Feature importances in binary classifier

gistfile1.txt

X_g = X[np.where(y==1)] 
X_b = X[np.where(y==0)]
M = X.shape[1]
ranges = []
for i in xrange(M):
    ranges.append((np.min(X[:,i]), np.max(X[:,i])))

importances = []
for i in xrange(M):
   g_dist = np.histogram(X_g[:,i],bins=50,density=True,range=ranges[i])[0]     
   b_dist = np.histogram(X_b[:,i],bins=50,density=True,range=ranges[i])[0]
   g_dist /= np.sum(g_dist)
   b_dist /= np.sum(b_dist)
   g_dist += 0.01
   b_dist += 0.01
   importances.append((i,scipy.stats.entropy(g_dist,b_dist)))
   
importances.sort(key=lambda x : x[1], reverse=True)