https://github.com/gangula-karthik/bank-transaction-classification

Classifying bank transactions with precision—your first step towards smarter finance management 💳🤖📊
https://github.com/gangula-karthik/bank-transaction-classification

finance machine-learning nlp scikit-learn

Last synced: 1 day ago
JSON representation

Classifying bank transactions with precision—your first step towards smarter finance management 💳🤖📊

• Host: GitHub
• URL: https://github.com/gangula-karthik/bank-transaction-classification
• Owner: gangula-karthik
• Created: 2022-08-04T14:15:07.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2024-02-19T07:42:29.000Z (10 months ago)
• Last Synced: 2024-11-05T11:37:26.131Z (about 2 months ago)
• Topics: finance, machine-learning, nlp, scikit-learn
• Language: HTML
• Homepage:
• Size: 5.39 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Bank-Transaction-Classification 💳🔍

Welcome to my project on bank transaction classification! This project aims to classify bank purchases into different categories such as entertainment, food and drinks, taxes, and more. 📊

## Overview ℹ️

This project was my very first venture into machine learning, and let's just say, it wasn't my finest hour. 😅 However, I've decided to give it another go 2 years later, revamping the code to hopefully yield better performance. 💻

## Approach 🛠️

Utilizing the transaction description column, I employed text classification techniques to categorize bank purchases. 📝 This involved leveraging word embeddings generated through word2vec and deploying various machine learning models for accurate predictions. 🤖 Due to the multitude of classes, I had to group some together, but I'm eyeing a multi-label text classification approach for future improvements. 🌟

## Performance 📈

The overall performance, particularly on the best baseline model (SVM), showed promising results. On stratified k-fold cross-validation, the model's performance was decent given the data that was provided.

Feel free to explore the code and contribute any insights or enhancements! Happy classifying! 🎉