class SilhouetteWindow(wx.Frame):
def __init__(self, pa, k, data, selected, p):
wx.Frame.__init__(self,
pa,
-1,
"Silhouette",
style=wx.DEFAULT_FRAME_STYLE ^ (wx.RESIZE_BORDER))
self.SetSize(600, 600)
self.pbutton = p
self.Maximize()
self.fig = None
self.Centre()
self.selceted = selected
self.k = k
self.db = []
for r in range(len(data)):
t = []
for c in range(len(data[r]) - 1):
t.append(data[r][c])
self.db.append(t)
self.fig, (self.ax1, self.ax2) = plt.subplots(1, 2)
self.pnl = wx.Panel(self,
style=wx.TAB_TRAVERSAL | wx.BORDER_SUNKEN,
size=(600, 600))
baseContainer = wx.BoxSizer(wx.HORIZONTAL)
lefSizer = wx.BoxSizer(wx.VERTICAL)
self.rightSizer = wx.BoxSizer(wx.VERTICAL)
infoLabel = wx.StaticText(self.pnl, label="Atributo")
self.opc1C = wx.ComboBox(self.pnl,
value=selected[0],
style=wx.CB_READONLY,
choices=selected)
self.opc2C = wx.ComboBox(self.pnl,
value=selected[1],
style=wx.CB_READONLY,
choices=selected)
applyButton = wx.Button(self.pnl, label="Aplicar")
applyButton.Bind(wx.EVT_BUTTON, self.update)
self.silhouette()
self.canvas = FigureCanvas(self.pnl, -1, self.fig)
lefSizer.Add(infoLabel, 0, wx.EXPAND | wx.ALL, 2)
lefSizer.Add(self.opc1C, 0, wx.EXPAND | wx.ALL, 2)
lefSizer.Add(self.opc2C, 0, wx.EXPAND | wx.ALL, 2)
lefSizer.Add(applyButton, 0, wx.EXPAND | wx.ALL, 2)
self.rightSizer.Add(self.canvas, 0, wx.EXPAND | wx.ALL, 2)
baseContainer.Add(lefSizer, 0, wx.EXPAND | wx.ALL, 0)
baseContainer.Add(self.rightSizer, 0, wx.EXPAND | wx.ALL, 0)
self.pnl.SetSizer(baseContainer)
self.Bind(wx.EVT_CLOSE, self.OnClose)
def OnClose(self, event):
self.pbutton.onCloseModule()
self.Hide()
def update(self, event):
o1 = self.opc1C.GetSelection()
o2 = self.opc1C.GetSelection()
self.canvas.ClearBackground()
self.rightSizer.Clear()
self.silhouette(o1, o2)
self.canvas.draw()
def silhouette(self, opc1=0, opc2=1):
db = np.array(self.db)
self.fig.clf()
self.fig.add_subplot(self.ax1)
self.fig.add_subplot(self.ax2)
self.fig.set_size_inches(13, 6)
n_clusters = len(self.k.clusters)
# The 1st subplot is the silhouette plot
# The silhouette coefficient can range from -1, 1 but in this example all
# lie within [-0.1, 1]
self.ax1.set_xlim([-0.1, 1])
# The (n_clusters+1)*10 is for inserting blank space between silhouette
# plots of individual clusters, to demarcate them clearly.
self.ax1.set_ylim([0, len(db) + (n_clusters + 1) * 10])
silhouette_avg = silhouette_score(db, self.k.labels)
# Compute the silhouette scores for each sample
sample_silhouette_values = silhouette_samples(db, self.k.labels)
y_lower = 10
for i in range(n_clusters):
# Aggregate the silhouette scores for samples belonging to
# cluster i, and sort them
ith_cluster_silhouette_values = \
sample_silhouette_values[self.k.labels == i]
ith_cluster_silhouette_values.sort()
size_cluster_i = ith_cluster_silhouette_values.shape[0]
y_upper = y_lower + size_cluster_i
color = cm.nipy_spectral(float(i) / n_clusters)
self.ax1.fill_betweenx(np.arange(y_lower, y_upper),
0,
ith_cluster_silhouette_values,
facecolor=color,
edgecolor=color,
alpha=0.7)
# Label the silhouette plots with their cluster numbers at the middle
self.ax1.text(-0.05, y_lower + 0.5 * size_cluster_i, str(i))
# Compute the new y_lower for next plot
y_lower = y_upper + 10 # 10 for the 0 samples
self.ax1.set_title("Visualización de silhouette.")
self.ax1.set_xlabel("Valor de silhouette")
self.ax1.set_ylabel("Etiqueta del cluster")
# The vertical line for average silhouette score of all the values
self.ax1.axvline(x=silhouette_avg, color="red", linestyle="--")
self.ax1.set_yticks([]) # Clear the yaxis labels / ticks
self.ax1.set_xticks([-0.1, 0, 0.2, 0.4, 0.6, 0.8, 1])
# 2nd Plot showing the actual clusters formed
dX = db[:, opc1]
dY = db[:, opc2]
aX = []
aY = []
for a in dY:
aY.append(float(a))
for a in dX:
aX.append(float(a))
colors = cm.nipy_spectral(self.k.labels.astype(float) / n_clusters)
self.ax2.scatter(aX,
aY,
marker='.',
s=30,
lw=0,
alpha=0.7,
c=colors,
edgecolor='k')
# Labeling the clusters
centers = self.k.clusters
#Draw white circles at cluster centers
self.ax2.scatter(centers[:, opc1],
centers[:, opc2],
marker='o',
c="white",
alpha=1,
s=200,
edgecolor='k')
for i, c in enumerate(centers):
self.ax2.scatter(c[opc1],
c[opc2],
marker='$%d$' % i,
alpha=1,
s=50,
edgecolor='k')
self.ax2.set_title("Visualización de los clusters")
self.ax2.set_xlabel("Primer Atributo")
self.ax2.set_ylabel("Segundo Atributo")
plt.suptitle(("Análisis de Silhouette para KMeans clustering "
"con %d clusters." % n_clusters),
fontsize=14,
fontweight='bold')
def ReDo(self, d, e):
pass
