Mushroom Classification

Developed a supervised machine learning application that classifies mushrooms as edible or poisonous through exploratory data analysis, feature preprocessing, model training, and web deployment.

Correctly distinguishing edible and poisonous mushrooms is challenging for non-experts and requires reliable classification based on observable characteristics.

Performed exploratory data analysis, feature preprocessing, and implemented a K-Nearest Neighbors (KNN) classifier before deploying the prediction model through a Flask-based web application.

• Conducted exploratory data analysis before model selection to better understand feature distributions.
• Selected KNN as a strong baseline due to the categorical nature of the dataset.
• Separated prediction logic from the user interface using Flask.

• Handling categorical feature encoding.
• Selecting appropriate hyperparameters for KNN.
• Designing a simple and intuitive prediction interface.

• Good data preprocessing significantly influences model performance.
• Model interpretability is valuable for educational machine learning projects.
• Deployment is an essential part of the machine learning lifecycle.

• Compare additional classification algorithms.
• Feature importance analysis.
• Image-based mushroom classification.
• Cloud deployment.