Data management"
In fHMM: Fitting Hidden Markov Models to Financial Data

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.dim = c(10,6),
  out.width = "80%",
  fig.align = 'center',
  fig.path = "fHMM-"
)

This vignette^[This vignette was build using R r paste(R.Version()[c("major", "minor")], collapse = ".") with the {fHMM} r utils::packageVersion("fHMM") package.] explains how to prepare or simulate data in {fHMM} for estimation.

library("fHMM")

Empirical data

Empirical data can be provided either as a data.frame or as a comma-separated values (.csv) file, see the vignette on specifying the controls for details.^[The download_data() function explained below provides a convenient tool for downloading stock data from https://finance.yahoo.com/ in csv-format.] The {fHMM} package comes with two datasets of the Deutscher Aktienindex and the VW stock for demonstration purpose that can be accessed as follows:

system.file("extdata", "dax.csv", package = "fHMM")
system.file("extdata", "vw.csv", package = "fHMM")

The prepare_data() function prepares the data based on the data controls specifications and returns an fHMM_data object that can be passed to the fit_model() function for model fitting.

controls <- list(
  states = 3,
  sdds   = "t",
  data   = list(file        = system.file("extdata", "dax.csv", package = "fHMM"),
                date_column = "Date",
                data_column = "Close",
                logreturns  = TRUE)
)
controls <- set_controls(controls)
data <- prepare_data(controls)
summary(data)

Download stock data

Daily stock prices listed on https://finance.yahoo.com/ can be downloaded directly via

download_data(symbol, from, to, file)

where

symbol is the stock's symbol that has to match the official symbol on https://finance.yahoo.com/,^[For example, "^GDAXI" is the symbol of the DAX and "^GSPC" the one of the S\&P 500.]
from and to define the time interval (in format "YYYY-MM-DD"),
file is the name of the file where the .csv-file is saved. Per default, it is saved in the current working directory under the name <symbol>.csv. If file = NULL (default), the data is not saved but returned as a data.frame.

For example, the call

dax <- download_data(symbol = "^GDAXI", from = "2000-01-01", to = Sys.Date())
head(dax)

downloads the 21st century daily data of the DAX into the current working directory.

Highlighting events

Historical events can be highlighted by specifying a named list events with elements dates (a vector of dates) and labels (a vector of labels for the events) and passing it to the plot method, for example:

events <- fHMM:::fHMM_events(
  list(
    dates = c("2001-09-11","2008-09-15","2020-01-27"),
    labels = c("9/11 terrorist attack","Bankruptcy of Lehman Brothers","First COVID-19 case in Germany")
    )
  )
print(events)
plot(data, events = events)

Simulated data

If the data parameter in the model's controls is unspecified, the model is fitted to simulated data from the model specification. This can be useful for testing the functionality or conducting simulation experiments. True model parameters can be specified by defining an fHMM_parameters-object via the fHMM_parameters() function and passing it to prepare_data().