Code and Results for Chapter 5:

Introduction:
These are results and code for the problems and examples found in Chapter 5 of this famous book.

• Utility Functions for Plotting Data
  • plot_labeled.m (plots the samples with colored symbols depending on the class label)
    
  • plot_two_labels.m (plots the samples with symbols depending on one class label and colors depending on another)
    
• prob_5_1.m (clustering the iris data)
  
• prob_5_2_geyser.m (clustering the geyser data)
  
• prob_5_2_leukemia.m (clustering the leukemia data)
  
• prob_5_2_lungB.m (clustering the lungB data)
  
• prob_5_2_oronsay.m (clustering the oronsay data)
  
• prob_5_2_singer.m (clustering the singer data)
  
• prob_5_2_skulls.m (clustering the skulls data)
  
• prob_5_2_spam_direct.m (clustering the spam data)
  
• prob_5_2_sparrow.m (clustering the sparrow data)
  
• prob_5_7.m (the gap statistic on uniform data)
  
• prob_5_8.m (the gap statistic on well separated clusters)
  
• prob_5_9.m (the gap-PCA procedure)
  
• prob_5_11_geyser.m (mojenaplots of the geyser data)
  
• prob_5_11_spam.m (mojenaplots of the spam data)
  
• prob_5_12.m (value of the Rand index)
  
• prob_5_14.m (k-means v.s. agglomerative clustering)
  
• prob_5_14_various_k.m (more k-means v.s. agglomerative clustering)
  
• prob_5_17.m (the gap-PC method on the lungB)
  
• prob_5_18.m (an example with the Calinski and Harabasz's cluster index)
  
• cal_har_k_index.m (implements the Calinski and Harabasz cluster index for various clusterings)
  
• prob_5_19.m (an example with the Hartigan's cluster index)
  
• hartigan_k_index.m (Implements the Hartigan cluster index for various clusterings)
  
• prob_5_20.m (an example with the Fowlkes-Mallows index)
  
• fowlkes_mallows_index.m (implements the Fowlkes-Mallows cluster index)
  
• mk_matching_matrix_N.m (computes a "matching matrix" from two clusterings)
  
• gap_pca.m (Returns the estimate of the number of statistically significant clusters based on the pca gap statistic)
  
• gap_uniform.m (Returns the estimate of the number of statistically significant clusters based on the uniform gap statistic)