def testCARTPrune(self):
numExamples = 500
X, y = data.make_regression(numExamples)
y = Standardiser().standardiseArray(y)
numTrain = numpy.round(numExamples * 0.33)
numValid = numpy.round(numExamples * 0.33)
trainX = X[0:numTrain, :]
trainY = y[0:numTrain]
validX = X[numTrain:numTrain+numValid, :]
validY = y[numTrain:numTrain+numValid]
testX = X[numTrain+numValid:, :]
testY = y[numTrain+numValid:]
learner = DecisionTreeLearner(pruneType="none", maxDepth=10, minSplit=2)
learner.learnModel(trainX, trainY)
learner = DecisionTreeLearner(pruneType="CART", maxDepth=10, minSplit=2, gamma=1000)
learner.learnModel(trainX, trainY)
self.assertTrue(learner.tree.getNumVertices() <= 1000)
predY = learner.predict(trainX)
learner.setGamma(200)
learner.learnModel(trainX, trainY)
self.assertTrue(learner.tree.getNumVertices() <= 200)
learner.setGamma(100)
learner.learnModel(trainX, trainY)
self.assertTrue(learner.tree.getNumVertices() <= 100)
learner = DecisionTreeLearner(pruneType="none", maxDepth=10, minSplit=2)
learner.learnModel(trainX, trainY)
predY2 = learner.predict(trainX)
#Gamma = 0 implies no pruning
nptst.assert_array_equal(predY, predY2)
#Full pruning
learner = DecisionTreeLearner(pruneType="CART", maxDepth=3, gamma=1)
learner.learnModel(trainX, trainY)
self.assertEquals(learner.tree.getNumVertices(), 1)
def testCvPrune(self):
numExamples = 500
X, y = data.make_regression(numExamples)
y = Standardiser().standardiseArray(y)
numTrain = numpy.round(numExamples * 0.33)
numValid = numpy.round(numExamples * 0.33)
trainX = X[0:numTrain, :]
trainY = y[0:numTrain]
validX = X[numTrain:numTrain+numValid, :]
validY = y[numTrain:numTrain+numValid]
testX = X[numTrain+numValid:, :]
testY = y[numTrain+numValid:]
learner = DecisionTreeLearner()
learner.learnModel(trainX, trainY)
error1 = Evaluator.rootMeanSqError(learner.predict(testX), testY)
#print(learner.getTree())
unprunedTree = learner.tree.copy()
learner.setGamma(1000)
learner.cvPrune(trainX, trainY)
self.assertEquals(unprunedTree.getNumVertices(), learner.tree.getNumVertices())
learner.setGamma(100)
learner.cvPrune(trainX, trainY)
#Test if pruned tree is subtree of current:
for vertexId in learner.tree.getAllVertexIds():
self.assertTrue(vertexId in unprunedTree.getAllVertexIds())
#The error should be better after pruning
learner.learnModel(trainX, trainY)
#learner.cvPrune(validX, validY, 0.0, 5)
learner.repPrune(validX, validY)
error2 = Evaluator.rootMeanSqError(learner.predict(testX), testY)
self.assertTrue(error1 >= error2)
def testPredict(self):
generator = ExamplesGenerator()
for i in range(10):
numExamples = numpy.random.randint(1, 200)
numFeatures = numpy.random.randint(1, 20)
minSplit = numpy.random.randint(1, 50)
maxDepth = numpy.random.randint(0, 10)
X, y = generator.generateBinaryExamples(numExamples, numFeatures)
y = numpy.array(y, numpy.float)
learner = DecisionTreeLearner(minSplit=minSplit, maxDepth=maxDepth)
learner.learnModel(X, y)
predY = learner.predict(X)
tree = learner.tree
for vertexId in tree.getAllVertexIds():
nptst.assert_array_equal(tree.getVertex(vertexId).getTrainInds(), tree.getVertex(vertexId).getTestInds())
#Compare against sklearn tree
regressor = DecisionTreeRegressor(min_samples_split=minSplit, max_depth=maxDepth, min_density=0.0)
regressor.fit(X, y)
sktree = regressor.tree_
#Note that the sklearn algorithm appears to combine nodes with same value
#self.assertEquals(sktree.node_count, tree.getNumVertices())
self.assertEquals(sktree.feature[0], tree.getRoot().getFeatureInd())
self.assertEquals(sktree.value[0], tree.getRoot().getValue())
self.assertAlmostEquals(sktree.threshold[0], tree.getRoot().getThreshold(), 3)
predY2 = regressor.predict(X)
#Note that this is not always precise because if two thresholds give the same error we choose the largest
#and not sure how it is chosen in sklearn (or if the code is correct)
self.assertTrue(abs(numpy.linalg.norm(predY-y)- numpy.linalg.norm(predY2-y))/numExamples < 0.05)
meanPenalties[k] += currentPenalties
meanTrainError += trainErrors
predY = bestLearner.predict(testX)
meanErrors[k] += bestLearner.getMetricMethod()(testY, predY)
#Compute ideal penalties and error on training data
meanIdealPenalities[k] += learner.parallelPenaltyGrid(trainX, trainY, testX, testY, paramDict)
for i in range(len(paramDict["setGamma"])):
allError = 0
learner.setGamma(paramDict["setGamma"][i])
for trainInds, testInds in idx:
validX = trainX[trainInds, :]
validY = trainY[trainInds]
learner.learnModel(validX, validY)
predY = learner.predict(trainX)
allError += learner.getMetricMethod()(predY, trainY)
meanAllErrors[i] += allError/float(len(idx))
k+= 1
numRealisations = float(numRealisations)
meanErrors /= numRealisations
meanPenalties /= numRealisations
meanIdealPenalities /= numRealisations
print(meanErrors)
plt.plot(sampleSizes, meanPenalties*numpy.sqrt(sampleSizes), label="Penalty")
plt.plot(sampleSizes, meanIdealPenalities*numpy.sqrt(sampleSizes), label="Ideal penalty")
validX = X[numTrainExamples:numTrainExamples+numValidExamples, :]
validY = y[numTrainExamples:numTrainExamples+numValidExamples]
testX = X[numTrainExamples+numValidExamples:, :]
testY = y[numTrainExamples+numValidExamples:]
learner = DecisionTreeLearner(minSplit=1, maxDepth=50)
learner.learnModel(trainX, trainY)
#Seem to be optimal
alphaThreshold = 100.0
learner.setAlphaThreshold(alphaThreshold)
learner.repPrune(validX, validY)
#learner.tree = learner.tree.cut(3)
predY = learner.predict(testX)
plt.figure(0)
plt.scatter(testX[:, 0], testX[:, 1], c=testY, s=50, vmin=0, vmax=1)
plt.colorbar()
plt.figure(1)
plt.scatter(testX[:, 0], testX[:, 1], c=predY, s=50, vmin=0, vmax=1)
plt.colorbar()
colormap = matplotlib.cm.get_cmap()
def displayTree(learner, vertexId, minX0, maxX0, minX1, maxX1, colormap):
vertex = learner.tree.getVertex(vertexId)
if learner.tree.isLeaf(vertexId):
p = mpatches.Rectangle([minX0, minX1], maxX0-minX0, maxX1-minX1, facecolor=colormap(vertex.getValue()), edgecolor="black")