Wine Quality Analysis

This project was done as a course project in the course Statistical Inference and Modelling. It is a team project. Two datasets containing assessments of Portuguese Vinho Verde red and white wines were examined [1]. Wine quality ranging from 1 to 10 (from lowest quality to highest quality) was determined by an average of atleast three expert evaluators. Additionally, eleven physicochemical tests were performed on individual wines and the values are reported in the dataset. For each type of wine, we predict its quality by fitting several models (linear, polynomial, spline­based GAM, Random Forest, and KNN) as a function of these physicochemical characteristics. We developed models using the quality as a qualitative variable (classification) as well as quantitative variable (regression). Through our models, we identified important predictors and suggested top performing model for each wine class. Detailed report is present in this github repository.

Languages Used: R