输出预测结果

main
鸽子 1 year ago
parent eb7f3ba5f6
commit c6b760d74b

@ -2,6 +2,7 @@ import pandas as pd
from prophet import Prophet from prophet import Prophet
import math import math
import matplotlib.pyplot as plt import matplotlib.pyplot as plt
import os
pd.set_option('display.width',None) pd.set_option('display.width',None)
def normal(x): def normal(x):
@ -16,58 +17,62 @@ df.drop(columns=['500kv(含330kv)及以上','220kv','110kv(含66kv)','20kv','pow
print(df.columns) print(df.columns)
# print(df.head()) # print(df.head())
print(dict(zip(df.columns,[(df[x]==0).sum()/len(df) for x in df.columns]))) print(dict(zip(df.columns,[(df[x]==0).sum()/len(df) for x in df.columns])))
for i in range(30):
df_ct = df[df['org_name']==df['org_name'].drop_duplicates().values[i]]
# print(df_ct.head())
df_ct['1-10kv'] /= 10000
df_ct['35kv'] /= 10000
df_ct['0.4kv及以下'] /= 10000
s1 = df_ct[['日期','1-10kv']]
s1.dropna(how='any',inplace=True)
s1 = s1.loc[normal(s1['1-10kv']).index]
# plt.plot(range(len(s1)),s1['1-10kv'])
# plt.show()
df_ct = df[df['org_name']==df['org_name'][0]] # 更改列名更改为Prophet指定的列名ds和y
print(df_ct.head()) dd = s1.rename(columns={'日期':'ds','1-10kv':'y'})
df_ct['1-10kv'] /= 10000 # 注意Prophet模型对于数据格式有要求日期字段必须是datetime格式这里通过pd.to_datetime来进行转换。
df_ct['35kv'] /= 10000
df_ct['0.4kv及以下'] /= 10000
s1 = df_ct[['日期','1-10kv']]
s1.dropna(how='any',inplace=True)
s1 = s1.loc[normal(s1['1-10kv']).index]
print(s1)
# plt.plot(range(len(s1)),s1['1-10kv'])
# plt.show()
# 更改列名更改为Prophet指定的列名ds和y
dd = s1.rename(columns={'日期':'ds','1-10kv':'y'})
# 注意Prophet模型对于数据格式有要求日期字段必须是datetime格式这里通过pd.to_datetime来进行转换。
dd['ds'] = pd.to_datetime(dd['ds']) dd['ds'] = pd.to_datetime(dd['ds'])
# 划分数据,划分为训练集和验证集,预测的数据设置为未来一个月 # 划分数据,划分为训练集和验证集,预测的数据设置为未来一个月
df_train = dd[:-3] df_train = dd[(dd['ds']>='2019-01-01')&(dd['ds']<='2023-10-31')][:-3]
df_test = dd[-3:] df_test = dd[(dd['ds']>='2019-01-01')&(dd['ds']<='2023-10-31')][-3:]
# 数据的变动会受到季节、周、天的影响存在一定的规律性因此我们将这三个参数设置为True # 数据的变动会受到季节、周、天的影响存在一定的规律性因此我们将这三个参数设置为True
model = Prophet(yearly_seasonality=True, weekly_seasonality=True, daily_seasonality=True) model = Prophet(yearly_seasonality=True, weekly_seasonality=True, daily_seasonality=True)
# 采用中国的假期模式,其余参数均保持默认 # 采用中国的假期模式,其余参数均保持默认
model.add_country_holidays(country_name="CN") model.add_country_holidays(country_name="CN")
model.fit(df_train) model.fit(df_train)
# make_future_dataframe: 作用是告诉模型我们要预测多长时间以及时间的周期是什么。这里设置为30即预测一个月时间的数据。 # make_future_dataframe: 作用是告诉模型我们要预测多长时间以及时间的周期是什么。这里设置为30即预测一个月时间的数据。
future = model.make_future_dataframe(periods=3, freq='D') future = model.make_future_dataframe(periods=3, freq='D')
# 进行预测返回预测的结果forecast # 进行预测返回预测的结果forecast
forecast = model.predict(future) forecast = model.predict(future)
# forecast['additive_terms'] = forecast['weekly'] + forecast['yearly'] # forecast['additive_terms'] = forecast['weekly'] + forecast['yearly']
# 有forecast['yhat'] = forecast['trend'] + forecast['additive_terms'] 。 # 有forecast['yhat'] = forecast['trend'] + forecast['additive_terms'] 。
# 因此forecast['yhat'] = forecast['trend'] +forecast['weekly'] + forecast['yearly']。 # 因此forecast['yhat'] = forecast['trend'] +forecast['weekly'] + forecast['yearly']。
# 如果有节假日因素那么就会有forecast['yhat'] = forecast['trend'] +forecast['weekly'] + forecast['yearly'] + forecast['holidays']。 # 如果有节假日因素那么就会有forecast['yhat'] = forecast['trend'] +forecast['weekly'] + forecast['yearly'] + forecast['holidays']。
# print(forecast) # print(forecast)
# 测试把ds列即data_series列设置为索引列 # 测试把ds列即data_series列设置为索引列
df_test = df_test.set_index('ds') df_test = df_test.set_index('ds')
# 把预测到的数据取出ds列预测值列yhat同样把ds列设置为索引列。 # 把预测到的数据取出ds列预测值列yhat同样把ds列设置为索引列。
forecast = forecast[['ds','yhat']].set_index('ds') forecast = forecast[['ds','yhat']].set_index('ds')
print(forecast)
# join:按照索引进行连接,
# dropna能够找到DataFrame类型数据的空值缺失值将空值所在的行/列删除后将新的DataFrame作为返回值返回。
df_all = forecast.join(dd.set_index('ds')).dropna()
print(df_all.loc['2023-10'])
# # 创建一个ExcelWriter对象
# with pd.ExcelWriter(r'C:\Users\鸽子\Desktop\output.xlsx',mode='a',if_sheet_exists='replace') as writer:
# # 将不同的子文件写入同一个Excel文件的不同工作表
# df_all.to_excel(writer, sheet_name=f'Sheet{i+1}')
# join:按照索引进行连接, # df_all.plot()
# dropna能够找到DataFrame类型数据的空值缺失值将空值所在的行/列删除后将新的DataFrame作为返回值返回。 # # 设置左上角小标
df_all = forecast.join(df_test).dropna() # plt.legend(['true', 'yhat'])
print(df_all) # plt.show()
df_all.plot()
# 设置左上角小标
plt.legend(['true', 'yhat'])
plt.show()

@ -116,7 +116,7 @@ if __name__ == '__main__':
print(target) print(target)
with open(fr'C:\Users\user\Desktop\电压等级电量预测.txt', 'a', encoding='utf-8') as f: with open(fr'C:\Users\user\Desktop\电压等级电量预测.txt', 'a', encoding='utf-8') as f:
f.write(f'{level}月底电量偏差率:{round(target, 5)}\n') f.write(f'{level}月底电量偏差率:{round(target, 5)}\n')
31

Loading…
Cancel
Save