ctsmr - Continuous Time Stochastic Modeling in R

TL;DR

ctsmr is an R package that enables the construction and estimation of continuous-time stochastic models for multivariate time series data, facilitating grey-box modeling of physical systems.

Contribution

The paper introduces ctsmr, a comprehensive R package for modeling and estimating continuous-time stochastic systems using maximum likelihood and Kalman filtering.

Findings

Successfully models physical phenomena with multivariate data

Demonstrates estimation of embedded parameters in stochastic models

Provides practical examples showcasing ctsmr's capabilities

Abstract

ctsmr is an R package providing a general framework for identifying and estimating partially observed continuous-discrete time gray-box models. The estimation is based on maximum likelihood principles and Kalman filtering efficiently implemented in Fortran. This paper briefly demonstrates how to construct a Continuous Time Stochastic Model using multivariate time series data, and how to estimate the embedded parameters. The setup provides a unique framework for statistical modeling of physical phenomena, and the approach is often called grey box modeling. Finally three examples are provided to demonstrate the capabilities of ctsmr.