Hyper_Spectral/HS_19_多元线性回归.py at master · highwy/Hyper_Spectral · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import datetime
import os
import argparse
import time
from scipy.optimize import curve_fit
from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_squared_error
from sklearn.metrics import r2_score
from sklearn.model_selection import train_test_split
from sklearn.model_selection import train_test_split, GridSearchCV
from pylab import mpl
import csv


plt.rc('font', family='Times New Roman',size=18)
mpl.rcParams['font.sans-serif'] = ['SimHei']  # 指定默认字体
mpl.rcParams['axes.unicode_minus'] = False  # 解决保存图像是负号'-'显示为方块的问题

print(__doc__)
start = time.clock()
starttime = datetime.datetime.now()
print('START RUNNING...')
print('Current time:%s' % time.ctime())
print('-' * 80)

data_path = r'D:\Data\Python\Hyper_Spectral\训练集\多变量\SPAD-多变量.xlsx'
full_path = os.path.split(data_path)[0]

HS_path = full_path

# 该函数用于在y_test与y_predict重新计算OLS回归，用于绘图展示
def reg(X, y):
    reg = LinearRegression().fit(X, y)
    coef = round(reg.coef_[0][0], 4)
    intercept = round(reg.intercept_[0], 3)
    model = str('y=') + str(coef) + 'x+' + str(intercept)
    y_predict = reg.predict(X)
    delta = abs(y - y_predict)
    data = pd.Series(delta.reshape(1, -1)[0])

    index = data.index
    return index, model, coef, intercept, y_predict


def main():
    # 对应各个生育期
    data1 = pd.read_excel(data_path, sheet_name='1st')
    data2 = pd.read_excel(data_path, sheet_name='2nd')
    data3 = pd.read_excel(data_path, sheet_name='3rd')
    data4 = pd.read_excel(data_path, sheet_name='4th')
    data5 = pd.read_excel(data_path, sheet_name='5th')
    data6 = pd.read_excel(data_path, sheet_name='All')
    l = [data1, data2, data3, data4, data5, data6]

    # Load dataset

    def fit_and_draw():
        fig, axes = plt.subplots(2, 3, figsize=(15, 10))
        periods = ['(a)', '(b)', '(c)', '(d)', '(e)', '(f)']
        for index, ax in enumerate(axes.ravel()):
            LR = LinearRegression()
            data = l[index]

            # X =np.vstack((data['光谱'], data['植被指数'], data['三边参数'])).transpose()
            # X=np.array(data['三边参数']).reshape(-1,1)
            X=data.iloc[:, 1:].values
            y = data['SPAD'].values

            X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=1/3, random_state=0)
            #k = 8
            reg_LR = LR.fit(X_train, y_train)

            #params = {'n_jobs': [1]}
            #reg_LR = GridSearchCV(LR, param_grid=params,cv=5)
            #reg_LR.fit(X_train,y_train)
            #建模步骤
            print(reg_LR.coef_)
            coef1= round(reg_LR.coef_[0], 2)
            coef2 = round(reg_LR.coef_[1], 2)
            coef3= round(reg_LR.coef_[2], 2)
            intercept = round(reg_LR.intercept_, 2)
            y_train_pre=reg_LR.predict(X_train)
            R_2 = r2_score(y_train,y_train_pre)
            RMSE1=round(np.sqrt(mean_squared_error(y_train,y_train_pre)),2)
            #计算验证集的R2和RMSE
            y_pre = reg_LR.predict(X_test)
            R2_2 = r2_score(y_test,y_pre)
            RMSE2 = round(np.sqrt(mean_squared_error(y_test, y_pre)), 3)
            # RE = round(np.mean(abs((y_pre - y_test)) / y_test), 3)

            print('Current parameter R2 is %s' % R_2)
            model = str(r'y = '+str(intercept)+'+ '+ str(coef1) + '*X1 +'+str(coef2) + '*X2 +'+str(coef3) + '*X3')
            #把方程式打印在excel里
            my_data = [{'periods': periods[index], 'model': model}, {'R2': R_2}, {'RMSE': RMSE1}]
            headers = ['periods', 'model', 'R2', 'RMSE']
            with open(os.path.join(HS_path, '_ols.csv'), 'a', newline='') as f:
                writer = csv.DictWriter(f, headers)
                writer.writeheader()
                writer.writerows(my_data)
            # plot

            #散点图的设置
            ax.scatter(y_test, y_pre, s=10, marker='o',color ='black')
            # 实测值与预测值再次建立回归模型，再次计算新模型的决定系数、rmse、绝对误差
            _, model, coef1, intercept1, y_pre_new = reg(np.array(y_test).reshape(-1,1), np.array(y_pre).reshape(-1,1))

            #验证集出图
            #line = np.linspace(25, 65, 5) * coef1 + intercept1
            #ax.plot(np.linspace(25, 65, 5), line, label=r'$R^2=$' + str(round(R_2,2)) + '\n' + 'RMSE=' + str(round(RMSE2, 2)),color='grey')
            line_2 = np.linspace(25, 65, 5)
            ax.plot(line_2,line_2,label=r'$R^2=$' + str(round(R2_2,2))+'\n'+'RMSE='+str(round(RMSE2,2)),linewidth=1,linestyle="--",color='grey')
            ax.legend(loc=0, edgecolor='white')
            ax.set_xlabel('Measured  LCC' + '\n' + '%s' % periods[index])
            ax.set_ylabel('Predicted  LCC')
            ax.set_xticks(np.linspace(25, 65, 5))
            ax.set_xticks(np.linspace(25, 65, 5))
            ax.set_xlim(25, 65)
            ax.set_ylim(25, 65)
            ax.set_aspect(1)
            # ax.spines['top'].set_visible(False)
            # ax.spines['right'].set_visible(False)
        plt.subplots_adjust(top=0.98,bottom=0.1,left=0.01,right=1,wspace=0.00, hspace=0.3)
        plt.savefig(os.path.join(HS_path, 'OLS' ),dpi=300)


    fit_and_draw( )


main()

end = time.clock()
print('Running:%fs' % (end - start))
print('DONE...')