A new generation of artificial intelligence technologies have emerged that hold considerable promise in helping improve the forecasting process including such applications as product demand, employee turnover, cash flow, distribution requirements, manpower forecasting, and inventory. These AI based systems are designed to bridge the gap between the two traditional forecasting approaches: managerial and quantitative.
Organizations develop forecasts to support planning and decision-making processes. Specific operations forecasting applications include product demand, inventory levels, manpower levels, scrap rates, and raw materials requirements. Forecasts can also be used as motivational tools. Technology based forecasts tend to focus on new product/service development. For example, how long will it take before the DVD is the primary media platform in the home? Economic forecasts deal with business parameters such as interest and inflation rates. (For a more detailed explanation of economic forecasting, see Economic Forecasting: How Pros Predict the Future in the Winter 2000 issue of the Graziadio Business Review.
Generally speaking, forecasts are based on quantitative analysis, qualitative analysis or a combination of both. Often quantitative forecasting is referred to as objective analysis while qualitative forecasting is called managerial or judgmental analysis. Typically, there is tension between these two approaches. Quantitative forecasts, which are often favored by operations, tend to be developed using a bottom up approach while managerial-based forecasts, usually preferred by the marketing group, are approached from a top down perspective. For example, a primary marketing goal is to insure adequate supply while operation’s focus is on minimizing inventory. The resolution of these two approaches is how forecasting errors occur and presents an opportunity for using artificial intelligence methods. Quantitative forecasting can be characterized by one of the two basic techniques:
- Time Series – The future will tend to look and behave like the past. For example, gasoline prices for the next six months will continue along the same lines as they have over the past six months.
- Relational – The future is dependent on the direction of a variety of factors. For example, new housing starts might be a function of interest rates and local weather conditions.
A time series is a set of data points recorded over successive time periods. Examples include monthly billables, weekly unit product demand and quarterly inventory levels and stock prices. A relational database consists of the recording of several variables for a number of observations. For example, a financial relational database could consist of revenues, earnings and assets for the Fortune 500.
The following graphic highlights the typical forecasting process. The resultant forecasts are evaluated by comparing predictions with actual results. This assessment is accomplished by examining the error terms. An error term is the difference between the prediction and the actual outcome. Based on an error assessment, the forecasting process is continually updated through the adjustment of model inputs.
Typical Forecasting Process
Typically, no one forecasting approach is best in all situations. Instead, it is most appropriate to use a combination of different forecasting techniques in arriving at composite estimates. Furthermore, it is usually a good idea to provide interval or range estimates as well as a single point forecast.
Timeframe and Data
Two major issues in the forecasting process are the time horizon and extent of data availability. The following graphic illustrates the relationship between qualitative and quantitative forecasting as a function of time horizon and data content.
Judgmental vs. Objective Forecasting
Often, objective approaches are used when there is sufficient supporting data and the horizon is relatively short. On the other hand, long-term forecasts tend to favor judgmental approaches since extrapolations based on historical data tend to break down over time. Furthermore, some forecast applications involve situations that do not have a history. For example, consider receiving the assignment, in 1985, of estimating the impact of the Internet on business by the year 2000. For more information on judgmental forecasting visit: http://www.ncedr.org/tools/tools/tool7/judgmental.htm
To the general public, artificial intelligence conjures up visions of science fiction as illustrated in films such as The Terminator, The Matrix, and A.I. In reality, AI has considerable potential for improving productivity throughout the organization.
Artificial Intelligence Use Expanding
What is AI? AI is generally defined as a computer-based analytical process that exhibits behavior and actions that are considered “intelligent” by human observers. AI attempts to mimic the human thought process including reasoning and optimization (http://ai-depot.com/). The overall market for AI related systems is growing rapidly. Presently, the United States accounts for over 60 percent of an estimated $900 million global AI market. One purpose of AI is to help organize and supply information for the management decision-making process in such a way as to improve overall efficiency and performance. Three of the more commonly used AI systems in forecasting are:
- Neural Nets emulate elements of the human cognitive process, especially the ability to recognize and learn patterns. The architecture consists of a large number of nodes that serve as calculators to process inputs and pass the results to other nodes in the network. These systems have the advantage of not requiring prior assumptions about possible relationships. One application of neural nets might be forecasting employee turnover by category based on such factors as tenure with the firm, managerial level, and gender.
- Expert Systems summarize the totality of available knowledge and rules. “Knowledge” is stored in a set of “if-then” rules. The knowledge base can be obtained by interviewing experts or integrating sets of data. For example, predicting upcoming weather conditions based on current temperatures, humidity levels, season of year, and geographical location.
- Belief Networks describe the database structure using a tree format. The nodes represent variables and the branches the conditional dependencies between variables. Belief nets generate conditional probabilities for a variety of future outcomes. For example, estimating the chances of various product sales levels based on such traditional factors such as marketing and R&D budgets as well as market signals like customer complaints.
These AI systems can be employed for both forecast classification (e.g., preferred customer vs. marginal customer) and prediction (e.g., annual sales). The following table provides a simple illustration of how AI could be used to refine a marketing strategy based on three customer behavior factors: profit margin, retention probability and potential long-term value to the firm.
Each of these factors is characterized as either low or high. In practice, a more complex characterization scheme with more factors and more levels (e.g., low, medium, high) can be used. This table shows the appropriate qualitative strategy given each set of circumstances. Customers would be characterized in terms of their demographics and prior purchasing behavior. Quantitative forecasts can also be developed along the same lines.
Long Term Value
|Reduce marketing resources|
|Market distinct product portfolio|
|Examine up-sale opportunities|
|Market missing products|
|Refocus marketing effort|
|Re-attract these customers|
|Increase marketing resources|
|Pursue these customers|
Used in this way, the system can automate the process of both qualifying and quantifying marketing prospects and forecasting demand. Other related AI capabilities include:
- Identify similar purchasing patterns within a given time frame.
- Segment databases into related factors.
- Detect relationships and sequential patterns.
- Develop categorization and estimation models.
Used in combination with traditional forecasting, AI can help ameliorate friction that may exist between objective and managerial oriented approaches. More specifically, it integrates the best features from both classical approaches in structuring a virtual forecasting system.
The human brain contains on the order of 1011 neurons. While this number is impressive the number of synapses, estimated at 1016, is truly unbelievable. This is equivalent to the number of printed characters in all of the books contained in the United States Library of Congress 300 times over! By contrast a typical forecasting application might contain a few thousand neurons.
The following list presents some examples where organizations have improved bottom line profitability by improving the forecasting process.
Literacy grew out of the collision of the steam engine and the printing press. What will the Internet’s linguistic impact be? We may be in for some real surprises. Will this process cause sophisticated artificial intelligence to finally burst onto the scene?
Michael Hawley – Technology Review
Improve Your Forecasts
In a turbulent business environment, forecasting can lead to significant competitive advantage as well as to costly mistakes. Forecasting errors impact organizations in two ways. The first is when faulty estimates lead to poor strategic choices, and the second is when inaccurate forecasts impair performance within the existing strategic plan. An example of the former would be to increase the level of vertical integration based on a forecast of stable demand when demand actually turned out to be highly unstable. An example of the latter would be to significantly increase facility capacity based on a forecast of strong demand when, in fact, demand turned out to be soft. Either way there will be a negative impact on profitability. The following process outlines a plan for improving forecast accuracy using artificial intelligence support systems:
1. Evaluate and characterize the current forecasting system.
2. Measure the current level of error.
3. Compare error levels with industry norms.
4. Specify new requirements.
5. Characterize the economic impact of improved forecasts.
6. Identify alternative AI forecasting options.
7. Select best approach(s).
8. Develop implementation schedule.
9. Identify potential bottlenecks and problem areas.
10. Implement new system and monitor performance.
A primary objective for using AI is to better integrate the managerial and quantitative estimates and thereby reducing the forecasting errors. Organizations that currently utilize AI may increase the accuracy of forecasts by improving data collection methods and expanding efforts to gather market intelligence. Past customer behavior is often a reliable data source of future behavior. Finally, as in other areas of knowledge management, it is critical to maintain strong support for the forecasting process from the entire management team. Visit the following site for more on how AI is impacting business: http://www.shai.com/ai_general/value.htm