MultivariateEDA: R6 class MultivariateEDA

In josephsdavid/tswgewrapped: Helpful wrappers for 'tswge', 'vars' and 'nnfor' time series packages

Description

R6 class MultivariateEDA

R6 class MultivariateEDA

Methods

Public methods

• MultivariateEDA$new()

• MultivariateEDA$get_data()

• MultivariateEDA$get_var_interest()

• MultivariateEDA$get_var_time()

• MultivariateEDA$get_data_var_interest()

• MultivariateEDA$set_verbose()

• MultivariateEDA$plot_data()

• MultivariateEDA$plot_scatterplots()

• MultivariateEDA$plot_ccf_analysis()

• MultivariateEDA$plot_lag_plots()

• MultivariateEDA$clone()

---

Method new()

Initialize an object to perform EDA on a Multivariate Time Series

Usage

MultivariateEDA$new(data = NA, var_interest = NA, var_time = NA, verbose = 0)

Arguments

data

The dataframe (or type that can be coerced to a dataframe) containing the time series realizations

var_interest

The output variable of interest (dependent variable)

var_time

If the dataframe has a time column, what is the name of this column?

verbose

How much to print during the model building and other processes (Default = 0)

Returns

A new 'MultivariateEDA' object.

---

Method get_data()

Returns the time series realization

Usage

MultivariateEDA$get_data(time = NA)

Arguments

time

NA will return the original data without the 'var_time' column 'original' will return the original data with any 'var_time' column (if applicable) 'sub' will return the original data without any 'var_time' column (if applicable) but with a substitute 'Time' column which is equal to the observation number

Returns

The Time Series Realization

---

Method get_var_interest()

Returns the dependent variable name

Usage

MultivariateEDA$get_var_interest()

Returns

The dependent variable name

---

Method get_var_time()

Returns the time variable

Usage

MultivariateEDA$get_var_time()

Returns

The time variable

---

Method get_data_var_interest()

Returns the dependent variable data only

Usage

MultivariateEDA$get_data_var_interest()

Returns

The dependent variable data only

---

Method set_verbose()

Adjust the verbosity level

Usage

MultivariateEDA$set_verbose(verbose = 0)

Arguments

verbose

0 = Minimal Printing only (usualy limited to step being performed) 1 = Basic printing of model builds, etc. 2 = Reserved for debugging mode. May slow down the run due to excessive printing, especially when using batches

---

Method plot_data()

Plots the time series with all the dependent variables

Usage

MultivariateEDA$plot_data(ncol = 1, scales = "free_y", ...)

Arguments

ncol

Number of columns to use to show the data

scales

The scales argument to be passed to ggplot facet_wrap layer (Default = 'free_y') Other appropriate options: 'fixed'

...

Arguments to pass to facet wrap. Example "ncol = 3, scales = 'free_y'"

---

Method plot_scatterplots()

Plots the scatterplots matric of all the variables in the dta

Usage

MultivariateEDA$plot_scatterplots()

---

Method plot_ccf_analysis()

Plots the CCF function for the dependent variable against all independent variables

Usage

MultivariateEDA$plot_ccf_analysis(lag.max = 12, negative_only = TRUE)

Arguments

lag.max

The maximum lag to evaluate

negative_only

Whether to take max cross correlation of only negative lags for the independent variables. Many times durign predictions, we dont have future values available for the independent variables In such cases, we can not use positive lag values for predictions. (Default = TRUE)

Returns

A dataframe containing (1) 'variable': the dependent variable name, (2) 'max_ccf_index': lag at which max cross correlation occurs (3) 'max_ccf_value': max cross correlation value (abs) (4) 'max_ccf_index_adjusted': adjusted index (if negative_only is FALSE, then this will show a value capped at lag = 0 for any positive lag index). User may then decide to use either this value or the 'max_ccf_index' dependig on if the positive lag values of the dependent variable will be available for the prediction

---

Method plot_lag_plots()

Usage

MultivariateEDA$plot_lag_plots()

---

Method clone()

The objects of this class are cloneable with this method.

Usage

MultivariateEDA$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

josephsdavid/tswgewrapped documentation built on July 31, 2020, 9:36 a.m.