Sentiment analysis of consumer tweets and reviews of Finnish retail companies

Abstract

The focus of this work has been on the defining of the approach through which sentiment analysis has been performed over the retail market of the Finnish. The reason behind the performing of the work is to understand the approaches that can be used for the sake of defining the root cause behind the reviews of customers. In terms of presenting the results, it can be noted that the four companies selected for the work are Alko Oy, FrennHelsinki, Kesko, and Lidl Suomi. To this end, three different sentiment analysis schemes have been defined in the work, including the implementation through deep learning, Two types of sentiment analysis: aspect-based and fine-grained. These techniques are included in order to get insight into potential approaches for the automation technique's development and to compare the outcomes in order to provide Sentiment analysis.

The accuracy of sentiment categorization is possible through this work also focuses on evaluating the model's performance. The work's outcomes show that the deep learning model's accuracy was 66.25%, the abstract-based sentiment analysis (ABSA) model's accuracy was 100%, and the fine-grained sentiment analysis model's accuracy was 83%. It suggests that ABSA is the best reported technique for the examination of review emotions. Furthermore, the path of the future might be described as using machine learning methods in conjunction with clustering techniques to increase the precision of conventional sentiment analysis schemes.