Reading OCDB data

The next code reads and treats OCDB data stored in an excel file with multiple sheets, as done in the Track 06, section "2.6. Gaussian Mixture on OCDB":

https://colab.research.google.com/drive/1XviDrKZ3RTBks8vEqqD7Fa6ED6XJ4wqF?usp=sharing

The code should be modified to separate data into three groups instead of four groups as done in the next code:

https://colab.research.google.com/drive/1bVAtSKvnkYOFoMp3ihcm8PbrvB4AnTaN?usp=sharing

Applying K-means to the data

Employing the previous notebook, the following code should be added to evaluate a K-means method for different values of  k centers.

import matplotlib.pyplot as plt

from sklearn.cluster import KMeans

# Using the elbow method to find the optimal number of clusters

wcss = []

for i in range(1, 11):

    kmeans = KMeans(n_clusters = i, init = 'k-means++', random_state = 42)

    kmeans.fit(x.reshape(-1, 1))

    wcss.append(kmeans.inertia_)

plt.plot(range(1, 11), wcss)

plt.title('The Elbow Method')

plt.xlabel('Number of clusters')

plt.ylabel('WCSS')

plt.show()

Drawing Data Labels from K-means

The next commands are useful to show the labeling made by the K-means.

list_x = [[] for d in range(n_clusters)]

list_y = [[] for d in range(n_clusters)]

k = 0

for elem in yhat:

  if (elem == 1):

    ind = 2

  elif (elem == 2):

    ind = 1

  else:

    ind = elem

  list_x[ind].append(x[k])

  list_y[ind].append(y[k])

  k = k+1

list_colors = ['red','orange','green']

#plt.hist(df_filtered_75p_lower['weight_kg'], bins=40)

for cluster,color in zip(range(n_clusters),list_colors):

  plt.scatter(list_x[cluster], list_y[cluster], color=color, label='N_1')

plt.show()

Comparing Gaussian Mixture and K-means labels

The next code compares data labels from Gaussian Mixture Method and K-means.

# Sequential numbering the classes given by GMM.

ygmm = target_class.copy()

for i in range(len(target_class)):

  value = ygmm[i]

  if (value == 2):

    elem = 0

  elif (value == 0):

    elem = 2

  else:

    elem = value

  ygmm[i] = elem

# Sequential numbering the classes given by K-Means.

ykme = yhat.copy()

for i in range(len(target_class)):

  value = ykme[i]

  if (value == 2):

    elem = 1

  elif (value == 1):

    elem = 2

  else:

    elem = value

  ykme[i] = elem

print('Classification using GMM: \n', ygmm)

print('Classification using Kmeans: \n', ykme)