def printAllOVATrees(data, max_depth=1): h = multiclass.OAA(len(data.labels), lambda: DecisionTreeClassifier(max_depth=max_depth)) h.train(data.X, data.Y) cnt = 0 for tree in h.f: print("Wine:: " + data.labels[cnt]) util.showTree(tree, data.words) raw_input() cnt = cnt + 1
from imports import * from sklearn.tree import DecisionTreeClassifier import multiclass import util from datasets import * h = multiclass.OAA(20, lambda: DecisionTreeClassifier(max_depth=1)) h.train(WineData.X, WineData.Y) P = h.predictAll(WineData.Xte) mean(P == WineData.Yte) mode(WineData.Y) WineData.labels[1] mean(WineData.Yte == 1) P = h.predictAll(WineData.Xte, useZeroOne=True) mean(P == WineData.Yte) h = multiclass.OAA(5, lambda: DecisionTreeClassifier(max_depth=3)) h.train(WineDataSmall.X, WineDataSmall.Y) P = h.predictAll(WineDataSmall.Xte) mean(P == WineDataSmall.Yte) mean(WineDataSmall.Yte == 1) The 1s mean "likely to be Sauvignon-Blanc" and the 0s mean "likely not to be". util.showTree(h.f[0], WineDataSmall.words) """ h = multiclass.OAA(20, lambda: DecisionTreeClassifier(max_depth=3)) h.train(WineData.X, WineData.Y) P = h.predictAll(WineData.Xte) mean(P == WineData.Yte)
from imports import * from sklearn.tree import DecisionTreeClassifier import multiclass import util import warnings from datasets import * warnings.filterwarnings("ignore") map = list() depth=1 for depth range(6): t=multiclass.OAA(5, lambda: DecisionTreeClassifier(max_depth=depth)) h.train(WineData.X, WineData.Y) P = h.predictAll(WineData.Xte) ((P == WineDataSmall.Yte),