def testReciprocalRankAtk(self):
m = 20
n = 50
r = 3
X, U, s, V, wv = SparseUtils.generateSparseBinaryMatrix((m, n),
r,
0.5,
verbose=True,
csarray=True)
k = 5
orderedItems = numpy.random.randint(0, n, m * k)
orderedItems = numpy.reshape(orderedItems, (m, k))
orderedItems = numpy.array(orderedItems, numpy.int32)
(indPtr, colInds) = X.nonzeroRowsPtr()
indPtr = numpy.array(indPtr, numpy.uint32)
colInds = numpy.array(colInds, numpy.uint32)
rrs = MCEvaluatorCython.reciprocalRankAtk(indPtr, colInds,
orderedItems)
rrs2 = numpy.zeros(m)
for i in range(m):
omegai = colInds[indPtr[i]:indPtr[i + 1]]
for j in range(k):
if orderedItems[i, j] in omegai:
rrs2[i] = 1 / float(1 + j)
break
nptst.assert_array_equal(rrs, rrs2)
#Test case where no items are in ranking
orderedItems = numpy.ones((m, k), numpy.int32) * (n + 1)
rrs = MCEvaluatorCython.reciprocalRankAtk(indPtr, colInds,
orderedItems)
nptst.assert_array_equal(rrs, numpy.zeros(m))
#Now, make all items rank 2
for i in range(m):
omegai = colInds[indPtr[i]:indPtr[i + 1]]
orderedItems[i, 1] = omegai[0]
rrs = MCEvaluatorCython.reciprocalRankAtk(indPtr, colInds,
orderedItems)
nptst.assert_array_equal(rrs, numpy.ones(m) * 0.5)
def testReciprocalRankAtk(self):
m = 20
n = 50
r = 3
X, U, s, V, wv = SparseUtils.generateSparseBinaryMatrix((m,n), r, 0.5, verbose=True, csarray=True)
k = 5
orderedItems = numpy.random.randint(0, n, m*k)
orderedItems = numpy.reshape(orderedItems, (m, k))
orderedItems = numpy.array(orderedItems, numpy.int32)
(indPtr, colInds) = X.nonzeroRowsPtr()
indPtr = numpy.array(indPtr, numpy.uint32)
colInds = numpy.array(colInds, numpy.uint32)
rrs = MCEvaluatorCython.reciprocalRankAtk(indPtr, colInds, orderedItems)
rrs2 = numpy.zeros(m)
for i in range(m):
omegai = colInds[indPtr[i]:indPtr[i+1]]
for j in range(k):
if orderedItems[i, j] in omegai:
rrs2[i] = 1/float(1+j)
break
nptst.assert_array_equal(rrs, rrs2)
#Test case where no items are in ranking
orderedItems = numpy.ones((m, k), numpy.int32) * (n+1)
rrs = MCEvaluatorCython.reciprocalRankAtk(indPtr, colInds, orderedItems)
nptst.assert_array_equal(rrs, numpy.zeros(m))
#Now, make all items rank 2
for i in range(m):
omegai = colInds[indPtr[i]:indPtr[i+1]]
orderedItems[i, 1] = omegai[0]
rrs = MCEvaluatorCython.reciprocalRankAtk(indPtr, colInds, orderedItems)
nptst.assert_array_equal(rrs, numpy.ones(m)*0.5)
def mrrAtK(positiveArray, orderedItems, k, verbose=False):
"""
Compute the mean reciprocal rank@k score for each row of the predicted matrix UV.T
using real values in positiveArray. positiveArray is a tuple (indPtr, colInds)
:param orderedItems: The ordered items for each user (users are rows, items are cols)
:param verbose: If true return mrr and first k recommendation for each row, otherwise just mrr
"""
if type(positiveArray) != tuple:
positiveArray = SparseUtils.getOmegaListPtr(positiveArray)
orderedItems = orderedItems[:, 0:k]
indPtr, colInds = positiveArray
mrr = MCEvaluatorCython.reciprocalRankAtk(indPtr, colInds, orderedItems)
if verbose:
return mrr, orderedItems
else:
return mrr.mean()
def mrrAtK(positiveArray, orderedItems, k, verbose=False):
"""
Compute the mean reciprocal rank@k score for each row of the predicted matrix UV.T
using real values in positiveArray. positiveArray is a tuple (indPtr, colInds)
:param orderedItems: The ordered items for each user (users are rows, items are cols)
:param verbose: If true return mrr and first k recommendation for each row, otherwise just mrr
"""
if type(positiveArray) != tuple:
positiveArray = SparseUtils.getOmegaListPtr(positiveArray)
orderedItems = orderedItems[:, 0:k]
indPtr, colInds = positiveArray
mrr = MCEvaluatorCython.reciprocalRankAtk(indPtr, colInds,
orderedItems)
if verbose:
return mrr, orderedItems
else:
return mrr.mean()
