Abstract:
The sentiment analysis is a cutting-edge technique for accessing internet data and these data
has been a growing discipline of the data mining and machine learning researchers and
academics for the last decades. Hence, sentiment analysis on employees Email data has not
been studied comprehensively. The main objective of the study to presents a method to email
sentiment analysis using an application that can spontaneously find out the conspiracy among
the employees by analysis their email records. In our study we used a popular TFIDF approach
to classify the conversion over email data. We evaluated the performance of a prominent
machine learning algorithm which is “Logistic Regression (LR)”. The performance of the
supervised-based techniques was examined with confusion matrix. In this experiment, our
model achieved the accuracy of 82.45% overall to classify the employee’s conversion in real
time. Our findings show that the Logistics Regression techniques outperformed to the detect
of email conversation of the employees. Therefore, our study has highlighted the research
studies and possibilities in the field of text data and sentiment study by machine learning
techniques.
