def test_set_alpha_with_array(self):
d = Dirichlet([0.1, 0.2, 0.3, 0.4])
expected = Series({'α1': 0.4, 'α2': 0.3, 'α3': 0.2, 'α4': 0.1})
d.alpha = [0.4, 0.3, 0.2, 0.1]
actual = d.alpha
self.assertTrue(expected.equals(actual))
def setUp(self) -> None:
self.prior_dirichlet = Dirichlet([1 + 70, 1 + 25, 1 + 5])
self.likelihood_dirichlet = Dirichlet([1 + 6, 1 + 3, 1 + 1])
self.likelihood_dirichlet_map = Series({
'$100':
Dirichlet([1 + 8, 1 + 1, 1 + 1]),
'$200':
Dirichlet([1 + 6, 1 + 3, 1 + 1])
})
self.prior_float_map = Series({'$100': 0.3, '$200': 0.2, '$300': 0.5})
self.likelihood_float_map = Series({
'$100': 0.1,
'$200': 0.2,
'$300': 0.3
})
def from_counts(
data: DataFrame,
prior_weight: float = 1.0
) -> 'MultipleBayesRule':
"""
Return a new BayesRule class using a DataFrame of counts.
:param data: DataFrame of counts where the index represents the
different states of the evidence B, and each column
represents the likelihood for one value of the prior A.
:param prior_weight: Proportion of the overall counts to use as the
prior probability.
"""
prior = Dirichlet(1 + data.sum() * prior_weight)
likelihood = Series({
key: Dirichlet(1 + row)
for key, row in data.iterrows()
})
return MultipleBayesRule(prior=prior, likelihood=likelihood)
def setUp(self) -> None:
self.b1 = Beta(700, 300)
self.b2 = Beta(600, 400)
self.b3 = Beta(500, 500)
self.d1 = Dirichlet([500, 300, 200])
self.d2 = Dirichlet({'x': 100, 'y': 200, 'z': 300})
self.b1__mul__b2 = self.b1 * self.b2
self.b3__mul__b1__mul__b2 = self.b3 * self.b1__mul__b2
self.b1__mul__comp__b1 = self.b1 * (1 - self.b1)
self.b_series = Series({'b1': self.b1, 'b2': self.b2, 'b3': self.b3})
self.b_frame = DataFrame({
'c1': {
'r1': self.b1,
'r2': self.b2
},
'c2': {
'r1': self.b2,
'r2': self.b3
}
})
self.float_series = Series({'$100': 0.8, '$200': 0.6})
def test_infer_posteriors(self):
expected = DataFrame(data=[(1, 1, 'e', Dirichlet({
1: 3,
2: 2,
3: 2
})), (2, 1, 'e', Dirichlet({
1: 2,
2: 3,
3: 2
})), (1, 2, 'e', Dirichlet({
1: 2,
2: 2,
3: 3
})), (2, 2, 'e', Dirichlet({
1: 3,
2: 2,
3: 2
})), (1, 1, 'f', Dirichlet({
2: 4,
3: 2
})), (2, 1, 'f', Dirichlet({
2: 3,
3: 3
})), (1, 2, 'f', Dirichlet({
2: 2,
3: 4
})), (2, 2, 'f', Dirichlet({
2: 4,
3: 2
}))],
columns=['a', 'b', 'prob_var', 'Dirichlet'])
actual = DirichletMultinomialConjugate.infer_posteriors(
data=self.multinomial_data,
prob_vars=['e', 'f'],
cond_vars=['a', 'b'])
for _, row in expected.iterrows():
actual_dirichlet = actual.loc[
(actual['a'] == row['a']) & (actual['b'] == row['b']) &
(actual['prob_var'] == row['prob_var']), 'Dirichlet'].iloc[0]
self.assertTrue(row['Dirichlet'] == actual_dirichlet)
def setUp(self) -> None:
self.alpha = Series({'a': 0.4, 'b': 0.3, 'c': 0.2, 'd': 0.1})
self.d_array = Dirichlet(alpha=self.alpha.values)
self.d_series = Dirichlet(alpha=self.alpha)
self.d_dict = Dirichlet(alpha=self.alpha.to_dict())
def plot_distribution():
ax = new_axes(width=10, height=10)
Dirichlet(alpha=[1, 2, 3]).plot(x=x, ax=ax)
plt.show()
def plot_pdf_simplex():
ax = new_axes(width=10, height=10)
Dirichlet(alpha=[1, 2, 3]).pdf().plot_simplex(ax=ax)
plt.show()
def test_infer_posterior(self):
expected = Dirichlet(alpha={'a': 5 + 1, 'b': 3 + 1, 'c': 2 + 1})
actual = DirichletMultinomialConjugate.infer_posterior(self.series)
self.assertEqual(expected, actual)