Sentiments Analysis for Afaan Oromoo Socio-Politics Text: Deep Learning Approach

Abstract

Sentiment analysis, also known as emotion analysis, has recently become one of the growing areas of research related to natural language processing and machine learning. It is a method to determine emotional feedback of people towards an event. Much sentiment about specific topics are available online; it allows several parties such as customers, companies and governments to explore these opinions. An opinion may be positive, negative or neutral depends on individual’s judgment or evaluation towards a topic. The analysis of natural language text for the identification of subjectivity and sentiment has been well studied in terms of the English language, Japanese and others. Conversely, the work that has been carried out in terms of Afaan Oromoo remains in its infancy; thus, more cooperation is required between research communities in order to offer a mature sentiment analysis system for Afaan Oromoo. This thesis addresses the problem of sentiments classification of Afaan Oromoo language reviews on Facebook social media and provides the rationale behind the proposed methods to enhance the performance of sentiment analysis in the Afaan Oromoo language. The first step is to increase the resources that help in the analysis process; the most important part of this task is to have annotated sentiment corpora. We also describes the work undertaken by the author to enrich sentiment analysis in Afaan Oromoo by building a new Afaan Oromoo Sentiment Corpus. The data is labeled not only with two polarities, but the neutral sentiment is also used during the annotation process. The second step includes apply preprocessing on textual data makes ready for learning features and classifying Afaan Oromoo text into different polarity. Generally the model is generated by a neural network variance called Convolutional Neural Network because it have shown great promise in the task of sentiment classification .The classifier model obtains a classification accuracy of 79.99%, which is encourage the researchers to continue in this direction of research.