woodminus/Kabyle-NLP-Plus

Improved natural language processing designed for the Kabyle language, incorporating new learning models and tokenization algorithms. This project is now maintained here, by woodminus # Kabyle-NLP-Plus Improved natural language processing designed for the Kabyle language, incorporating new learning models and tokenization algorithms. Main Algorithms: - The first will only generate the learning model, allowing it to be re-run to generate several models from several manually tagged corpora and save them on the disk. These models can then be shared online for solution needs. - The second algorithm will load a model to tag a new text. - Corpus correction and improvement. - Improvement of the integrated tokenization algorithm when running the tagging algorithm. - Update of the affix corpus for tokenization needs. When POS tagging: From the raw_text.txt file, the tagging algorithm will generate two files: A non-tagged segmented file and another tagged one. Contribution: Download the corpus, make your modifications, test and then send.