Wine Quality Predictor

ML Project

Machine Learning | Chemical Analysis to Quality Score

Overview

Built a machine learning model to predict wine quality scores (3-9) based on physicochemical properties. Used the UCI Wine Quality dataset with 4,898 samples of Portuguese wines.

ML Pipeline

Raw Data ──► Clean & Normalize ──► Feature Engineer ──► Train Model │ │ └──────────────────────────────────────────────────────┘ ▼ Deploy (API) ◄── Evaluate (Metrics) ◄── Tune (GridSearchCV)

Model Performance

87%

Accuracy

0.84

F1 Score

11

Features

4.9K

Samples

Key Features

Alcohol

Strongest predictor

Volatile Acidity

Negative correlation

Sulphates

Preservative levels

Citric Acid

Freshness indicator

Technologies

Pythonscikit-learnPandasNumPy Random ForestXGBoostMatplotlibSeaborn

← Back to Portfolio