Project Descriptions

This repository contains machine learning and data preprocessing projects created for different purposes, including internships and hackathons. Below is a detailed overview of each project:

1. Breast Cancer Prediction

Purpose: Developed for an internship at ADVERK Technologies, this notebook focuses on predicting the presence of breast cancer using Naive Bayes Classifier.

Key Features:

• Libraries Used:
  • numpy, pandas, seaborn, matplotlib.pyplot
  • sklearn modules: preprocessing, model_selection, metrics, naive_bayes
• Workflow:
  1. Load breast cancer data using datasets.
  2. Preprocess numerical features.
  3. Split the dataset into training and testing subsets.
  4. Apply RobustScaler for scaling.
  5. Train and evaluate a Gaussian Naive Bayes classifier.

---

2. Product Confirmer ML

Purpose: Developed for the Indian Government's "Dark Patterns Buster Hackathon," this notebook predicts the usability or validation of a product based on various attributes.

Key Features:

• Libraries Used:
  • pandas, category_encoders, joblib, graphviz, pickle
  • sklearn modules: ensemble, metrics, model_selection, tree
• Workflow:
  1. Load and preprocess product data, filling missing values with appropriate defaults.
  2. Encode categorical features using OneHotEncoder.
  3. Train a Random Forest Classifier.
  4. Test the model and evaluate its accuracy.
  5. Save the trained model using pickle and joblib for reuse.

---

3. Stroke Prediction

Purpose: Developed for an internship at ADVERK Technologies, this project aims to predict the likelihood of a stroke based on patient data using Decision Tree.

Key Features:

• Libraries Used:
  • pandas, numpy, category_encoders, graphviz
  • sklearn modules: model_selection, metrics, tree
• Workflow:
  1. Load healthcare dataset.
  2. Handle missing values and preprocess categorical data using OrdinalEncoder.
  3. Train a Decision Tree Classifier with Gini Impurity.
  4. Evaluate the model and visualize decision trees using graphviz.

---

Repository Structure

• Breast cancer.ipynb: Notebook for breast cancer prediction.
• ProductconfirmerML.ipynb: Notebook for product confirmation model.
• Strokefinding.ipynb: Notebook for stroke prediction.

---

Requirements

Install the necessary libraries using the following command: For Windows Command:

pip install numpy pandas matplotlib seaborn scikit-learn category_encoders graphviz joblib

For Jupyter Notebook:

!pip install numpy pandas matplotlib seaborn scikit-learn category_encoders graphviz joblib

---

How to Use

1. Clone this repository:

   git clone <repository-url>

2. Open the desired notebook using Jupyter Notebook or JupyterLab:

   jupyter notebook <notebook-name>.ipynb

3. Follow the workflow described within each notebook.

---

Acknowledgments

• ADVERK Technologies: For providing the opportunity to work on real-world machine learning problems.
• Indian Government: For organizing the "Dark Patterns Buster Hackathon," inspiring innovative solutions like Product Confirmer ML.

---

Feel free to explore and modify the projects as needed. Contributions are welcome!