Sentiment Analysis Perspective using Supervised Machine Learning Method

Customer feedback play an important role in assisting the organization. Sentiment analysis of the customer feedback takes the business to next level of profit making. Based on the sentiment or the opinion of the customer organization can identify the needs of the customer. Sentiment analysis is applicable not only for the customer feedback analysis, it can also be used employee review system. To create the sentiment analysis model supervised machine learning algorithm can be used extensively. Supervised learning algorithm is used for training the model for classification and regression problem. In this article various learning algorithm like linear regression, KNN, SVM, Random Forest, Bagging and Gradient Boosting are considered for developing and testing the model. The dataset was divided into training and test dataset. The model was developed using the training dataset and tested on the test dataset. The data was collected from 884 employees working in different organizations. The reliability control has shown that 1.1 % of respondents were unreliable, as some questions were left unattended. The final data of 802 employees was taken for the study. The different lexical resources like Afinn, VADER and sentiment from textblob were considered for determining the sentiment related to all the three dimensions of positive leadership those are strength, perspective and recognition. The result shows that random forest supervised machine learning algorithm gives the highest accuracy of 0.711 whereas the KNN algorithm gives the accuracy of 0.515.