validate_input: Validate Input Data Structure

Description

Validates that input data meets the requirements for SignalY analysis. Performs comprehensive checks on data types, dimensions, missing values, and numeric properties.

Usage

validate_input(
  y,
  X = NULL,
  time_index = NULL,
  na_action = c("fail", "omit", "interpolate"),
  min_obs = 10,
  verbose = FALSE
)

Value

A list with validated and potentially transformed data:

y: Validated numeric vector of target signal
X: Validated matrix of predictors (or NULL)
time_index: Vector of time indices
n_obs: Number of observations
n_vars: Number of predictor variables (or 0)
var_names: Names of predictor variables (or NULL)
na_indices: Indices of removed observations (if any)

Arguments

y: Numeric vector representing the target signal. Must be a numeric vector with no infinite values. Missing values (NA) are handled according to the na_action parameter.
X: Optional matrix or data frame of candidate predictors. If provided, must have the same number of rows as length(y).
time_index: Optional vector of time indices. If NULL, sequential integers will be used.
na_action: Character string specifying how to handle missing values. Options are "fail" (stop with error), "omit" (remove observations with NA), or "interpolate" (linear interpolation).
min_obs: Minimum number of observations required. Default is 10.
verbose: Logical indicating whether to print diagnostic messages.

Details

This function implements defensive programming principles to ensure data integrity before computationally intensive analyses. The validation process includes:

Type checking: Ensures y is numeric and X (if provided) is numeric matrix/data.frame
Dimension checking: Verifies compatible dimensions between y and X
Missing value handling: Processes NA values according to specified action
Finiteness checking: Removes or flags infinite values
Minimum sample size: Ensures sufficient observations for analysis