def multiFisher():
filePath = "iris.data.txt"
bX, bY = loadData.DataLoader(filePath).multiClassifyGetData()
mX = np.array(bX)
mY = np.array(bY).reshape(150, 1) # 用reshape函数修改Y矩阵的形式
trainX, trainY, testX, testY = generateData.DataGenerator(mX,
mY).randDivide()
def binaryFisher():
filePath = "iris.data.txt"
bX, bY = loadData.DataLoader(filePath).binaryClassifyGetData()
bX = np.array(bX)
bY = np.array(bY).reshape(100, 1) # 用reshape函数修改Y矩阵的形式
trainX, trainY, testX, testY = generateData.DataGenerator(bX,
bY).randDivide()
trainX = trainX.astype(float)
trainY = trainY.astype(float)
testX = testX.astype(float)
testY = testY.astype(float)
# print(trainX)
# print(trainY)
# print(testX)
# print(testY)
# print(np.shape(trainX), np.shape(trainY), np.shape(testX), np.shape(testY))
miu1, length1 = fisher.FisherLinearDiscriminant(trainX, trainY, testX,
testY).getMiu(1)
miu2, length2 = fisher.FisherLinearDiscriminant(trainX, trainY, testX,
testY).getMiu(2)
# # print(miu1)
# # print(miu2)
# S1 = np.zeros((4, 4))
# S2 = np.zeros((4, 4))
# for i in range(len(trainX)):
# if int(trainY[i]) == 1:
# xMat = (trainX[i] - miu1).reshape((4, 1))
# s = np.dot(xMat, xMat.T)
# # print(s)
# S1 = S1 + s
# else:
# xMat = (trainX[i] - miu2).reshape((4, 1))
# s = np.dot(xMat, xMat.T)
# # print(s)
# S2 = S2 + s
# # print(S1)
# # print(S2)
S1 = fisher.FisherLinearDiscriminant(trainX, trainY, testX,
testY).getSw(miu1, 1)
S2 = fisher.FisherLinearDiscriminant(trainX, trainY, testX,
testY).getSw(miu2, 2)
Sw = np.mat(S1 + S2)
miu = (miu2 - miu1).reshape(4, 1)
w = np.dot(Sw.I, miu)
# print(w)
fisher.FisherLinearDiscriminant(trainX, trainY, testX, testY).plot(w)
miuAll = (length1 * miu1 + length2 * miu2) / (length1 + length2)
fisher.FisherLinearDiscriminant(trainX, trainY, testX,
testY).modelTest(w, miuAll)
def getaccuracy(ytest, predictions):
correct = 0
for i in range(len(ytest)):
if ytest[i] == predictions[i]:
correct += 1
return (correct/float(len(ytest)))*100.0
def geterror(ytest, predictions):
return (100.0 - getaccuracy(ytest, predictions))
if __name__ == '__main__':
dataFile = 'dataset/bank.csv'
dataloader = ld.DataLoader(dataFile)
numruns = 1
# Follw the same testing format as in Assignments
classalgs = {'Random': al.Classifier(), # Baseline Algorithm
'Linear SVM': al.SVMClassifier(), # Linear SVM
'Logistic Regression L2 regularizer': al.LogisticRegressionClassifier({'regularizer':'l2', 'regularizerValue':0.01}),
'Logistic Regression No regularizer': al.LogisticRegressionClassifier(),
'Neural Network': al.NeuralNetwork(),
}
numalgs = len(classalgs)
parameters = (
{'regwgt':0.0, 'nh':(50, ), 'regularizerValue': 0.01 },
{'regwgt':0.0, 'nh':(100, ), 'regularizerValue': 0.1 },
import numpy as np
def trans_label(label_matrix):
temp_list = []
for label in label_matrix.T:
if label.getA()[0][1] == 1.0:
temp_list.append(0)
else:
temp_list.append(1)
return temp_list
if __name__ == "__main__":
file_path = r'iris.txt'
x_train, x_test, y_train, y_test = loadData.DataLoader(
file_path).get_train_test_data()
y_train_list = trans_label(y_train)
y_test_list = trans_label(y_test)
#
logistic_model = logistic_bayes_model.LogisticBayesClassification(
x_train,
np.mat(y_train_list).T, x_test,
np.mat(y_test_list).T)
weight = logistic_model.get_weight()
# logistic_model.load_model()
logistic_model.save_model()
for i in range(x_test.shape[1]):
print(logistic_model.predict(x_test[:, i:i + 1]))
print(y_test_list[i])
print('\n')
print("accuracy:")
#!/usr/bin/env python
# -*- coding: utf-8 -*-
import loadData
import dataGenerate
import AIC_Logistic
import showPlt
if __name__ == '__main__':
lD = loadData.DataLoader("iris.data.txt")
dataX, dataY = lD.loadData()
dG = dataGenerate.DataGenerator(dataX, dataY)
trainX, trainY, testX, testY = dG.randDivide(testRatio=0.3)
# print(trainX, "\n", trainY, "\n", testX, "\n", testY)
aicL = AIC_Logistic.AicLogistic(trainX, trainY, testX, testY)
p = aicL.phi()
weights = aicL.logisticTrain(p, numIter=150)
sP = showPlt.ShowPlt(trainX, trainY, weights)
sP.plotMap()
aicL.logisticTest(weights)
import generateData
import linearRegression
def get_RMSE(data_Y, pre_Y):
m = len(data_Y)
loss = 0.0
for i in range(m):
loss += pow((data_Y[i] - pre_Y[i]), 2)
return pow(loss / m, 0.5)
if __name__ == "__main__":
file_path = 'data.txt'
# 从文件中加载数据
X, Y = loadData.DataLoader(file_path).get_data()
# 划分训练集测试集
train_data_X, train_data_Y, test_data_X, test_data_Y = generateData.DataGenerator(
X, Y).rand_divide()
# 训练模型
liner_model = linearRegression.LinearRegression(train_data_X, train_data_Y)
liner_model.train_model()
# 预测
pre_Y = []
f1 = open('pre.txt', 'w')
for i in test_data_X:
pre = liner_model.predict_model(i)
f1.write(str(pre) + '\n')
