Cabotaje Opman

Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 10

Steps in the Forecasting Process

1. Determine the purpose of the forecast- Define the goals of the forecast, including
specific questions to answer and how the results will be used in decision-making.
2. Establish a time horizon- Decide on the forecast's time frame—short-term (up to a
year), medium-term (one to three years), or long-term (three years or more). This choice
affects methods and data collection.
3. Select a forecasting technique- Choose an appropriate forecasting method based on the
purpose, data availability, and required accuracy. Techniques can be qualitative or
quantitative
4. Obtain, clean, and analyze appropriate data- Collect relevant historical data, clean it
to remove inaccuracies, and analyze it to identify patterns and trends. Use visualization
tools for better understanding.
5. Make the forecast- Generate forecasts using the selected model and prepared data.
Document assumptions made during this process, as they impact results.
6. Monitor the forecast- Continuously evaluate forecast accuracy against actual outcomes.
Regular reviews help refine models and improve future forecasting accuracy.

Forecast Accuracy and Control


Forecast accuracy is a critical aspect of decision-making in various fields, particularly in supply
chain management and inventory planning. It refers to how closely a forecast aligns with actual
outcomes, providing insights into the reliability of predictions and the potential deviations from
expected values.
 Forecasters want to minimize forecast errors
 Forecasting is an essential practice in many fields (e.g., business, economics, weather,
supply chain management), but it is inherently uncertain. Forecasters aim to make
predictions that are as close as possible to actual future values. However, due to the
complexity of real-world systems and numerous unpredictable variables, it is
difficult, if not impossible, to forecast with complete accuracy on a regular basis.
 Because errors are inevitable in forecasting, it is crucial not only to acknowledge that
errors will happen but also to quantify and communicate how much a forecast might
deviate from the actual outcome. This is done by providing an indication of the
extent of the forecast's deviation from the actual value. This can be presented
through error margins, confidence intervals, or other statistical measures that express
the level of uncertainty in the forecast.
 Forecast accuracy should be an important forecasting technique selection criterion
 In forecasting, different techniques can be applied depending on the problem at
hand. The accuracy of the forecast should play a key role in deciding which
technique to use. The primary purpose of forecasting is to make informed
decisions based on expected future events. If the chosen method consistently
produces inaccurate results, decision-makers might make flawed plans. Accuracy
shouldn't be the only factor when choosing a forecasting method, but it should be
a significant one. Balancing accuracy with other factors like simplicity, data
availability, and cost is crucial.
 Example: For instance, if an inventory management system uses a forecasting
technique that underestimates demand by 20%, it could lead to stockouts, lost
sales, and customer dissatisfaction. In contrast, overestimating demand might lead
to excess inventory, increased costs, and waste.

 Minimizing Forecast Errors


Forecasting is an essential practice in fields such as business planning, production, inventory
management, and financial predictions, but it’s nearly impossible to predict real-world values
(like sales, stock prices, or demand) with perfect accuracy. There are always variables,
uncertainties, and unpredictable factors that influence the outcomes. These include changing
market conditions, human behaviour, technological advancements, or external events like natural
disasters, which make it challenging to forecast perfectly on a regular basis.
 Minimizing Errors: While perfection is impossible, forecasters strive to minimize
errors. The goal is to make the forecast as close to the actual value as possible. A
well-calibrated forecasting system, even with some error, can still provide value by
helping organizations plan and make informed decisions.
 Indicating the Extent of Deviation. Since perfect accuracy isn’t achievable, it’s
important to provide a range or margin of error around forecasts. This helps users of
the forecast understand how much the predicted value might differ from the actual
result. For example, instead of forecasting that sales will be exactly 10,000 units, a
better forecast might be that sales will be 10,000 units ± 500 units, which means the
forecast is accurate within 5% of the target.

Uncertainty Communication: Indicating the extent of forecast deviation—either through


confidence intervals, standard deviations, or error margins—helps stakeholders prepare for
possible variances. This transparency helps mitigate risks and uncertainties, enabling businesses
to make better-informed decisions, such as adjusting inventory levels or budgets.

 Forecast Accuracy as a Criterion for Technique Selection


 Forecast accuracy is essential for effective decision-making, as accurate forecasts
enable better planning, resource allocation, and risk management. However, since no
forecast is perfect, monitoring and addressing forecast errors is equally important.
Accuracy should be a critical selection criterion because a technique that consistently
provides closer estimates to actual outcomes will be more reliable and valuable.

 Forecast errors should be monitored


 If errors fall beyond acceptable bounds, corrective action may be necessary
 Error Formula: Error = Actual – Forecast: This basic formula calculates the
difference between what actually occurred and what was predicted. It’s a starting
point for evaluating forecast performance
 Types of Forecast Errors: Positive Error: The actual value is higher than the
forecast, indicating underestimation. Negative Error: The actual value is lower
than the forecast, indicating overestimation.
Estimating slope and intercept
• Slope and intercept can be estimated from historical data. Estimating the slope (b) and
intercept (a) is essential in linear regression analysis, which is used to model the relationship
between two variables. The goal is to predict or describe how one dependent variable changes in
response to another independent variable.

Slope (b):

 The slope represents the rate of change of the dependent variable (y) with respect to the
independent variable (x). In simple terms, it tells you how much the dependent variable
increases or decreases as the independent variable changes.
 For example, if you're predicting sales based on advertising spend, the slope tells you
how much sales will increase for every additional dollar spent on advertising.

Intercept (a):

 The intercept is the value of the dependent variable (y) when the independent variable (x)
is zero. It represents the starting point or baseline value of the dependent variable.
 For example, in the same sales prediction, the intercept would represent the baseline sales
when no money is spent on advertising.

Example:

As a sales manager, you have gathered monthly sales data over the past five months and aim to
estimate the trend in sales. The objective is to use this historical data to predict future sales. The
following data represents sales figures for the last five months:

Month Sales (y)


1 100
2 110
3 130
4 150
5 160

Steps:

1. Identify the Data:

n = Number of data points (5 months in this case)


t = Month (1, 2, 3, 4, 5)
y = Sales (100, 110, 130, 150, 160

2. Calculate the summations:


∑ t=1+2+3+ 4+5=15
∑ y=100+110+130+150+ 160=650
2
∑ t =1²+2²+3² +4²+5²=1+ 4+9+16 +25=55
∑ ty=(1 ×100)+(2× 110)+(3 × 130)+(4 ×150)+(5 ×160)=100+220+390+600+ 800=2110

3. Calculate the slope (b): Substitute the values

5 ( 2110 )−( 15 ) (650)


b=
5 (55 )−¿ ¿

10550−9750
b=
275−225

800
b= =16
50

4. Calculate the intercept (a): Substitute the values

650−16 (15)
a=
5
650−240 410
a= = =82
5 5

5. Final equation:
Slope (b) = 16 y=16 x +82
Intercept (a) = 82
6. Graphing

The actual sales data points (blue dots) show the sales performance for each of the five months.

The regression line (red dashed line) provides the estimated trend in sales, which increases by 16
units per month, reflecting the underlying pattern in the data.

As the sales data points follow a linear increasing trend, the regression line provides a good fit to
predict future sales based on this historical pattern.

Trend-Adjusted Exponential Smoothing

Trend-Adjusted Exponential Smoothing is a forecasting method that incorporates both a


smoothed forecast and a trend factor to better predict data with trends.

TAF t +1=St +T t

 TAF t +1 :Trend-Adjusted Forecast for the next period is the sum of:

 St : The smoothed forecast for the current period.


 T t : The trend adjustment or trend factor for the current period.

Components

1. St (Smoothed Error):
 Represents the forecast value based on the data, but without considering any
trend. It combines the previous forecast and observed data to "smooth" out
fluctuations.
 TAF t : The previous trend-adjusted forecast
 At :The actual observed value
 ( A ¿ ¿ t−TAF t ):¿ The forecast error, which is smoothed using the smoothing constant
α (alpha).
 a : A smoothing constant

2. T t (Trend Factor):
 Represents the current trend estimate. It adjusts the smoothed forecast by
incorporating the rate at which the data is trending upward or downward.

T t=T t −1 + β ( TAF t −TAF t −1−T t −1)

 T t−1: The previous trend value.


 TAF t −TAF t −1: The change in forecast between two consecutive time
periods.
 β: The smoothing constant for the trend

Example:

You own a coffee shop and are forecasting tomorrow's sales. You already know that sales have
been increasing by 5 cups per day on average. Today's actual sales were 120 cups, and
yesterday's forecast was 100 cups. What is the forecast for tomorrow?

Given:

 At : Actual coffee cups sold today = 120 cups.


 TAF t Forecasted cups for today (calculated yesterday) = 100 cups
 St : Smoothed baseline (average cups sold daily).
 T t: Trend (how sales are increasing or decreasing daily).
 α : 0.3
 β : 0.2
Steps:

1. Solve for St

 Substitute the values


St =100+0.3 ( 120−100 )=100+ 6=106
The updated baseline is 106 cups/day.
2. Solve for the T t

 Substitute the values


T t=5+ 0.2 ( 120−100−5 )=5+0.2 ( 15 )=5+3=8

The new trend is 8 cups/day

3. Forecast for tomorrow TAF t +1

 Substitute the values


TAF t +1=106+ 8=114

Forecast:

Predicted cups of coffee that will be sold tomorrow is 114 based on the updated baseline and
trend.

Techniques for Cycles

Cycles are long-term variations in data caused by external factors, such as economic trends,
industry dynamics, or global events. Unlike seasonal variations, which are predictable within a
year, cycles are irregular and can last years, making them more complex to forecast.

The explanatory approach for forecasting cycles involves identifying a leading variable—a
variable that occurs before the event you're forecasting and is strongly correlated with it. This
allows you to use the leading variable as a predictor for future changes. A leading variable is
something that happens earlier than the variable of interest and is linked to it. Use historical data
to find a strong relationship between the leading variable and the variable of interest. If a high
correlation can be established with a leading variable, it can develop an equation that describes
the relationship, enabling forecasts to be made.

Associative Forecasting Techniques

Associative forecasting is a method that uses predictor variables—factors that are known to
influence the variable of interest—to make forecasts. This approach focuses on identifying
cause-and-effect relationships and developing a mathematical equation that predicts the target
variable based on its relationship with the predictors.
 Predictor Variables: These are independent variables that influence or predict the
dependent variable (the variable of interest).
 Example: Home size (square footage), Property size (lot size), Location, Number of
bedrooms or bathrooms

 Target Variable: This is the dependent variable you want to predict.


 Example: The price of a house or future sales of a product.

 Mathematical Model: An equation that quantifies the relationship between predictor


variables and the target variable.
References

1.6 The basic steps in a forecasting task | Forecasting: Principles and Practice. (n.d.). In

otexts.com. https://otexts.com/fpp2/basic-steps.html

A Comprehensive Guide to Accurate Forecasting [ With Formulas] | My Hours. (n.d.).

Myhours.com. https://myhours.com/articles/guide-to-accurate-forecasting

Armstrong, J. S. (2009). Selecting Forecasting Methods. SSRN Electronic Journal.

https://doi.org/10.2139/ssrn.1941247

Arnaud Belinga. (2023). Breakcold.com. https://www.breakcold.com/explain/forecast-accuracy

Exploring bivariate numerical data | Khan Academy. (2019). Khan Academy.

https://www.khanacademy.org/math/statistics-probability/describing-relationships-

quantitative-data

Forecast Accuracy | The Who, What, Why, and How of Forecast Accuracy. (n.d.). Ninetailed.io.

https://ninetailed.io/glossary/forecast-accuracy/

Groebner, D., Shannon, P., Fry, P., & Smith, K. (2022). Search. Pearson.com.

https://www.pearson.com/store/p/business-statistics/P100000678449

Lai, H., & Ng, E. C. Y. (2020). On business cycle forecasting. Frontiers of Business Research in

China, 14(1). https://doi.org/10.1186/s11782-020-00085-3

Measuring Forecast Accuracy: The Complete Guide. (2017, November 3). RELEX Solutions.

https://www.relexsolutions.com/resources/measuring-forecast-accuracy/

Newbold, William L. Karla, and Betty Thorne, P., Karla, W., & Betty, B. (2022). Search.

Pearson.com. https://www.pearson.com/store/p/statistics-for-business-and-economics/

P100000635422
Oyibo, C. (2024). 3: Forecasting. Oyibo.com.

https://legacy2.oyibo.com/notes/IDS355/3Forecasting.htm

Render, B., Munson, C., & Heizer, J. (2017). Operations management: Sustainability and supply

chain management. Pearson.com. https://www.pearson.com/store

Ross, S. (2014). Introduction to Probability and Statistics for Engineers and Scientists. Elsevier

Shop. https://www.elsevier.com/books/introduction-to-probability-and-statistics-for-

engineers-and-scientists/ross/978-0-12-394811-3

Schmidt, J. (2023, March 9). Forecasting. Corporate Finance Institute.

https://corporatefinanceinstitute.com/resources/valuation/forecasting/

Steps of Forecasting. (2023, May 8). GeeksforGeeks. https://www.geeksforgeeks.org/steps-of-

forecasting/

Stevenson, W. J. (2018). McGraw-Hill Education. McGraw Hill Education.

https://www.mheducation.com/

The art and science of forecasting in operations management. (2019). UniversalClass.com.

https://www.universalclass.com/articles/business/the-art-and-science-of-forecasting-in-

operations-management.htm

What is financial forecasting + how to do it [7 Steps]. (n.d.). Www.paddle.com.

https://www.paddle.com/resources/financial-forecasting

What is forecasting? (Definition, key steps and methods). (2024). Indeed Career Guide.

https://uk.indeed.com/career-advice/career-development/what-is-forecasting

Zarnowitz, V. (1992, January). Business cycles: Theory, history, indicators, and forecasting.

University of Chicago Press.

https://www.nber.org/system/files/chapters/c10381/c10381.pdf

You might also like