cluster.py 文件源码

def calc_agg_coeff(self):
        """ Returns the agglomerative coefficient, measuring the clustering
        structure of the linkage matrix. Because it grows with the number of 
        observations, this measure should not be used to compare datasets of 
        very different sizes.

        For each observation i, denote by m(i) its dissimilarity to the first 
        cluster it is merged with, divided by the dissimilarity of the merger 
        in the final step of the algorithm. The agglomerative coefficient is 
        the average of all 1 - m(i).
        """

        # Turn into pandas DataFrame for fancy indexing
        Z = pd.DataFrame(self.linkage, columns = ['obs1','obs2','dist','n_org'] )

        # Get all distances at which an original observation is merged 
        all_dist = Z[ ( Z.obs1.isin(self.leafs) ) | (Z.obs2.isin(self.leafs) ) ].dist.values

        # Divide all distances by last merger
        all_dist /= self.linkage[-1][2]

        # Calc final coefficient
        coeff = np.mean( 1 - all_dist )

        return coeff