| Abstract: | Sentiment analysis (SA) is an ongoing research field in the field of text mining. SA is the calculation and processing of the opinions, emotions, and subjectivity of the text. The comments given by viewers of the program reflect whether the program is positive (positive increment) or negative (negative decrement) or neutral. SA can analyze a given text into predefined categories based on emotional terms that appear in self-righteous documents, such as positive, incremental positive, negative, reduced negative, or neutral. These opinions need to be explored, analyzed, and organized in order to make better decisions.
Early related researchers did not fully consider sentiment analysis in Tigrigna which is very important for identifying the polarity of emotions. They also did not consider the irony and ladder of expressions. And they only considered positive and negative polarity, but it is important to consider inverter words that change polarity. In this study, these gaps are attempted using NLP technology.
The sentiment analysis system uses rule-based and dictionary-based methods to resolve polarity. The questionnaire we used to do this study was to prepared and collect comments from Facebook and the website. Audience/non-audience comments were collected from website/Facebook pages, focus group discussions, and distribution of open-ended questioners.
The experiment uses 1633 (one thousand six hundred thirty-three) sentiment comments and four target research fields. The average accuracy, precession, recall, and f-score are 0.84, 0.94, 0.84, and 0.87, respectively. The experimental results using the comment viewer show the effectiveness of the system and the main limitation of this study was our inability to collect sufficient data. Hence, further research needs to be done to prepare a standardized data set that canusedable for experimentation and following the progress of the study. |
