ms1034 / document-classification-using-knn

Documents classification using KNN Algorithm a graph based approach along with scrapped data

Graph-Based Document Classification using KNN's

Description

This project aims to implement a document classification system using graph theory principles. By representing documents as graphs and leveraging graph-based features, the system can categorize documents into predefined topics with improved accuracy compared to traditional vector-based models.

Table of Contents

• Installation
• Features
• Contributing
• License
• Credits
• Contact

Installation

To install and set up the project, follow these steps:

1. Clone the repository to your local machine.

Features

• Representation of documents as directed graphs.
• Extraction of graph-based features using common subgraph identification techniques.
• Classification of documents using the K-Nearest Neighbors (KNN) algorithm based on graph similarity measures.

Contributing

Contributions to the project are welcome! If you'd like to contribute:

1. Fork the repository.
2. Create a new branch for your feature or bug fix.
3. Make your changes and commit them with clear messages.
4. Push your changes to your fork.
5. Submit a pull request, clearly describing the changes implemented.

License

This project is licensed under the MIT License.

Credits

Sir Waqas Ali

Contact

For any inquiries or feedback, please contact:

• [email protected]

document-classification-using-knn's Issues

Evaluation Metrics

• Assess classification performance using accuracy, precision, recall, and F1-score.
• Plot confusion matrix to visualize classification results.

KNN Implementation

Implement the KNN algorithm.
Define distance measure based on maximal common subgraph.
Classify test documents based on k-nearest neighbors.

Report Preparation

Compile a comprehensive report detailing methodology, results, and reflections.
Discuss challenges encountered and potential improvements.

Graph Representation

Represent each document as a directed graph.
Define nodes and edges based on term relationships.

Data Collection

Collect or create text data for each assigned topic.

• Sports
• Food
• Business & Finance
  Note : Ensure each page contains approximately 300 words

Dataset Preparation

• Preprocess the data (tokenization, stop-word removal, stemming).
• Divide the dataset into training and test sets.

Subgraph Identification

Mine frequent subgraphs within the training set graphs.
Identify common subgraphs as features for classification.