SpasMilenkov / Document-Classification-Algorithm

The repository contains both single-process and Open MPI-based parallel implementations of a document classification algorithm. Developed as part of the TU Sofia Parallel Processing of Information course, this project showcases the use of C++ and Open MPI to efficiently classify documents based on a predefined catalog of topics and key words.

Document-Classification-Algorithm

This repository contains implementations of a document classification algorithm using both single-process and MPI-based parallel processing techniques. This project is part of the TU Sofia Parallel Processing of Information course.

Table of Contents

• Description
• Technologies Used
• Installation
• Usage
• License

Description

The Document-Classification-Algorithm project demonstrates two different approaches to document classification:

1. Single-Process Implementation: A straightforward implementation that processes documents one by one.
2. Open MPI-Based Parallel Implementation: A parallelized version that uses MPI to distribute the workload among multiple processes, improving performance and scalability.

Both versions classify documents based on a catalog of topics and identifiers, reading from text files and determining the relevance of each document to different topics.

Single-Process Implementation

• Reads a catalog of topics and identifiers.
• Tokenizes and processes each document sequentially.
• Outputs the classification results to a CSV file.

MPI-Based Parallel Implementation

• Utilizes MPI for parallel processing.
• Distributes document classification tasks across multiple processes.
• Aggregates and outputs the results.

Technologies Used

• C++: The core programming language used for both implementations.
• Open MPI (Message Passing Interface): Used for parallel processing in the MPI-based implementation.
• Standard Template Library (STL): Utilized for data structures and algorithms.
• Filesystem Library: Used for directory and file handling.
• Chrono Library: Used for measuring execution time.

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/Document-Classification-Algorithm.git

2. Navigate to the project directory:

   cd Document-Classification-Algorithm

For Single-Process Implementation

3. Compile the single-process code:

   g++ ./main.cpp -o single_classification

For MPI-Based Parallel Implementation

4. Compile the MPI-based code:

     mpicxx -I/usr/lib/x86_64-linux-gnu/openmpi/include main.cpp -o parallel

Usage

Single-Process Implementation

1. Run the single-process classification:

   ./single_classification

MPI-Based Parallel Implementation

2. Run the MPI-based classification (example with 4 processes):

   mpirun -np 4 ./mpi_classification

   Where 4 is the amount of processes you want Open MPI to spawn.

License

This project is licensed under the MIT License - see the LICENSE file for details.

This project is part of the TU Sofia Parallel Processing of Information course.