Classic Time Series Model
Uploaded by
ScribdTranslationsClassic Time Series Model
Uploaded by
ScribdTranslationsCLASSIC
SERIES OF
INFERENTIAL STATISTICS II 41V INDUSTRIAL ENGINEERING
-= - - —= - - ..... -- - _ — —
• GUTI ER REZ CASTELLANOS VICENTE
• MENDIOLA SOTELO OSCAR DAMIAN
• VARGAS TORRES LAURA YURITZI
~ ~ . --------------
In Statistics, this is the name given to a set of values observed during a
series of sequentially ordered time periods, such periods can be
weekly, monthly, quarterly or annually .
A time series is a set of numerical values obtained in periods
By time series we refer to statistical data that is collected, observed or
recorded at regular time intervals (daily, weekly, semi-annual, annual,
among others).
The term time series is applied, for example, to data recorded periodically
that show, for example, total annual warehouse sales, the quarterly value
total GDP quarterly contracts. ados, the value
Graph 1 Production vs
3
Years
1 -I---------------1--------------t--------------t---------------1-------------,
1970 1975 1980 1985 1990 1995
Years
AIM
The fundamental assumption of time series analysis is that factors that
have influenced activity patterns in the past and present will have more or
less the same influence in the future.
Analyzing a time series has the following objectives, among others:
■ Determine if certain patterns or non-random patterns occur
■ Isolate and then study its components in order to provide clues for future
movements.
■ It makes it possible to forecast future movements as well as other
aspects that are synchronized.
EXAMPLE GRAPHICS
TIME SERIES
It is represented by a line graph on whose horizontal axis the periods are
represented and on whose vertical axis the values of the time series are
represented.
ANNUAL GASOLINE SALES IN THE CITY
FROM TOLUCA
D 900 i Evolution of Number of Users with narrow band connection
(Connection speed less than 1Mbps)
ME 850 - , 1 800 -
L
L 750 - -----
Or 1 700 '
N T 650 ■ „*
R
E 600 -
S 0 550 -
-S
500 -I---------1--------1--------1--------r--------r--------1---------1---------1-------
1995 1998 1997 1998 1999 2000 2001 2002 2003
YEARS
CHARACTERISTICS
► To carry out an analysis of this type, the components of the time
series must first be identified , then statistical techniques must be
applied for its analysis and, finally, projections or forecasts of future
events must be made.
► In this way, time series analysis is the procedure by which time-
related factors that influence the values observed in the time
series are identified and isolated so that once identified , these
factors can contribute to the interpretation. of historical time series
values and until then forecast future time series values.
COMPONENTS OF THE
TIME SERIES ". 1
The classical method identifies four influences or components:
• Trend (T)
• Cyclical fluctuations (C)
• Seasonal variations (E)
• Irregular variations (I)
Which have a multiplicative relationship that gives shape to the classic time
series model, that is, for any designated period in the time series, the value
of the variable is determined by the four components in the following way:
It is the general long-term movement of the values of the time
series (y) over an extended period of years .
They are long-term trends in sales, employment, stock prices, and
other economic and business series (without alterations of a time
series). The secular movement presents smooth long-term
movements, which are dominated by economic factors.
Long-term sales, job offers, section prices are examples of secular trend.
They are measured in years, some move continuously upward, others decline, others
remain unchanged.
Figure 6.
CTUATIONS
CYCLIC (C)
Recurrent upward and downward movements with
respect to the trend lasting for years.
It is the rise and fall of a time series in periods greater
than one year. The cyclical component is the wave-like
fluctuation around the trend, thus regularly affecting the
general economic conditions.
• Employment, production, stock prices, are examples of cyclical fluctuations.
• They are measured in years.
• They represent rise and fall in periods greater than one year.
They suffer from a period of prosperity followed by
recovery.
I• 11 o -11 a'e
VARIATIONS
Upward and downward movements with respect to the trend that are
completed within one year and repeated annually.
The seasonal component refers to a pattern of change that repeats
itself year after year. In the case of monthly series, the seasonal
component measures the variability of the series, for example, for
January, February, etc. In the quarterly series there are four
seasonal elements, one for each quarter.
High sales at Christmas and low sales afterwards, Consumption related to the
seasons of the year are examples of seasonal variations.
They are only appreciated if you have quarterly or monthly data.
The pattern of the collection is repeated every year, and I agree that this is an
annual series.
Stciorlid=d
Erratic variations from the trend that cannot be attributed to cyclical or
seasonal influences. The random component measures the variability
of the time series after the other components are removed . Accounts
for random variability in a time series caused by unforeseen and non-
occurring factors. Most irregular components are made up of random
variability, however, unpredictable events can cause irregularity in a
variable.
Wars, Strikes and Natural Disasters are examples of irregular variations.
These variations can be predicted and measured but not with certainty.
(d)
Time in years
A Time Series is called a set of measurements of a certain phenomenon or
experiment recorded sequentially in time , for example every hour, monthly,
quarterly, semi-annually, etc.
Discrete time series, equally spaced in which case it is assumed that: {x(t1), x(t2), ...,
x(tn)}= {x(1), x(2), ..., x (n)}. Due to the introductory nature, it was restricted to the case
of univariate time series.
When analyzing a time series, the first thing to do is graph the series. This allows us to
detect the essential components of the series . The series graph will allow: detect
Outlier , detect trends, seasonal variation, irregular variations (or random component
).
A classic model for a time series can be expressed as addition either
product of three
components: trend, seasonal and a random error term . There are three time series
models. These are:
1 . Additive: X(t) = T(t) + E(t) + A(t)
2 . Multiplicative: X(t) = T(t) · E(t) · A(t)
3 . Mixed: X(t) = T(t) · E(t) + A(t)
In order to obtain a model, it is necessary to estimate the trend and seasonality . To
estimate the trend, it is assumed that the seasonal component is not present. The
estimation is achieved by fitting a time function to a polynomial or smoothing the series
through moving averages. To estimate seasonality, it is necessary to have decided on
the model to use I "NA NI-IINN III- AeI I - m - I - AANI- I -
í vIM- \f
Time series can be cited in different areas:
eeme ■ x ■
CAricc AcAnAricAc
Economic series
Physical
series
Geophysi
cs
4 ' II' Demographic series
time series
Marketing series
Telecommunication
series
Transport series
TIME SERIES EXAMPLES
- Prices of an item
- Unemployment rates
- Rate of inflation
1. Economic series: - Price index, etc.
- Meteorology
- Amount of water fallen
2. Physical Series: - Daily maximum temperature
- Wind speed (wind energy)
- Solar energy, etc.
3. Geophysics: - Seismology series
- Population growth rates
4. Demographic series: - Birth rate, mortality
- Population census results
5. Marketing Series: - Demand series, expenses, offers
6. series of - Signal analysis
telecommunication:
