def coursera_LR(filename):
f_LR = C1.load_tonumpy(filename)
m, n = shape(f_LR)
#get X and the lable Y,
#attention please the features X, first column is all 1,so we need to make change to fit this
X_feature = zeros([m, n])
X_feature[:, 0] = 1
X_feature[:, 1:-1] = f_LR[:, 0:-2]
y_lable = f_LR[:, -1]
#get LR model from sklearn.linear model
LR = linear_model.LinearRegression()
#training the LR model to get the weights of the LR
LR.fit(X_feature, y_lable)
weight_LR = LR.coef_ #coef_ is weight of the LR
print("the training weights of the LR model is:")
print weight_LR
print LR.predict
#ploting for better viewing
#ploting together:scatter and plot_surface
ax1 = plt.subplot(111, projection='3d') #build 3d project
ax1.scatter(X_feature[:, 1], X_feature[:, 2], y_lable)
#second subplot,showing the surface of the regression
#ax2=plt.subplot(111,projection='3d') #build 3d project
X = arange(500.0, 5000.0, 500.0)
Y = arange(0.0, 9.0, 1.0)
X, Y = meshgrid(X, Y)
Z = X * weight_LR[1] + Y * weight_LR[2]
ax1.plot_surface(X, Y, Z)
#showing
plt.show()
def Logistic_sklearn_prepare(filename):
f_not_scaling = C1.load_tonumpy(filename)
#need scaling
f_scaling = C2.Logistic_scaling(f_not_scaling)
#the first column should be all 1
f_scaling_universe = C2.Logistic_universe_form(f_scaling)
return f_scaling_universe
def view_before(filename):
f = C1.load_tonumpy(filename)
m, n = shape(f)
feature_0, feature_1 = devide(f)
#data have devided into two parts!
#ploting!
#y=0
plt.scatter(feature_0[:, 0], feature_0[:, 1], s=m, c='r', marker='*')
#y=1
plt.scatter(feature_1[:, 0], feature_1[:, 1], s=m, c='b', marker='o')
#showing()
plt.show()
def view_after_scaling(filename):
f = C1.load_tonumpy(filename)
m, n = shape(f)
#scaling
f_scaling = Logistic_scaling(f)
#devide into two parts
feature_0, feature_1 = devide(f_scaling)
#ploting!
#y=0
plt.scatter(feature_0[:, 0], feature_0[:, 1], s=m, c='r', marker='*')
#y=1
plt.scatter(feature_1[:, 0], feature_1[:, 1], s=m, c='b', marker='o')
#showing()
plt.show()
def coursera_LR(filename):
f_LR=C1.load_tonumpy(filename)
m,n=shape(f_LR)
#get X and the lable Y,
#attention please the features X, first column is all 1,so we need to make change to fit this
X_feature=zeros([m,n])
X_feature[:,0]=1
X_feature[:,1:-1]=f_LR[:,0:-2]
y_lable=f_LR[:,-1]
#get LR model from sklearn.linear model
LR=linear_model.LinearRegression()
#training the LR model to get the weights of the LR
LR.fit(X_feature,y_lable)
weight_LR=LR.coef_#coef_ is weight of the LR
print("the training weights of the LR model is:" )
print weight_LR
def Logistic_prepare(filename):
#not scaling!
f_not_scaling=C1.load_tonumpy(filename)
#devide for ploting
feature_not_scaling_0,feature_not_scaling_1=C2.devide(f_not_scaling)
#the first column should be all 1
m0,n0=shape(f_not_scaling)
fprepare_not_scaling=zeros([m0,n0+1])
fprepare_not_scaling[:,0]=1.0
fprepare_not_scaling[:,1:]=f_not_scaling
#scaling
f_scaling=C2.Logistic_scaling(f_not_scaling)
#devide for ploting
feature_scaling_0,feature_scaling_1=C2.devide(f_scaling)
m1,n1=shape(f_scaling)
#the first column should be all 1
fprepare_scaling=zeros([m1,n1+1])
fprepare_scaling[:,0]=1.0
fprepare_scaling[:,1:]=f_scaling
return fprepare_not_scaling,feature_not_scaling_0,feature_not_scaling_1,fprepare_scaling,feature_scaling_0,feature_scaling_1
def coursera_LR(filename):
f_LR = C1.load_tonumpy(filename)
m, n = shape(f_LR)
#get X and the lable Y,
#attention please the features X, first column is all 1,so we need to make change to fit this
X_feature = zeros([m, n])
X_feature[:, 0] = 1
X_feature[:, 1] = f_LR[:, 0]
y_lable = f_LR[:, 1]
#get LR model from sklearn.linear model
LR = linear_model.LinearRegression()
#training the LR model to get the weights of the LR
LR.fit(X_feature, y_lable)
weight_LR = LR.coef_ #coef_ is weight of the LR
print("the training weights of the LR model is:")
print weight_LR
#ploting for better viewing
#scatter and plot is different
plt.scatter(X_feature[:, 1], y_lable, color='black')
plt.plot(X_feature[:, 1], LR.predict(X_feature), color='blue')
plt.show()
def Logistic_prepare(filename):
#not scaling!
f_not_scaling=C1.load_tonumpy(filename)
#devide for ploting
feature_not_scaling_0,feature_not_scaling_1=C2.devide(f_not_scaling)
#feature_mapping
f_mapping=C2.feature_mapping(f_not_scaling)
#the first column should be all 1
m,n=shape(f_mapping)
fprepare_not_scaling=zeros([m,n+1])
fprepare_not_scaling[:,0]=1.0
fprepare_not_scaling[:,1:]=f_mapping
##scaling
#f_scaling=C2.Logistic_scaling(f_not_scaling)
##devide for ploting
#feature_scaling_0,feature_scaling_1=C2.devide(f_scaling)
#m1,n1=shape(f_scaling)
##the first column should be all 1
#fprepare_scaling=zeros([m1,n1+1])
#fprepare_scaling[:,-1]=1.0
#fprepare_scaling[:,0:-1]=f_scaling
return fprepare_not_scaling,feature_not_scaling_0,feature_not_scaling_1