Unsupervised Learning - Kmeans Week 11 Challenge

1.

We can classify measures in several ways, based on: (i) the manner in which they fill the entries of the similarity matrix, (ii) whether more weight is a function of the frequency of the attribute values, (iii) the arguments used to propose the measure (probabilistic,information-theoretic, etc.). we will describe the measures by classifying them as follows:
• those that fill the diagonal entries only. These are measures that set the off-diagonal entries to 0 (mismatches are uniformly given the minimum value)and give possibly different weights to matches.
• those that fill the off-diagonal entries only. These measures set the diagonal entries to 1 (matches are uniformly given the maximum value) and give possibly different weights to mismatches.
• those that fill both diagonal and off-diagonal entries. These measures give different weights to both matches and mismatches.

2.

import os
import numpy as np
import pandas as pd
import numpy as np, pandas as pd
import matplotlib.pyplot as plt
from sklearn.cluster import KMeans

data = pd.read_csv('data/car_quality/car.data',names=['buying','maint','doors','persons','lug_boot','safety','class'])
data.head()

data.info()

data['class'],class_names = pd.factorize(data['class'])


print(class_names)
print(data['class'].unique())

data['buying'],_ = pd.factorize(data['buying'])
data['maint'],_ = pd.factorize(data['maint'])
data['doors'],_ = pd.factorize(data['doors'])
data['persons'],_ = pd.factorize(data['persons'])
data['lug_boot'],_ = pd.factorize(data['lug_boot'])
data['safety'],_ = pd.factorize(data['safety'])
data.head()

X = data
y = data.iloc[:,-1]


model = KMeans(n_clusters=4,random_state=123)
model.fit(X)

labels = model.labels_
# check how many of the samples were correctly labeled
correct_labels = sum(y == labels)
print("Result: %d out of %d samples were correctly labeled." % (correct_labels, y.size))


from mpl_toolkits.mplot3d import Axes3D

def elbow_plot(data, maxK=40, seed_centroids=None):
    """
        parameters:
        - data: pandas DataFrame (data to be fitted)
        - maxK (default = 10): integer (maximum number of clusters with which to run k-means)
        - seed_centroids (default = None ): float (initial value of centroids for k-means)
    """
    sse = {}
    for k in range(1, maxK):
        print("k: ", k)
        if seed_centroids is not None:
            seeds = seed_centroids.head(k)
            kmeans = KMeans(n_clusters=k, max_iter=500, n_init=100, random_state=0, init=np.reshape(seeds, (k,1))).fit(data)
            data["clusters"] = kmeans.labels_
        else:
            kmeans = KMeans(n_clusters=k, max_iter=300, n_init=100, random_state=0).fit(data)
            data["clusters"] = kmeans.labels_
        # Inertia: Sum of distances of samples to their closest cluster center

sse[k] = kmeans.inertia_
    plt.figure()
    plt.plot(list(sse.keys()), list(sse.values()))
    plt.show()
    return

elbow_plot(X, maxK=10)


model = KMeans(n_clusters=2,random_state=123)
model.fit(X)
labels = model.labels_
# check how many of the samples were correctly labeled
correct_labels = sum(y == labels)

print("Result: %d out of %d samples were correctly labeled." % (correct_labels, y.size))
print("correct %.02f classification " % (correct_labels/float(y.size)))