Wine Review PT1 — ML
Introduction
Sommeliers are trained and knowledgeable wine professional. Master sommeliers can identify wines through blind tasting and it is an exceptional skill not everyone can do it.
This skill required a lot of experiences in tasting variety of wines and have knowledge or connection between grapes and wines as a wine is produced from certain grape.’
Machine can not taste wine. However, we can train machine to learn from descriptions that was given by sommeliers and then to identify wines.
Once machine is trained then we can give description of wine and tell machine to identify wines for us.
Dataset
We will use winemag-data-130k-v2.csv dataset for machine learning.
Goal
Our goal is to train a machine learning model that can take descriptions and then output name of wines that are related to descriptions.
Moreover, we can build a service on top of the model for other applications such as mobile app, web app.
Problem definition
Input description of taste for a wine into model, then output all related wine in their name.
Base on our dataset, we can categorize the problem as classification. Specifically, it is a multi-classification problem, as model is tasked to predict grape that used to make wine cross multiple type of grapes.
Dataset that was given include features and labels, thus we can train our model in supervised learning.
Our journey has began
