The study aims to predict wine quality by applying machine learning techniques. We develop models that can highly accurately predict the quality ratings provided by experts by means of thorough evaluation of the physicochemical characteristics of wines.
Methodology:
- Data collecting included a dataset including wines with qualities including acidity, pH, amount of alcohol, and quality ratings.
- Cleaning the dataset, correcting in missing values, and feature normalisation help to prepare the data for model training.
- Resolve the most relevant features for wine quality prediction in feature
- Applying linear regression, decision tree, random forest, and gradient boosting among other regression techniques helps to build models.
- Model evaluation uses Root Mean Squared Error (RMSE) and Mean Absolute Error (MAE) to evaluate a model.
The result for wine quality, gradient boosting models produce most exact predictions. This work offers a structured approach to developing predictive models for evaluating the quality of wine and demonstrates the application of machine learning in the discipline of oenology.