I'm retiring from SAS (effective November 1), but not from the forecasting community, To continue the conversation, please find me on LinkedIn where I'll be happy to connect (if we aren't already).

No more bad forecasting!

The post The last of the BFD appeared first on The Business Forecasting Deal.

[Sidenote: Join me live in New York City, December 6-7, to hear from Spyros and Nassim at the M5 Conference. I'm looking forward to an amazing event, with two days of presentations and discussion of the M5 Forecasting Competition. Check out the link for agenda and incredible list of speakers providing results, analysis, and commentary.]

Intermittent Demand Forecasting (IDF)

For me, intermittent demand is not so much a forecasting problem as an inventory management problem. As Makridakis and Taleb point out, just because you can't solve a problem with forecasting, doesn't mean there aren't other ways to address it. This same kind of mindset -- this openness to look beyond just throwing models at a problem -- is displayed in the new book Intermittent Demand Forecasting: Content, Methods and Applications by John Boylan and Aris Syntetos. I also appreciate their recognition of the meaningful environmental benefit of better managing our inventories.

From the description:

The first text to focus on the methods and approaches of intermittent, rather than fast, demand forecasting

Intermittent Demand Forecasting is for anyone who is interested in improving forecasts of intermittent demand products, and enhancing the management of inventories. Whether you are a practitioner, at the sharp end of demand planning, a software designer, a student, an academic teaching operational research or operations management courses, or a researcher in this field, we hope that the book will inspire you to rethink demand forecasting. If you do so, then you can contribute towards significant economic and environmental benefits.

No prior knowledge of intermittent demand forecasting or inventory management is assumed in this book. The key formulae are accompanied by worked examples to show how they can be implemented in practice. For those wishing to understand the theory in more depth, technical notes are provided at the end of each chapter, as well as an extensive and up-to-date collection of references for further study. Software developments are reviewed, to give an appreciation of the current state of the art in commercial and open source software.

Thanks to John and Aris for this very important (and much needed!) contribution to our handling of the IDF problem. Find my full review of the book in a forthcoming issue of Journal of Business Forecasting.

The post Intermittent Demand Forecasting (new book by Boylan and Syntetos) appeared first on The Business Forecasting Deal.

Preview of Foresight #63 (2021-Q4)

FORESIGHT HALL OF FAME

Adopting the idea from other journals that recognize outstanding contributions to the field through best paper awards, we are pleased to announce that Foresight has established its own Hall of Fame to honor standout articles printed in our pages. The initial selection features papers from our first 12 years, 2005-2017. Nominations of several dozen articles came from the IIF membership, followed by a vote of the HoF selection committee. Ultimately, we inducted five articles from this period, as well as one discrete series of articles. The HoF authors were also honored with a cash stipend of $500.

We’re delighted to preface this current issue of Foresight with a listing of the winners along with a reproduction of the key points of those papers. Of course, you are welcome to read the HoF articles in our back issues, available for free download.

Foresight Hall of Fame Award Winners

• Good and Bad Judgment in Forecasting: Lessons from Four Companies Robert Fildes
  and Paul Goodwin, Issue 8 (2007)
• Can We Obtain Valid Benchmarks from Published Surveys of Forecast Accuracy?
  Stephan Kolassa, Issue 11 (2008)
• Choosing Levels of Aggregation for Supply-Chain Forecasts John Boylan, Issue 18 (2010)
• Why Should I Trust Your Forecasts? M. Sinan Gönül, Dilek Önkal, and Paul Goodwin, Issue
  27 (2012)
• FVA: A Reality Check on Forecasting Practices Mike Gilliland, Issue 29 (2013)
• Series of five articles on Forecast Quality Steve Morlidge (2013-2015)

BEHAVIORAL ECONOMICS

Behavioral economics has infused the forecasting profession, playing an enlightening role in recognizing “misbehaviors” in the forecasting process. A perpetually cited example is the silo mentality manifest in functional areas of the business. In Issue 13 (2009) of Foresight, John Mello described the dysfunctional “games” played and their detrimental effect on supply chains in “The Impact of Sales Forecasting Game Playing on Supply Chains”; these include Enforcing, Hedging, Sandbagging, and Filtering. Issue 5 (2006) featured an article by Rogelio Oliva and Noel Watson, “Managing Functional Biases in Organizational Forecasts.” In Issue 2 (2005), Elaine Deschamps recommended “Six Steps to Overcome Bias” with examples from the public and private sectors.

At the individual forecaster level, Foresight articles have addressed concerns about biases, both conscious and subconscious. Paul Goodwin summarizes the “hindsight bias” in his Issue 17 (2010) column “Why Hindsight Can Damage Foresight.” Oliva and Watson returned in Issue 25 (2012) to examine sources of bias such as “blind spots” and “incentive misalignment.” And most recently (Issue 61, 2021), Jon Karelse analyzes his survey results for identifying and “Mitigating Unconscious Bias in Forecasting.”

In his comprehensive examination of individual biases, Nobel laureate Daniel Kahneman, author of Thinking, Fast and Slow among other works, helped us come to grips with the shortcuts – heuristics – people use to make rapid decisions, and how these too often lead to misjudgment and error-prone predictions. Now, a decade later, Kahneman has teamed with Olivier Sibony and Cass Sunstein to write Noise: A Flaw in Human Judgment. The authors emphasize in this new work the distinction between bias and noise. Where bias is the systematic tendency to misjudge in the same way (e.g. to underforecast most of the time), noise – unwanted variability – is a phenomenon due to “the luck of the draw” in selections of the decision maker; for example, which judge is assigned to a case involving sentencing of a criminal. Noise aggravates uncertainty, undermines efficiency, and creates distortions and inequities.

Foresight Associate Editor Stephan Kolassa and I collaborated to write the book review of Noise that begins this issue of Foresight. It is followed by a commentary from Christopher Plummer, who takes issue with Jon Karelse’s characterization of forecasting biases as “unconscious”:

When we make choices, perform judgments, and evaluate competing propositions, we are engaged in completely cognitive functions. Our cognitive biases are not unconscious, as has long been recognized by behavioral scientists.

RISK AND UNCERTAINTY

Our previous issue (2021:Q3) presented a variety of perspectives on the distinction between risk and uncertainty and the question of whether we can assign probabilities to uncertain scenarios of the future. Yes, we can and  should, argued Peter Scoblic and Philip Tetlock in their article “A Better Crystal Ball: The Right Way to Think About the Future.” Not so fast was the reaction of most of our featured commentators: scenarios are more about plausibility, and we can’t really think probabilistically about uncertainty.

Now, in this issue, Scoblic and Tetlock respond to these critiques, saying we need to balance…

…the risks of two opposing mistakes: the false-positive error of concluding something is forecastable when it is not versus the false-negative error of concluding something cannot be forecast when it can. We can now ask (a) which mistake is likelier in a given setting, and (b) which mistake should we worry more about in each setting?

Presenting a middle ground is the first of a two-part article from Steve Morlidge and Paul Goodwin, “Into the (Largely) Unknown: A Simple Way to Handle Uncertainty.” The authors introduce the concepts of Possibility Theory and offer possibility distributions as a way to characterize the plausibility of alternative futures. Part 2 will follow in the 2022:Q1 issue.

LONG-RANGE FORECASTING OF COVID'S IMPACT

Ira Sohn examines “The Impact of COVID-19 on the Economy and Strategic Environment of the United States” through a review of reports from the U.S. Department of Labor’s Bureau of Labor Statistics (BLS) and the U.S. National Intelligence Council (NIC), both released earlier in 2021. The BLS report projects the pandemic’s impact on the U.S. labor market to 2029, while the NIC presents keypresents key trends and uncertainties that will shape the U.S. strategic environment out to 2040.

FORECAST PERFORMANCE MEASUREMENT

Elizabeth Yardley and Fotios Petropoulos criticize the exclusive use of forecast error metrics to evaluate forecast methods, arguing for consideration of forecast value, a construct that also accounts for how the forecast is used in decision making as well as the costs incurred in computation. Their piece is called “Beyond Error Measures to the Utility and Cost of the Forecasts.”

FORECASTING AND PLANNING PERSPECTIVES

This issue concludes with “Integrated Business Planning: A New Narrative for an Old Process,” as authors Niels van Hove and Hein Regeer give us their insights on the future of business planning. They write that it will be a thoroughly new look, one in which decision making is supported by Wave 3 technology, enabling widespread automation, knowledge augmentation, and tighter integration of planning into day-to-day operations.

FORESIGHT ADVISORY BOARD

The FAB welcomes newest member Chris Turner, Principal of Stratabridge (www.stratabridge.com), a boutique management consulting firm in the UK. Chris has decades of experience working with companies to leverage their culture, process, and capabilities to develop effective strategies. See his visionary article in the previous issue of Foresight (2021:Q3), “Strategy in Uncertain Times: Lenses to Approach Decision Making, Forecasting, and Planning.”

The post Preview of Foresight #63 (2021-Q4) appeared first on The Business Forecasting Deal.

The Makridakis Open Forecasting Center has announced the M6 Financial Forecasting Competition, to begin in February 2022. This will be a "live" competition running through February 2023, with a focus on forecasts of stock price (returns) and risk, and on investment decisions based on the forecasts. This forecasting and investing "duathlon" will draw substantial coverage from the forecasting community and from the public and mass media, with a prize pool targeted at $300,000.

As in all M competitions, the M6 is designed as a research endeavor. Its purpose is to empirically investigate an apparent paradox: The efficient market hypothesis posits that share prices reflect all relevant information, so consistently outperforming the market is not feasible. The EMH is supported by considerable evidence that active investment managers do not beat, on average, random stock selections. However, certain investors such as Warren Buffet have achieved long-term market beating returns impossible to justify by mere chance. The M6 is expected to shed new light on the EMH.

Find details on entering the competition and M6 guidelines on the competition website.

M5 Conference (December 6-7)

The M6 follows 2020's M5 and 2018's M4, which showed an increasingly valuable role of machine learning methods to augment and improve upon traditional time series methods. The M5 Conference, currently scheduled for December 6-7 in New York City, will provide a stellar lineup of speakers to review, analyze, and discuss the results and implications of the M5 competition. This includes remarks by M-competition founder Spyros Makridakis, and keynotes by Nassim Nicholas Taleb of NYU and Chris Fry of Google.

The post M6 Financial Forecasting Competition announced appeared first on The Business Forecasting Deal.

For the eighteenth year, the IIF, in collaboration with SAS®, is proud to announce financial support for research on improving forecasting methods and business forecasting practice. The award for this year will be (2) $10,000 grants. The deadline date for applications is October 1, 2021.

Forecasting research has seen major changes in the theoretical ideas underpinning forecasting effectiveness over the last 40 years. However, there has been less impact on forecasting practice. This grant award will offer financial support for research on how to improve forecasting methods, and business forecasting practice, including organizational aspects of management of the forecasting process.

This grant was created in 2003 by the IIF, with financial support from the SAS® Institute, in order to promote research on forecasting principles and practice. The fund is divided to support research on (1) how to improve forecasting methods and (2) business forecasting practice and applications. To learn more and to apply for this grant, visit our website.

Application Details

For more details, click here IIF-SAS award. Applications must include:

• Description of the project (4 page max)
• Brief c.v., including references (4 page max)
• Budget and work-plan for the project (1 page max)

All applications must be in pdf format and sent to IIF Business Director

Criteria for the award of the grant will include likely impact on forecasting methods and business applications.

Consideration will be given to new researchers, and whether supplementary funding is likely to be gained. It is also expected that the research supported by the SAS/IIF grant be presented in an International Symposium on Forecasting (ISF) organized by the IIF. The applications will be assessed through a committee appointed by the IIF directors. The results of the evaluation will be announced to the applicants within 12 weeks of the closing date.

Grant recipients are also required to author a paper reporting on their research for possible publication in the International Journal of Forecasting (IJF). Therefore it is useful to keep in mind the IJF suggestions for authors.

In addition to the advances made at our academic research institutions, there is also considerable innovation coming from practitioners -- as we've seen in the M4 and M5 Forecasting Competitions. Practitioners are encouraged to submit proposals.

Learn about previous recipients, on the IIF website. Last year's recipients:

• Hussain Syed Kazmi, KU Leuven; and Maria Paskevich, King (Activision-Blizzard), for the project proposal in the category of Business Applications, Incorporating Downstream, Task-Specific Information in Forecasting Models.
• Ahmed Aziz​ Ezzat, Rutgers University, for the project proposal in the category of Methodology, Forecasting in Unknown Territory: Towards Physically Motivated Learning for Local Wind Fields.

Application deadline is October 1, 2021.

The post SAS/IIF Research Grants (proposals due October 1) appeared first on The Business Forecasting Deal.

Jonathon Karelse CEO, NorthFind Management

On September 2 (3pm UTC / 11am EDT), I'll be joining Jonathon Karelse, CEO of NorthFind Management, for an interactive "fireside chat" on the application of Behavioral Economics in demand planning. This is part of the Foresight Webinar Series, and registration is free.

Since we first met at an Institute of Business Forecasting conference ~2006, Jonathon has been a longtime friend and collaborator. Back then he was an early adopter of FVA analysis while working at Yokohama Tire (Canada). More recently, and after founding global management consultancy NorthFind Management in 2012, Jonathon has become a go-to guy on the role of behavioral economics, and the topic of unconscious biases in judgmental forecasting.

For a preview of the topics, see Jonathon's article "Mitigating Unconscious Bias in Forecasting" in the 2021:Q2 issue of Foresight: The International Journal of Applied Forecasting, and his piece "Solving for the Irrational: Why Behavioral Economics is the Next Big Idea in Demand Planning" in Business Forecasting: The Emerging Role of Artificial Intelligence and Machine Learning (Wiley, 2021). You can also check out my "FVA: A Reality Check on Forecasting Practices," recently inducted into the Foresight Hall of Fame, and available for free download.

To see Jonathon in action, take 30 minutes to watch his brilliant and entertaining International Symposium on Forecasting presentation, "Trust but Verify: Using Behavioral Economics to Improve the Integration of Judgment in Forecasting." Jonathon was also a recent guest on the IBF Podcast hosted by Eric Wilson, on topic of "Bias and Behavioral Heuristics in Forecasting."

Join Jonathon and me for what we plan to be a candid and highly audience-interactive discussion. (register here)

The post Behavioral economics in demand planning (webinar September 2) appeared first on The Business Forecasting Deal.

Rob was recently named a Fellow of the International Institute of Forecasters, and is someone who is known (or should be known) by everyone in the forecasting profession. As a young statistician in 1998, Rob co-authored (with Makridakis and Wheelwright) the classic Forecasting: Methods and Applications (3rd edition) -- the first forecasting book I owned. A few years later, as still a young researcher, he became Editor-in-Chief of the International Journal of Forecasting, which he ran from 2005-2018. On a personal note, Rob graciously contributed original pieces to both Business Forecasting: Practical Problems and Solutions (Wiley, 2015) and Business Forecasting: The Emerging Role of Artificial Intelligence and Machine Learning (Wiley, 2021).

Hyndman's 5 Conditions for Easy (or Difficult) Forecasting

In the podcast (starting at approximately 9:00), Rob defines forecasting as "the estimation of a probability distribution of a variable to be observed in the future." This, as he notes, is a very statistical definition -- which is exactly what we would expect (and want) from a statistician. [I was pleased Rob's definition is not inconsistent with my definition of a forecast as a "best guess at what is going to happen in the future." Only his definition is a lot more sophisticated!]

By either definition anything is forecastable, but not everything can be forecasted easily or accurately. That's where Rob's five conditions come into play, describing something that is easy to forecast:

1. You understand the factors that contribute to its variation.
2. There is a lot of data available.
3. The forecasts cannot affect the thing you are trying to forecast (so there is no feedback -- your forecasts do not make a difference to the outcome of the thing you are forecasting).
4. The future is similar to the past.
5. There is relatively little natural or unexplainable variation.

I'd like to compare these conditions to my treatment of the topic in The Business Forecasting Deal (p.7):

...the best we can ever do is discover the underlying structure or rule guiding the behavior that is being forecast, to find a model that accurately represents the pattern of behavior, and then pray the behavior pattern doesn't change in the future.

Assume we do discover the underlying structure of the behavior, we correctly model that structure in our forecasting software, and the structure does not change in the future. Should we then be able to achieve perfect forecasts? Unfortunately the answer is no. In any complex business or social system (including things like the buying behavior of customers), there remains an element of randomness. Even though we know the underlying structure and model the behavior correctly, our forecast accuracy will still be limited by the amount of randomness, and no further improvement in accuracy will be possible.

Again, I'm happy to find general consistency between these two characterizations of the conditions for easy / accurate forecasting. Rob's # 1, 4, 5 correspond directly to points in my characterization, and his point #3 is an important addition -- something I overlooked. For example, a forecast for low demand might impact sales and marketing policy decisions, which then create an increase in demand, resulting in poor accuracy of the original forecast.

Although having a lot of data available (Rob's #2) is always a good thing, in some cases we may have other ways of knowing what drives the variation in behavior. For example, in games of chance (such as rolling dice or tossing a fair coin), we don't need data on past rolls or tosses to estimate the probability distribution of the future outcomes. But games of chance are just a tiny subset of the kinds of real-life forecasting challenges we face, and for which lots of data will be helpful.

Check out Rob's full Forecasting Impact interview for more interesting backstory on his career, and for a humorous account of how he became co-author with Makridakis and Wheelwright of the classic text.

The post Hyndman's 5 Conditions for Easy Forecasting appeared first on The Business Forecasting Deal.

Preview of Foresight #62 (2021:Q3)

This 62nd issue of Foresight has been heavily “infected” by the COVID pandemic.

Stephan Kolassa’s book review of Resurrecting Retail by Doug Stephens raises the question of whether the COVID disruptions decisively shifted the entire game of retail – as the author contends – or merely accelerated existing trends at play in the industry.

George Monokroussos and colleagues at Wayfair detail the company’s fundamental changes in forecasting methodology to adapt to the uncertainty and pattern disruption caused by the pandemic. Wayfair developed a structured approach to forecast its business drivers on the basis of scenarios for how the pandemic could evolve.

Chris Turner offers a strategic perspective on how businesses can survive and thrive during periods of extreme volatility and instability. His four lenses urge organizations to carefully balance their desire for control of existing processes with the need to develop resilient processes that will enable growth and prosperity in the long run.

Forecasters are an integral part of charting the map for an uncertain future, understanding how the demands from decision makers will change, developing new capabilities, and demonstrating the value that forecasting can bring to strategy.

The feature section in this issue is entitled A Better Crystal Ball. In their article of that name in the journal Foreign Affairs, reprinted with permission here, J. Peter Scoblic and Philip Tetlock propose a synthesis of scenario forecasting – the creation of plausible futures – with probabilistic judgments about the likelihood of different scenarios.

Scenario planners maintain that there are so many possible futures that one can imagine them only in terms of plausibility, not probability. By contrast, forecasters believe it is possible to calculate the odds of possible outcomes, thereby transforming amorphous uncertainty into quantifiable risk. Because each method has its strengths, the optimal approach is to combine them.

Commentaries on the “better crystal ball” clarify the objectives of scenario development, examine the distinction between risk and uncertainty, and lay out important obstacles to the attachment of probabilities to uncertain events. [Commentaries provided by Paul Goodwin, Steve Morlidge, Roy Batchelor, Stefan de Kok, Mike Trembley, and Robert Fildes.] The Scoblic-Tetlock response to the commentaries will be printed in the next issue of Foresight, 2021:Q4.

Also upcoming in the Q4 issue are:

• a tutorial on random forests, an ensemble method of classification and forecasting that has achieved wide attention in data science;
• a perspective on “one-number forecasting” and whether this is the best way to achieve one mindset in an organization;
• an assessment of forecasting performance that goes “beyond forecast errors” to encompass considerations of utility and cost;
• an introduction to possibility theory and how possibility distributions can be applied to forecasts in the absence of historical data;
• a review article assessing the impacts of COVID-19 on the economy and strategic environment of the United States.

The post Preview of Foresight #62 (2021-Q3) appeared first on The Business Forecasting Deal.

The 41st International Symposium on Forecasting will be virtual again this year, and begins Sunday June 27. SAS will have a huge presence -- as event sponsor, sponsor of the IIF/SAS Research Grants, and with 19 individual presentations.

ISF registration is free to members of the International Institute of Forecasters, and it is easy to Join the IIF. A one-year Premium membership costs $145 and includes hardcopy subscriptions (plus online access) to both IIF journals: the International Journal of Forecasting and Foresight: The International Journal of Applied Forecasting. 

Students can join for just $25 per year, which includes online-only access to both journals.

As a truly international event, presentations run literally around-the-clock from Monday June 28 through mid-day on Wednesday June 30. Special half-day Workshops run on Sunday June 27, as well as Thursday July 1 and Friday July 2. (Workshops require a $50 additional fee.)

In addition to the SAS presentations and the Practitioner Track listed below, here are several other sessions I'm looking forward to (all dates/times EDT). Be ready to get up early and stay up late:

• Workshop: Deep Learning for Forecasting - Tim Januschowski & Lorenzo Stella (Sunday June 27, 14:00-18:00)
• Accuracy, Explainability, and Trust in Business Forecasting - Simon Spavound (Monday June 28, 04:00-04:20)
• Re-Analysis of Intermittent Demand Forecasting Methods - John Boylan (Monday June 28, 08:40-09:00)
• Opening Welcome / Members Meeting - George Athanasopoulos & IIF Board of Directors (Monday June 28, 09:00-10:00)
• A Picture is Worth a Thousand Data Points: An Image-Based Time Series Forecasting Approach - Artemios-Anargyros Semenoglou (Monday June 28, 15:00-15:20)
• Size Does Matter: Timer Series Augmentation for Enhanced Cross-Learning - Evangelos Spiliotis (Monday June 28, 15:20-15:40)
• Forecasting Uncertainty: The Quest for Quantification - Steve Morlidge (Monday June 28, 16:00-16:20)
• Keynote: Humachine: Humankind, Machines, and the Future of the Enterprise - Nada Sanders (Monday June 28, 18:00-19:00)
• Probabilistic Ensemble Forecasting of Australian COVID-19 Cases - Rob Hyndman (Monday June 28, 21:40-22:00)
• Some Theoretical Ways in which FVA Analysis Can Be Misleading and How This Can Be Remedied - Paul Goodwin (Tuesday June 29, 05:20-05:40)
• Stylised Facts of FVA, a Meta-Analysis Where Do Judgmental Adjustments Improve Accuracy? - Robert Fildes (Tuesday June 29, 05:40-06:00)
• Algorithm Aversion or Algorithm Appreciation? - Shari De Baets (Tuesday June 29, 06:00-06:20)
• Using Judgmental Forecasting and Scenario Thinking for Anticipating the Future - George Wright (Tuesday June 29, 06:20-06:40)
• Judgmental Interventions: Model Tuning and Forecast Adjustments in a Retailing Case Study - Anna Sroginis (Tuesday June 29, 06:40-07:00)
• Evaluating the Impact of Business Practices on Inventory Performance - Evangelos Theodorou (Tuesday June 29, 08:20-08:40)
• New Product Life-Cycle Forecasting with Temporal Hierarchies - Oliver Schaer (Tuesday June 29, 08:20-08:40)
• Demand Forecasting Under Lost-Sales Stock Policies - Juan Trapero (Tuesday June 29, 08:40-09:00)
• Fast and Frugal Time Series Forecasting - Fotios Petropoulos (Tuesday June 29, 11:40-12:00)
• Prediction Intervals: Neglected Diagnostics? - Keith Ord (Tuesday June 29, 16:00-16:20)
• Estimating Interval Forecasts using Pruned Ensembles - Erick Meira (Tuesday June 29, 16:20-16:40)
• General NN Forecaster - Slawek Smyl (Tuesday June 29, 16:20-16:40)
• Forecast Combinations, Pooling, and Hierarchies: How do They Combine? - Nikos Kourentzes (Tuesday June 29, 17:40-18:00)
• Forecasting for Social Good - Bahman Rostami-Tabar (Wednesday June 30, 08:40-09:00)
• Keynote: Forecasting Climate Change, Pandemics and Econometrics - David Hendry (Wednesday June 30, 10:00-11:00)
• Workshop: Forecasting to Meet Demand - Stephan Kolassa & Roland Martin (Thursday July 1, 14:00-18:00)
• Workshop: Business Forecasting: Techniques, Application and Best Practices - Eric Stellwagen & Sarah Darin (Thursday July 1, 14:00-18:00)
• Workshop: Evaluating Forecasting Performance - Evangelos Spiliotis (Friday July 2, 14:00-18:00)

SAS Presentations (date/time is EDT)

Monday June 28

Session: Demand Forecasting 2 (06:00-07:00)

• Demand Forecasting in Times of COVID - Michel Kurcewicz (06:40-07:00)

Session: Open Source Forecasting in SAS (11:00 - 12:20)

• Accelerate Open Source Forecasting with SAS (Part 1) - Jessica Curtis
• Accelerate Open Source Forecasting with SAS (Part 2) - Andrea Moore
• Deep Learning for Retail Sales Forecasting - Szymon Haponiuk
• Major Paradigm Shifts in Modern Forecasting Methodology - Russ Wolfinger

Practitioner Track: Large-Scale New Product Forecasting: A ML-Based Approach - Nitzi Roehl (13:00-13:30)

Session: COVID-19 Forecasting: Exploring and Modeling Data (14:00-15:20)

• Location Network Analysis and Supervised ML Models to Identify VIrus Spread Trends - Carlos Pinheiro
• Evaluation of Statistical Models for Producing Weekly COVID-19 Forecast - Ran Bi
• Representing and Forecasting COVID-19 Pandemic Using Differential Equation Models - Marc Kessler
• Visualization by Pattern Similarity for COVID-19 Data Set - Youngjin Park

Session: Forecasting and Uncertainty (16:00-16:40)

• The Forecaster's Predicament: Issues with Communicating Uncertainty - Mike Gilliland (16:20-16:40)

Session: Forecasting and Software (17:00-18:00)

• Scalable Cloud-Based Automatic Time Series Imputation - Thiago Quirino
• Forecasting Software Trends for the Next Decade - Michele Trovero
• Using Open Source Machine Learning Algorithms in SAS Visual Forecasting - Javier Delgado

Tuesday June 29

Session: Time Series Clustering for Forecasting (02:00-03:00)

• Time Series Segmentation Using Two-Stage Clustering Approach - Sagar Mainkar (02:00-02:20)

Session: Retail Forecasting 2 (09:00-10:00)

• Enhancing Short-Term Demand Sensing Using Machine Learning - Charles Chase (09:20-09:40)

Session: Automated Forecasting (11:00-12:00)

• Monitoring Forecast Model Fitness Using Control Charts - Joe Katz (11:20-11:40)

Session: Neural Networks 2 (16:00-17:00)

• Time Series Forecasting with Time Series Plot and Computer Vision - Taiyeong Lee (16:40-17:00)

Friday July 2

Workshop: Formulating State Space Models - Rajesh Selukar (14:00-18:00)

Practitioner Track

Again this year the ISF will feature a special Practitioner Track of 30-minute presentations from ten top contributors to the field (see the link for session times, titles, and abstracts):

• Alla Anashenkova, ThroughPut Inc
• Patrick Bower, Combe International
• Simon Clarke, Argon & Co
• Ran Ding, Google
• Stefa Etchegaray Garcia, IBM
• Jonathon Karelse, NorthFind Management
• Sara Park, Coca-Cola
• Nitzi Roehl, SAS
• Niels van Hove, Aera Technology
• Dawn Woodard, Uber

Register here to attend (virtual) ISF 2021. As you can see, there are a lot more than 19 good reasons to attend!

The post 19 reasons to attend ISF 2021 (virtual - starting June 27) appeared first on The Business Forecasting Deal.

Vandeput's book is based on a compelling premise: Using data science to solve a problem requires a scientific mindset more than coding skills. I share this viewpoint. FVA is essentially the application of basic scientific method to the forecasting process, starting with a null hypothesis,

Ho: The forecasting process has no effect

As a forecaster / data scientist, your aim is to determine whether Ho can be rejected, thereby concluding your forecasting process does have an effect. The effect can be either positive (the process improves accuracy) or negative (your process is just making the forecast worse). Of course, if you cannot reject the null hypothesis that your forecasting process has no effect, is the process even worth executing?

Find more discussion of this topic in Changing the Paradigm for Business Forecasting (Part 8 of 12).

Data Science for Supply Chain Forecasting

Vandeput published a first edition of this book in 2018, with extensive coverage of traditional statistical / time series forecasting methods as well as the more recently popular machine learning methods. It included do-it-yourself sections illustrating the implementation of these methods in Python (and Excel for the statistical models).

This 2nd edition begins with a new Foreword by Spyros Makridakis. It includes considerable "how to" material on statistical and machine learning forecasting methods, along with much new content: adding an introduction to neural networks and an all-new Part III discussing demand forecasting process management. This new Part III features an entire chapter on Forecast Value Added, from which we share the following excerpt:

What Is a Good Forecast Error?

Throughout this book, we created forecasting models. Initially, we focused on statistical models, and later, we focused on machine learning models. Usually, you will try these models against your own dataset, first trying a simple model and then moving on to more complex ones. As soon as you get the first results from your model, you will ask yourself this very question: Are these results good? How do you know if an MAE of 20% is good? What about an MAE of 70%?

The accuracy of any model depends on the demand's inner complexity and random nature. Forecasting car sales in Norway per month is much easier than predicting the sales of a specific smartphone, in a particular shop, during one specific day. It is exceedingly difficult to estimate what is a good forecast error for a particular dataset without using a benchmark.

Benchmarking

To know if a certain level of accuracy is good or bad, you must compare it against a benchmark. As a forecast benchmark, we will use a naive model. We will then compare any forecast against this naive forecast, and see by how much extra accuracy (or error reduction) our forecasting model will beat it.

...

Process and Forecast Value Added

Forecast value added is not only meant to measure the added value of a model compared to a (naive) benchmark, but also to track the efficiency of each step in the forecasting process. By performing FVA analysis of the whole forecasting process, you will be able to highlight steps that are efficient (i.e., reducing the forecast error while not consuming too much time) and those that both consume resources and do not bring any extra accuracy. FVA is the key to process excellence. For each team (or stakeholder) working on the forecasting process, you will need to track two aspects:

• • their FVA compared to the previous team in the forecasting process flow
  • their time spent working on the forecast

Those teams can be demand planners, the sales team, senior management, and so on. (You can even track the FVA of each individual separately.) As you will track the FVA and time spent by each step in the process, you will have the right tools to reduce most judgmental bias, either intentional or unintentional.

Best Practices

Let's review the best practices when using the forecast value added framework.

• • FVA process analysis should be performed over multiple forecast cycles -- anyone can be (un)lucky from time to time
  • If you want to push FVA further and focus on the most critical items, you should use it together with weighted KPIs. Demand planners should focus on the SKUs for which the forecast model got the highest weighted error over the last periods, or on the items for which they think the model will lack insights.
  • If you are in a supply chain with different sales channels (or business units) implying different information, teams, and buying behavior, it is advised to track FVA separately for each business unit.

As a final piece of advice, we can review the conclusion drawn by Fildes and Goodwin (2007) who looked at the forecast adjusting process at four British companies. They saw that planners were making many small adjustments to the forecasts, bringing nearly no added value and consuming time. They also noted that larger adjustments were more likely to improve accuracy. This is due to the fact that they require more explanations from senior management, as well as higher (personal) risk if they are wrong. Finally, they saw that planners tend to be overly optimistic (a usual cognitive bias), resulting in too many positive adjustments. This is so much of an issue that in order to improve forecast added value, the authors provocatively suggest banning positive adjustments.

Process Efficiency

With the help of FVA, you will quickly realize that the marginal improvement of each new team working on the forecast is decreasing. It might be easy to improve the most significant shortcomings of a forecast model. But it is much more challenging to improve a forecast that has already been reviewed by a few professional teams relying on multiple sources of information. That is fine, as FVA is here to help businesses allocate the appropriate resources to forecast edition. Past a certain point, working more on the forecast will not be worth it.

The post Vandeput's Data Science for Supply Chain Forecasting (book excerpt) appeared first on The Business Forecasting Deal.