A Python Package for Computing Error Metrics for Observed and Predicted Time Series

Keywords

Hydrology, Water Resource Engineering, Python, Error Metrics, Statistics

Start Date

25-6-2018 10:40 AM

End Date

25-6-2018 12:00 PM

Abstract

Error metrics are statistical measures used to quantify the error or bias of forecasted model data compared to observed data. Error metrics are used extensively in water resource engineering when evaluating hydrologic models to determine the accuracy and applicability of the model. The literature reports a large number of error metrics, however, it is not always clear which metric to use and which metrics are applicable to time series data as different metrics highlight different biases or errors. We created a Python package for hydrologic time series data with over 50 different commonly used error metric functions as well as visualization and data management tools. The functions include error checks to make sure that the input data meets requirements and will return real values The package includes references, explanations, and source code. In this paper we provide an introduction the package, including descriptions of the error metrics implemented and recommendations for use along with examples of use with sample data.

Stream and Session

Stream A: Advanced Methods and Approaches in Environmental Computing

A1: Towards More Interoperable, Reusable and Scalable Environmental Software

This document is currently not available here.

Share

COinS
 
Jun 25th, 10:40 AM Jun 25th, 12:00 PM

A Python Package for Computing Error Metrics for Observed and Predicted Time Series

Error metrics are statistical measures used to quantify the error or bias of forecasted model data compared to observed data. Error metrics are used extensively in water resource engineering when evaluating hydrologic models to determine the accuracy and applicability of the model. The literature reports a large number of error metrics, however, it is not always clear which metric to use and which metrics are applicable to time series data as different metrics highlight different biases or errors. We created a Python package for hydrologic time series data with over 50 different commonly used error metric functions as well as visualization and data management tools. The functions include error checks to make sure that the input data meets requirements and will return real values The package includes references, explanations, and source code. In this paper we provide an introduction the package, including descriptions of the error metrics implemented and recommendations for use along with examples of use with sample data.