手把手教你用Prophet快速进行时间序列预测（附Prophet和R代码）

导入所需要的包并读入数据集
#Importing datasets
import pandas as pd
import numpy as np
from fbprophet import Prophet
#Read train and test
train = pd.read_csv('Train_SU63ISt.csv')
test = pd.read_csv('Test_0qrQsBZ.csv')


#Convert to datetime format
train['Datetime'] = pd.to_datetime(train.Datetime,format='%d-%m-%Y %H:%M') 
test['Datetime'] = pd.to_datetime(test.Datetime,format='%d-%m-%Y %H:%M')
train['hour'] = train.Datetime.dt.hour

# Calculate average hourly fraction
hourly_frac = train.groupby(['hour']).mean()/np.sum(train.groupby(['hour']).mean())
hourly_frac.drop(['ID'], axis = 1, inplace = True)
hourly_frac.columns = ['fraction']


# convert to time series from dataframe
train.index = train.Datetime
train.drop(['ID','hour','Datetime'], axis = 1, inplace = True)


daily_train = train.resample('D').sum()

y -> 目标（Target）
ds -> 时间（Datetime）

daily_train['ds'] = daily_train.index
daily_train['y'] = daily_train.Count
daily_train.drop(['Count'],axis = 1, inplace = True)

m = Prophet(yearly_seasonality = True, seasonality_prior_scale=0.1)
m.fit(daily_train)
future = m.make_future_dataframe(periods=213)
forecast = m.predict(future)

m.plot_components(forecast)

# Extract hour, day, month and year from both dataframes to merge
for df in [test, forecast]:
    df['hour'] = df.Datetime.dt.hour
    df['day'] = df.Datetime.dt.day
    df['month'] = df.Datetime.dt.month
    df['year'] = df.Datetime.dt.year


# Merge forecasts with given IDs
test = pd.merge(test,forecast, on=['day','month','year'], how='left')
cols = ['ID','hour','yhat']
test_new = test[cols]


# Merging hourly average fraction to the test data
test_new = pd.merge(test_new, hourly_frac, left_on = ['hour'], right_index=True, how = 'left')
# Convert daily aggregate to hourly traffic
test_new['Count'] = test_new['yhat'] * test_new['fraction']
test_new.drop(['yhat','fraction','hour'],axis = 1, inplace = True)
test_new.to_csv('prophet_sub.csv',index = False)

library(prophet)
library(data.table)
library(dplyr)
library(ggplot2)


# read data
train = fread("Train_SU63ISt.csv")
test = fread("Test_0qrQsBZ.csv")


# Extract date from the Datetime variable
train$Date = as.POSIXct(strptime(train$Datetime, "%d-%m-%Y"))
test$Date = as.POSIXct(strptime(test$Datetime, "%d-%m-%Y"))


# Convert 'Datetime' variable from character to date-time format
train$Datetime = as.POSIXct(strptime(train$Datetime, "%d-%m-%Y %H:%M"))
test$Datetime = as.POSIXct(strptime(test$Datetime, "%d-%m-%Y %H:%M"))


# Aggregate train data day-wise
aggr_train = train[,list(Count = sum(Count)), by = Date]


# Visualize the data
ggplot(aggr_train) + geom_line(aes(Date, Count))


# Change column names
names(aggr_train) = c("ds", "y")


# Model building
m = prophet(aggr_train)
future = make_future_dataframe(m, periods = 213)
forecast = predict(m, future)


# Visualize forecast
plot(m, forecast)


# proportion of mean hourly 'Count' based on train data
mean_hourly_count = train %>%
group_by(hour = hour(train$Datetime)) %>%
summarise(mean_count = mean(Count))


s = sum(mean_hourly_count$mean_count)
mean_hourly_count$count_proportion = mean_hourly_count$mean_count/s


# variable to store hourly Count
test_count = NULL


for(i in 763:nrow(forecast)){
test_count = append(test_count, mean_hourly_count$count_proportion * forecast$yhat[i])
}


test$Count = test_count