Scribd
Upload
36 views17 pages

Linear Trend Projection in Time Series Analysis

Uploaded by

kobeandreicrufo
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
36 views17 pages

Linear Trend Projection in Time Series Analysis

Uploaded by

kobeandreicrufo
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

CA5103 – Management Science

Linear Trend Projection


Learning Objectives:
At the end of the lesson, the students should be able to:
• apply the linear trend projection method for time series exhibiting
trend patterns.
Note: Forecasting of time series with seasonality will be discussed in
the course Operations Management.
Concept of Time Series
Singh and Singhal reiterated in their paper presentation the conceptual
description of time series by Ya Lun Chau (2014)

Time series data is a two variable data wherein the one variable is
the time element or factor associated with the other variable of
relevant interest.
Example: Year and Sales; Days and Stock Price; Months and Cost of
Goods Sold, etc.
LINEAR TREND PROJECTION
The time series techniques of moving averages and exponential
smoothing related a single variable being forecasted (such as demand or
sales) to time. This techniques are appropriate when a time series exhibit a
horizontal pattern.
In contrast, regression is a forecasting technique that is appropriate for
time series exhibiting trend patterns. Recall the bicycle sales time series
presented in Module 2 Part 2.
LINEAR TREND PROJECTION
The simplest form of regression is linear regression. It is most
often thought of as a causal method of forecasting in which a
mathematical relationship is developed between a dependent
variable and some other factor that affects that variable. This
relationship is usually found in a manner that minimizes the sum of
squared errors.
When the dependent variable displays an obvious trend over
time, a least squares regression line, or linear trend line, can be used
to forecast the dependent variable.
LINEAR TREND PROJECTION
A linear trend line is a linear regression model that relates a
dependent variable to time.
Least squares, also used in regression analysis, determine the
unique trend line forecast or linear trend projection which minimizes
the mean square error between the trend line forecasts and the
actual observed values for the time series.
The independent variable is the time period, and the dependent
variable is the actual observed value in the time series.
Linear Trend Projection (Anderson et al., 2024)

▪ Using the method of least squares, the formula for


the trend projection is:

intercept of the linear trend line at t = 0


Linear Trend Projection (Anderson et al., 2024)

where: t = the time period


Yt = actual value of the time series in period t
n = number of periods in the time series
Linear Trend Projection
Example: Bicycle Sales
Consider the data for bicycle sales over the past 10 years. The linear trend
line provides a reasonable approximation of the long-run movement of the
series. Use linear trend projection to forecast the sales for years 11 and 12.
Linear Trend Projection (Anderson et al., 2024)
Solution: Year, t
Sales,
(1000s of units)
1
2
3
4
5
6
7
8
9
10
Total
Linear Trend Projection (Anderson et al., 2024)
Solution:
Linear Trend Projection (Anderson et al., 2024)

The linear regression equation for the time series is:

The forecasts for the sales of years 11 and 12 are:


Note that the forecasted value
increases by 1,100 bicycles in
each year.
Linear Trend Projection (Anderson et al., 2024)
Forecast accuracy of the linear trend projection method for the bicycle time
series

Note that minimizing the sum of


squared errors also minimizes MSE.
Seasonality
Developing forecasts for a time series that has a seasonal
pattern (either with or without a trend) will be discussed in
the course Operations Management.
Assignment: Answer the given below problems (Anderson et
al., 2024, p. 700).
Assignment: Answer the following problems.
(Anderson et al, (2024), page 701)
References:

Anderson, D. R., Sweeney, D. J., Williams, T. A., Camm, J. D., & Cochran, J. J.
(2019). An introduction to management science: quantitative approach.
Cengage learning.

Cengage Learning Asia Philippines (2016). Chapter 15, Part B


Time Series Analysis and Forecasting [PowerPoint slides].

Ya-Lun-Chau,"Time series establish relation between “cause” & “Effects”.


(Feb.19, 2014)

Common questions

Powered by AI

Discussing forecasting of time series with seasonality separately is essential because seasonality introduces periodic fluctuations that are not captured by linear trend methods, which focus on long-term directional movement only. Properly accounting for seasonality requires specific models or adjustments to avoid forecasting errors and accurately reflect the cyclical nature of the data .

The method of least squares optimizes the accuracy of a linear trend projection by minimizing the sum of squared differences between the observed values and the values predicted by the trend line. This minimization also reduces the mean squared error (MSE), which enhances the forecast accuracy for the dependent variable over time .

To forecast a time series with both a consistent trend and cyclical seasonal variations, a combination of techniques should be used. Initially, one could use linear trend projection to model the long-term trend. For the seasonal component, decomposing the time series into seasonal indices can help adjust the forecasts to account for regular periodic shifts, thus enhancing forecast accuracy for such complex datasets .

Linear trend projection may not be suitable for forecasting time series with seasonality because it does not account for periodic fluctuations present in seasonal data. This method focuses solely on trend patterns over time and ignores regular intervals of variation, leading to inaccurate forecasts if seasonality is significant .

The linear trend line is considered a causal forecasting method because it establishes a mathematical relationship between a dependent variable and time, indicating the effect of time on changes in this variable. This causal link is utilized to forecast future values by projecting the established trend into the future based on past data .

The intercept of the linear trend line is calculated using the method of least squares, representing the predicted value of the dependent variable when the time period (independent variable) is zero. It acts as a starting point for the trend line and influences the placement of the trend within the data range .

The primary difference lies in the data patterns each method is suitable for. Linear trend projection is used when a time series exhibits trend patterns, as it uses linear regression to relate a dependent variable to time and forecasts future values. Conversely, moving averages and exponential smoothing are appropriate for time series with horizontal patterns, predicting future values based directly on past observations .

Minimizing mean squared error (MSE) enhances forecast accuracy by ensuring the forecasted values are as close as possible to the observed values in a time series. It quantifies the average of the squares of the errors, which are the differences between observed and predicted values. Reducing MSE thus directly improves the precision of the forecasts generated by the linear trend projection .

The statement implies a constant increase in the forecasted values, indicating a consistent upward trend in the time series data. For each subsequent year, the linear trend projection predicts an increase of 1,100 units, reflecting a steady growth pattern over time .

Understanding the concept of 'time series' facilitates the effective use of linear trend projection by providing insights into the systematic segmentation of data into time and dependent variables. Recognizing trends, cycles, and patterns within this framework allows for more accurate application of linear regression to predict future data points, making adjustments for time-based changes .

You might also like