I am trying to find multiple seasonality in a stock price time series (GOOG). Converting the double closing price column to an integer complains also. There are no NA's and all of the numbers are positive. Ho do I get rid of this error?
Edit 1: Not using a pipe gets the same warning:
fitClose <- model(close,TSLM(Close ~ trend() + season())) # complains
Original code:
rm(list = ls())
require(fpp3)
goog<-gafa_stock|>filter(Symbol=="GOOG",year(Date)==2018)
close<-goog|>select(Close) # just get close
sum(is.na(close)) # no na's
min(close$Close) # min is positive
fitClose <- close |> model(TSLM(Close ~ trend() + season())) # complains
# Warning message:
# 1 error encountered for TSLM(Close ~ trend() + season())
# [1] argument must be coercible to non-negative integer
c<-close |> mutate(Close=as.integer(Close)) # convert to integer
#autoplot(c)
#close|>mutate(Close = Close / first(Close))
fit <- c |> model(TSLM(Close ~ trend() + season())) # still complains
report(fit)
The tsibble has an irregular time index, as you can see below with the "[!]"
library(fpp3)
gafa_stock
#> # A tsibble: 5,032 x 8 [!]
#> # Key: Symbol [4]
#> Symbol Date Open High Low Close Adj_Close Volume
#> <chr> <date> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 AAPL 2014-01-02 79.4 79.6 78.9 79.0 67.0 58671200
#> 2 AAPL 2014-01-03 79.0 79.1 77.2 77.3 65.5 98116900
#> 3 AAPL 2014-01-06 76.8 78.1 76.2 77.7 65.9 103152700
#> 4 AAPL 2014-01-07 77.8 78.0 76.8 77.1 65.4 79302300
#> 5 AAPL 2014-01-08 77.0 77.9 77.0 77.6 65.8 64632400
#> 6 AAPL 2014-01-09 78.1 78.1 76.5 76.6 65.0 69787200
#> 7 AAPL 2014-01-10 77.1 77.3 75.9 76.1 64.5 76244000
#> 8 AAPL 2014-01-13 75.7 77.5 75.7 76.5 64.9 94623200
#> 9 AAPL 2014-01-14 76.9 78.1 76.8 78.1 66.1 83140400
#> 10 AAPL 2014-01-15 79.1 80.0 78.8 79.6 67.5 97909700
#> # ℹ 5,022 more rows
You can re-index it to be daily data (with missing values on non-trading days) using the as_tsibble() function with regular = TRUE:
gafa_stock |>
filter(Symbol=="GOOG", year(Date)==2018) |>
as_tsibble(index = Date, regular = TRUE) |>
model(TSLM(Close ~ trend() + season())) |>
report()
#> Series: Close
#> Model: TSLM
#>
#> Residuals:
#> Min 1Q Median 3Q Max
#> -136.531 -53.911 -7.886 54.088 151.789
#>
#> Coefficients:
#> Estimate Std. Error t value Pr(>|t|)
#> (Intercept) 1113.25932 12.02183 92.603 <2e-16 ***
#> trend() 0.01593 0.04114 0.387 0.699
#> season()week2 -2.82771 13.45490 -0.210 0.834
#> season()week3 NA NA NA NA
#> season()week4 NA NA NA NA
#> season()week5 -7.58925 13.67159 -0.555 0.579
#> season()week6 -2.88701 13.45333 -0.215 0.830
#> season()week7 -1.56729 13.52010 -0.116 0.908
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> Residual standard error: 67.93 on 245 degrees of freedom
#> Multiple R-squared: 0.001919, Adjusted R-squared: -0.01845
#> F-statistic: 0.0942 on 5 and 245 DF, p-value: 0.99306
Created on 2023-06-12 with reprex v2.0.2
season() here has automatically selected weekly seasonality, but due to the missing weekends, two of the seasonal components can't be estimated.
In any case, there is no seasonality of any kind in stock prices in an efficient market.