Skip to content

Latest commit

 

History

History
19 lines (16 loc) · 921 Bytes

README.md

File metadata and controls

19 lines (16 loc) · 921 Bytes
Raw

Wine Quality

This is a project based on Suport Vector Machines (SVM)

First step

This code will try to predict the quality of a wine. Into the winequality-red.csv you can find wines each one with it's own specific characteristics. One of them is the quality, given by the taster. This file will be our training-test with analogy 75%-25% to mesure the performance of our model. After we "train" our code we will try to predict the quality of a wine using Suport Vector Machines (SVM) and we are going to mesure the outcome of the model using f1 score, precision and recall.

Second step

The second part of the code will first remove the 33% of the items in the pH column. After that, we have 4 different methods to handle the empty fields:

  1. Empty the whole column.
  2. Fill the columnt with the average of the remaining items.
  3. Fill the column using Logistic-Regression.
  4. Fill the column using K-means.