Miscellaneous (misc)¶
CostMatrix¶
CostMatrix is an object that stores costs of (mis)classifications. Costs can be either negative or positive.
- class Orange.misc.CostMatrix¶
- class_var¶
The (class) attribute to which the matrix applies. This can also be None.
- dimension(read only)¶
Matrix dimension, ie. number of classes.
- CostMatrix(dimension[, default cost])¶
Constructs a matrix of the given size and initializes it with the default cost (1, if not given). All elements of the matrix are assigned the given cost, except for the diagonal that have the default cost of 0. (Diagonal elements represent correct classifications and these usually have no price; you can, however, change this.)
import Orange cm = Orange.misc.CostMatrix(3) print "classVar =", cm.classVar for pred in range(3): for corr in range(3): print cm.getcost(pred, corr), print
This initializes the matrix and print it out:
0.0 1.0 1.0 1.0 0.0 1.0 1.0 1.0 0.0
- CostMatrix(class descriptor[, default cost])
Similar as above, except that classVar is also set to the given descriptor. The number of values of the given attribute (which must be discrete) is used for dimension.
data = Orange.data.Table("iris") cm = Orange.misc.CostMatrix(data.domain.classVar, 2)
This constructs a matrix similar to the one above (the class attribute in iris domain is three-valued) except that the matrix contains 2s instead of 1s.
- CostMatrix([attribute descriptor], matrix)
Initializes the matrix with the elements given as a sequence of sequences (you can mix lists and tuples if you find it funny). Each subsequence represents a row.
cm = Orange.misc.CostMatrix(data.domain.classVar, [(0, 2, 1), (2, 0, 1), (2, 2, 0)])
If you print this matrix out, will it look like this:
0.0 2.0 1.0 2.0 0.0 1.0 2.0 2.0 0.0
- setcost(predicted, correct, cost)¶
Set the misclassification cost. The matrix above could be constructed by first initializing it with 2s and then changing the prices for virginica’s into 1s.
cm = Orange.misc.CostMatrix(data.domain.classVar, 2) cm.setcost("Iris-setosa", "Iris-virginica", 1) cm.setcost("Iris-versicolor", "Iris-virginica", 1)
- getcost(predicted, correct)¶
Returns the cost of prediction. Values must be integer indices; if class_var is set, you can also use symbolic values (strings). Note that there’s no way to change the size of the matrix. Size is set at construction and does not change. For the final example, we shall compute the profits of knowing attribute values in the dataset lenses with the same cost-matrix as printed above.
data = Orange.data.Table("lenses") meas = Orange.feature.scoring.Cost() meas.cost = ((0, 2, 1), (2, 0, 1), (2, 2, 0)) for attr in data.domain.attributes: print "%s: %5.3f" % (attr.name, meas(attr, data))
As the script shows, you don’t have to (and usually won’t) call the constructor explicitly. Instead, you will set the corresponding field (in our case meas.cost) to a matrix and let Orange convert it to CostMatrix automatically. Funny as it might look, but since Orange uses constructor to perform such conversion, even the above statement is correct (although the cost matrix is rather dull, with 0s on the diagonal and 1s around):
meas.cost = data.domain.classVar
SymMatrix¶
SymMatrix implements symmetric matrices of size fixed at construction time (and stored in SymMatrix.dim).
- class Orange.misc.SymMatrix¶
- dim¶
Matrix dimension.
- matrix_type¶
Can be SymMatrix.Lower (0), SymMatrix.Upper (1), SymMatrix.Symmetric (2, default), SymMatrix.LowerFilled (3) or SymMatrix.Upper_Filled (4).
If the matrix type is Lower or Upper, indexing above or below the diagonal, respectively, will fail. With LowerFilled and Upper_Filled, the elements upper or lower, respectively, still exist and are set to zero, but they cannot be modified. The default matrix type is Symmetric, but can be changed at any time.
If matrix type is Upper, it is printed as:
>>> import Orange >>> m = Orange.misc.SymMatrix( ... [[1], ... [2, 4], ... [3, 6, 9], ... [4, 8, 12, 16]]) >>> m.matrix_type = m.Upper >>> print m (( 1.000, 2.000, 3.000, 4.000), ( 4.000, 6.000, 8.000), ( 9.000, 12.000), ( 16.000))
Changing the type to LowerFilled changes the printout to
>>> m.matrix_type = m.LowerFilled >>> print m (( 1.000, 0.000, 0.000, 0.000), ( 2.000, 4.000, 0.000, 0.000), ( 3.000, 6.000, 9.000, 0.000), ( 4.000, 8.000, 12.000, 16.000))
- __init__(dim[, value])¶
Construct a symmetric matrix of the given dimension.
Parameters: - dim (int) – matrix dimension
- value (double) – default value (0 by default)
- __init__(data)
Construct a new symmetric matrix containing the given data. These can be given as Python list containing lists or tuples.
The following example fills a matrix created above with data in a list:
import Orange m = [[], [ 3], [ 2, 4], [17, 5, 4], [ 2, 8, 3, 8], [ 7, 5, 10, 11, 2], [ 8, 4, 1, 5, 11, 13], [ 4, 7, 12, 8, 10, 1, 5], [13, 9, 14, 15, 7, 8, 4, 6], [12, 10, 11, 15, 2, 5, 7, 3, 1]] matrix = Orange.data.SymMatrix(m)
SymMatrix also stores diagonal elements. They are set to zero, if they are not specified. The missing elements (shorter lists) are set to zero as well. If a list spreads over the diagonal, the constructor checks for asymmetries. For instance, the matrix
m = [[], [ 3, 0, f], [ 2, 4]]
is only OK if f equals 2. Finally, no row can be longer than matrix size.
- get_values()¶
Return all matrix values in a Python list.
- get_KNN(i, k)¶
Return k columns with the lowest value in the i-th row.
Parameters:
- avg_linkage(clusters)¶
Return a symmetric matrix with average distances between given clusters.
Parameters: clusters (list of lists) – list of clusters
Indexing¶
For symmetric matrices the order of indices is not important: if m is a SymMatrix, then m[2, 4] addresses the same element as m[4, 2].
>>> import Orange
>>> m = Orange.misc.SymMatrix(4)
>>> for i in range(4):
... for j in range(i+1):
... m[i, j] = (i+1)*(j+1)
Although only the lower left half of the matrix was set explicitely, the whole matrix is constructed.
>>> print m
(( 1.000, 2.000, 3.000, 4.000),
( 2.000, 4.000, 6.000, 8.000),
( 3.000, 6.000, 9.000, 12.000),
( 4.000, 8.000, 12.000, 16.000))
Entire rows are indexed with a single index. They can be iterated over in a for loop or sliced (with, for example, m[:3]):
>>> print m[1]
(2.0, 4.0, 6.0, 8.0)
>>> m.matrix_type = m.Lower
>>> for row in m:
... print row
(1.0,)
(2.0, 4.0)
(3.0, 6.0, 9.0)
(4.0, 8.0, 12.0, 16.0)
Random number generator¶
Random uses the Mersenne twister algorithm to generate random numbers.
>>> import Orange
>>> rg = Orange.misc.Random(42)
>>> rg(10)
4
>>> rg(10)
7
>>> rg.uses # We called rg two times.
2
>>> rg.reset()
>>> rg(10)
4
>>> rg(10)
7
>>> rg.uses
2
- class Orange.misc.Random(seed)¶
Parameters: initseed (int) – Seed used for initializing the random generator. - __call__(n)¶
Return a random integer R such that 0 <= R < n.
- reset([seed])¶
Reinitialize the random generator with initseed. If initseed is not given use the existing value of attribute initseed.
- uses¶
The number of times the generator was called after initialization/reset.
- initseed¶
Random seed.
Two examples or random number generator uses found in the documentation are Orange.evaluation.testing and Orange.data.Table.