Project Workflow

1. Data Inspection & Cleaning:
   Thorough review of the entire dataset and removal of attributes that were clearly redundant or non-informative for the classification task.

2. Exploratory Data Analysis (EDA):
   Visualization of data distributions, analysis of outliers, and identification of anomalies to understand the nature of the features.

3. Clustering Analysis:
   Application of K-Means and Agglomerative Clustering algorithms to explore whether the classes (hazardous vs. non-hazardous) could be naturally separated based on specific features.

4. Model Training & Hyperparameter Tuning:
   Training of multiple machine learning models and conducting experiments with various hyperparameters to find the best configuration for maximum performance.

5. Evaluation & Results:
   Analysis of final metrics and comparison of model behaviors. The complete step-by-step process, visualizations, and comprehensive results are fully documented inside the Jupyter Notebook (main.ipynb).

Tech Stack

• Python: pandas kagglehub ipywidgets numpy matplotlib seaborn scikit-learn imblearn scipy