Tensor factorizations for knowledge base construction

This is a project which evolves Turkish text classification of two different classes (Financial texts or Random Wikipedia texts) and their prediction results. As a classifier, Logistic regression is used. The intention of this project is to successfully identify the texts which contain financial news or data about companies from the other texts which are in random topics. The importance of the financial news documents can be understood if we look at the correlation between the stock price of a company, and the publicly published news about this company.