Wine Quality Dataset

Wine Quality Dataset Description

This dataset includes both red and white wine samples, commonly found in such datasets. The attributes mentioned typically reflect chemical properties and quality measures. Below is a more comprehensive explanation of each column, assuming they follow the format of well-known wine quality datasets like the UCI Machine Learning Repository’s Wine Quality datasets:

1. fixed acidity: Measured in g/dm³. It mainly consists of tartaric acid, which affects the wine’s taste and stability.
2. volatile acidity: Measured in g/dm³. Primarily acetic acid, high levels can lead to an unpleasant vinegar taste. Acceptable levels vary with wine type.
3. citric acid: Measured in g/dm³. Adds freshness and flavor, acting as a preservative. Lower levels can make wine taste flat.
4. residual sugar: Measured in g/dm³. The amount of sugar remaining after fermentation; higher levels result in sweeter wines.
5. chlorides: Measured in g/dm³. Represents salt content, which can affect taste and preservation. High levels may indicate contamination.
6. free sulfur dioxide: Measured in mg/dm³. Free SO₂ helps prevent microbial growth and oxidation. It’s a key parameter for wine preservation.
7. total sulfur dioxide: Measured in mg/dm³. The total amount of SO₂ in both free and bound forms. Important for wine preservation but should be controlled to avoid negative sensory impacts.
8. density: Measured in g/cm³. Closely related to alcohol and sugar content. Typically ranges from 0.990 to 1.010 for wines.
9. pH: A scale from 0 to 14 indicating acidity/basicity. Most wines fall between 3 and 4. Affects taste, color, and stability.
10. sulphates: Measured in g/dm³. Contributes to SO₂ levels. Enhances antimicrobial and antioxidant properties.
11. alcohol: Percentage of alcohol by volume (%). Higher alcohol levels can enhance body, sweetness, and warmth of the wine.
12. quality: An integer score between 0 and 10. Represents the sensory quality of the wine as evaluated by experts. Often the target variable for predictive models.

Additionally, datasets might include the following metadata:

• wine_type: Categorical variable indicating whether the wine is red or white.
• sample_id: Unique identifier for each wine sample.

Example Row from the Dataset:

Usage:

• Exploratory Data Analysis (EDA): Understanding the distribution and relationship between variables.
• Predictive Modeling: Using machine learning algorithms to predict wine quality based on chemical properties.
• Quality Control: Identifying key chemical properties that influence wine quality for better quality control.

Assignment

Wine Quality dataset consists of various chemical properties of wine and a quality rating, making it suitable for predicting wine quality based on its chemical attributes. First 11 columns define physicochemical properties of wine and 12th column indicates the quality  of the wine. You have to develop (i) Multiclass classification algorithm and (ii) Regression algorithm to maintain and considering the following properties:

(i) Preprocessing technique
(ii) Feature selection technique
(iii) k-fold cross validation technique
(iv) Spot-check of Linear, Non-linear machine and ensemble learning techniques
(v) Parameter tunning of the selected algorithm(s)
(vi) Report the performance of the selected algorithm according to performance matrix