Statistical language recognition

This work deals with the application that uses the machine-learning methods for the automatic language and encoding recognition. Various topics related to the project are analysed here one by one. Brief descriptionof the application should give the reader an idea about the features andthe usage of the program. The n-gram language model, the EM-smoothing algorithm and the entropy is explained in the chapter about the statistical methods. Some implementation matters are subject of analysis, for example the data file format and the inner word representation using the trie. The interpretation of a vast set of experiments evaluating the precision of the application is included as well as a short summary of the possible usage of the software