Predicting the rating of scientific papers

This bachelor thesis deals with the prediction of scientific publication ratings using machine learning methods. The main aim of this bachelor thesis is to find out what success can be achieved in predicting the peer review ratings of scientific publications. A search of studies dealing with the prediction of scientific publication reviews is conducted to achieve the main objective. Software tools and libraries for applying machine learning methods are described. Data collection and description of datasets are also described. Evaluation metrics for assessing model performance are also described, as well as machine learning methods. In the experimental part, traditional machine learning methods such as logistic regression, linear regression, decision trees, the AutoGluon-Tabular interface, and the SciBERT language model are applied. Then, the results of the models from the experimental part are compared.