def predict(self, x_label, y_class):
pyx = []
likelihood = make_likelihood_table(x_label, y_class)
Y = np.unique(y_class)
X = np.unique(x_label)
for j in range(len(Y)):
total = 0
for i in range(len(X)):
if (likelihood[i][j] == 0):
continue
total += math.log(likelihood[i][j])
y_sum = (y_class == Y[j]).sum()
if y_sum:
total += math.log(y_sum / len(y_class))
pyx.append([total, X[i], Y[j]])
prediction = max(pyx)
return [prediction[1], prediction[2]]
def predict(self, x_label, y_class):
"""
Naive Bayes Model to predict the
class given the label.
PARAMETERS
==========
x_label: ndarray(dtype=int,ndim=1,axis=1)
Array of labels.
y_class: ndarray(dtype=int,ndim=1,axis=1)
Array of classes.
RETURNS
=======
Most probable output or prediction, as list
of the label and class name.
"""
pyx = []
likelihood = make_likelihood_table(x_label, y_class)
Y = np.unique(y_class)
X = np.unique(x_label)
for j in range(len(Y)):
total = 0
for i in range(len(X)):
if (likelihood[i][j] == 0):
continue
total += math.log(likelihood[i][j])
y_sum = (y_class == Y[j]).sum()
if y_sum:
total += math.log(y_sum / len(y_class))
pyx.append([total, X[i], Y[j]])
prediction = max(pyx)
return [prediction[1], prediction[2]]
