Skip to content

pmcelroy4/Multi-Class-Classification---Dermatology-Disease-Prediction

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

21 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Dermatology Disease Classification

Overview

This project classifies six different dermatological diseases using machine learning. The dataset was explored using data visualization and exploratory data analysis (EDA) to better understand feature distributions and relationships.

Two models were built and compared:

  • Decision Tree
  • Random Forest

Dataset

The dataset is located in the data folder:

  • dermatology_database_1.csv

It contains clinical and histopathological features used to predict six dermatological disease classes.


Methods

The following steps were performed:

  • Data cleaning and preprocessing
  • Exploratory Data Analysis (EDA)
  • Data visualization of features
  • Model training using Decision Tree and Random Forest
  • Model evaluation using classification reports
  • Visualization of feature importances

Model Evaluation

Models were evaluated using classification reports including:

  • Precision
  • Recall
  • F1-score
  • Accuracy

Feature Importance

Feature importance plots were generated to identify the most influential features, mainly from the Random Forest model.


Project Structure

.
├── data
│   └── dermatology_database_1.csv
├── models
│   └── derma_classification_final.ipynb
├── old
│   └── Derma_Classification.ipynb
└── README.md

Tools and Libraries

  • Python
  • Pandas
  • NumPy
  • Matplotlib / Seaborn
  • Scikit-learn
  • Jupyter Notebook

Notes

  • The models folder contains the final notebook.
  • The old folder contains an earlier version of the notebook kept for reference.

Author

Patrick McElroy


License

For educational and research purposes.

About

Utilizing various ensemble machine learning models to accurately predict dermatology based on family history and clinical samples

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors