43 lines
2.5 KiB
Markdown
43 lines
2.5 KiB
Markdown
# Veracity_AI
|
|
|
|
## Overview
|
|
|
|
Veracity_AIn is designed to analyze articles and determine their authenticity using Natural Language Processing (NLP) techniques and machine learning models. The application consists of two main components: a user interface for inputting URLs or text, and a backend system that processes the data and makes predictions about the veracity of the article.
|
|
|
|
## User Interface
|
|
|
|
|
|
|
|
The user interface is built using CustomTkinter, a modern and customizable Python UI library. The main screen is divided into three sections:
|
|
|
|
1. **Input Section**: This section allows users to enter the URL of an article or paste the text directly into the input box.
|
|
2. **Result Section**: Displays the result of the analysis (Real or Fake) and the confidence level in percentage.
|
|
3. **Leaderboard Section**: Shows a list of news providers along with their fake news percentages, sorted by the highest fake news rate.
|
|
|
|
## Backend System
|
|
|
|
The backend system is responsible for processing user input, communicating with the database, and making predictions using the VeraMind model. Here's an overview of its components:
|
|
|
|
### Data Models
|
|
- **TextData**: Stores the URL, text content, provider, result, confidence, and a flag indicating if the news is fake.
|
|
- **Provider**: Represents a news provider with attributes for name, total articles, fake articles count, and a list of associated TextData objects.
|
|
|
|
### Database
|
|
The application uses a DuckDB database to store analyzed data. The `FakeNewsChecker` class manages database operations such as inserting new data and fetching existing data.
|
|
|
|
### Machine Learning Model
|
|
- **VeraMindInference**: An inference engine for the VeraMind model, which is used to predict whether an article is real or fake news based on its text content.
|
|
- **ArticleRater**: A class that uses the Large Language Model (LLM) to generate a response based on the analyzed text data.
|
|
|
|
## Usage
|
|
|
|
To use the Fake News Checker application, follow these steps:
|
|
|
|
1. Enter an article URL or paste the text directly into the input box.
|
|
2. Click on the "Check" button to initiate the analysis process.
|
|
3. Once the analysis is complete, the result (Real or Fake) and confidence level will be displayed in the result section.
|
|
4. The leaderboard section will automatically update with the latest news providers' fake news percentages.
|
|
|
|
## License
|
|
|
|
This application is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International. See the [LICENSE](LICENSE) file for more details. |