Forecasting and Productivity Lecture Notes

A set of vocabulary flashcards covering qualitative and quantitative forecasting methods, time series components, error metrics, and productivity formulas.

Qualitative Forecasting

A primary category of forecasting methods that includes techniques like jury of executive opinion, the Delphi method, and sales force composite.

Quantitative Forecasting

A primary category of forecasting methods subdivided into Time Series and Regression Methods.

Jury of executive opinion

A qualitative forecasting method involving a group of high-level experts or managers.

Delphi method

A qualitative forecasting method using a series of questionnaires to achieve a consensus among experts.

Sales force composite

A qualitative forecasting method relying on estimates provided by sales personnel regarding their specific regions.

Time series

A forecasting technique that uses a series of past data points to predict the future.

Regression Methods

Quantitative forecasting techniques including linear regression and multiple regression.

Trend

The underlying pattern of growth or decline in a time series.

Cyclical patterns

Time series components describing regular patterns in data that occur over long periods of time.

Random variation (noise)

The unexplained deviation from a predictable pattern in time series data.

Seasonal patterns

Time series components consisting of repeatable periods of ups and downs over short periods of time.

Naïve forecast

A simple forecasting method where the forecast for any period equals the previous period's actual value.

Moving Average ($MA$)

A short-term forecasting method calculated using the formula: $MA(n) = \frac{\sum \text{Demand in previous } n \text{ periods}}{n}$.

Weighted Moving Average (WMA)

A forecasting technique that assigns more weight to specific values, usually the most recent past data points, using the formula: $\text{WMA} = \frac{\sum (\text{Weight for period } n)(\text{Demand in period } n)}{\sum \text{Weights}}$.

Exponential Smoothing

A weighted moving average technique where data points are weighted by an exponential function using the formula: $F_t = F_{t-1} + \alpha(A_{t-1} - F_{t-1})$.

Smoothing constant ($\alpha$)

A parameter in exponential smoothing where $0 \le \alpha \le 1$. A value closer to zero results in slower adjustments and greater smoothing.

Error term

In forecasting, the value represented by $(A - F)$, or Actual minus Forecast.

Mean Absolute Deviation (MAD)

A forecast error metric calculated as: $\text{MAD} = \frac{\sum |\text{Actual} - \text{Forecast}|}{n}$.

Mean Squared Error (MSE)

A forecast error metric calculated as: $\text{MSE} = \frac{\sum (\text{Actual} - \text{Forecast})^2}{n}$.

Mean Absolute Percentage Error (MAPE)

A metric used in forecast error analysis to determine the accuracy of a forecasting model.

Labor Productivity

A productivity measure calculated as: $\text{Labor Productivity} = \frac{\text{Units of Output}}{\text{Labor Input (e.g., man-hours)}}$.

Machine Productivity

A productivity measure calculated as: $\text{Machine Productivity} = \frac{\text{Units of Output}}{\text{Machine Time}}$.

Multifactor Productivity

A productivity measure calculated as: $\text{Multifactor Productivity} = \frac{\text{Output}}{\text{Labor} + \text{Materials} + \text{Overhead}}$.

Productivity Growth

The rate of change in productivity calculated as: $\text{Productivity Growth} = \frac{\text{Current Productivity} - \text{Previous Productivity}}{\text{Previous Productivity}}$.

Mission

The element at the top of the organizational hierarchy that guides goals.

Organizational Hierarchy

The top-to-bottom order of planning: Mission, Goals, Organizational strategies, and Tactics.

Time Series Decomposition

Analyzing a data series by breaking it into components such as trend, cycle, noise, and seasonality to project them forward.

What are the two primary categories of basic forecasting methods?

Qualitative and Quantitative methods.

Which qualitative forecasting method involves a group of high-level experts or managers?

Jury of executive opinion.

Which qualitative forecasting method uses a series of questionnaires to achieve a consensus among experts?

Delphi method.

What qualitative forecasting method relies on the estimates provided by sales personnel regarding their specific regions?

Sales force composite.

What are the two sub-categories of quantitative forecasting methods?

Time Series and Regression Methods.

List five examples of time series forecasting methods?

Naïve, Moving averages, Weighted averages, Exponential smoothing, and Trend projection.

What are the two common types of regression methods used in forecasting?

Linear regression and Multiple regression.

The forecasting technique that uses a series of past data points to predict the future is known as _____.

Time series

What are the four components typically identified in a time series decomposition?

Trend, Cyclical patterns, Random variation (noise), and Seasonal patterns.

In time series analysis, what is defined as the underlying pattern of growth or decline?

Trend

What time series component describes regular patterns in data that occur over long periods of time?

Cyclical patterns.

What is the term for the unexplained deviation from a predictable pattern in time series data?

Random variation (or noise)

What time series component consists of repeatable periods of ups and downs over short periods of time?

Seasonal patterns.

According to the common seasonality patterns table, how many "seasons" are in a month if the season length is a week?

4

How many "seasons" are identified in a year when the period length is a year and the season length is a month?

12

Why can statistical methods alone fail to produce accurate practical forecasts?

They cannot account for external factors like sales promotions, natural disasters, or competitive strategies.

What is the first step in developing a practical forecast?

Understand the purpose, time horizon, and level of aggregation.

According to the Naïve forecast method, the forecast for any period equals the .

previous period's actual value

What is one primary benefit of using a Naïve forecast?

It is simple and quick to prepare with virtually no cost.

Moving Average (MA) methods are most effective for what type of planning horizon?

Short planning horizons.

In a Moving Average forecast, how does increasing the number of periods ($n$) affect the forecast's response to recent data changes?

The forecast reacts more slowly to recent changes.

What is the formula for a Moving Average ($MA$) for $n$ periods?

MA(n) = $\sum$ Demand in previous n periods / n

What is the primary difference between a simple Moving Average and a Weighted Moving Average (WMA)?

WMA assigns more weight to specific values (usually recent ones) rather than treating them equally.

In a Weighted Moving Average, which data points are generally given higher weighting?

The most recent past data points.

What is the formula for calculating a Weighted Moving Average?

WMA = $\sum$ (Weight for period n)(Demand in period n) / $\sum$ Weights

How is Exponential Smoothing defined as a forecasting technique?

A weighted moving average technique where data points are weighted by an exponential function.

What is the formula for Exponential Smoothing to find the new forecast for period $t$ ($F_t$)?

F_t = F{t-1} + $\alpha$(A{t-1} - F{t-1})

In the exponential smoothing formula, what does the symbol $\alpha$ represent?

The smoothing constant.

In forecasting, what does the term $(A - F)$ represent?

The error term (Actual minus Forecast).

If a starting forecast is not provided for exponential smoothing, what is the standard assumption for $F_1$?

F_1 = A_1 (the first period's actual demand).

What happens to the exponential smoothing forecast if the smoothing constant $\alpha$ is set to 0?

The forecast does not reflect recent data and stays equal to the previous forecast (F_t = F{t-1}).

What happens to the exponential smoothing forecast if the smoothing constant $\alpha$ is set to 1?

The forecast becomes identical to the previous period's actual demand (F_t = A{t-1}).

When choosing a smoothing constant $\alpha$, what is the result of a value closer to zero?

The forecast adjusts slower to errors, providing greater smoothing.

What is the general objective when selecting a forecasting model or technique?

To obtain the most accurate forecast, typically identified by the lowest forecast error.

What does the metric MAD stand for in forecast error analysis?

Mean Absolute Deviation

How is Mean Absolute Deviation (MAD) calculated?

MAD = $\sum$ |Actual - Forecast| / n

What does the metric MSE stand for in forecast error analysis?

Mean Squared Error

How is Mean Squared Error (MSE) calculated?

MSE = $\sum$ (Actual - Forecast)^2 / n

What does the metric MAPE stand for in forecast error analysis?

Mean Absolute Percentage Error

How is labor productivity calculated?

Labor Productivity = Units of Output / Labor Input (e.g., man-hours)

How is machine productivity calculated?

Machine Productivity = Units of Output / Machine Time

What is the formula for Multifactor Productivity?

Multifactor Productivity = Output / (Labor + Materials + Overhead)

What is the formula for Productivity Growth?

Productivity Growth = (Current Productivity - Previous Productivity) / Previous Productivity

According to the organizational hierarchy in Chapter 2, what sits at the top and guides goals?

Mission

Place the following in the correct hierarchical order from top to bottom: Tactics, Goals, Mission, Organizational strategies.

Mission, Goals, Organizational strategies, Tactics.

Concept: Time Series Decomposition

Definition: Analyzing a data series by breaking it into components (trend, cycle, noise, seasonality) and projecting them forward.

In the context of Moving Averages, what does the variable $n$ represent?

The number of periods or data points included in the average.

What is the range of possible values for the smoothing constant $\alpha$ in exponential smoothing?

0 $\le$ $\alpha$ $\le$ 1

What is the primary goal of using 'Weights' in a Weighted Moving Average forecast?

To make more recent data more significant in the calculation.

If 4 workers installed 720 square yards of carpet in 8 hours, what is the labor productivity per worker-hour?

22.5 square yards per worker-hour (720 / (4 x 8)).