README.md

In dpweix/mlmcusum: Detects Changes in Covariance for Multivariate TS

README

Set Up

This repository contains an R package called mlmewma (pairing Machine Learning models with multivariate exponentially weighted moving average (MEWMA) control charts). The functions in this package are used in the paper Monitoring Covariance in Multivariate Time Series: Comparing Machine Learning and Statistical Approaches.

Note that to use the gated recurrent unit (GRU) that the reticulate package must be installed, along with Python, and the modules tensorflow and numpy. Basic usage of the GRU is accomplished through the train_fd function; however, in depth fine tuning of the GRU model will require editing the gru_functions.py file.

library("reticulate")
# Custom package https://github.com/dpweix/mlmewma.git
library("mlmewma")

# Load GRU functions
path_py <- "~/git/mlmewma/inst/python/gru_functions.py"
source_python(path_py)

Functions

| Function name | Description | |---------------|------------------------------------------------------------------------------------------------------------------| | train_fd | Used to train models for fault detection and apply fault detection methods to training data. | | predict_fd | Uses the output of train_fd to make predictions on new data and apply fault detection methods to testing data. |

dpweix/mlmcusum documentation built on July 31, 2023, 10:13 a.m.