Automatic Amharic News Text Classification with C# and Python

Machine learning approach for text classification is widely used and have shown a promising accuracy. This approach is also applied for Amharic text and have shown an acceptable result. my objective is to apply machine learning approach using naive Bayes to the Amharic news text classification problem, and demonstrate how the resulting model from this approach can be used in a real world environment.300 files with 45,195 words were used; divided into 100 per class category (Politics, Social and Economy/Business). The files were preprocessed and features were extracted out of them for training and testing the naive Bayes classifier. Finally the designed model is implemented using C# and Python.