def readQuinTreeRec(fle, domain):
node = orange.TreeNode()
line = fle.readline()
match = re_node.search(line)
if match:
splitattr, thresh = match.group("splitattr", "thresh")
node.branchSelector = orange.Classifier(lambda ex, rw, attr=domain[
splitattr], thr=float(thresh): ex[attr] > thr)
node.branchDescriptions = ["<= " + thresh, "> " + thresh]
node.branches = [
readQuinTreeRec(fle, domain),
readQuinTreeRec(fle, domain)
]
node.branchSizes = [
node.branches[0].nExamples, node.branches[1].nExamples
]
else:
match = re_constraint.search(line)
if not match:
raise "Cannot read line '%s'" % line
attrs, signs, cov, amb, xmpls = match.group("attrs", "signs", "cov", "amb",
"xmpls")
mqc = dict(zip(attrs.split(","), signs.split(",")))
node.nodeClassifier = orange.DefaultClassifier(defaultVal=int(
"".join([sidx[mqc.get(attr.name, None)]
for attr in domain.attributes]), 3))
node.setattr("coverage", cov == "**" and -1 or float(cov[:cov.index("%")]))
node.setattr("ambiguity", float(amb))
node.setattr("nExamples", float(xmpls))
return node
def __call__(self, example, result_type=orange.GetValue):
classifier = None
for r in self.rules:
# r.filter.domain = example.domain
if r(example) and r.classifier:
classifier = r.classifier
classifier.defaultDistribution = r.classDistribution
break
if not classifier:
classifier = orange.DefaultClassifier(example.domain.classVar,
self.prior.modus())
classifier.defaultDistribution = self.prior
if result_type == orange.GetValue:
return classifier(example)
if result_type == orange.GetProbabilities:
return classifier.defaultDistribution
return (classifier(example), classifier.defaultDistribution)
print
impdata = imputer(data)
for i in range(20, 25):
print data[i]
print impdata[i]
print
print "\n*** CUSTOM IMPUTATION BY MODELS ***\n"
imputer = orange.Imputer_model()
imputer.models = [None] * len(data.domain)
imputer.models[data.domain.index("LANES")] = orange.DefaultClassifier(2.0)
tord = orange.DefaultClassifier(orange.Value(data.domain["T-OR-D"], "THROUGH"))
imputer.models[data.domain.index("T-OR-D")] = tord
import orngTree
len_domain = orange.Domain(["MATERIAL", "SPAN", "ERECTED", "LENGTH"], data.domain)
len_data = orange.ExampleTable(len_domain, data)
len_tree = orngTree.TreeLearner(len_data, minSubset=20)
imputer.models[data.domain.index("LENGTH")] = len_tree
orngTree.printTxt(len_tree)
spanVar = data.domain["SPAN"]
def computeSpan(ex, rw):
if ex["TYPE"] == "WOOD" or ex["PURPOSE"] == "WALK":
return orange.Value(spanVar, "SHORT")