setup_model: Set initial parameters for MVAD.

In lanl-ansi/MVAD: MultiVariate Anomaly Detection

Description

setup_model returns the initial mvad object created be initializing parameters.

Usage

setup_model(
  num_variables = 4,
  tau = 10 * 60 * 30,
  data_resolution = 15,
  num_lag_terms = 1
)

Arguments

num_variables	Integer - defining ncol of training/testing data tables.
tau	Integer - refering to number of data points to train model over. i.e., 10 mins @ 30Hz = 10 * 60 * 30.
data_resolution	Integer - number of terms to be 'smoothed' for per-point by the model estimation.
num_lag_terms	Integer - number of lag terms to include in the model. TODO expand from 1 to k

Value

An S3 object of MVAD type.

Examples

dt.ad <- setup_model(4, 10 * 60 * 30, 15, 1)
dt.ad <- setup_model() ## defaults to above

lanl-ansi/MVAD documentation built on July 24, 2020, 3:45 a.m.