Wine Dataset with PCA and Linear Regression

Use of Wine dataset (red+white) using PCA and Logistic Regression in order to predict "quality".

Data Distribution of quality Feature after joining white and red wines

Variance distribution after PCA

After PCA we can choose to use the first 8 components only, which explain about 90% of the variance

Linear Regression

• Training set: 70%
• Test set: 30%

Conclusions

Not excellent results for quality prediction on this dataset, with a score of 50%.