This project demonstrates the implementation and evaluation of custom classifiers alongside a Random Forest classifier for a thyroid dataset. The custom classifiers include a Bernoulli Naive Bayes classifier and a K-Nearest Neighbors classifier.

Dataset

The dataset used in this project is Thyroid_Diff.csv, which contains various features related to thyroid conditions. The target variable is whether the condition recurred (Recurred).

Requirements

To run the code in this project, you'll need:

• Python 3.x
• Required libraries: numpy, pandas, scikit-learn, matplotlib, seaborn

Installation

1. Clone this repository:

   git clone https://github.com/Sasi-Praneeth-Reddy/predicting-thyroid-cancer-reccurence.git

2. Install the required libraries using pip:

   pip install -r requirements.txt

Usage

1. Ensure you have the dataset Thyroid_Diff.csv placed in the data/ directory.

2. Run the Python script thyroid cancer.py located in the directory.

3. The script will generate various evaluation metrics and plots, including confusion matrices, ROC curves, accuracy comparison, and AUC comparison.

4. The predictions made by each classifier will be saved in the outputs/predictions.txt file.

Directory Structure

• data/: Contains the dataset file Thyroid_Diff.csv.
• models/: Contains the Python script thyroid cancer.py for implementing and evaluating custom classifiers.
• outputs/: Contains output files such as confusion matrices, ROC curves, and comparison plots.
• tests/: Directory for storing any test scripts or test data.

Results

• The outputs/ directory contains various evaluation plots and files, including confusion matrices, ROC curves, accuracy comparison, and AUC comparison.