Demonstration of an open source platform for reproducible comparison of predictive models

,
  Department of Electronics and Communication, Visvesvaraya National Institute of Technology, Nagpur, India, Nagpur, Maharashtra, India
DOI
10.7287/peerj.preprints.2251v1
Published
Accepted
Subject Areas
Data Mining and Machine Learning, Data Science, Scientific Computing and Simulation, Software Engineering, Visual Analytics
Keywords
Predictive analysis, time series prediction, R Package, Performance evaluation, Predictive models, Test bench
Copyright
© 2016 Bokde et al.
Licence
This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, reproduction and adaptation in any medium and for any purpose provided that it is properly attributed. For attribution, the original author(s), title, publication source (PeerJ Preprints) and either DOI or URL of the article must be cited.
Cite this article
Bokde N, Kulat K. 2016. Demonstration of an open source platform for reproducible comparison of predictive models. PeerJ Preprints 4:e2251v1

Abstract

This paper discusses about a tool PredictTestbench, which is an R package which provides a testbench to do comparison of prediction methods. This package compares a proposed time series prediction method with other default methods like Autoregressive integrated moving average (ARIMA) and Pattern Sequence based Forecasting (PSF). The testbench is not limited to these methods. It allows user to add or remove multiple numbers of methods in the existing methods in the study. By default, testbench compares different imputation methods considering different error metrics RMSE, MAE or MAPE. Along with this, it facilitates user to add new error metrics as per requirements. The simplicity of the package usage and significant reduction in efforts and time consumption in state of art procedure, adds valuable advantage to it. The aim of the testbench is reduce the efforts for coding, experiments on output visualization and time for different steps involved in such study. This paper explains the use of all functions in PredictTestbench package with the demonstration of examples.

Author Comment

This is a preprint submission to PeerJ Preprints. This manuscript is submitted to a peer reviewed journal.