def testFactorsAreCorrectlyUpdated(self):
pmf = Pmf(numTopics=3, lamb=0.5, gamma=0.25)
factors1 = np.array([4, 2, 3])
factors2 = np.array([5, 2, 1])
updatedFactors = pmf._updateFactors(factors1, factors2, 3)
self.assertListEqual([7.25, 3.25, 3.375], updatedFactors.tolist())
def testFactorsRemainsUnchangedIfGammaIsZero(self):
pmf = Pmf(numTopics=3, lamb=0.5, gamma=0)
factors1 = np.array([1, 2, 3])
factors2 = np.array([4, 5, 6])
updatedFactors = pmf._updateFactors(factors1, factors2, -5)
self.assertListEqual([1, 2, 3], updatedFactors.tolist())
def testFactorsAreCorrectlyUpdatedWhenLambdaIsZero(self):
pmf = Pmf(numTopics=3, lamb=0, gamma=0.5)
factors1 = np.array([1, 2, 3])
factors2 = np.array([2, 1, 2])
updatedFactors = pmf._updateFactors(factors1, factors2, 2)
self.assertListEqual([3, 3, 5], updatedFactors.tolist())
def MakeCdfFromList(seq, name=''):
"""Creates a CDF from an unsorted sequence.
Args:
seq: unsorted sequence of sortable values
name: string name for the cdf
Returns:
Cdf object
"""
hist = Pmf.MakeHistFromList(seq)
return MakeCdfFromHist(hist, name)
import numpy as np from pmf import Pmf rows = 15 cols = 15 x = np.random.random_integers(1, 5, size=(rows, cols)) y = np.random.uniform(0, 1, size=(rows, cols)) # sparsity z = y < 0.10 user_rating_mat = np.multiply(x, z) t = Pmf() t.x = user_rating_mat t.k = 10 t.pmf() print(t.P) print(t.x)
def testBiasIsCorrectlyUpdated(self):
pmf = Pmf(numTopics=1, lamb=0.25, gamma=0.5)
updatedBias = pmf._updateBias(8, 4)
self.assertEqual(9, updatedBias)
def testBiasIsCorrectlyUpdatedWhenLambdaIsZero(self):
pmf = Pmf(numTopics=1, lamb=0, gamma=0.5)
updatedBias = pmf._updateBias(10, 4)
self.assertEqual(12, updatedBias)
def testBiasRemainsUnchangedIfGammaIsZero(self):
pmf = Pmf(numTopics=1, lamb=0.5, gamma=0)
updatedBias = pmf._updateBias(12, -5)
self.assertEqual(12, updatedBias)
def setUp(self):
pmf = Pmf(numTopics=4, lamb=0.1, gamma=0.2)
self.model = pmf.train(ratings=None,
numRows=2,
numCols=3,
maxIterations=0)