170720 混淆矩阵绘制+pandas读取数据（有点乱，后面抽空再整理）

Python 数据分析包：pandas 基础

E:\Backup\validation confusion matrix_final2

# -*- coding: utf-8 -*-
"""
Created on Fri May 19 11:17:12 2017

@author: Bruce Lau
"""
#%%
import itertools
import numpy as np
import matplotlib.pyplot as plt
from scipy import stats

from sklearn.metrics import confusion_matrix
from p4_1_ds_competition import idscnn2
# load the data
def reverse_onehot(onehot):
    y = []
    for i in onehot:
        i = i.tolist()
        y.append(i.index(max(i)))
    return y
#%%
def cal_acc(m1,m2):
    a1 = np.array(reverse_onehot(m1))
    re = sum(a1==m2)/len(m2)
#    print(re)
    return re
#%%

def plot_confusion_matrix(cm, classes,
                          normalize=False,
                          title='Confusion matrix',
                          cmap=plt.cm.Blues):
    """
    This function prints and plots the confusion matrix.
    Normalization can be applied by setting `normalize=True`.
    """
#    plt.figure(facecolor='w')

    im = plt.imshow(cm, interpolation='nearest', cmap=cmap)
    plt.title(title)
#    plt.colorbar()
    plt.colorbar(im,fraction=0.046, pad=0.04)

    tick_marks = np.arange(len(classes))
    plt.xticks(tick_marks, classes) #, rotation=45
    plt.yticks(tick_marks, classes)

    if normalize:
        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]
        print("Normalized confusion matrix")
    else:
        print('Confusion matrix, without normalization')

#    print(cm)

    thresh = cm.max() / 2.
    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):
        cm[i,j]=round(cm[i,j],3)
        plt.text(j, i, cm[i, j],
                 horizontalalignment="center",
                 color="white" if cm[i, j] > thresh else "black")

#    plt.tight_layout()
    plt.ylabel('True fault type')
    plt.xlabel('CNN prediction fault type')


#%%
def cm_plot(y_,y,name,idx):
    class_names=np.array(['0','1','2','3','4','5','6','7','8','9'])
    cnf_matrix = confusion_matrix(y_, y)
    np.set_printoptions(precision=2)

    # Plot non-normalized confusion matrix
    #plt.figure(facecolor='w')
    #plot_confusion_matrix(cnf_matrix, classes=class_names,
    #                      title='Confusion matrix, without normalization')

    # Plot normalized confusion matrix
    plt.subplot(1,3,idx)
    plot_confusion_matrix(cnf_matrix, classes=class_names, normalize=False,
                          title='Normalized confusion matrix')
    plt.title(name)
    if idx == 3:
        plt.savefig(str(idx)+'.png',dpi=300)
        plt.show()

#%%
def papershow(de_c,fe_c):   
    # load the prediction data 9-17
    labels = np.load('labels.npy')
    # de_accuracy and fe_accuracy
    de_acc = cal_acc(de_c,labels)
    fe_acc = cal_acc(fe_c,labels)
    # ids fusion process
    me = np.ones((2500,10,2))
    me[:,:,0]=de_c
    me[:,:,1]=fe_c
    re = np.ones((2500,10))
    for i in range(2500):
        stack = me[i,:,:]
        re[i]=idscnn2(stack.T)
    # ids result
    fusion_result = cal_acc(re,labels)
    return np.array([de_acc, fe_acc, fusion_result]), re

#%%

def save_cm(path1,path2):
    de = np.load(path1)
    fe = np.load(path2)
    acc, me = papershow(de,fe)
    y1_ = de
    y2_ = fe
    y3_ = me
    y4_ = np.load('labels.npy')
    y1_ = reverse_onehot(y1_)
    y2_ = reverse_onehot(y2_)
    y3_ = reverse_onehot(y3_) 
    y4_ = y4_ 
    #
    plt.figure(facecolor='w',figsize=(16,4))
    cm_plot(y4_,y1_,'# 7 CNN model @ drive end\n accuracy=83.8%',1)

    cm_plot(y4_,y2_,'# 25 CNN model @ fan end\n accuracy=79.2%',2)

    cm_plot(y4_,y3_,'Fused model for # 7 and # 25\n accuracy=92.4%',3)

#109
#112
#%%
accuracy =  np.ones((20,3))
for i in np.arange(1,21):
    print(i)
    de = np.load('cnn_pre_pro2/pre_pro_'+str(i)+'/CA/107_de.npy')
    fe = np.load('cnn_pre_pro2/pre_pro_'+str(i)+'/CA/109_fe.npy')
    acc,re =  papershow(de,fe)
    accuracy[i-1,:] = acc
#%%
path1 = 'cnn_pre_pro2/pre_pro_1//CA/107_de.npy'
path2 = 'cnn_pre_pro2/pre_pro_1//CA/109_fe.npy'
#
save_cm(path1,path2)
#%% t-test
t_de = accuracy[:,0]
t_fe = accuracy[:,1]
t_me = accuracy[:,2]

t_de_me = stats.ttest_ind(t_de,t_me,equal_var=False)
t_fe_me = stats.ttest_ind(t_fe,t_me,equal_var=False)

print('t-test significant difference between de and me is: %f'%t_de_me[1])
print('t-test significant difference between fe and me is: %f'%t_de_me[1])
print("The averages are: ",np.mean(accuracy,axis=0))
avg = np.mean(accuracy,axis=0)
#%%
import pandas as pd
data = pd.read_excel('statistical-analysis2.xlsx',sheetname=1,skiprows=1)
data_array = data.values
print("The mean values of DS and IDS are \n", data.mean())
print('\n')
print("The std values of DS and IDS are \n", data.std())