automatic forecasting systems

Why Autobox?

Autobox Functionality

Autobox will AUTOMATICALLY...

Identify intermittent demand and use a special modeling approach to forecast the "lumpy" demand.

Forecasting

Create Customized Models for each Data Set. No "Pick Best" Here!

Tweaking the Model

Weight observations based on variance, to fix historical "volatility" (i.e. Variance Change detection via the Tsay Test)

Identifying the Relationship

Fundamental Differences in Methodology

DATA

Identify and Correct Outliers and for the Causal Variables to keep the model in the forecast unaffected.

Ignore older data that is part of a different "model" i.e. a parameter change (via the Chow Test)

Identify and incorporate local time trends to predict an accurate LEAD or LAG relationship between the causal variables to be predicted.

We then forecast the possibility we are inaccurate within a certain standard deviation (user set)

To Fix the Model, we check for any thing out of the ordinary.

In order to identify the relationship we use customizable settings

Autobox's all-in-one solution to your statistical needs

To Synthesize Data, we focus heavily on historical data to predict changes in values.

Forecast different

scenarios to

evaluate upside and downside scenarios using different future

values of causals

EARLY WARNING SYSTEM

• Autobox will identify the lead/lag relationship in the causals

• Autobox will remove unimportant Causals

• Autobox will potentially identify and Add 4 Types of Outliers providing “Early Warning Detection”

• Check for Constancy of Parameters

DATA CLEANSING

• Autobox Algorithm (Can be customized by user)

• Autocorrelation Function

• Partial Autocorrelation Function

• Cross Correlation Function

• ARIMA models

• Transfer Function models

FORECASTING

• Historical Data

• Historical Causal Data(i.e. Price, Customer Insight, GNP, Unemployment, Population, etc.)

• Historical Knowledge of Events (outliers, mergers, promotions, holidays etc.)

• Future Values of Causals

• Future Values of

Promotions

WHAT-IF ANALYSIS

Applications of Autobox

IDENTIFY DRIVERS

Still Not Enough?

Our Pricing Varies on Need

We are highly acclaimed

A Brief Introduction to AFS

Talk to us about your needs, and we will recommend the software that will fit your analytical requirements.

Featured in:

After Improving the Pack Program...

• Journal of Forecasting

• Journal of Business Forecasting

• North American Actuarial Journal

• Forest Research and Management Institute

• Environmental and Resource Economics

• Technological Forecasting and Social Change

• Fraud Magazine

• Canadian Journal of Forest Research

• Applied Economics

• Journal of Applied Pharmacology

• Journal of Endocrinology and Metabolism

• Journal of Urban Studies

From Humble Beginnings...

Automatic Forecasting Systems (AFS) was founded in 1976 to launch and market radical improvements in forecasting software. By incorporating causal factors into the forecast such as sales of related products (e.g., shampoo and conditioner), cannibalization effects and even the price of competitors products, a forecasting software with accuracy never before seen hit the market. Over time, events like advertising and promotions, including lead (sales arising ahead of holidays) and lag effects (e.g., reduced sales the week after a promotion ends) were added to further improve the efficacy of Autobox.

Awards:

Beginning in 1968, David Reilly, then at the University of Pennsylvania, developed a rule-based heuristics system that is today the Autobox approach to ARIMA modeling. The main thrust was to provide a mechanized or automated option to the Box-Jenkins (ARIMA) approach to modeling a time series process. AFS, SAS and BMDP were early source code customers in 1972 for the Pack Program, the basic "skeleton" which would give way to forecasting programs. However, the Pack Program was far from perfect.

• Named the "Best Dedicated Forecasting Package" in "Principles of Forecasting" by J. Scott Armstrong

• Placed 1st Among Automated Software for Daily Data in the 2008 NN5 Forecasting Competition

• Placed 2nd in the 2007 NN3 Competition for Monthly Data, and 1st Among More Difficult Data Sets

Looking to the Future..

In the thirty -eight years since its launch AFS continues to focus on that core goal while adding statistical features that are hot out of the statistical journals. AFS's products, across platforms and interfaces, continue to be based on a forecast engine that has been programmed and tuned to do what only the best of forecasters can do; build Box-Jenkins ARIMA and Transfer Function models.

Forecasting Differently

An Automatic Forecasting Systems Presentation