Do exercises 8.1, 8.5, 8.6, 8.7, 8.8, 8.9 in Hyndman. Please submit both the link to your Rpubs and the .pdf file with your run code
library("fpp3")
## Registered S3 method overwritten by 'tsibble':
## method from
## as_tibble.grouped_df dplyr
## ── Attaching packages ──────────────────────────────────────────── fpp3 1.0.1 ──
## ✔ tibble 3.2.1 ✔ tsibble 1.1.6
## ✔ dplyr 1.1.4 ✔ tsibbledata 0.4.1
## ✔ tidyr 1.3.1 ✔ feasts 0.4.1
## ✔ lubridate 1.9.3 ✔ fable 0.4.1
## ✔ ggplot2 3.5.1
## ── Conflicts ───────────────────────────────────────────────── fpp3_conflicts ──
## ✖ lubridate::date() masks base::date()
## ✖ dplyr::filter() masks stats::filter()
## ✖ tsibble::intersect() masks base::intersect()
## ✖ tsibble::interval() masks lubridate::interval()
## ✖ dplyr::lag() masks stats::lag()
## ✖ tsibble::setdiff() masks base::setdiff()
## ✖ tsibble::union() masks base::union()
library("tidyverse")
## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ forcats 1.0.0 ✔ readr 2.1.5
## ✔ purrr 1.0.2 ✔ stringr 1.5.1
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ tsibble::interval() masks lubridate::interval()
## ✖ dplyr::lag() masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors
library("ggplot2")
Question 8.1
Consider the the number of pigs slaughtered in Victoria, available in the aus_livestock dataset.
a.) Use the ETS() function to estimate the equivalent model for
simple exponential smoothing. Find the optimal values of
α and ℓ0, and generate forecasts for the next four months.
data("aus_livestock")
aus_livestock
## # A tsibble: 29,364 x 4 [1M]
## # Key: Animal, State [54]
## Month Animal State Count
## <mth> <fct> <fct> <dbl>
## 1 1976 Jul Bulls, bullocks and steers Australian Capital Territory 2300
## 2 1976 Aug Bulls, bullocks and steers Australian Capital Territory 2100
## 3 1976 Sep Bulls, bullocks and steers Australian Capital Territory 2100
## 4 1976 Oct Bulls, bullocks and steers Australian Capital Territory 1900
## 5 1976 Nov Bulls, bullocks and steers Australian Capital Territory 2100
## 6 1976 Dec Bulls, bullocks and steers Australian Capital Territory 1800
## 7 1977 Jan Bulls, bullocks and steers Australian Capital Territory 1800
## 8 1977 Feb Bulls, bullocks and steers Australian Capital Territory 1900
## 9 1977 Mar Bulls, bullocks and steers Australian Capital Territory 2700
## 10 1977 Apr Bulls, bullocks and steers Australian Capital Territory 2300
## # ℹ 29,354 more rows
help("aus_livestock")
pig_meat = aus_livestock %>%
filter(State == "Victoria", Animal == "Pigs")
pig_meat |>
autoplot(Count) +
labs(y = "Pigs Slaughtered", title = "Pig Production in Victoria")
# Use the ETS function to find the optimal values of alpha and level values , # SES = Simple Exponential Smoothing
pig_fit <- pig_meat |>
model(SES = ETS(Count ~ error("A") + trend("N") + season("N")))
# ^ help us find the model
# use report to get the values from the model
model_report = pig_fit %>%
report()
## Series: Count
## Model: ETS(A,N,N)
## Smoothing parameters:
## alpha = 0.3221247
##
## Initial states:
## l[0]
## 100646.6
##
## sigma^2: 87480760
##
## AIC AICc BIC
## 13737.10 13737.14 13750.07
In the report we find the optimal valiues for alpha is 0.3221247 and l[0] is 100646.6
To find the next four months or four periods
pig_forecast = pig_fit %>%
forecast( h = 4) # four months ahead
# we are looking for the next four months so i can begin on later data like 2016
pig_forecast %>%
autoplot(pig_meat %>% filter(Month >= yearmonth("2016 Jan"))) %>%
labs(y = "Pigs Slaughtered", title = "Pig Production in Victoria")
## [[1]]
##
## $y
## [1] "Pigs Slaughtered"
##
## $title
## [1] "Pig Production in Victoria"
##
## attr(,"class")
## [1] "labels"
b.) Compute a 95% prediction interval for the first forecast using ^ y ± 1.96s where s is the standard deviation of the residuals. Compare your interval with the interval produced by R.
# first need get the residuals to calculate the standard deviation
pig_residuals = residuals(pig_fit)
pig_sd = sd(pig_residuals$.resid,na.rm = TRUE)
num1_pig_forecast = pig_forecast$.mean[1] # get the first pig forecast, in the first row
# we can can find the plus and minus of the 1.96 confidence interval , and multiplying it by the standard deviations
upper_bound_pig = pig_forecast$.mean[1] - (pig_sd * 1.96)
lower_bound_pig = pig_forecast$.mean[1] + (pig_sd * 1.96)
print(upper_bound_pig)
## [1] 76871.01
print(lower_bound_pig)
## [1] 113502.1
Now we have to compare prediction intervals
# The hilo() function converts the forecast distributions into intervals (Chapter 5.5)
pig_pred_interval = hilo(pig_forecast, level = 95) # confidence level in 95 percent
In the pig meat production in the state of Victoria, in Jan 2019, the production of pig within the percent confidence interval is 76871 to 113502 pig produced.
Question 8.5
Data set global_economy contains the annual Exports from many countries. Select one country to analyse.
data("global_economy")
global_economy
## # A tsibble: 15,150 x 9 [1Y]
## # Key: Country [263]
## Country Code Year GDP Growth CPI Imports Exports Population
## <fct> <fct> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 Afghanistan AFG 1960 537777811. NA NA 7.02 4.13 8996351
## 2 Afghanistan AFG 1961 548888896. NA NA 8.10 4.45 9166764
## 3 Afghanistan AFG 1962 546666678. NA NA 9.35 4.88 9345868
## 4 Afghanistan AFG 1963 751111191. NA NA 16.9 9.17 9533954
## 5 Afghanistan AFG 1964 800000044. NA NA 18.1 8.89 9731361
## 6 Afghanistan AFG 1965 1006666638. NA NA 21.4 11.3 9938414
## 7 Afghanistan AFG 1966 1399999967. NA NA 18.6 8.57 10152331
## 8 Afghanistan AFG 1967 1673333418. NA NA 14.2 6.77 10372630
## 9 Afghanistan AFG 1968 1373333367. NA NA 15.2 8.90 10604346
## 10 Afghanistan AFG 1969 1408888922. NA NA 15.0 10.1 10854428
## # ℹ 15,140 more rows
I will be using the country of Ghana
ghana_export <- global_economy |>
filter(Country == "Ghana")
a.) Plot the Exports series and discuss the main features of the data.
head(ghana_export)
## # A tsibble: 6 x 9 [1Y]
## # Key: Country [1]
## Country Code Year GDP Growth CPI Imports Exports Population
## <fct> <fct> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 Ghana GHA 1960 1217230038. NA NA 35.4 28.2 6652287
## 2 Ghana GHA 1961 1302674264. 3.43 NA 36.5 26.1 6866539
## 3 Ghana GHA 1962 1382515590. 4.11 NA 28.5 24.2 7085464
## 4 Ghana GHA 1963 1540797517. 4.41 NA 27.5 21.2 7303432
## 5 Ghana GHA 1964 1731296119. 2.21 0.00131 23.9 19.9 7513289
## 6 Ghana GHA 1965 2053462872. 1.37 0.00166 26.7 17.1 7710549
ghana_export %>%
autoplot(Exports) +
labs(title = 'Ghana Export Trends: From 1960 to 2017', subtitle = "Annual Exoport Value Over Time")
This graph shows the Ghana exports form 1960 to 2000. The exports started around 28 units and steadily dropped throughout the years until the 1980s. From the mid 1980s to the year 2000, exports grew steadily, where it hit its peak of 50 units, with some fluctuations throughout the growth. The Ghana exports relies most on commodities like gold, oil and cocoa. These commodities are heavily impact by global market polices like tariffs and interest rates. After the year 2000, there was sharp decline, with the 2008 recession impact the economy. However, there there was been slow recovery on the export of commodities.
b.) Use an ETS(A,N,N) model to forecast the series, and plot the forecasts.
# using the ANN model to forecast the series
gh_fit = ghana_export %>%
model(SES = ETS(Exports ~ error("A") + trend("N") + season("N")))
gh_ex_report = gh_fit %>%
report()
## Series: Exports
## Model: ETS(A,N,N)
## Smoothing parameters:
## alpha = 0.9755569
##
## Initial states:
## l[0]
## 28.14357
##
## sigma^2: 20.1376
##
## AIC AICc BIC
## 413.6206 414.0650 419.8019
gh_forecast = gh_fit %>%
forecast( h = 8) # 8 years forecasted
gh_forecast %>%
autoplot(ghana_export) + labs(title = 'Ghana Export Forecast For 8 Years', subtitle = "Annual Exoport Value Over Time")
b.) Compute the RMSE values for the training data.
gh_accuracy =accuracy(gh_fit) # We use accuracy function RSME values
print(gh_accuracy) # RSME to help find how far the predicted values are from the actual vlaues
## # A tibble: 1 × 11
## Country .model .type ME RMSE MAE MPE MAPE MASE RMSSE ACF1
## <fct> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 Ghana SES Training 0.124 4.41 3.10 -2.10 16.6 0.986 0.991 0.000306
gh_accuracy %>%
pull(RMSE) # pull out the RMSE value
## [1] 4.409445
d.) Compare the results to those from an ETS(A,A,N) model. (Remember that the trended model is using one more parameter than the simpler model.) Discuss the merits of the two forecasting methods for this data set.
gh_aan_model = ghana_export %>%
model( Additive = ETS(Exports ~ error("A") + trend("A") + season("N")))
accuracy(gh_aan_model) # The ANN model is the additive used for constant variance, also called the Holts Linear Trend Model
## # A tibble: 1 × 11
## Country .model .type ME RMSE MAE MPE MAPE MASE RMSSE ACF1
## <fct> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 Ghana Additive Training 0.0154 4.41 3.12 -2.80 16.8 0.993 0.991 -0.000769
gh_aan_accuracy = accuracy(gh_aan_model)
Comparing them both
print(gh_accuracy)
## # A tibble: 1 × 11
## Country .model .type ME RMSE MAE MPE MAPE MASE RMSSE ACF1
## <fct> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 Ghana SES Training 0.124 4.41 3.10 -2.10 16.6 0.986 0.991 0.000306
print(gh_aan_accuracy)
## # A tibble: 1 × 11
## Country .model .type ME RMSE MAE MPE MAPE MASE RMSSE ACF1
## <fct> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 Ghana Additive Training 0.0154 4.41 3.12 -2.80 16.8 0.993 0.991 -0.000769
The Simple Exponential Smoothing Model has a lower RSME (4.4094) than our Additive Model of 4.4106 indicating it will have more accurate prediction of the export trend for the next 8 years. The ANN model observes trend for constant increase and decrease, however our original data show huge magnitude of fluctuations throughout the data, an AAN will not be useful for forecasting. Our SES model does perform better than the ANN model , because of our weak and strong our trend develop on sharp and low levels.
e.) Compare the forecasts from both methods. Which do you think is best?
gh_forecast_sse = gh_fit %>%
forecast(h = 8) # Using 8 years in the future for forecast
gh_forecast_aan = gh_aan_model %>% # forecast for addtive model
forecast(h = 8)
autoplot(ghana_export, Exports) +
autolayer(gh_forecast_sse, series = "Simple Exponential Smoothing (ANN Model)", color = "blue") +
autolayer(gh_forecast_aan, series = "Additive Model (AAN)", color = "red") +
labs(
title = "Ghana Export Forecast in 8 Years)",
subtitle = " SES vs Holts Additive Models",
y = "Exports"
)
## Warning in ggdist::geom_lineribbon(without(intvl_mapping, "colour_ramp"), :
## Ignoring unknown parameters: `series`
## Warning in geom_line(mapping = without(mapping, "shape"), data =
## unpack_data(object[single_row[["FALSE"]], : Ignoring unknown parameters:
## `series`
## Warning in ggdist::geom_lineribbon(without(intvl_mapping, "colour_ramp"), :
## Ignoring unknown parameters: `series`
## Warning in geom_line(mapping = without(mapping, "shape"), data =
## unpack_data(object[single_row[["FALSE"]], : Ignoring unknown parameters:
## `series`
## Scale for fill_ramp is already present.
## Adding another scale for fill_ramp, which will replace the existing scale.
f.) Calculate a 95% prediction interval for the first forecast for each model, using the RMSE values and assuming normal errors. Compare your intervals with those produced using R.
num1_sse_forecast = gh_forecast %>%
slice(1) %>% # get the first row in the SES Model
pull(.mean) # get the mean value
nuum1_add_forecast = gh_forecast_aan %>%
slice(1) %>% # get the first row for additive model
pull(.mean)
rmse_sse = gh_accuracy %>%
pull(RMSE)
rmse_add = gh_aan_accuracy %>%
pull(RMSE)
# Simple Smoothing Prediction Interval in the 95 percent confidence interval
lower_sse <- num1_sse_forecast - 1.96 * rmse_sse
upper_sse <- num1_sse_forecast+ 1.96 * rmse_sse
# Holts Linear Prediction Interval
lower_add <- nuum1_add_forecast - 1.96 * rmse_add
upper_add <- nuum1_add_forecast + 1.96 * rmse_add
# we can use the "cat" to show the results
cat("Simple Exponential Smoothing ETS Method 95% Confidence Interval: [", lower_sse, ",", upper_sse, "]\n")
## Simple Exponential Smoothing ETS Method 95% Confidence Interval: [ 26.53294 , 43.81796 ]
cat("Holt's Additive Trend in 95% Confidence Interval: [", lower_add, ",", upper_add, "]\n")
## Holt's Additive Trend in 95% Confidence Interval: [ 26.65874 , 43.94842 ]
Question 8.6 Forecast the Chinese GDP from the global_economy data set using an ETS model. Experiment with the various options in the ETS() function to see how much the forecasts change with damped trend, or with a Box-Cox transformation. Try to develop an intuition of what each is doing to the forecasts.
china_gdp = global_economy %>%
filter(Country == 'China')
china_gdp %>% autoplot(GDP) +
labs(title = 'China Gross Domestic Product')
Experiment with the various options in the ETS() function to see how
much the forecasts change with damped trend
china_sse_fit = china_gdp %>%
model(
SSE = ETS(GDP ~ error("A") + trend("N") + season("N")))
china_holt_fit = china_gdp %>%
model(Holt = ETS(GDP ~ error("A") + trend("A") + season("N")))
china_damped_fit = china_gdp %>%
model(
Damped = ETS(GDP ~ error("A") + trend("Ad") + season("N"))
)
# Forecasting for the next 15 years
china_forecast_sse = china_sse_fit %>%
forecast(h = 15)
china_forecast_holt = china_holt_fit %>%
forecast(h = 15)
china_forecast_damp = china_damped_fit %>%
forecast(h = 15)
china_forecast_sse %>%
autoplot(china_gdp) + autolayer(china_forecast_holt, series = "Holt", colour = "yellow") + autolayer(china_forecast_damp, series = "Damped", color = "red") +
theme_update() + labs(title = "China GDP Forecast For 8 Years", subtitle = "Comparison of ETS Model: Simple Exponential Smoothing, Holt's Linear Trend, and Damped Models ")
## Warning in ggdist::geom_lineribbon(without(intvl_mapping, "colour_ramp"), :
## Ignoring unknown parameters: `series`
## Warning in geom_line(mapping = without(mapping, "shape"), data =
## unpack_data(object[single_row[["FALSE"]], : Ignoring unknown parameters:
## `series`
## Scale for fill_ramp is already present.
## Adding another scale for fill_ramp, which will replace the existing scale.
## Warning in ggdist::geom_lineribbon(without(intvl_mapping, "colour_ramp"), : Ignoring unknown parameters: `series`
## Ignoring unknown parameters: `series`
## Scale for fill_ramp is already present.
## Adding another scale for fill_ramp, which will replace the existing scale.
Question 8.7
Find an ETS model for the Gas data from aus_production and forecast the next few years. Why is multiplicative seasonality necessary here? Experiment with making the trend damped. Does it improve the forecasts?
data("aus_production")
aus_production
## # A tsibble: 218 x 7 [1Q]
## Quarter Beer Tobacco Bricks Cement Electricity Gas
## <qtr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 1956 Q1 284 5225 189 465 3923 5
## 2 1956 Q2 213 5178 204 532 4436 6
## 3 1956 Q3 227 5297 208 561 4806 7
## 4 1956 Q4 308 5681 197 570 4418 6
## 5 1957 Q1 262 5577 187 529 4339 5
## 6 1957 Q2 228 5651 214 604 4811 7
## 7 1957 Q3 236 5317 227 603 5259 7
## 8 1957 Q4 320 6152 222 582 4735 6
## 9 1958 Q1 272 5758 199 554 4608 5
## 10 1958 Q2 233 5641 229 620 5196 7
## # ℹ 208 more rows
gas_prod = aus_production %>%
select(Quarter,Gas)
gas_prod %>%
autoplot()
## Plot variable not specified, automatically selected `.vars = Gas`
gas_fit = gas_prod %>%
model(ETS(Gas))
report(gas_fit)
## Series: Gas
## Model: ETS(M,A,M)
## Smoothing parameters:
## alpha = 0.6528545
## beta = 0.1441675
## gamma = 0.09784922
##
## Initial states:
## l[0] b[0] s[0] s[-1] s[-2] s[-3]
## 5.945592 0.07062881 0.9309236 1.177883 1.074851 0.8163427
##
## sigma^2: 0.0032
##
## AIC AICc BIC
## 1680.929 1681.794 1711.389
gas_fit_forecast = gas_fit %>%
forecast(h = 12)
gas_fit_forecast %>%
autoplot(gas_prod %>% filter(Quarter >= yearquarter("2005 Q1")))
Now with multiplicative seasonality ^ same model from before ( Rewrote code here)
gas_multi_fit = gas_prod %>%
model( Mulit = ETS(Gas ~ error("M") + trend("A") + season("M")))
gas_multi_fit_forecast = gas_multi_fit %>%
forecast(h = 12)
gas_multi_fit_forecast %>%
autoplot(gas_prod %>% filter(Quarter >= yearquarter("2005 Q1"))) + labs(title = "Multiplicative Seasonality Model for Gas Production Forecasting ")
Experimenting with the damped additive model
gas_damped_fit = gas_prod %>%
model( Damped = ETS(Gas ~ error("A") + trend("Ad") + season("M"))) # Ad for damped, trend can gradually decrease or become stable
gas_forecast_damped <- gas_damped_fit %>%
forecast(h = 12)
gas_forecast_damped %>%
autoplot(gas_prod %>% filter(Quarter >= yearquarter("2005 Q1"))) + labs(title = "Damped Additive Model for Gas Production Forecasting")
Multiplicative Seasonality is necessary here due the seasonal peak seen through out the quarters. The demand for gas is most dependent on the month the consumers are using their gas powered product. The seasonal fluctuations are observed to increased proportionally in the series, and multiplicative model will capture trends increasing in a proportional level.However, using the damped additive with a multiplicative seasonal component shows that we expect the demand for gas to stay stable overtime. This approach to observing the data, is not useful as in the future demand for gas been influenced by factors such as rise in population, construction of new houses, and technoligcal advances..
Question 8.8
Recall your retail time series data (from Exercise 7 in Section 2.10).
set.seed(12345678)
myseries <- aus_retail |>
filter(`Series ID` == sample(aus_retail$`Series ID`,1))
myseries %>%
autoplot()
## Plot variable not specified, automatically selected `.vars = Turnover`
a.) Why is multiplicative seasonality necessary for this series?
We use multiplicative seasonality due to the seasonal turnover fluctuation increasing throughout the years. There are peaks in which people are hired for seasonal positions which a retial locaiton expects an high increase of demand. This months are summer and the winter, combined with popular holidays.
b.) Apply Holt-Winters’ multiplicative method to the data. Experiment with making the trend damped.
retail_multi_fit = myseries %>%
model( Holt_Winters = ETS(Turnover ~ error("M") + trend("A") + season("M")),
Damped_Winters = ETS(Turnover ~ error("M") + trend("Ad") + season("M"))
)
retail_holt_forecast = retail_multi_fit %>%
forecast(h = 12)
retail_holt_forecast %>%
autoplot(myseries) + labs(title = " Turnover in Retail", subtitle = "Holt-Winters vs Damped Winters Models" )
c.) Compare the RMSE of the one-step forecasts from the two methods. Which do you prefer?
retail_accuracy = accuracy(retail_multi_fit)
view(retail_accuracy)
Comparing our RMSE values in the one step forecast with the Holt Winters and Damped Winters, we get the value 0.6130408 for Holts Winter and 0.6155067 for Damped Winters model. The Holts Winter model has lower RMSE, indicating a better fit for accuary in the forecasting data.
d.) Check that the residuals from the best method look like white noise.
retail_multi_fit %>%
select(Holt_Winters) %>%
gg_tsresiduals()
e.) Now find the test set RMSE, while training the model to the end of 2010. Can you beat the seasonal naïve approach from Exercise 7 in Section 5.11?
myseries2010 = myseries %>%
filter(year(Month) < 2011)
myseries2010_fit = myseries2010 %>%
model( Holt_Winters = ETS(Turnover ~ error("M") + trend("A") + season("M")),
Damped_Winters = ETS(Turnover ~ error("M") + trend("Ad") + season("M")),
Seasonal_Naive = SNAIVE(Turnover) # seasonal naive method from chapter 5
)
myseries2010_forecast = myseries2010_fit %>%
forecast(new_data = anti_join(myseries, myseries2010)) # using anti join to filer out rows that been in myseries already, make the data much cleaner
## Joining with `by = join_by(State, Industry, `Series ID`, Month, Turnover)`
myseries2010_forecast %>%
autoplot(myseries %>% filter(Month >= yearmonth("2010 Jan")), level = NULL) +
labs(y = "Turnover", title = "2010 and Beyond Retial Turnover Forecasting ", subtitle = "Holt_Winters vs Damped_Winters vs Seasonal_Naive ") +
guides(color = guide_legend(title = "Model"))
Question 8.9 For the same retail data, try an STL decomposition applied to the Box-Cox transformed series, followed by ETS on the seasonally adjusted data. How does that compare with your best previous forecasts on the test set?
series_lambda <- myseries2010 |>
features(Turnover, features = guerrero) |> # function to find lambda
pull(lambda_guerrero)
series_boxcox = myseries2010 %>%
mutate(
Boxcox = box_cox(Turnover,series_lambda)
) # make row for the boxcox of each turnover, box used for normal distribtion of data
series_boxcox_fit <- series_boxcox |>
model(
ETS = ETS(Boxcox), # find the best compoenents
Seasonal_Component = STL(Boxcox ~ season(window = "periodic")) # for decomposing the compoents
)
# Graph looked funky b4, fixed now adjusted with element text(horizonal adjustment)
series_boxcox_fit |>
components() |>
autoplot() +
labs(
title = "Decomposition of Box-Cox Transformed Series",
subtitle = "Trend, Seasonal, and Remainder Components of Boxcox ",
y = "Turnover Boxcox",
x = "Month"
) +
theme_minimal() + # Clean theme
theme(
plot.title = element_text(hjust = 0.5), # bring the title to middle of the graph
plot.subtitle = element_text(hjust = 0.5),
legend.position = "bottom" # move legend from the side to the bottom to looker better
) +
scale_color_brewer(palette = "Set2") #ggplot has lots fo color palette so this looks
## Warning: Removed 12 rows containing missing values or values outside the scale range
## (`geom_line()`).
box_cox_hw = myseries2010 %>%
model(
MAN_model = ETS(Turnover ~ error("M") + trend("A") + season("N")))
box_cox_hw_forecast = box_cox_hw %>%
forecast(h = 12)
box_cox_hw_forecast %>%
autoplot(myseries2010) + labs(title = "Retial Turnover Boxcox Forecast for 12 Months")