-
Notifications
You must be signed in to change notification settings - Fork 44
/
basico_comparacion.R
163 lines (106 loc) · 7 KB
/
basico_comparacion.R
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
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
###############################Modelo aditivo prophet#####################################
library(prophet)
setwd("C:/Users/Kuko/Desktop/Series temporales/Pruebas R")
df <- read.csv('C:/Users/Kuko/Desktop/Series temporales/Pruebas R/example_wp_peyton_manning.csv')
plot(1:length(df$ds),df$y,type="l")
df$y <- log(df$y)
plot(1:length(df$ds),df$y,type="l")
m <- prophet(df)
future <- make_future_dataframe(m, periods = 365)
tail(future)
forecast <- predict(m, future)
tail(forecast[c('ds', 'yhat', 'yhat_lower', 'yhat_upper')])
plot(m, forecast)
prophet_plot_components(m, forecast)
plot(1:length(df$ds),df$y,type="l",xlim=c(min(time(forecast$ds)),max(time(forecast$ds))),ylim=c(min(df$y,forecast$yhat),max(df$y,forecast$yhat)))
lines(1:length(forecast$ds),forecast$yhat,type="l",col="red")
library(forecast)
accuracy(df$y,forecast$yhat[1:length(df$y)])
###########################Alisado exponencial doble stat###################################
df <- read.csv('C:/Users/Kuko/Desktop/Series temporales/Pruebas R/example_wp_peyton_manning.csv')
plot(1:length(df$ds),df$y,type="l")
df$y <- log(df$y)
plot(1:length(df$ds),df$y,type="l")
j <- ts(df$y,frequency=365)
aliHW <- HoltWinters(j)
pred.aliHW <- predict(aliHW, n.ahead=365, prediction.interval=FALSE)
#plot(aliHW, pred.aliHW)
plot(1:length(df$ds),df$y,type="l",xlim=c(min(time(df$ds)),max(time(df$ds))+365),ylim=c(min(df$y,aliHW$fitted[,1],pred.aliHW),max(df$y,aliHW$fitted[,1],pred.aliHW)))
lines(366:length(df$y),aliHW$fitted[,1],col="blue")
lines((length(df$ds)+1):(length(df$ds)+365),pred.aliHW,type="l",col="blue")
accuracy(df$y[366:length(df$y)],aliHW$fitted[,1])
#################################red neuronal forecast######################################
df <- read.csv('C:/Users/Kuko/Desktop/Series temporales/Pruebas R/example_wp_peyton_manning.csv')
plot(1:length(df$ds),df$y,type="l")
df$y <- log(df$y)
plot(1:length(df$ds),df$y,type="l")
#df$y = (df$y-min(df$y))/(max(df$y)-min(df$y))
df$y <- scale(df$y)
j <- ts(df$y,frequency=365)
fit <- nnetar(j)
f.nn <- forecast(fit,h=365)
plot(f.nn)
plot(1:length(df$ds),df$y,type="l",xlim=c(min(time(df$ds)),max(time(df$ds))+365),ylim=c(min(df$y,fit$fitted,f.nn$mean,na.rm=TRUE),max(df$y,fit$fitted,f.nn$mean,na.rm=TRUE)))
lines(1:length(df$y),fit$fitted,col="green")
lines((length(df$ds)+1):(length(df$ds)+365),f.nn$mean,col="green")
accuracy(df$y[366:length(df$y)],fit$fitted)
##########################################ARIMA forecast##################################
df <- read.csv('C:/Users/Kuko/Desktop/Series temporales/Pruebas R/example_wp_peyton_manning.csv')
plot(1:length(df$ds),df$y,type="l")
df$y <- log(df$y)
plot(1:length(df$ds),df$y,type="l")
j <- ts(df$y,frequency=365)
h <- auto.arima(j,stepwise=FALSE,approximation=FALSE)
arima <- Arima(j,model=h)
f <- forecast(arima,h=365)
plot(f)
plot(1:length(df$ds),df$y,type="l",xlim=c(min(time(df$ds)),max(time(df$ds))+365),ylim=c(min(df$y,f$fitted,f$mean,na.rm=TRUE),max(df$y,f$fitted,f$mean,na.rm=TRUE)))
lines(1:length(df$y),f$fitted,col="orange")
lines((length(df$ds)+1):(length(df$ds)+365),f$mean,col="orange")
accuracy(df$y,f$fitted)
par(mfrow=c(2,2))
plot(1:length(df$ds),df$y,type="l",xlim=c(min(time(forecast$ds)),max(time(forecast$ds))),ylim=c(min(df$y,forecast$yhat),max(df$y,forecast$yhat)),main="Prophet (MAE=0.33 RMSE=0.48)",xlab="time",ylab="y")
lines(1:length(forecast$ds),forecast$yhat,type="l",col="red")
plot(1:length(df$ds),df$y,type="l",xlim=c(min(time(df$ds)),max(time(df$ds))+365),ylim=c(min(df$y,aliHW$fitted[,1],pred.aliHW),max(df$y,aliHW$fitted[,1],pred.aliHW)),main="Holt Winter (MAE=0.37 RMSE=0.54",xlab="time",ylab="y")
lines(366:length(df$y),aliHW$fitted[,1],col="blue")
lines((length(df$ds)+1):(length(df$ds)+365),pred.aliHW,type="l",col="blue")
plot(1:length(df$ds),df$y,type="l",xlim=c(min(time(df$ds)),max(time(df$ds))+365),ylim=c(min(df$y,f$fitted,f$mean,na.rm=TRUE),max(df$y,f$fitted,f$mean,na.rm=TRUE)),main="auto.arima (MAE=0.28 RMSE=0.42)",xlab="time",ylab="y")
lines(1:length(df$y),f$fitted,col="orange")
lines((length(df$ds)+1):(length(df$ds)+365),f$mean,col="orange")
plot(1:length(df$ds),df$y,type="l",xlim=c(min(time(df$ds)),max(time(df$ds))+365),ylim=c(min(df$y,fit$fitted,f.nn$mean,na.rm=TRUE),max(df$y,fit$fitted,f.nn$mean,na.rm=TRUE)),main="nnetar (MAE=0.73 RMSE=1.01)",xlab="time",ylab="y")
lines(1:length(df$y),fit$fitted,col="green")
lines((length(df$ds)+1):(length(df$ds)+365),f.nn$mean,col="green")
#########################################COMPARACIÓN####################################################
library(forecast)
library(prophet)
plot(AirPassengers)
AirPassengers <- log(AirPassengers)
df <- data.frame(time(AirPassengers),AirPassengers[1:length(AirPassengers)])
colnames(df) <- c("ds","y")
y <- (AirPassengers-min(AirPassengers))/(max(AirPassengers)-min(AirPassengers))
m <- prophet(df)
aliHW <- HoltWinters(AirPassengers)
h <- auto.arima(AirPassengers,stepwise=FALSE,approximation=FALSE)
arima <- Arima(AirPassengers,model=h)
fit <- nnetar(y)
future <- make_future_dataframe(m, periods = 20, freq = 'm')
forecast <- predict(m, future)
pred.aliHW <- predict(aliHW, n.ahead=20, prediction.interval=FALSE)
f <- forecast(arima,h=20)
f.nn <- forecast(fit,h=20)
accuracy(df$y,forecast$yhat[1:length(df$y)])
accuracy(AirPassengers[13:length(AirPassengers)],aliHW$fitted[,1])
accuracy(AirPassengers[1:length(AirPassengers)],f$fitted)
accuracy(y[13:length(y)],fit$fitted[13:length(fit$fitted)])
par(mfrow=c(2,2))
plot(1:length(df$ds),df$y,type="l",xlim=c(min(time(forecast$ds)),max(time(forecast$ds))),ylim=c(min(df$y,forecast$yhat),max(df$y,forecast$yhat)),main="Prophet (MAE=0.026 RMSE=0.033)",xlab="time",ylab="y")
lines(1:length(forecast$ds),forecast$yhat,type="l",col="red")
plot(1:length(AirPassengers),AirPassengers[1:length(AirPassengers)],type="l",xlim=c(0,length(AirPassengers)+20),ylim=c(min(AirPassengers[1:length(AirPassengers)],aliHW$fitted[,1],pred.aliHW),max(AirPassengers[1:length(AirPassengers)],aliHW$fitted[,1],pred.aliHW)),main="Holt Winter (MAE=0.029 RMSE=0.039)",xlab="time",ylab="y")
lines(13:length(AirPassengers),aliHW$fitted[,1],col="blue")
lines((length(AirPassengers)+1):(length(AirPassengers)+20),pred.aliHW,type="l",col="blue")
plot(1:length(AirPassengers),AirPassengers[1:length(AirPassengers)],type="l",xlim=c(0,length(AirPassengers)+20),ylim=c(min(AirPassengers[1:length(AirPassengers)],f$fitted,f$mean),max(AirPassengers[1:length(AirPassengers)],f$fitted,f$mean)),main="auto.arima (MAE=0.026 RMSE=0.035)",xlab="time",ylab="y")
lines(1:length(df$y),f$fitted,col="orange")
lines((length(AirPassengers)+1):(length(AirPassengers)+20),f$mean,col="orange")
plot(1:length(y),y[1:length(y)],type="l",xlim=c(0,length(AirPassengers)+20),ylim=c(min(y[1:length(y)],fit$fitted,f.nn$mean,na.rm=TRUE),max(y[1:length(y)],fit$fitted,f.nn$mean,na.rm=TRUE)),main="nnetar (MAE=0.023 RMSE=0.029)",xlab="time",ylab="y")
lines(1:length(y),fit$fitted,col="green")
lines((length(y)+1):(length(y)+20),f.nn$mean,col="green")