import numpy as np import matplotlib.pyplot as plt from sklearn.datasets import load_wine 

In [2]:

# In this assignment, we use the wine data set to apply PCA in a classification problem. # Check the page modules to read an example that uses PCA. X = load_wine() X.data.shape, X.target.shape 

Out[2]:

((178, 13), (178,))

In [3]:

# The data set contains data on three types of wines from the same Italian region. # The classes (i.e. the target) are wines 1, 2, and 3. # The features are 13 chemical properties that can be relatively easily measured. # Uncomment the following line to read more about the data set. # print(X.DESCR) 

In [4]:

# We extract the training set. X_train = X.data y_train = X.target + 1 X_train.shape, y_train.shape 

Out[4]:

((178, 13), (178,))

In [5]:

# question 1. # Apply PCA on the training set, and store the result in X_train_pca. X_train_pca.shape 

Out[5]:

(178, 13)

In [6]:

# question 2. # Plot the first two columns in X_train_pca, i.e. the x axis is column 0, and the y axis is column 1, # and show the three wines in different colors and symbols. # Use the function plt.plot(x_value, y_value, parameters, label) 3 times, # where parameters are 'yo', 'bx', and 'g+', for wines 1, 2, and 3 respectively. plt.legend(loc="upper left") 

Out[6]:

In [7]:

# We just bought 4 new bottles of Italian wine, however we do not know their wine type, # and we want to use data science to find the answer. # We measure all 13 chemical properties for each bootle, and obtain the following data. # [[ 14, 2, 2, 16, 127, 3, 3, 0, 2, 6, 1, 4, 1065] # [ 13, 2, 3, 19, 101, 3, 3, 0, 3, 6, 1, 3, 1185] # [ 14, 2, 2, 17, 113, 4, 3, 0, 2, 8, 1, 3, 1480] # [ 14, 2, 3, 20, 98, 2, 3, 0, 2, 6, 1, 3, 1180]] 

In [8]:

# question 3. # Using the PCA previously trained on X_train, transform the data of the 4 new wine bottles, and plot the results along with the training data set. # For the new wine, use the following call to the function plot. # plt.plot(x_value, y_value, 'rs', label='new wine') plt.legend(loc="upper left") 

Out[8]:

In [9]:

# question 4. # Answer in a comment. Using the previous plot, what is the predicted class of the four new wine bottles? 

In [ ]: