SVR预测电力需求

import pandas as pd 
from sklearn.svm import SVR


# 读取数据
data = pd.read_excel(r'F:\桌面\论文831电力预测\所有数据\筛选5变量电力需求预测.xlsx')

# 数据归一化\
# 数据划分
from sklearn.preprocessing import StandardScaler
standardScaler = StandardScaler()
X_train = standardScaler.fit_transform(data.iloc[0:49,0:4])

X_test = standardScaler.fit_transform(data.iloc[49:61,0:4])
#data = pd.DataFrame(df_standard)
# data_origin = standardScaler.inverse_transform(df_standard) 数据还原

# print(data_origin)

y_train = data.iloc[0:49,4:5]
y_test = data.iloc[49:61,4:5]

from sklearn.model_selection import GridSearchCV
import warnings 
warnings.filterwarnings('ignore')

#径向基核函数
model = SVR(kernel='rbf',C=100000,gamma=10)
c_can = np.arange(10,100000,200)
gamma_can = np.arange(0,2,0.2)
rbf_svr = GridSearchCV(model,param_grid={'C': c_can,'gamma' : gamma_can},cv=5)
rbf_svr.fit(X_train,y_train)
print('最优参数是：',rbf_svr.best_params_)
rbf_svr_pred=rbf_svr.predict(X_test)

print(rbf_svr_pred)

最优参数是： {'C': 99810, 'gamma': 1.2000000000000002}
[16305.98465365 13339.98964102  3709.62603183  2419.98410149
  5276.2118453   8142.66785956  8665.65691092  9710.09770166
 12451.35186377 16205.8231178  17800.5526973  18500.90244179]

df = pd.DataFrame(rbf_svr_pred)
df.to_excel('SVR预测数据.xlsx')