Sustainability and Social Justice

Development of a Method to Improve Statistical Forecasts Using Interpolation and Cluster Analysis

Document Type

Article

Abstract

In a previous chapter (Klimberg, Ratick, & Smith, 2018), we introduced a novel approach in which cluster centroids were used as input data for the predictor variables of a multiple linear regression (MLR) used to forecast fleet maintenance costs. We applied this approach to a real data set and significantly improved the predictive accuracy of the MLR model. In this chapter, we develop a methodology for adjusting moving average forecasts of the future values of fleet service occurrences by interpolating those forecast values using their relative distances from cluster centroids. We illustrate and evaluate the efficacy of this approach with our previously used data set on fleet maintenance.

Publication Title

Advances in Business and Management Forecasting

Publication Date

1-1-2021

Volume

14

First Page

71

Last Page

85

ISSN

1477-4070

DOI

10.1108/S1477-407020210000014006

Keywords

cluster analysis, fleet maintenance data, forecasting, interpolation, regression, time series data

Link to Full Text

Share

COinS