PARTE 1: LEYENDO LOS DATOS - Indice de produccion industrial

library(ggplot2)
library(forecast)
library(readxl)
AGSyP_ts <- read_excel("C:/Users/Economia/Desktop/AGS y P.xlsx", 
                     col_types = c("skip", "skip", "numeric"))
data=AGSyP_ts %>% ts(start = c(2005,1),frequency = 12)->AGSyP
print(AGSyP)
##         Jan    Feb    Mar    Apr    May    Jun    Jul    Aug    Sep    Oct
## 2005  92.64  68.41  60.10  65.43 137.50 141.97  71.05 128.73  97.27  71.26
## 2006  86.67  69.31  62.80  65.16 136.68 143.43  70.70 137.40 107.81  74.25
## 2007  94.30  73.70  65.83  69.07 146.57 156.23  72.82 153.89 117.05  76.11
## 2008  91.09  70.96  64.06  67.95 150.52 165.58  75.11 154.51 114.93  74.25
## 2009  86.82  70.17  65.24  67.36 137.69 153.16  75.56 136.56 109.59  74.07
## 2010  97.58  82.25  68.26  68.26 153.40 149.50  75.17 147.21 103.37  74.22
## 2011  92.58  68.70  68.64  68.33 145.21 149.11  69.83 131.08 101.34  75.92
## 2012  88.55  62.62  61.39  65.20 149.91 183.43  71.85 146.35 116.73  73.90
## 2013  90.05  66.16  62.83  63.18 124.26 164.76  68.33 125.32 117.47  72.17
## 2014  97.83  75.14  61.11  67.20 122.92 154.11  74.61 124.65 118.20  74.18
## 2015  91.59  67.18  66.22  69.67 112.99 142.87  73.00 116.07 112.53  71.90
## 2016  82.95  70.20  66.91  66.61 118.81 168.59  74.25 132.16 126.73  75.35
## 2017  87.09  64.26  70.53  63.94 125.82 178.89  76.25 125.83 121.14  73.12
## 2018  94.25  71.82  71.18  68.57 127.33 164.70  71.65 126.29 116.30  71.36
## 2019  92.88  72.59  70.73  68.54 128.30 168.92  75.81 130.51 121.34       
##         Nov    Dec
## 2005 115.84 123.22
## 2006 124.54 127.28
## 2007 131.05 135.13
## 2008 127.42 132.96
## 2009 122.34 122.11
## 2010 125.23 137.40
## 2011 126.99 130.58
## 2012 131.53 130.47
## 2013 124.56 110.55
## 2014 124.93 105.12
## 2015 112.19  97.74
## 2016 123.86 104.69
## 2017 119.79 106.42
## 2018 121.08 108.99
## 2019

PARTE 2: PROYECCION (A 6 MESES)

library(forecast)
modelo<-auto.arima(y = AGSyP)
summary(modelo)
## Series: AGSyP 
## ARIMA(3,0,2)(0,1,2)[12] 
## 
## Coefficients:
##          ar1      ar2     ar3     ma1     ma2     sma1     sma2
##       0.0966  -0.5712  0.4551  0.2145  0.7144  -0.2315  -0.1695
## s.e.  0.1352   0.1736  0.0777  0.1591  0.1973   0.0943   0.0663
## 
## sigma^2 estimated as 47.5:  log likelihood=-550.34
## AIC=1116.68   AICc=1117.61   BIC=1141.53
## 
## Training set error measures:
##                     ME     RMSE      MAE        MPE     MAPE     MASE
## Training set 0.2220619 6.511608 4.442784 0.01846539 4.259824 0.772935
##                     ACF1
## Training set -0.02842029

PARTE 2: PRONOSTICO ARIMA

pronosticos<-forecast(modelo,h = 6) 
autoplot(pronosticos)+xlab("Años")+ylab("indice")+theme_bw()

## PARTE 2: AJUSTE SARIMA

library(forecast)
autoplot(pronosticos$x,series = "IPI")+autolayer(pronosticos$fitted,series = "Pronostico")+ggtitle("Ajuste SARIMA")

## PARTE 3: Serie ampliada

AGSyP_h<-ts(as.numeric(rbind(as.matrix(pronosticos$x),as.matrix(pronosticos$mean))),start = c(2005,1),frequency = 12)
tail(AGSyP_h,75)
##            Jan       Feb       Mar       Apr       May       Jun       Jul
## 2014  97.83000  75.14000  61.11000  67.20000 122.92000 154.11000  74.61000
## 2015  91.59000  67.18000  66.22000  69.67000 112.99000 142.87000  73.00000
## 2016  82.95000  70.20000  66.91000  66.61000 118.81000 168.59000  74.25000
## 2017  87.09000  64.26000  70.53000  63.94000 125.82000 178.89000  76.25000
## 2018  94.25000  71.82000  71.18000  68.57000 127.33000 164.70000  71.65000
## 2019  92.88000  72.59000  70.73000  68.54000 128.30000 168.92000  75.81000
## 2020  92.34674  71.27294  70.17367                                        
##            Aug       Sep       Oct       Nov       Dec
## 2014 124.65000 118.20000  74.18000 124.93000 105.12000
## 2015 116.07000 112.53000  71.90000 112.19000  97.74000
## 2016 132.16000 126.73000  75.35000 123.86000 104.69000
## 2017 125.83000 121.14000  73.12000 119.79000 106.42000
## 2018 126.29000 116.30000  71.36000 121.08000 108.99000
## 2019 130.51000 121.34000  74.17889 122.24368 108.89578
## 2020

PARTE 3.1: DESCOMPOSICION DE LA SERIE TEMPORAL

library(stats)
fit<-stl(AGSyP_h,"periodic")
autoplot(fit)+theme_bw()

## PARTE 3.1: FIT DE LA SERIE TEMPORAL

TC<-fit$time.series[,2]
print(TC)
##            Jan       Feb       Mar       Apr       May       Jun       Jul
## 2005  99.30646  98.96314  98.61982  98.36323  98.10664  97.92288  97.73913
## 2006  98.05989  98.43906  98.81824  99.22775  99.63727 100.07406 100.51085
## 2007 104.27961 104.96769 105.65577 106.17197 106.68817 107.01153 107.33489
## 2008 107.69477 107.78734 107.87991 107.82202 107.76412 107.69743 107.63073
## 2009 104.79565 104.12677 103.45789 103.00564 102.55338 102.50319 102.45299
## 2010 105.32099 105.53291 105.74483 105.79917 105.85351 105.83374 105.81397
## 2011 104.40080 103.89881 103.39683 102.93896 102.48108 102.13335 101.78563
## 2012 103.45495 104.54178 105.62861 106.21143 106.79426 107.14076 107.48726
## 2013 103.35562 102.47504 101.59446 100.99105 100.38764 100.12589  99.86415
## 2014 100.42543 100.33188 100.23834 100.15238 100.06642  99.82817  99.58993
## 2015  97.98251  97.29118  96.59985  95.99082  95.38180  94.78287  94.18394
## 2016  96.17369  97.15048  98.12727  99.13544 100.14360 100.71021 101.27682
## 2017 101.52769 101.50575 101.48381 101.51721 101.55061 101.73214 101.91367
## 2018 101.64549 101.45494 101.26439 101.21410 101.16380 101.08608 101.00836
## 2019 101.59115 102.00219 102.41324 102.72356 103.03389 102.94467 102.85544
## 2020 102.33280 102.19633 102.05986                                        
##            Aug       Sep       Oct       Nov       Dec
## 2005  97.58468  97.43024  97.48151  97.53279  97.79634
## 2006 101.02328 101.53572 102.18422 102.83271 103.55616
## 2007 107.37674 107.41859 107.43641 107.45423 107.57450
## 2008 107.49144 107.35214 106.85108 106.35002 105.57283
## 2009 102.89061 103.32823 103.89876 104.46929 104.89514
## 2010 105.59514 105.37631 105.18243 104.98855 104.69467
## 2011 101.47462 101.16361 101.31702 101.47043 102.46269
## 2012 107.51365 107.54004 106.70075 105.86146 104.60854
## 2013 100.09613 100.32810 100.48115 100.63420 100.52981
## 2014  99.42943  99.26894  99.12744  98.98595  98.48423
## 2015  93.96142  93.73891  94.10217  94.46542  95.31956
## 2016 101.31418 101.35154 101.32941 101.30729 101.41749
## 2017 102.09371 102.27375 102.22224 102.17073 101.90811
## 2018 100.91452 100.82068 100.89901 100.97734 101.28425
## 2019 102.77756 102.69967 102.63149 102.56331 102.44806
## 2020

PARTE 4: CALCULO DE LAS TASAS (CENTRADA)

library(dplyr)
library(zoo)
TC %>% as.numeric() %>% as.data.frame()->TC_df
names(TC_df)<-c("TC")
TC_df %>% mutate(T_1_1=(TC/dplyr::lag(TC,n=1)-1)*100,
                 T_1_12=(TC/dplyr::lag(TC,n=12)-1)*100,
                 T_12_12=(rollapply(TC,12,mean,align='right',fill=NA)
                          /rollapply(dplyr::lag(TC,n=12),12,mean,align='right',fill=NA)-1)*100) %>%
  

    
    #REALIZANDO CENTRADO DE TASAS
  
  

          mutate(T_1_12C=dplyr::lead(T_1_12,n = 6),
                 T_12_12C=dplyr::lead(T_12_12,n = 12)) %>% ts(start = c(2005,1),frequency = 12)->tabla_coyuntura
print(tail(tabla_coyuntura,n=12))
##                TC       T_1_1     T_1_12     T_12_12    T_1_12C T_12_12C
## Apr 2019 102.7236  0.30301617  1.4913614 -0.37330363  1.7170487       NA
## May 2019 103.0339  0.30210076  1.8485755 -0.18859088  1.5706239       NA
## Jun 2019 102.9447 -0.08659896  1.8386187  0.01653896  1.1490550       NA
## Jul 2019 102.8554 -0.08667402  1.8286465  0.24224916  0.7300332       NA
## Aug 2019 102.7776 -0.07572131  1.8461575  0.49218953  0.1903276       NA
## Sep 2019 102.6997 -0.07577869  1.8637010  0.76659926 -0.3450458       NA
## Oct 2019 102.6315 -0.06638765  1.7170487  1.01881935         NA       NA
## Nov 2019 102.5633 -0.06643176  1.5706239  1.24869637         NA       NA
## Dec 2019 102.4481 -0.11237592  1.1490550  1.39662567         NA       NA
## Jan 2020 102.3328 -0.11250235  0.7300332  1.46227263         NA       NA
## Feb 2020 102.1963 -0.13335783  0.1903276  1.43253188         NA       NA
## Mar 2020 102.0599 -0.13353591 -0.3450458  1.30757644         NA       NA

PARTE 5: ANALISIS DE RESULTADOS

La distribución resultante una vez procesada toda la muestra es la utilizada para generar las previsiones bajo los distintos escenarios. Tal estimación contiene una frecuencia semetral del período 2005 (enero), al 2019 (septiembre)

tabla_coyuntura %>% as.data.frame() %>% select(T_12_12) %>% ts(start = c(2005,1),frequency = 12) %>% autoplot()

Como la distribución se han eliminado los componentes estacionales e irregulares la tendencia no se encuentra sucia y presenta una tendencia decadente.