CHAPTER 11Summary: A Best Practice Blueprint for Using Your Software
11.1 INTRODUCTION
Demand forecasting is challenging and the performance of your forecasts depends largely on having the right facilities available in your software to support you in this task. In this last chapter, we will draw up a shopping list of desirable features of forecasting software products. We will then summarize the key messages in the book to provide a blueprint for best practice.
11.2 DESIRABLE CHARACTERISTICS OF FORECASTING SOFTWARE
In addition to obvious features – such as ease-of-use, speed, and accuracy of calculation, a wide range of available methods, the ability to interface with other systems, and the availability of training and support—there are a number of more specific characteristics to look for when choosing forecasting software. We discuss these here, and they are summarized in Table 11.1.
Table 11.1 Desirable Features of Forecasting Software
Desirable Feature | ✓ |
Data Preparation | |
Handles missing values | |
Handles outliers | |
Allows trading-day adjustments | |
Allows transformations to meet model assumptions | |
Graphical Displays | |
Clear displays of sales history | |
Seasonal-cycle graphs | |
Scattergraphs | |
Residual plots | |
Method Selection | |
Automatic method selection available | |
Explanations for selection of methods | |
Flagging of forecasts that need attention | |
Separates fitting and out-of-sample periods | |
Reports multiple measures of fit, accuracy, and bias | |
Allows rolling-origin evaluation ... |
Get Profit From Your Forecasting Software now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.