Text categorization with text mining

After the invention and development of computers, we have been living in a more different and developing world. Computers make our life easier and improve the quality of our life by providing beter results with easier ways. Since computers make the same work automatically and effectively, human sourced errors become less. In the same way with the development of computers, human had the facility of access to too much data, and up to now, new systems, that is ?databases?, have been formed and these database systems have been getting bigger and bigger as the time passes. There are different kinds of databases. Text categorization methods are used in order to get the information from the databases which includes text type data in. With the increase of the number of documents, classification has been being made automatically, not by humans. For this purpose, with the help of the keywords of which categories are determined firstly, text type data can be classified. In that way, during the researching of this thesis, text categorization techniques which are used in text type data classification (Naive Bayes, K-NN) and various weightening methods are examined, then a text categorization programme has been done by using these techniques and VisualBasic.NET programming language, and at the same time exact classification probabilities of these techniques have been compared with each other. This thesis presents compilation of a Turkish dataset, called Anadolu Agency Newsgroup in order to study in Text Categorization. Key Words: Text categorization, naive bayes and k-nn algorithms, text mining, classification, wild card method.