Module #02.

1: Time Series Analysis and

Forecasting Accuracy
Time Series and Forecasting: Overview
A. Past
 Historical Information
B. Present
 Analysis of the Past information to uncover design
or pattern or model.
C. Future
 Forecasting or predicting possible emerging event in
the future based on the model made at present based
on the analyzed past information.

Major Area of Forecasting

A. Demand Forecasting
 Predicts timing and quantity of demand of firm's
commodities
B. Economic Forecasting
 Predicts future business conditions with reference
to the underlying economic factors such as
inflation, GDP, GNP, etc.
C. Technology Forecasting
 Predicts possible technological advancements in the
future such latest digital devices and software and
systems, future, and hardware, etc.

Forecasting Approaches (Reid & Sanders, 2012)

Classification of Forecasting Approaches
1. Qualitative
2. Quantitative
Qualitative Approach Basically
a. Subjective, makes use of professional prudence in
making forecasts.
b. Suitable if the variable involved does NOT have
available historical information or is NOT applicable.
Quantitative Approach Essentially
a. Objective, which makes use of quantitative or
numerical data to make a forecast.
b. Suitable when variable involved has an available
historical quantitative relevant data.

According to Ross (2015):

Concept of Time Series Analyis

Singh and Singhal reiterated in their paper presentation
the conceptual of time series by Ya Lun Chau (2014):
A time series is...

 A set of data depending on the time

 A series of values over a period of time
 Collection of magnitudes belonging to different time
periods of some variable or composite of variables
such as production of steel, per capita income,
gross national income, price tobacco, index of
industrial production.
 (According to YouTube Video) Time series data is
collected from a real life thing that we are
interested in. The data is analyzed using a computer
to give graphic and numeric output. The output of
the analysis tells us more about the real life
condition. It can be used to make informed
predictions of future values.
 (According to YouTube Video) A time series is a set
of numerical measurements of the same entity taken
at equally spaced intervals over time.
  Time is act as a device to set of common stable
reference point.
 In time series, time act as an independent variable
to estimate dependent variables
 (According to Book) a sequence of observations on a
variable measured at successive points in time or
over successive periods of time.

Time Series Plot (According to Book)

 First step to identify the underlying pattern in the

data.
 a graphical presentation of the relationship between
time and the time series variable, time is represented
on the horizontal axis and values of the time series
variable are shown on the vertical axis.

Examples of Time Series

1. Stock price, Sensex Exchange rate, interest rate,
inflation rate, national GDP
2. Retail sales
3. Electric power consumption
4. Number of accident fatalities
5. Average monthly temperature
6. Annual profit
7. Daily petrol price
8. Hourly electricity consumption
9. Quarterly house sales

Time Series Patterns

Anderson et al. (2018) stated the common types of data
patterns that can be identified when examining a time
series plot include:
A. Horizontal
B. Trend
C. Seasonal or Seasonality
D. Trend & Seasonal
E. Cyclical or Cycles
 A. Horizontal Pattern
 A horizontal pattern exists when the data fluctuate
randomly around a constant mean over time or
Stationary.
These data show the number of gallons of gasoline (in
1000s) sold by a gasoline distributor in Bennington,
Vermont, over the past 12 weeks. The average value
Horizontal Pattern or mean for this time series is 19.25 or
19,250 gallons per week.

Figure shows a time series plot for these data. Note how
the data fluctuate around the sample mean of 19,250
gallons. Although random variability is present, we would
say that these data follow a horizontal pattern.

(According to Book)
The term stationary time series is used to denote a time
series whose statistical properties are independent of
time. In particular this means that
1. The process generating the data has a constant mean.
 2. The variability of the time series is constant over
time.
A time series plot for a stationary time series will always
exhibit a horizontal pattern random fluctuations.

B. Trend Pattern
 According Doane and Seward (2016), Trend (T) is the
general movement over all the years (t = 1, 2, ...,
n).
 A mathematical trend can be fitted to any data but
may or may not be useful for predictions.
 Trends may be steady and predictable, increasing,
decreasing, or staying the same.
 (According to Book) Although time series data
generally exhibit random fluctuations, a time series
may also show gradual shifts or movements to
relatively higher or lower values over a longer
period of time. If a time series plot exhibits this
type of behavior, we say that a trend pattern
exists.
 (According to Book) trend is usually the result of
long-term factors:
a. population increases or decreases
b. shifting demographic characteristics of the
population
c. improving technology, and/or
d. changes in consumer preferences.

C. Seasonal Pattern
 exists when a series is influenced by seasonal
factors or in regular interval. (e.g., the quarter
of the year, the month, or day of the week).
  The trend of a time series can be identified by
analyzing movements in historical data over multiple
years.
 Seasonal patterns are recognized by observing
recurring patterns over successive periods of time.

D. Trend and Seasonal Pattern

 Some time series include a combination of a trend
and seasonal pattern.
 In such cases we need to use a forecasting method
that has the capability to deal with both trend and
seasonality.
 Time series decomposition can be used to separate or
decompose a time series into trend and seasonal
components.
 Doane and Seward (2016) said:
a. Seasonal (S) is a repetitive cyclical pattern
within a year (or within a week, day, or other
time period).
b. Over a small number of time periods, cycles are
undetectable or may resemble a trend.
c. By definition, annual data have no seasonality.

E. Cyclical Pattern
 exists when the data exhibit rises and falls that
are not of a fixed period.
 In addition, Doane and Seward (2016) illustrated
that:
a. Cycle (C) is a repetitive up-and down movement
about a trend that covers several years.
b. Over a small number of time periods, cycles are
undetectable or may resemble a trend.
  A cyclical pattern exists if the time series plot
shows an alternating sequence of points below and
above the trend line lasting more than one year.
 Often, the cyclical component of a time series is
due to multiyear business cycles.
 Business cycles are extremely difficult, if not
impossible, to forecast.
 In this chapter we do not deal with cyclical effects
that may be present in the time series.
 A cyclical pattern exists if the time series plot
shows an alternating sequence of points below and
above the trend line that lasts for more than one
year.
 Many economic time series exhibit cyclical behavior.

Rahman (2015) reported the following:

Uses of Time Series
a. To study the past behavior of the variable
b. To formulate policy decisions and planning of future
operations.
c. To predict or estimate or forecast the behavior of the
phenomenon in future which is very essential for
business planning
d. To compare the changes in the values of different
phenomenon at different times

Anderson et al., (2018) averred that...

a. The underlying pattern in the time series is an
important factor in selecting a forecasting method.
 b. Hence, if we observe a trend in the data, then we need
to use a method that has the capability to handle
trend effectively.
c. To select which forecasting method (which will be
discussed in the following chapters) is appropriate or
effective, we need to measure the accuracy of the
forecast. Thus, the following slides will focus on
various ways of evaluating forecast accuracy with
examples.
Forecast Accuracy: Concepts, Terms and Definitions
Anderson (2018) defined the following:
a. Measures of forecast accuracy are used to determine
how well a particular forecasting method is able to
reproduce the time series data that are already
available.
b. Measures of forecast accuracy are important factors in
comparing different forecasting methods.
c. By selecting the method that has the best accuracy for
the data already known, we hope to increase the
likelihood that we will obtain better forecasts for
future time periods.

Forecast Bias

 persistent tendency for forecasts to be greater or

less than the actual values of a time series
Forecast Error

 difference between the actual value and the value

forecasted for a given period.
 The key concept associated with measuring forecast
accuracy

Actual Value - Forecast = Forecast Error

A positive forecast error indicates the forecasting method
underestimated the actual value.
A negative forecast error indicates the forecasting method
overestimated the actual value.
Forecasting Accuracy: General Idea

Several measures of forecasting accuracy follow:

1. Mean Absolute Deviation (MAD)
 a MAD of 0 indicates the forecast exactly
predicted demand.
2. Mean Absolute Percentage Error (MAPE)
 provides a perspective of the true magnitude of
the forecast error.
3. Mean Squared Error (MSE)
 analogous to variance, large forecast errors are
heavily penalized

How to measure forecast accuracy? (Aderson et. al, 2018)

 To demonstrate the computation of these measures of

forecast accuracy we will introduce the simplest of
forecasting methods.
 The naïve forecasting method uses the most recent
observation in the time series as the forecast for the
next time period.

𝑌𝑡+1 = Actual Value in Period t

For Example:
 a. Rosco Drugs Sales of Comfort brand headache medicine
for the past 10 weeks at Rosco Drugs are shown below.
If Rosco uses the naïve forecast method to forecast
sales for weeks 2 - 10, what are the resulting MAE,
MSE, and MAPE values?

Formula:
Forecast Error = Actual – Naïve Forecast
Squared Forecast Error = (Absolute Value of Forecast Error)2
Percentage Error = (Forecast Error / Actual)(100)

Naive Forecast Accuracy

*Add all the variables and divide by the number of items

Some Notes About Measures Of Forecast Accuracy:

a. forecast accuracy may not assure of good forecasts.
 b. forecast accuracy gives an idea of how close the
forecasted values, predicted by the estimated model,
or be alike with the actual values of the
corresponding historical past data.
c. MAD computes absolute difference.
d. MAPE computes relative to the totality of the data and
commonly used when comparing cases of different size
or degree/amount. (e.g case 1 in millions; case 2 in
tens or units)
e. ** the lower the value of any of the 3 measures, the
better quality of forecast accuracy.

Two basic types of Quantitative Forecasting Methods:

a. Naïve
b. Non-Naive