Arabic-news-classification

Arabic news classification

Introduction

In this project we Collect Arabic Sentence and Tried to make a Arabic word Embedding, as a result I tried to build Arabic news classification.Arabic news classification is a webapp that detects news type in arabic language, trained in 4 different algorithms SVM, Naive Bayes, CNN and RNN. The Project was aimed to study the difference between them.

Install

• pip install -r requirements.txt

• python3 main.py

• in firefox go to : Local_host

• Download the pretrained models : pretrained_model

• also : you can download the dataset : dataset

start

• Sign in with the user name and password :

• user_name : mohammed@gmail.com

• Passport : 12345678

• then choose the Algorithm you want to classifie

• Upload the arabic news.

TODO

• make it more than two classes

colab Links :

• text_similarity
• svm_and_nb
• precision_recall_fscore
• LSTM Cell
• CNN