cc

- Two datasets (one for red wines and one for white wines), 6497 total combined entries

- Relatively clean, no preprocessing required

- Output values (quality) are discrete and naturally ordered (e.g. 7 represents better quality than 5)

Expected Performance:

- Research done on the subject by the University of Minho: - Tested:

- Multiple regression(MR)

- Neural Networks(Multilayer Perceptron) - Support Vector Machines(SVM)

- All yelded very high accuracy, from 84.3% (MR) to 86.8% (SVM) with a tolerance T=1

- Based on the results of the research, we expect good accuracy from Linear Regression, while Softmax Regression, not taking into account the ordering of the classes, should perform worse

Metrics:

- Mainly accuracy and precision, measured at different levels of tolerance, from T=0.25 to T=1 - REC curve (accuracy for varying values of T)