def colicTest(trainPath, testPath):
# read data from training dataset and test dataset
trainingInputs, trainingLabels = util.loadDataSet(trainPath)
testInputs, testLabels = util.loadDataSet(testPath)
logRegModel = HorseLogReg(0.01, 500)
trainTheta = logRegModel.fit(trainingInputs, trainingLabels)
errorCount = 0
numTestVec = len(testLabels)
for i in range(numTestVec):
if int(util.classifyVector(testInputs[i],
trainTheta)) != testLabels[i]:
errorCount += 1
errorRate = float(errorCount) / float(numTestVec)
print("the error rate of this test is: %f" % errorRate)
return errorRate
kNN = KNeighborsClassifier()
kNN.fit(x, y)
#test
testX = testingSet[:, :-1]
testY = testingSet[:, -1]
score = kNN.score(testX, testY)
print("KNN score is : ", score)
if __name__ == "__main__":
np.set_printoptions(threshold=np.nan)
trainingSet = util.loadDataSet(constants.CONST_TRAINING_FILENAME)
testingSet = util.loadDataSet(constants.CONST_TESTING_FILENAME)
features = util.loadDataSet(constants.CONST_FEATURE_FILENAME)
#print(features)
sklearnDt(trainingSet, features, testingSet)
sklearnSVC(trainingSet, testingSet)
sklearnKNN(trainingSet, testingSet)
#calcEntropy(trainingSet)
for cluster in clusters:
vecR = cluster[1]
for i in range(2, len(cluster)):
vecR = list(map(lambda x: x[0] + x[1], zip(vecR, cluster[i])))
vec = [a / (len(cluster) - 1) for a in vecR]
Z.append(vec)
if vec != cluster[0]:
isEnd = False
count += 1
print("第{}轮分类，是否结束:{},聚类中心:{}".format(count, isEnd, Z))
if isEnd:
break
return clusteri, clusters
if __name__ == "__main__":
#dataSet = [[0,0],[1,0],[0,1],[1,1],[2,1],[1,2],[2,2],[3,2],[6,6],[7,6],[8,6],[6,7],[7,7],[8,7],[9,7],[7,8],[8,8],[9,8],[8,9],[9,9]]
dataSet = util.loadDataSet("EEG_feature.txt")
labels = util.loadDataSet("valence_arousal_label.txt")
data = util.loadDataSet("EEG_pca_feature.txt")
Z = [dataSet[0], dataSet[1], dataSet[2], dataSet[7]]
K = 4
clusteri, clusters = kMeans(dataSet, K, Z)
util.plotFeature(data, clusteri)
# for i in range(len(clusteri)):
# print("分类:{:d}".format(i))
# for inx in clusteri[i]:
# print("编号:{:d},标签:{}".format(inx,labels[inx]))
ax = plt.axes()
ax.set_xticks(pos + (width / 2)) # center the ticks
ax.set_xticklabels(X_labels)
plt.bar(pos, frequencies, width, color='r')
plt.show()
# TODO the histogram of predictions that predicted too many tags
def testSVM(X_train, Y_train, testingSet):
print "TESTING SVM"
classifier = multiLabelClassifier(X_train, Y_train)
X_test, Y_test = mergeTitlesAndBodies(testingSet)
print "Parsed the testing data"
classifier.fit(X_train, Y_train)
print "Fit the training data"
predicted = classifier.predict(X_test)
#printPrediction(X_test, predicted)
return predicted
if __name__ == '__main__':
trainingSet = util.loadDataSet('out_1000_0')
testingSet = util.loadDataSet('out_2000_2000_0')
#trainingSet = util.loadDataSet('out0')
#testingSet = util.loadDataSet('out1')
#my_nb = nb.NaiveBayes(trainingSet, 100)
#my_nb.train()
#my_nb.test(testingSet)
my_base = bp.BaselinePredictor(trainingSet, 100)
my_base.train()
my_base.test(testingSet)
#coding=utf-8
import numpy as np
from sklearn.decomposition import PCA
import util
dataSet = util.loadDataSet("EEG_feature.txt")
pca = PCA(n_components=2) #降到2维
pca.fit(dataSet) #训练
newX = pca.fit_transform(dataSet) #降维后的数据
# PCA(copy=True, n_components=2, whiten=False)
# print(pca.explained_variance_ratio_) #输出贡献率
for a in newX:
print('{}\t{}'.format(a[0], a[1]))
def formatData(str): #load features from file features = util.loadDataSet(constants.CONST_FEATURE_FILENAME) return util.formatDescToArray(util.preprocessDesc(str), features)