Sentiment analysis is an increasingly important part of data mining, especially in the age of social media and social networking where there is endless opinion and commentary that could be of use to a wide range of stakeholders in commerce, other businesses, and even politics.
Now, an innovative and efficient method of sentiment analysis of comments on the microblogging platform, Twitter, is reported in the International Journal of Data Mining, Modelling and Management by a team from India. Hima Suresh of the School of Computer Sciences, at Mahatma Gandhi University, in Kottayam, Kerala and Gladston Raj. S of the Department of Computer Science, Government College, also in Kerala explain how sentiment analysis centres on analysing attitudes and opinions revealed in a data set and pertaining to a particular topic of interest. The analysis exploits machine learning approaches, lexicon-based approaches and hybrid approaches that splice both of the former.
“An efficient approach for predicting sentiments would allow us to bring out opinions from the web contents and to predict online public choices,” the team suggests. They have now demonstrated a novel approach to sentiment analysis surrounding the discussion of a commercial brand on Twitter using data collected over a fourteen-month period. Their method has an unrivalled accuracy for gleaning the true opinion almost 87% of the time in their tests using a specific smart phone model as the target brand being studied. They suggest that accuracy could be improved still further by incorporating a wider lexicon that included Twitter slang, for instance.
Suresh, H. and Raj. S, G. (2019) ‘An innovative and efficient method for Twitter sentiment analysis‘, Int. J. Data Mining, Modelling and Management, Vol. 11, No. 1, pp.1-18.