ML-Fun

Overview

ML-Fun is a curated portfolio of machine learning projects organized by domain. Each project is contained in its own directory with its notebooks, code, data assets, and supporting documentation.

Project Index

Classification

• Fake Currency Detection: Deep learning image classification for real vs. fake currency using MobileNetV2 transfer learning, two-branch embeddings, fine-tuning, and attention (CBAM).
  • Project: classification/fake-currency-detection
  • Key notebooks: 01_Base_Model_Training.ipynb, 02_Two_Branch_Model_Training.ipynb, 03_feature-enhancements.ipynb, 04_attention-enhanced-model.ipynb
  • Models: Models

Natural Language Processing

• News Article Summarizer with Sentiment Analysis: Streamlit app that extracts article metadata, summarizes content, and classifies sentiment using standard NLP tooling.
  • Project: nlp/news-article-summarizer-sentiment-analysis
  • App entry point: app.py
  • Notebook: Cleaned_News_Article_Summarizer.ipynb
• SMS Spam Detection: TF-IDF based spam classifier with Streamlit UI and serialized model artifacts.
  • Project: nlp/sms-spam-detection
  • App entry point: app.py
  • Notebook: SMS_spam_detection.ipynb

Clustering and Segmentation

• Customer Segmentation using KMeans (RFM): Unsupervised clustering based on Recency, Frequency, and Monetary features with elbow and silhouette evaluation.
  • Project: clustering/customer-segmentation-kmeans
  • Notebook: customer-segmentation.ipynb
  • Dataset output: final_customer_segments_original_values.csv
• Image Compressor using KMeans: Flask web app that compresses images via KMeans color clustering and a reusable pipeline.
  • Project: clustering/image-compressor-kmeans
  • App entry point: app.py
  • Notebook: image_compressor.ipynb

Structure

ML-Fun/
  classification/
    fake-currency-detection/
  clustering/
    customer-segmentation-kmeans/
    image-compressor-kmeans/
  nlp/
    news-article-summarizer-sentiment-analysis/
    sms-spam-detection/