step_ma: Extract moving average features
In bagasbgy/quantrecipes: Extra recipes for quantitative analysis

Description Usage Arguments Details Value Examples

View source: R/step_ma.R

step_ma creates a specification of a recipe step that will extract moving average features from an asset price historical data.

step_ma(recipe, ..., ma_fun = TTR::SMA, n = 10, weights = NULL,
  ma_options = list(), state = FALSE, ratio = TRUE, prefix = "ma",
  prices = NULL, role = "predictor", trained = FALSE, skip = FALSE,
  id = rand_id("ma"))

## S3 method for class 'step_ma'
tidy(x, info = "terms", ...)

`recipe`	A recipe object. The step will be added to the sequence of operations for this recipe.
`...`	One or more selector functions to choose which variables are affected by the step. See selections (from recipes package) for more details.
`ma_fun`	A `function` to extract moving average, or a `character` vector of length one which specify a moving average function. Defaults to TTR::SMA.
`n`	A `numeric` vector of length one which specify the moving average window.
`weights`	A `character` vector of length one that specify a column name, or a `numeric` vector for `ma_fun` that has `wts` or `volume` argument. See details for more information.
`ma_options`	A `list` of additional argument(s) that would be passed to `ma_fun` function.
`state`	An option to specify whether to return the current states of the calculated moving averages. See details for more informations. Defaults to `FALSE`.
`ratio`	Whether to return the moving average `spread` as ratio or absolute difference. See details for more informations. Defaults to `TRUE`.
`prefix`	A `character` vector of length one that would be used as a prefix to the created columns.
`prices`	A container for selected prices columns. Leave to `NULL` as it will be populated by prep() function.
`role`	For model terms created by this step, what analysis role should they be assigned? By default, the function assumes that the created columns will be used as `"predictors"` in a model.
`trained`	A logical to indicate if the necessary informations for preprocessing have been estimated.
`skip`	A logical. Should the step be skipped when the recipe is baked by bake()? While all operations are baked when prep() is run, some operations may not be able to be conducted on new data (e.g. processing the outcome variable(s)). Care should be taken when using `skip = TRUE` as it may affect the computations for subsequent operations
`id`	A character string that is unique to this step to identify it.
`x`	A `step_ma` object.
`info`	Options for `tidy()` method; whether to return tidied information for used `"terms"` or `"params"`

The output from this step are several new columns which contains the extracted moving average features.

For basic output, this step will produces:

value: the moving average value

If state argument is TRUE, it will also produces:

spread: current spread to the prices variable
state: current state of the spread; "bullish" or "bearish"

Note that if ratio argument is TRUE, the spread value would be returned as a ratio to prices variable instead of an absolute difference.

An updated version of recipe with the new step added to the sequence of existing steps (if any).

# import libs
library(quantrecipes)

# basic usage
rec <- recipe(. ~ ., data = btcusdt) %>%
  step_ma(close) %>%
  step_naomit(all_predictors()) %>%
  prep()

# get preprocessed data
juice(rec)

# using state argument
rec <- recipe(. ~ ., data = btcusdt) %>%
  step_ma(close, state = TRUE) %>%
  step_naomit(all_predictors()) %>%
  prep()

# get preprocessed data
juice(rec)

# using pass-through options
rec <- recipe(. ~ ., data = btcusdt) %>%
  step_ma(close,
    ma_fun = TTR::EMA,
    n = 10,
    ma_options = list(wilder = TRUE),
    state = TRUE
  ) %>%
  step_naomit(all_predictors()) %>%
  prep()

# get preprocessed data
juice(rec)

# using custom weights for weighted moving average
rec <- recipe(. ~ ., data = btcusdt) %>%
  step_ma(close,
    ma_fun = TTR::WMA,
    n = 10,
    weights = c(rep(1, 9), 10),
    state = TRUE
  ) %>%
  step_naomit(all_predictors()) %>%
  prep()

# get preprocessed data
juice(rec)

# using volume-based moving average
rec <- recipe(. ~ ., data = btcusdt) %>%
  step_ma(close,
    ma_fun = TTR::VWMA,
    n = 10,
    weights = "volume",
    state = TRUE
  ) %>%
  step_naomit(all_predictors()) %>%
  prep()

# get preprocessed data
juice(rec)