Tim Rey

Applied Data Mining for Forecasting Using SAS


Скачать книгу

Creating Time Series Data Using Accumulation Methods

       Creating Data Hierarchies Using Aggregation Methods

       10.2 Statistical Forecast Reconciliation

       10.3 Intermittent Demand

       10.4 High-Frequency Data and Mixed-Frequency Forecasting

       High-Frequency Data

       Mixed-Interval Forecasting

       10.5 Holdout Samples and Forecast Model Selection in Time Series

       Introduction

       10.6 Planning Versus Forecasting and Manual Overrides

       10.7 Scenario-Based Forecasting

       10.8 New Product Forecasting

       Chapter 11 Model Building: Alternative Modeling Approaches

       11.1 Nonlinear Forecasting Models

       11.1.1 Nonlinear Modeling Features

       11.1.2 Forecasting Models Based on Neural Networks

       11.1.3 Forecasting Models Based on Support Vector Machines

       11.1.4 Forecasting Models Based on Evolutionary Computation

       11.2 More Modeling Alternatives

       11.2.1 Multivariate Models

       11.2.2 Unobserved Component Models (UCM)

       Chapter 12 An Example of Data Mining for Forecasting

       12.1 The Business Problem

       12.2 The Charter

       12.3 The Mind Map

       12.4 Data Sources

       12.5 Data Prep

       12.6 Exploratory Analysis and Data Preprocessing

       12.7 X Variable Imputation

       12.8 Variable Reduction and Selection

       12.9 Modeling

       12.10 Summary

       Appendix A

       Appendix B

       References

       Index

      Preface

       It is utterly impossible that a mathematical formula should make the future known to us, and those who think it can would once have believed in witchcraft.

      Jacob Bernoulli, in Ars Conjectadi, 1713

      Curiosity about “what will happen next” is part of human nature, and thus the first attempts at forecasting are found rooted in history. In the ancient and medieval times, prophets like the Oracle of Delphi or Nostradamus had the status of demigods. The situation is significantly different in the 21st century, though, when predicting the future is not divine magic anymore but a necessity in contemporary business. Thousands of professionals are building forecasts in almost all areas of human activity. Since the global recession of 2008–2009, it has been much more widely understood that reliable forecasting is necessary.

      The increased demand for forecasting triggered the development of new methods in addition to the “classical” time series statistical approaches, such as exponential smoothing and the Box-Jenkins AutoRegressive Integrated Moving-Average (ARIMA) models. One fruitful direction of development is that of nonlinear time series modeling, based on various computational intelligence methods, such as neural networks, support vector machines, and genetic programming. Other developments, of special importance to industrial applications, are the efforts for improving the time series forecasts by selecting the best potential drivers using various data mining methods. A short list of such methods includes but is not limited to the following: similarity analysis, sequential pattern matching, Principal Component Analysis (PCA), decision trees, co-integration analysis, variable cluster analysis, stepwise regression, and genetic programming.

      Unfortunately, the available literature for integrating data mining methods in forecasting is very limited. The existing books on the market are either focused on forecasting methods or on data mining approaches. In addition, there are very few references that discuss the numerous practical issues of applying forecasting in a business setting. The practitioner needs a book that addresses the issues of applied industrial forecasting, gives a framework for integrating data mining and time series forecasting, and gives a methodology for large-scale multivariate industrial forecasting.

      Applied Data Mining for Forecasting Using SAS is one of the first books on the market that fills this need.

      Purpose of the Book

      The purpose of the book is to give the reader an industrial perspective concerning applying data mining for forecasting different business activities using some of the most popular software—SAS Institute's range of SAS products including Base SAS, SAS Enterprise Guide, SAS Enterprise Miner, and SAS Forecast Server. The key topics of the book are as follows:

      1 What a practitioner needs to know to successfully apply data mining for forecasting – The first main topic of the book focuses on the ambitious task of giving guidelines to practitioners about building the necessary framework for effective forecasting in a business setting. It covers the issues of justifying the need for industrial forecasting, offering a work process within the popular Six Sigma platform, and discussing the necessary infrastructure and application issues.

      2 How data mining improves forecasting – The second key topic of the book clarifies the important question of using data mining for forecasting. Its main focus is on presenting the key data mining methods for variable reduction and selection and their implementation in SAS.

      3 How to apply data mining for forecasting in practice – The third key topic of the book covers the central point of interest: the application strategy for business forecasting. It includes a short survey of the key contemporary forecasting methods based on time series and illustrates them with appropriate examples from business practices.

      Who This Book Is For

      The targeted audience is much broader than the existing scientific communities in forecasting and data mining. The readers who can benefit from this book are described below:

       Industrial practitioners – This group includes forecasters in a number of different traditional company departments, such as strategy, sales, marketing, finance, supply-chain, purchasing, and so on. They will benefit from the book by understanding the impact of data mining on forecasting and using the discussed forecasting methods and application methodology to broaden and improve their forecast's performance.

       Data miners and modelers – This group consists of the large professional community of users of data mining technologies in different industries. This book will introduce them to contemporary forecasting methods and will demonstrate how they can leverage their data mining skills in the area of industrial forecasting.

       Econometricians – This group includes the key community driving the demand for development and application of time series statistical methods, which is at the basis of industrial forecasting. The book will give them substantial information about data mining methods related to time series forecasting, as well as important feedback from industry