Text mining in education: Dictionary-based in Turkish texts sentiment analysis

Information, whose adventure started with transferring texts, has reached enormous dimensions with the effect of developments in the field of technology. Therefore, nowadays, the ability to process and evaluate information has become much more important than the ability to access information. Being one of the various sources of information and frequently preferred as an educational tool, texts are fundamental materials that need to be discovered, examined and analyzed for specific purposes. The field of Text Mining in Education proves helpful in realising these objectives. This study aims to apply dictionary-based sentiment analysis, which is a method of text mining, on Turkish texts and determine dominant sentiment poles based on emotion words on its content. The material of this descriptive qualitative study consists of 165 short stories of Ömer Seyfettin. Analysis of data is performed by a dictionary-based sentiment analysis program which is developed on R, which is a statistical computation and graphics software, with means of programming dictionary-based sentiment analysis. In this two-staged research, the analysis is firstly applied on the text of 51 short stories to ensure a more balanced and consistent comparison. Afterwards, the short stories that were not included are also added, and the analysis is reapplied on a total of 165 short stories. In order to determine the effects of the fundamental components of the analysis, such as the size and word diversity of sentiment dictionaries, on the result of analysis two sentiment dictionaries are found; and these dictionaries are seperately used on two study groups. As a result of the dictionary-based sentiment analysis application, it is concluded that the polarity-based general sentiment state is positive on the texts of Ömer Seyfettin; and emotion-category-based general sentiment state on the same is excited. Accordingly, a judgement is reached that Ömer Seyfettin's works in short story genre constitute an important source of education and transfer, considering that they improve the permanancy, efficacy, and success of education and teaching. Keywords: Text mining, Sentiment analysis, Sentiment dictionary, Text analysis, R