lwazidladlaa/HonoursProject

IsiZulu Genre Classification Software ###HonoursProject ##This is an implemenation of IsiZulu Genre Classification model In this project we have used FastText and Word2Vec word embeddings In order to run the software you need the isiZulu dataset, Download the dataset from the SaDiLAR website. Here's the link to the dataset, repo.sadilar.org , but I have attached the folder together with the sofware. The pre-trained fastText word vectors for isiZulu are from this link fasttext.cc, but I have attached the folder together with the sofware. We have also implemented our own word embeddings : Word2vec vectors for isiZulu language Here's the link to the folder with the embedding models. stuukznac-my.sharepoint.com , but I have attached the folder together with the sofware. The setup for the FastText word Vectors Data pre-processing. Install FastText pip install Fasttext import all necessary libraries Numpy, Pandas, Matplot, for visualization of dataset TensorFlow Sklearn import the classifiers Nltk Load using KeyedVectors.load_format method from gensim word vectors Text cleaning download the english nltk stopwords from nltk import nltk ##ALSO NOTE THAT I HAVE MODIFIED THE DATASET DIRECTORY BY COMBINING THE TEST DATA WITH THE TRAINING DATA INTO ONE GENRE FOLDER WHEN YOU DOWNLOAD THE DATASET IT WILL HAVE SOMETHING LIKE THIS : C:\Users\lwazi\OneDrive\Documents\COMP700\zu\zu\zu\NF-NEU\Test SO I MODIFIED IT SO THAT I CAN DO THE SPLIT THE DATA TO TRAIN AND TEST USING train_test_split Split the data to train and test data with a test size of 0.2 Using train_test_split Download nltk from punkt from nltk Convert text data into vector representation using fasttext pre-trained model Cl …