Skip to content

Latest commit

 

History

History
72 lines (36 loc) · 1.42 KB

README.md

File metadata and controls

72 lines (36 loc) · 1.42 KB

Welcome to Fake news

(Desktop client: https://github.com/ncrypt96/Fake-news-desktop)

This project aims to detect Fake news

We take the URL from the user-> Extract Keywords-> Crawl the web for related articles-> Compare the similarity with each article

Installation

git clone https://github.com/ncrypt96/Fake-news.git
cd Fake-news
pip install -r requirements.txt

Also

import nltk
nltk.download('punkt')
nltk.download('stopwords')
nltk.download('averaged_perceptron_tagger')
nltk.download('wordnet')

Usage

Go to https://newsapi.org/ and get the free API key under the Developer Plan

make a python file inside the cloned folder Fake-news

for example some_file.py

now run python3 some_file.py with the following code in it

Set a thershold value to the score returned for classification

the reccomended threshold is 0.4

from abstractor2 import with_keywords,without_keywords

print("Enter the URL")

url = input()

results = without_keywords(url,"<YOUR API KEY HERE>")

if(results["status"]=="fail"):

    print("Enter the keywords seperated by ',' \n")
    print("Suggested keywords:")
    print(results["suggestions"])

    keywords = input().split(',')

    print(keywords)


    print(with_keywords(url,keywords,"<YOUR API KEY HERE>")["similarity"])

else:

    print(results["similarity"])