services through social networks channel provides a wealth of information understanding perception. Therefore, this paper, hybrid framework that integrated sentiment analysis machine learning techniques is developed to analyze interactive conversations among customers service providers order identify change polarity such conversation. This aims detect switch as well predict end provider. would help companies improve enhance engagement. The effectiveness proposed measured by extracting real dataset expresses more than 5000 conversational threads between agent an online retail provider (AmazonHelp) different using retailer’s twitter public account duration one month. Different classical ensemble classifiers were applied, results showed decision trees outperformed all other techniques.