Anetta Zielińska
Akademia Ekonomiczna we Wrocławiu, Polska
An application of
multidimensional comparative analyses for the estimation of environment
condition
1. Introductory problems
Humans are strongly affiliated with natural
environment and it is not possible to overcome those bonds. The influence of
man on the nature (antropo – pressure) has disturbed the equality between
different forms of economical and social activities of man and the quality of
life.
Development
of human civilization, connected among other things with producing different
sorts of goods, leads on one side into
improvement of quality of life but on the other side into creating all sorts of
wastes, depredating natural surroundings.
The
environment plays a lot of different functions, beginning from biological,
constituting life sustaining settings, through resources – creating function
and producing one, ending at cultural and civilization function.
Environmental
protection in the side of the environmental law means undertaking or giving up
those actions which are helpful in protecting or restoring natural equality.
Those efforts depend on a rational action towards environment modeling
according to the principles of sustainable development and they counteract
pollution and can help to restore those natural elements to their original
state[1].
Environmental protection includes water, soil, and air.
All
people are responsible for environmental protection which comes out straight
from Polish Constitution and other legal sources. There are some series of laws
regulating the way of executing designated treatments to protect the
environment. The accomplishment of those tasks should assure ecological
safeness and sustain natural equality. So the environmental questions should be
taken into consideration with different realms of life such us society, economy
and they should play important role in different strategies and programs of
development at local, national, regional development.
Trying
to diagnose the condition of environmental protection for different enquiring
objects (eg. household, company, local units of administration and government)
we face some problems with establishing collective measurement for quality of
environment. It is expected to consider what kind of methods should be taken to estimate the quality of
environment; for what enquiring objects the investigation should be undertaken
and what describing environment feature should be appointed.
To
answer those questions the multidimensional
comparative analyses (VAP) should be taken under the consideration. The method
is helpful in estimating the level of environmental quality for investigating
objects. The knowledge should be applied in the process of making proper
economical – social – environmental decisions.
The
aim of the article is to illustrate possibilities for the application of
multidimensional comparative analyses in the estimation of environment
condition for different objects according to meaningful amounts of features,
which play rational role on natural environment.
2. Environmental data
Estimating environmental condition does not
belong to an easy task and it requires familiarity of some details concerning
environment: water, biota, quality of air and some environmental determiners
such as: waste, business, agriculture, forestry, tourism and so on. Vast range
of information concerning environment exclude possibilities of getting
knowledge about all components. That is why recognition and estimation of
values describing environment should be limited to some chosen elements which
could be presented in a figure of environmental data.
Information
of environmental data include, among others, the following things:
·
Natural
conditions (geographic, hydrographic);
·
Condition
and changes in exploiting resources of the earth, threats and protection of the
soil;
·
Protection of nature, landscape and
biodiversity;
·
Waste
products created by industry and community;
·
Jon
radiation and noise;
·
Economical
aspects of environmental protection[2].
To overhaul many-sided environment
analyzes it should be taken variables in three aspects:
·
Environmental,
·
Social,
·
Economical.
Therefore those three areas are
taking into consideration on measuring the natural environment and they are
presented by table 1.
Taken into consideration mentioned above areas
have strong impact on environment, or in
other words environmental antropopression which takes place in all possible
objects ( eg. industry, units of territorial administration government). They
do not cover all aspects connected with economy society and natural
environment.
Table 1. Fields having meaningful impact on a
natural environment
|
NATURAL ENVIRONMENT |
||
|
ENVIRONMENTAL ASPECT |
SOCIAL ASPECT |
ECONOMICAL ASPECT |
|
Climate |
Demography |
Material
supply |
|
Noise |
Education |
Resource
absorbency of economical growth |
|
Wastes |
Health
and its protection |
Development
of economical activities |
|
Water
resources, its consumption |
Quality
of life |
Sustainability
of economical growth |
|
Water
quality |
Economical
activity of people |
Economy
openness |
|
Soil |
Security |
|
|
Air quality |
Communication |
|
|
Structure
of soil usage |
||
|
Usage of
pesticides |
||
|
Nature
protection and forestry resources |
||
|
Biodiversity
protection |
||
|
FIELDS HAVING IMPACT ON NATURAL ENVIRONMENT |
||
The source: own elaboration.
3. Multidimensional
comparative analyses
Differentiation of available data in
comparative analyzes describing natural environment determines necessity of
using some various measurements with different extends. It makes impossible to
compare situations on different levels, and sometimes on this same level when
single types events happen changeable.
Advisable in this case it would be to find such a method which enables natural
environment estimating process objective [3].
It is possible to choose multidimensional comparative analyses method, which
aims at multidimensional comparative research (WAP). It is a tool that helps in
establishing statistical regularity
around investigated aggregation where single units are described by relatively
representative set of environmental data.
Advantage
of aggregative measures makes available surveying multidimensional occurrences
as well as settling objects and qualification their position according to
established criteria. Essential domain of aggregative measures is composed
qualification by means of one value of number – it makes possible comparative
analyzes and it systemizes partial figures, which might not be enough readable
in univocal estimations. Those measures are also limited: they lead into
simplified establishments and sometimes it is not possible to settle down their
direct interpretation[4].
The
analytical problem, which refers to the condition of natural environment, is a
complex phenomena. Classified objects are characterized by many features, which
makes difficulties of those objects qualifications by formal procedures, which
are very helpful for objective research of the problem in complex phenomena.
First
step for multidimensional comparative analyses consist in standardization of
variables (quality of elements having impact on the environment). Drawing up of
standardization of variables must not limit to the situation where all possible
variables takes form of stimulants, that means they have got positive influence
on investigated natural environment. In case of variables, happening in the
setout, which remind destymulants or noninants, there is a necessity to fix
them up as stimulants[5].
Next
operation in VAP depends on eliminating names (values of elements having impact
on environment) within investigated variables and standardizing the orders of
magnitude for making them comparable (nominalization). The following, most
often, forms of nominalization are[6]:
where x
- arithmetical mean j environmental data (feature),
S j – standard deviation j environmental data (feature),
xoj – the base of normalization j environmental
data (feature), which might be represented by Sj, Rj, Xj, max (xij), min (xij).
After
using one of the normalized forms for all environmental variabilities we
receive standardized matrix of data (Z), which is valuable for next
estimations.
In
case non-standard formula, as pi takes form of normalized mean of environmental
values (features) for particular objects. For “the best” object it is chosen
that one which demonstrates max value pi, on the other hand “the worst” one is
that with – min pi:
Subsequent
to those operations we can count chosen measure of VAP. There are two types of
classification in multidimensional comparative analyses:
·
hierarchical
classification,
·
nonhierarchical
classification.
Hierarchical classification uses linear ordering methods of objects
setting. The aim of those methods depends on arranging (pointing sequences) of
objects or their settings. Those methods can be used only in situations, when a
certain superior criteria is possible to established for sequencing objects
from “the best” to “the worst”. The tool for the method of linear ordering is
synthetic measure of development (SMR), which plays certain function of
aggregating particular information contracted in each variables and appointed
for all objects. Aggregating formula of variables might be divided on standard
and non – standard once[7]
.
Standard formula depends on
inquiring of particular objects distances from standard object, which in
general is lower or upper pole of development[8]:
·
upper
pole of development which includes the most advantageous values of particular
environmental data (features); “the best” object occurs that one which assume minimal
value pi, but “the worst” that one with max pi;
·
lower
polar of development (including less advantageous values of particular
environmental data (features); “the best” object occurs that one which assume
maximal value pi, but “the worst” that one with min pi
where:
pi – synthetic measure of development
for i object;
m – number of data describing environment (feature);
zij – standardized value of j environment data (feature) in i object;
z0j – coordinate j environmental data (feature) of object pattern.
In case non-standard formula, as pi takes form
of normalized mean of environmental values (features) for particular objects.
For “the best” object it is chosen that one which demonstrates max value pi, on
the other hand “the worst” one is that with – min pi:
Non – hierarchical classification depends on
partitioning of heterogeneous set of the objects among certain number of
classes, where similar objects are located, but in different classes the objects
are not similar to each other because of established environmental data values.
The
object similarity extend is described by the measure of similarity, among which
the measure of distance is used the most often. The measure of distance is
regulated in <0;1> interval. The measure of distance values closer to 1
point that objects are not much similar
to each other because of established environmental data values, but numbers
closer to 0 show high level of similarity among objects.
The
most frequent measure of distance (dij) is used Hamminga “municipal” measure[9]:
where:
zrj, zsj – normalized values of environmental data for
objects “r” and “s”; j = 1,2,…,m.
The
measure suppose equal scales for all enquiring environmental data, that means
they are attributed to equal importance to investigated environmental objects.
In practice such a procedure is often questioned and it is suggested to use
“scaled” measures of distance.
Among
them, the most known measures are:
·
·
Bowl
method (cluster method).
Bowl method supposes
that classified objects represent set of points in the pole of m – dimentional. For each of them it
should be created a bowl with a radius “R” and assigned number of objects
located inside each of them. Those objects are treated as “similar” to the
objects composing center of the bowl. Procedure
of objects set division depends on establishing radius of the bowl assigning
numbers of points inside of each bowl according do the matrix. The bowl consist
the biggest number of points constituted in the bowl. Each enquired object has
to be qualified into certain class. There are different procedures of
establishing radius of the bowl - R[10]:
1) The interval should limit radius and
it might bi the number established arbitrary:
2) The radius appointed according to statistic
formula:
4. Resumption
Methods of multidimensional comparative
analyses constitute valuable source of information, because using meaningful
set of data describing objects it is possible to make classification, appointing
the best object or choose sets of objects which are equal to each other
compatible with the condition of environment.
Applied
characteristic of WAP leads to following conclusions:
·
There
is increasing interest with environmental problems, so the application for
comprehensive information becomes necessary. There is a big need of measuring
and enquiring the state of environment;
·
The
important objective depends on selection features in proper way, what is
connected with the access to the environmental information. It is important to
consider either positive and negative information. Insightful and essential
analyzes require comprehensive acquaintance of problem.
·
It
is necessary to establish proper ranges for particular features because there
are not all enquired elements which could be considered as having impact on
environment.
·
Considered
measures might be helpful in case of making decision concerning environment.
Literature
1.
Gatnar
E., Walesiak M. (red.), Metody
statystycznej analizy wielowymiarowej w badaniach marketingowych, Wydawnictwo
AE, Wrocław 2004.
2.
Grabiński
T., Wielowymiarowa analiza
porównawcza w badaniach dynamiki zjawisk ekonomicznych, Zeszyty
Naukowe AE w Krakowie, Seria specjalna monografie nr 61, Kraków 1984.
3. Kozłowski S., Czy transformacja polskiej gospodarki zmierza w kierunku rozwoju zrównoważonego, [w:] Mechanizmy i uwarunkowania ekorozwoju, t. I, Wydawnictwo Politechniki Białostockiej, Białystok 1996.
4. Mazurski K. R., Nasilenie zanieczyszczenie środowiska Polski w ocenie ilościowej, [w:] Problemy terenów zanieczyszczonych w Europie Środkowej i Wschodniej, Race News Special Issue, Katowice 1999.
5.
Ochrona środowiska 2005,
Główny Urząd Statystyczny, Warszawa 2006.
6. Pluta W., Wielowymiarowa analiza porównawcza w modelowaniu ekonometrycznym, Państwowe Wydawnictwo Naukowe, Warszawa 1986.
7. Strahl D. (red.), Metody oceny rozwoju regionalnego,
Wydawnictwo Akademii Ekonomicznej, Wrocław 2006.
8. Ustawa Prawo ochrony środowiska
z dnia 27 kwietnia 2001, Dz.U. Nr 62, poz. 627.
9.
Zielińska
A., Sej-Kolasa M., Excel w statystyce, materiały do ćwiczeń,
Wydawnictwo Akademii Ekonomicznej im. Oskara Langego we Wrocławiu,
Wrocław 2004.