def plot_figures(som):
#BMU map
vhts = BmuHitsView(12, 12, "Hits Map", text_size=12)
#U matrix
u = sompy.umatrix.UMatrixView(50,
50,
'umatrix',
show_axis=True,
text_size=8,
show_text=True)
UMAT = u.build_u_matrix(sm, distance=1, row_normalized=False)
#Cluster map
sm.cluster(6)
hits = HitMapView(10, 10, "Clustering", text_size=12)
#Show factor influence
view2D = View2D(15, 15, "time-series", text_size=10, names=names)
#Show plots
view2D.show(sm, col_sz=4, which_dim="all", denormalize=True)
vhts.show(sm,
anotate=True,
onlyzeros=False,
labelsize=12,
cmap="Greys",
logaritmic=False)
UMAT = u.show(sm,
distance2=1,
row_normalized=False,
show_data=True,
contooor=True,
blob=False)
a = hits.show(sm)
plt.show()
- random
:param name: name used to identify the som
:param training: Training mode (seq, batch)
"""
from sompy.visualization.mapview import View2DPacked
view2D = View2DPacked(10,10,"", text_size=7)
view2D.show(sm, col_sz=5, what = 'codebook',)#which_dim="all", denormalize=True)
plt.show()
#These maps' scale is between -1 and 1 because we did the normalization
# Interpretation at the end should be the same
# SOM: less sensitive to outliers
from sompy.visualization.mapview import View2D
view2D = View2D(10,10,"", text_size=7)
view2D.show(sm, col_sz=5, what = 'codebook',)#which_dim="all", denormalize=True)
plt.show()
# Number of people in each neuron
from sompy.visualization.bmuhits import BmuHitsView
vhts = BmuHitsView(12,12,"Hits Map",text_size=7)
vhts.show(sm, anotate=True, onlyzeros=False, labelsize=10, cmap="autumn", logaritmic=False)
# Apply k- means over SOM
# K-Means Clustering
from sompy.visualization.hitmap import HitMapView
sm.cluster(3)# <n_clusters>
hits = HitMapView(10,10,"Clustering",text_size=7)
a=hits.show(sm, labelsize=12)
verbose='info',
train_rough_len=35,
train_finetune_len=80)
# 'final_clusters_consump' is a dataframe similar to df but including a column 'Labels' which indicates the closest neuron to each obs
final_clusters_consump = pd.DataFrame(sm_consump._data, columns=Consumption.columns).set_index(Consumption.index)
my_labels_c = pd.DataFrame(sm_consump._bmu[0], columns=['Labels']).set_index(Consumption.index)
final_clusters_consump = pd.concat([final_clusters_consump, my_labels_c], axis=1)
# Plot the number of observations associated to each neuron
vhts_c = BmuHitsView(12, 12, "Hits Map", text_size=7)
vhts_c.show(sm_consump, anotate=True, onlyzeros=False, labelsize=10, cmap="summer", logaritmic=False)
plt.show()
# Visualization of the value of the grid neurons in each variable
view2D_c = View2D(9, 9, "", text_size=7)
view2D_c.show(sm_consump, col_sz=5, what='codebook')
# 14.1.1. Hierarchical Clustering on top of SOM
# Create a dataframe that specifies the coordinates of the neurons
grid_consump = final_clusters_consump.groupby(by='Labels').mean()
labels_som_cons = grid_consump.index
# Plot a dendrogram to choose the number of clusters
Z = linkage(grid_consump, method='ward') # method='single', 'complete', 'ward'
dendrogram(Z, p=30)
# Aplly hierarchical clustering with k clusters defined and best method (ward)
k = 3
Hclustering = AgglomerativeClustering(n_clusters=k, affinity='euclidean', linkage='ward')
my_HC_consump = Hclustering.fit(grid_consump)
print(Z)
print('\n\n')
sm = SOMFactory().build(Z,
normalization='var',
initialization='random',
component_names=gunn)
sm.train(n_job=1, verbose=False, train_rough_len=2, train_finetune_len=5)
topographic_error = sm.calculate_topographic_error()
quantization_error = np.mean(sm._bmu[1])
print("Topographic error = %s; Quantization error = %s" %
(topographic_error, quantization_error))
vhts = BmuHitsView(10, 10, 'Hits Map', text_size=7)
vhts.show(sm,
anotate=True,
onlyzeros=False,
labelsize=12,
cmap='Greys',
logaritmic=False)
pl.show()
view2D = View2D(10, 10, 'rand data', text_size=10)
view2D.show(sm, col_sz=4, which_dim='all', denormalize=True)
pl.show()
sm = sompy.SOMFactory().build(colors, [20, 30],
normalization='var',
initialization='random',
neighborhood='gaussian',
component_names=['R', 'G', 'B'])
sm.train(n_job=4,
verbose=False,
train_len_factor=0.5,
train_rough_len=2,
train_finetune_len=100)
# In[113]:
from sompy.visualization.mapview import View2D
view2D = View2D(50, 50, "rand data", text_size=10)
view2D.show(sm, col_sz=4, which_dim="all", denormalize=True)
# In[114]:
codebook = sm._normalizer.denormalize_by(sm.data_raw, sm.codebook.matrix)
mp = codebook.reshape(sm.codebook.mapsize[0], sm.codebook.mapsize[1],
codebook.shape[1])
mp.shape
# Definiendo un SOM con una cuadrícula de 20x40 neuronas de salida y 400 iteraciones.
# In[115]:
import matplotlib.pyplot as plt
#som = SimpleSOMMapper((20, 30), 400, learning_rate=0.05)
sm = SOMFactory().build(data.values,
normalization='var',
initialization='pca',
component_names=names)
sm.train(n_job=1, verbose=False, train_rough_len=2, train_finetune_len=5)
# The quantization error: average distance between each data vector and its BMU.
# The topographic error: the proportion of all data vectors for which first and second BMUs are not adjacent units.
topographic_error = sm.calculate_topographic_error()
quantization_error = np.mean(sm._bmu[1])
print("Topographic error = %s; Quantization error = %s" %
(topographic_error, quantization_error))
# component planes view
from sompy.visualization.mapview import View2D
view2D = View2D(10, 10, "rand data", text_size=12)
view2D.show(sm, col_sz=4, which_dim="all", desnormalize=True)
# U-matrix plot
from sompy.visualization.umatrix import UMatrixView
umat = UMatrixView(width=10, height=10, title='U-matrix')
umat.show(sm)
# do the K-means clustering on the SOM grid, sweep across k = 2 to 20
from sompy.visualization.hitmap import HitMapView
K = 20 # stop at this k for SSE sweep
K_opt = 18 # optimal K already found
[labels, km, norm_data] = sm.cluster(K, K_opt)
hits = HitMapView(20, 20, "Clustering", text_size=12)
a = hits.show(sm)
mapSize = [20, 20]
sm = SOMFactory().build(training_split_0.values,
mapSize,
normalization="var",
lattice="rect",
initialization="random",
component_names=training_split_0.columns)
sm.train(
n_job=1, verbose=None, train_rough_len=2,
train_finetune_len=100) # I left some of the codes as the example provided
# plot the results, components map
from sompy.visualization.mapview import View2D
view2D = View2D(20, 20, "", text_size=12)
view2D.show(sm, col_sz=3, which_dim="all", denormalize=False)
# Hit maps
from sompy.visualization.bmuhits import BmuHitsView
vhts = BmuHitsView(15, 10, "Hits Map", text_size=12)
vhts.show(sm,
anotate=False,
onlyzeros=False,
labelsize=12,
cmap="jet",
logaritmic=False)
# U martix
from sompy.visualization.umatrix import UMatrixView
dif_hiper_features=list((dif_SA_hiper,dif_dyn_hiper,dif_fj_hiper,dif_granfuzzy_hiper)) features_dif_oclusao=prepara_features_p_som(dif_oclusao_features) features_dif_hiper=prepara_features_p_som(dif_hiper_features) ## SELF ORGANIZED MAP mapsize=[40,40] som_oclusao = sompy.SOMFactory.build(features_dif_oclusao, mapsize, mask=None, mapshape='planar', lattice='rect', initialization='pca', neighborhood='gaussian', training='batch') som_oclusao.train(n_job=1,train_rough_len=3, train_finetune_len=2,verbose='info') som_recov = sompy.SOMFactory.build(features_dif_hiper, mapsize, mask=None, mapshape='planar', lattice='rect', initialization='pca', neighborhood='gaussian', training='batch') som_recov.train(n_job=1,train_rough_len=3, train_finetune_len=2,verbose='info') sa_som_oclusao=extrai_plano_componente(som,0,40) sa_som_recov=extrai_plano_componente(som_recov,0,40) dyn_som_oclusao=extrai_plano_componente(som,1,40) dyn_som_recov=extrai_plano_componente(som_recov,1,40) from sompy.visualization.mapview import View2D,View2DPacked from matplotlib import pyplot as plt v1=View2D(100,100,"Data Map",text_size=14) v1.show(som,col_sz=4,desnormalize=True) v2=View2D(10,10,'SOM',text_size=10) v2.show(som_recov,what='codebook',which_dim='all',cmap='jet',col_sz=4,desnormalize=True)
##Investigating df.info() som = None # create the SOM network and train it. You can experiment with different normalizations and initializations som = SOMFactory().build(df.values, mapsize=[50,50], normalization = 'var', initialization='pca', component_names=names) som.train(n_job=1, verbose=False, train_rough_len=2, train_finetune_len=5) # The quantization error: average distance between each data vector and its BMU. # The topographic error: the proportion of all data vectors for which first and second BMUs are not adjacent units. topographic_error = som.calculate_topographic_error() quantization_error = np.mean(som._bmu[1]) print "Topographic error = %s; Quantization error = %s" % (topographic_error, quantization_error) from sompy.visualization.mapview import View2D view2D = View2D(4,4,"rand data",text_size=16) view2D.show(som, col_sz=2, which_dim="all", desnormalize=True) # U-matrix plot from sompy.visualization.umatrix import UMatrixView umat = UMatrixView(width=10,height=10,title='U-matrix') umat.show(som) from sompy.visualization.hitmap import HitMapView K=10 Kluster = som.cluster(K) hits = HitMapView(20,20,"K-Means Clustering",text_size=16) a=hits.show(som)
def draw_input_weights(self):
from sompy.visualization.mapview import View2D
view2D = View2D(10, 10, "rand data", text_size=10)
view2D.show(self.sm, col_sz=4, which_dim="all", desnormalize=True)
plt.show()
def extrai_plano_componente(objeto_som, indice_variavel, N_som):
plano = np.reshape(objeto_som.codebook.matrix[:, indice_variavel],
(N_som, N_som))
return plano
from sompy.visualization.hitmap import HitMapView
som.cluster(3)
hits = HitMapView(10, 10, "Clustering", text_size=12)
a = hits.show(som)
from sompy.visualization.mapview import View2D, View2DPacked
v = View2D(10, 10, 'SOM', text_size=10)
v.show(som,
what='codebook',
which_dim='all',
cmap='jet',
col_sz=4,
desnormalize=True)
from sompy.visualization.bmuhits import BmuHitsView
vhts = BmuHitsView(10, 10, "Hits Map", text_size=5)
vhts.show(som,
anotate=True,
onlyzeros=False,
labelsize=12,
cmap="Greys",
def prototype_visualization(sm):
view2D = View2D(4, 4, "", text_size=7)
view2D.show(sm, col_sz=5, which_dim="all", denormalize=True)
plt.show()
def som_component_planes(som):
# component planes view
view2D = View2D(15, 15, "rand data", text_size=12)
view2D.show(som, col_sz=4, which_dim="all", desnormalize=True)
plt.show()
