Predicting poetry category using natural language processing methods

Natural language processing is a field of science that combines methods from computer science and linguistics to enable computers to understand, interpret and respond to human language (natural language). Natural language processing has found many applications to improve human-computer interaction and to make it easier to analyse and understand natural language data, especially for organisations and researchers working with large data sets. One such application is text classification. In this thesis, a study was conducted on the classification of poems in text format and written in different categories according to pre-labelled categories. The study used a dataset of 4198 poems obtained by web scraping. Thirteen different natural language processing steps were applied to the dataset. The Zemberek library was used to perform these operations. Six different machine learning algorithms were used for classification, the results obtained were evaluated and hyperparameter analysis was performed to improve model performance. The methods GridSearchCV and RandomizedSearchCV were used for the hyperparameter analysis. When the results of the classification algorithms were compared, it was found that the Random Forest and SVM algorithms gave the highest accuracy rate.