Wine variety prediction
The model used is a 3-Layer Neural Network: 1st Layer - 100 dimensional FC layer with ReLU activation 2nd Layer - 100 dimensional FC layer with ReLU activation 3rd Layer - 29 dimensional FC layer with Sigmoid activation Adam optimizer with categorical cross entropy was used
The features extracted were words from the reviews generated by the sklearn countTokenizer
Training Accuracy = 70%
Price of different Varieties of Wine
Popular Wines
Country
Distribution of Wine
Varieties of Wine across the world