https://github.com/fatchur/news-censorship-engine
This is an engine for censoring news in bahasa Indonesia with machine learning. Up to now, the validation accuracy is 90.4% for 1000 samples
https://github.com/fatchur/news-censorship-engine
bahasa-indonesia neural-network news-censorship nlp opinion-mining python tensorflow
Last synced: about 1 month ago
JSON representation
This is an engine for censoring news in bahasa Indonesia with machine learning. Up to now, the validation accuracy is 90.4% for 1000 samples
- Host: GitHub
- URL: https://github.com/fatchur/news-censorship-engine
- Owner: fatchur
- Created: 2017-05-18T13:22:02.000Z (about 9 years ago)
- Default Branch: master
- Last Pushed: 2017-05-18T13:40:09.000Z (about 9 years ago)
- Last Synced: 2025-08-01T14:51:54.048Z (11 months ago)
- Topics: bahasa-indonesia, neural-network, news-censorship, nlp, opinion-mining, python, tensorflow
- Language: Jupyter Notebook
- Homepage:
- Size: 44.6 MB
- Stars: 0
- Watchers: 1
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: readme.txt
Awesome Lists containing this project
README
Requirements:
1. install tensorflow
2. install flask
3. python3 and other imported modules (numpy, scikitlearn, scipy etc)
-- file train.ipynb
contains our training method and steps.
Make sure you have installed all imported modules.
During training process, our model reached 90.4% of accuracy
--file engine.py
article category predictor based on model developed from train.ipynb file
--file classifier.html
an html file for submitting a will predicted article
--the model folder
a folder containing neural net biases and weight
---- how to run ----
1. run engine.py (python3 engine.py)
2. open classifier.html in web browser, paste your article here, and press "klasifikasi" button.
3. the result will appear (sensitive or nonsensitive)