UNIVERSITY OF BENIN, NIGERIA
DEPT: GEOGRAPHY AND REGIONAL PLANNING
NAME: OSAZUWA OSAZE SUNDAY
MATRICULATION NUMBER: SSC1205599
LEVEL: 300L
COURSE CODE: GEO 326
COURSE TITLE: ADVANCE QUANTITATIVE
TECHNIQUES
TITLE: NEAREST NEIGHBORS ANALYSIS
20, JULY
INTRODUCTIONGeography deals with the study of spatial distribution in space. This is to say that
although phenomena occur in individual location they are distributed in space forming a
pattern. Phenomena do not only locate in space but also interacts with one another.
Knowing the pattern of occurrence that some particular phenomena in space is a major
geographical problem which required analysis. Thus to know the pattern of occurrence of
phenomena in space we study the relationship between individual phenomena using
various methods. An example of the method used in the search for pattern in landscape of
geographical area is the nearest neighbor analysis. The question then arise 'what is nearest
neighbor analysis?'.
NEAREST NEIGHBOR ANALYSISGeographers for a long time have been interest in how features are distributed in
space and this interest made them study the patterns of the phenomena in space which is
known as pattern analysis. Pattern analysis in this case may be defined or described as the
mean distance or distances between an array of points in space. The aspect of pattern
analysis which is usually used un the point pattern analysis of which the nearest neighbor
is a part. Different scholars has defined nearest neighbor analysis and one of this
definition was given by Upton. Nearest Neighbor Analysis according to Upton (1985), is
a method of exploring pattern in locational data by comparing graphically the observed
distribution of functions of event-to-event or random point-to-event nearest neighbor
distances, either with each other or with those that may be theoretically expected from
various hypothesized models, in particular that of spatial randomness. That is Nearest
Neighbor Analysis uses the distance between each location (points) in a layer to
determine if the point pattern are clustered, random or regular.
The Nearest Neighbor Analysis technique was devised by a botanist who wished
to describe and provide a quantitative description of the patterns of plant distribution
especially the distribution of trees. The early beginnings of nearest neighbor analysis can
be attributed to the pioneering works of P.J Clark and F.C Evans in 1954 is their attempt
to describe and analyze the pattern and distribution of trees and other plants in the forest.
However since geographers are interested in the study of the pattern of distribution of
phenomenon over space the techniques of the nearest neighbor has since been adapted for
geographical studies. As such the nearest neighbor analysis has since evolved to it been
used to identify a tendency towards or calculate the degree of nucleation (clustering) or
dispersion of phenomena in space. The nearest neighbor analysis can be used to analyze
the distribution of schools, hospitals, buildings, settlement and a myriad of physical
features such as wells, springs mountains, hills etc on the earth surface.
The nearest neighbor analysis since its introduction into geographical studies has
found more application in settlement system analysis. On maps settlements often appears
as dots. Dot patterns are difficult to describe even though they are highly useful in
portraying distribution of phenomenon in geographical studies. Nearest neighbor analysis
attempt to find the order of arrangement of this settlements on the map. This attempts to
measure the distributions according to whether they are clustered, random or regular. As
such settlement are expected to be either clustered, random or regular. Nearest neighbor
analysis allows us to compare and contrast between distribution and also help to monitor
and compare changes occurring over a period of time.
DESCRIPTION OF NEAREST NEIGHBOR STATISTICS
As I have said before nearest neighbor analysis attempts to measure distribution
according to whether they are clustered, random or regular but these pattern appear on
map in such a way that they are judged or analyze subjectively using guess work in the
course of analyzing the map manually. These is where the nearest neighbor statistic
comes in. The nearest neighbor statistics is the method or formula used in nearest
neighbor analysis and is a means of standardizing the assessment of a distribution either
on map or ground objectively that is using the scientific method and calculations to
determine whether phenomenon are clustered, random or regular.
The nearest neighbor statistics produce a figure (expressed as Rn) which measures
the extent to which a particular distribution on map or land is clustered, random or
regular. After calculations the nearest neighbor formula will produce results between
zero(0) and 2.15 where the patterns of the distribution forms a continuum. These pattern
are explained below for clarity.
CLUSTERED PATTERN: A cluster patterns occur when all distribution (in this
case represented by dot) converge closely to a point. Another name for clustered
distribution is nucleated distribution. For example in settlements, nucleated
settlements are made up of buildings clustered together in an area. These areas are
usually the meeting point of transport routes, such as rivers, roads and railway lines.
On the continuum a perfectly clustered distribution will have an Rn value of zero (0)
but since there are no perfectly clustered distribution the Rn values tend to be above 0
but below one(1).
RANDOM PATTERN: A distribution is termed random when it does not show a
pattern at all. In this situation the Rn value will be equal to 1.0. Another name for
random pattern is dispersed pattern. For example in settlement analysis a
random/dispersed settlement patterns are made up of individual buildings scattered
over a wide area. These buildings may be separated by large open spaces, farmlands,
forests or grasslands. The usual pattern of settlement is one which is predominantly
random with a tendency either towards clustering or regularity.
REGULAR PATTERN: When distribution exist in a perfectly uniform manner in
space or on the map we say the distribution has a regular pattern. Although a perfectly
regular pattern is not possible in reality but assuming it exist in reality, the Rn value
on the continuum will be 2.15 which will then be explained that each dot (settlement)
are equidistant from all its neighbors.
On the nearest neighbor statistics continuum the patterns above are arrange on the
continuum in such a way as shown below with zero(0) representing the clustered
pattern, one(1) representing the random pattern while 2.15 on the continuum
represents the regular patter
.
From the graph above the value of Rn=1.27 shows a distribution that is random but
tending toward the regular.
METHODOLOGY FOR CALCULATIONS OF THE NEAREST NEIGHBOR STATISTICS
The nearest neighbor formula used to calculate the nearest neighbor is given as:
Rn = 2đ√n/A
Where
Rn = is the nearest neighbor value describing the distribution.
đ = this is the mean distance between the nearest neighbors.(km)
n = this represent the total number of distribution(points) in the study area.
A = This represents the study area(km²).
The procedure to follow when calculating the nearest neighbor statistics is as follows
The first thing to do if one wants to calculate the nearest neighbor statistics is to first
locate the area where the distribution (houses in a settlement or settlements in a town
or any other phenomenon of interest) or get the map showing the distribution.
After getting the location or map of the distribution the next step is to find the mean
distance đ which is usually a straight line measure of the distance between each
phenomenon or settlement and its nearest neighbors. The mean distance is gotten by
calculating the sum if all the distance gotten and dividing it by the total number of the
distribution (points) under study. Using the formular below:
đ =Σd/nWhere:
Σ = sigma means to sum up the values of distance.
d = distance of phenomenon from the nearest neighbor.
n = total number of phenomenon.
The third procedure in computing nearest neighbor statistics is to determine the area
of the map or the area the phenomenon is occupying by multiplying the length by the
breath of the location in km².
The last step is to calculate the nearest neighbor statistics by inputting or substituting
all the elements unto the nearest neighbor formula and computing to the the nearest
neighbor value.
LIMITATIONS AND PROBLEMS OF NEAREST NEIGHBOR ANALYSIS
LIMITATIONS
Although the nearest neighbor has been variously applied and found to be useful in
geographical studies, care must be taken when using the technique. However the
following limitations affects the usefulness of the technique.
The first of the limitations is that of the size of the area chosen. If two regions are
under observation in nearest neighbor analysis, the two regions must have similarly
size if the comparison between the two regions is to be valid.
The size of the area of the chosen distribution must not be too large as this will reduce
or lower the Rn values(as it exaggerates the degree of clustering of the distributions)
nor must the area be be to small as this will increases the Rn values (as it exaggerates
the level of regularity).
The boundaries of the area under study is also critical most especially where the area
under study is part of a larger area or region. Hence delineating or choosing of such
boundaries must be done carefully. This kind of problem is called the area modification problem.
PROBLEMSAs useful as nearest neighbor analysis is its has some major problems of which the
following are common.
When carrying out a nearest neighbor analysis of a valley where the nearest neighbors
are separated by a river, distortions are likely to occur in such analysis.
The choice of distribution or features to be included in the nearest neighbor a
statistics race rises a lot of questions for example what size of area a acceptable, are
we to includes farmsteads or hamlets.
The nearest of a point or features maybe outside a study area.
How to determine the center of a settlement is a subjective decisions.
REFERENCESWaug .D (1995), geography an integrated approach, Thomas and son's limited
http://www.hbp.usm.my/Thesis/HeritageGIS/master%5Cthesis%5C3-Nearest
%20neighbor%20analysis.htm
http://ceadserv1.nku.edu/longa//geomed/ppa/doc/html/ppa.html
https://en.m.wikipedia.org/wiki/Nearest_neighbor_search?
_e_pi_=7%2CPAGE_ID10%2C5837470691
http://geographyfieldwork.com/nearest_neighbour_analysis.htm