TNM108 - Machine Learning for Social Media

This repository is a part of the course TNM108 Machine Learning for Social Media at Linköping University 2024. It contains both labs and a project.

Project - Chatbot

This project is a simple conversational chatbot implemented in Python, which runs directly in the terminal. It allows users to engage in text-based conversations, using pre-processed conversational datasets for its responses.

Datasets

The chatbot is powered by multiple large conversational datasets, including:

• 3K Conversations Dataset for ChatBot
• Human Conversation training data
• Chatbot Dataset Topical Chat

How It Works

The chatbot uses TF-IDF (Term Frequency-Inverse Document Frequency) and cosine similarity to process user input and select the most relevant response from the pre-processed dataset. Here's an overview:

• Preprocessing: Datasets' are lemmatized and features are extracted using TF-IDF.
• Similarity Matching: User input is compared with the dataset using cosine similarity to find the closest matching response.
• Response: The chatbot returns the most relevant response based on the similarity score.

Installation

Clone the repository:

git clone https://github.com/rasmussvala/TNM108-Machine-Learning-for-Social-Media.git

Install required dependencies (bonus point if you do it in a virtual environment):

pip install -r requirements-chatbot.txt

Usage

Run the chatbot in the terminal:

py chatbot.py