Sentiment Analysis Model for Opinionated Awngi Text
Mihret Melese, Muluneh Atinaf · 2019
Sentiments are very imperative to decide for both individuals and organizations. Due to the rapid growth of Awngi text on the web, there is no sufficiently available sentiment corpus to be used for research. We developed our corpus by collecting around one thousand five hundred posts from online sources. Even infrequent texts available on the web mostly transliterated in Latin due to lack of accessibility and convenience to typing. This paper proposed a feature-level sentiment analysis method based on machine learning for opinionated Awngi music sentiments. Thus, pre-processing techniques have been employed to clean the data, to convert transliterations to the native Ethiopic script, and to change the words to their base form by removing the inflectional morphemes. To improve the calculation method of feature selection and weighting and proposed a more suitable sentiment analysis algorithm for feature extraction named CHI and weight calculation called TF IDF, increasing the proportion and weight of sentiment words in the feature words. The experiment results show that, among the two learning setups, the accuracy of the SVM is found to be promising.