Topic 6 DQ 1 May 9-11, 2024Describe the different forecasting methods and provid

Topic 6 DQ 1
May 9-11, 2024Describe the different forecasting methods and provide an example of when each is most applicable.
Submitted on:
May 10, 2024, 4:38 PM.
Summer Riddle
May 10, 2024, 4:05 PM
Unread
Prof. Eixenberger and Classmates,
Forecasting Defined
Forecasting is a science but also an art. Sales needs to forecast demand so that production can begin to plan for what they need to meet the forecast goal. Their are many different forecasts such as strategic forecasts which are medium and long -term forecasts that are used related to strategy and aggregate demand. Tactical forecasts focus on the day-to-day operations of an organization and are short term. Both qualitative and quantitative techniques are used to evaluate what forecasting process should be used.
Forecasting is classified into four different models such as time series analysis stating that past demand can predict future demand. Past demand includes, trends, seasonal, and cyclical influences. Causal forecasting assumes that demand is caused by an underlying factor and uses linear regression. Simulation forecasting runs through a range of different assumptions of the forecast. Qualitative and time series are the most commonly used for decision making.
Demand Components
Demand components are broken down into six components which are average demand for the period, a trend, a seasonal element, cyclical elements, random variation, and auto-correlation. Cyclical factors are difficult to determine and influences come from political elections, war, economic conditions or sociological pressures. Random variations are just that random, by chance events. When all the other components of demand are subtracted random variations are left because they are based on chance and the unexplained portion of why demand did or did not occur. Auto-correlation is represented by how often occurrences happen. Trend lines remind me of trajectories in health care. They follow a specific pattern. As a hospice nurse we know that if someone has COPD that they have an up and down trajectory that slowly declines over time based on exacerbations of the disease process unlike Alzheimers or dementia which has a slow decline consistent but slow along with MS and Parkinson’s.
Forecasting Methods
One method of forecasting plots data points and then searches for the trajectory of the points that fit best. Time series forecasting has different lengths of time in which someone calculates the future demand with. For example short term forecasting reviews under three months, medium forecasting is three months to two years and long-term forecasting is greater than two years. Short term models measure current variability in demand which works for having inventory back up just in case. Medium term captures the affects of seasons and holidays on demand and long-term models detect general trends and can see major events in the behavior of demand.
Certain forecasting methods require a minimum of historical data before being represented well. For example a simple moving average requires 6-12 months of weekly data. A weighted moving average and simple exponential smoothing and exponential smoothing with trend requires 5 to 10 observations needed to start. Linear regression requires 10-20 observations and trend and seasonal models require 2 to 3 observations per season. When choosing a method an organization considers time horizon to forecast, data availability, accuracy required, size of forecasting budget, and availability of qualified personnel.
Monthly forecasts are appropriate for budgets and weekly forecasts are appropriate for inventory. In a department store a weighted average with percentages taken on current and past months and combined to give a forecast of the next month. With the most recent past being the indicator that has the most weight, lessening the further back one goes. Seasonal trends must be taken into account based on service or product providing. This concept makes exponential smoothing the most used in forecasting. They are accurate, easy to calculate, small storage requirements, and it can be tested for accuracy. Smoothing takes into account previous forecast, current data and then determines what the reaction rate should be. The more growth a company is experiencing the higher the reaction rate should be. Smoothing constants reduce the impact of the error of the forecast and actual demand creating an appropriate demand for the next time period. (MY HEAD HURTS UUUUGGHH!!)
Linear regression analysis explores relationships between two or more correlated variables created from observed data.
Linear regression is used for time series forecasting and causal relationship forecasting. The number of deaths from lung cancer would increase with smokers would be an example of a causal relationship (Jacobs & Chase, 2020).
I need to keep studying this its always the math that I struggle with but enjoy if that makes any sense.
Thanks for reading my post.
Reference
Jacobs, F. R., & Chase, R. B. (2020). Operations and supply chain management (16th
Ed.). New York, NY: McGraw-Hill. ISBN-13: 9781264091676