Judgmental forecasts

In many cases, judgmental forecasting is the only option, such as when there is a complete
lack of historical data, or when a new product is being launched, or when a new competitor
enters the market, or during completely new and unique market conditions.
There are also situations where the data are incomplete, or only become available after
some delay.
Accuracy of judgmental forecasting improves when the forecaster has
(i) important domain knowledge, and
(ii) (ii) more timely, up-to-date information.
A judgmental approach can be quick to adjust to such changes, information or events.
Over the years, the acceptance of judgmental forecasting as a science has increased, as has
the recognition of its need. More importantly, the quality of judgmental forecasts has also
improved, as a direct result of recognising that improvements in judgmental forecasting can
be achieved by implementing well-structured and systematic approaches. It is important to
recognise that judgmental forecasting is subjective and comes with limitations. However,
implementing systematic and well-structured approaches can confine these limitations and
markedly improve forecast accuracy.
There are three general settings in which judgmental forecasting is used:
(i) there are no available data, so that statistical methods are not applicable and
judgmental forecasting is the only feasible approach;
(ii) (ii) data are available, statistical forecasts are generated, and these are then
adjusted using judgement; and
(iii) (iii) data are available and statistical and judgmental forecasts are generated
independently and then combined.
We should clarify that when data are available, applying statistical methods is preferable and
should always be used as a starting point. Statistical forecasts are generally superior to
generating forecasts using only judgement

Artificial Intelligence forecasting

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.
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.

The overall market for AI related systems is growing rapidly. 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.
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.

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
MEASURING FORECAST ACCURACY

Based on the accumulated forecast errors over time, the two methods look equally
good. But,most observers would judge that Method 1 is generating better forecasts
than Method 2 (i.e.,smaller misses).
Mean Absolute Deviation (MAD): To eliminate the problem of positive errors canceling
negative errors, a simple measure is one that looks at the absolute value of the error (size of the
deviation, regardless of sign). When we disregard the sign and only consider the size of the error,
we refer to this deviation as the absolute deviation. If we accumulate these absolute deviations
over time and find the average value of these absolute deviations, we refer to this measure as the
mean absolute deviation (MAD). For our hypothetical two forecasting methods, the absolute
deviations can be calculated for each year and an average can be obtained for these yearly
absolute deviations, as follows:
The smaller misses of Method 1 has been formalized with the calculation of the
MAD. Method 1 seems to have provided more accurate forecasts over this six year
horizon, as evidenced by its considerably smaller MAD.
Mean Squared Error (MSE): Another way to eliminate the problem of positive errors
canceling negative errors is to square the forecast error. Regardless of whether the forecast error
has a positive or negative sign, the squared error will always have a positive sign. If we
accumulate these squared errors over time and find the average value of these squared errors, we
refer to this measure as the mean squared error (MSE). For our hypothetical two forecasting
methods, the squared errors can be calculated for each year and an average can be obtained for
these yearly squared errors, as follows:

Method 1 seems to have provided more accurate forecasts over this six year
horizon, as evidenced by its considerably smaller MSE.
The Question often arises as to why one would use the more cumbersome MSE
when the MAD calculations are a bit simpler (you don’t have to square the
deviations). MAD does have the advantage of simpler calculations. However, there
is a benefit to the MSE method. Since this method squares the error term, large
errors tend to be magnified. Consequently, MSE places a higher penalty on large
errors. This can be useful in situations where small forecast errors don’t
cause much of a problem, but large errors can be devastating.
Mean Absolute Percent Error (MAPE):
A problem with both the MAD and MSE is that their values depend on the magnitude of the
item being forecast. If the forecast item is measured in thousands or millions, the MAD and MSE
values can be very large. To avoid this problem, we can use the MAPE. MAPE is computed as
the average of the absolute difference between the forecasted and actual values, expressed as a
percentage of the actual values. In essence, we look at how large the miss was relative to the size
of the actual value. For our hypothetical two forecasting methods, the absolute percentage error
can be calculated for each year and an average can be obtained for these yearly values, yielding
the MAPE, as follows:

Method 1seems to have provided more accurate forecasts over this six year
horizon, as evidenced by the fact that the percentages by which the forecasts miss
the actual demand are smaller with Method 1 (i.e., smaller MAPE).
ILLUSTRATION OF THE FOUR FORECAST ACCURACY MEASURES
Here is a further illustration of the four measures of forecast accuracy, this time
using
hypothetical forecasts that were generated using some different methods than the
previous illustrations (called forecasting methods A and B; actually, these forecasts
were made up for purposes of illustration). These calculations illustrate why we
cannot rely on just one measure of forecast accuracy.

You can observe that for each of these forecasting methods, the same MFE
resulted and the same MAD resulted. With these two measures, we would have no
basis for claiming that one of these forecasting methods was more accurate than
the other. With several measures of accuracy to consider, we can look at all the
data in an attempt to determine the better forecasting method to use. Interpretation
of these results will be impacted by the biases of the decision maker and the
parameters of the decision situation. For example, one observer could look at the
forecasts with method A and note that they were pretty consistent in that they were
always missing by a modest amount (in this case, missing by 20 units each year).
However, forecasting method B was very good in some years, and extremely bad
in some years (missing by 60 units in years 5 and 6).
That observation might cause this individual to prefer the accuracy and consistency
of forecasting method A. This causal observation is formalized in the calculation of
the MSE.
Forecasting method A has a considerably lower MSE than forecasting method B.
The squaring magnified those big misses that were observed with forecasting
method B. However, another individual might view these results and have a
preference for method B, for the sizes of the misses relative to the sizes of the
actual demand are smaller than for method A, as indicated by the MAPE
calculations.