Eye Detection Classification

Overview

This project builds and compares multiple machine learning models to classify eye detection status (EYEDETECTION) using data stored in Snowflake. Models evaluated:

• K-Nearest Neighbors (KNN)
• Naive Bayes
• Logistic Regression

Data Source

• Data is fetched from a Snowflake database using snowflake-connector-python.
• Table queried: EYE (schema: PUBLIC, database: EYE)

Tech Stack

• Python
• Snowflake Connector
• Pandas, NumPy
• Scikit-learn
• Matplotlib

Workflow

1. Connect to Snowflake and load data into a Pandas DataFrame

2. Split features and target (EYEDETECTION)

3. Apply appropriate feature scaling:

   • Normalizer for KNN
   • MinMaxScaler for Naive Bayes
   • No scaling for Logistic Regression

4. Train and evaluate models using accuracy

5. Compare model performance using a bar chart

Models & Preprocessing

• KNN

  • Normalizer
  • n_neighbors = 5

• Naive Bayes

  • MinMax Scaling
  • GaussianNB

• Logistic Regression

  • Default configuration

Evaluation

• Metric used: Accuracy
• Train/Test split: 80/20
• Random state: 42

Output

• Accuracy score for each model
• Bar chart comparing model accuracies

How to Run

1. Install dependencies:

   pip install snowflake-connector-python pandas numpy scikit-learn matplotlib

2. Update Snowflake credentials in the script

3. Run the notebook or Python script