api:dk11:delphi:gisclassification.tgis_classificationmethod

DK11 for Delphi
| GisClassification.**TGIS_ClassificationMethod**
| Enums

Enumeration of classification methods.

// Delphi type TGIS_ClassificationMethod = ( DefinedInterval, EqualInterval, GeometricalInterval, KMeans, KMeansSpatial, NaturalBreaks, Quantile, Quartile, StandardDeviation, StandardDeviationWithCentral, Unique, Manual );

// C++ Builder enum DECLSPEC_DENUM TGIS_ClassificationMethod unsigned short { DefinedInterval, EqualInterval, GeometricalInterval, KMeans, KMeansSpatial, NaturalBreaks, Quantile, Quartile, StandardDeviation, StandardDeviationWithCentral, Unique, Manual };

Name | Value | Description | |
---|---|---|---|

DefinedInterval | The interval between class breaks is defined by the user through the Interval property. Number of classes is determined automatically and can be gotten from the NumClasses property. | ||

EqualInterval | The range of attribute values is partitioned into equal-sized subranges by defining the number of classes through the NumClasses property. | ||

GeometricalInterval | Class breaks are based on class intervals that have a geometrical series. User specifies the number of class intervals through the NumClasses property. | ||

KMeans | Class breaks are determined using the K-Means Clustering Algorithm. All values from an attribute are divided into N=NumClasses clusters so the within-cluster sum of squares is minimized. | ||

KMeansSpatial | K-Means algorithm is used to partition data into N=NumClasses spatial clusters based on objects centroids. Available only for vector layers. New field named 'KMEANS_ID' will be added or updated if it exists. | ||

NaturalBreaks | Class breaks are determined using the Fisher's Exact Optimization algorithm. Dataset is divided into N=NumClasses clusters so that the between-cluster sum of squares is maximized. This is default method. | ||

Quantile | Class intervals are created to have an equal number of observations. Number of classes can be defined through the NumClasses property. This method is also known as Equal Count. | ||

Quartile | Quantile method is applied with fixed property NumClasses=4. This method creates class breaks at the first, second, and third quartiles i.e. at 25-th, 50-th, and 75-th percentiles respectively. | ||

StandardDeviation | Class breaks are set above and below the mean of the attribute values at intervals, for example at 1, 1/2, 1/3, or 1/4 standard deviations until all the data values are contained within these ranges. Interval is specified through the Interval property. This method produces even number of classes. | ||

StandardDeviationWithCentral | Same as StandardDeviation but class breaks are shifted by a half of the interval to get central class in range [Mean-Interval/2..Mean+Interval/2]. For example, if Interval equals 1 Std Dev, central class is limited by breaks -0.5 Std Dev and 0.5 Std Dev. This method produces odd number of classes. | ||

Unique | Class breaks are set for unique values. Number of classes can be limited through the NumClasses property. | ||

Manual | Class breaks are set manually by the user through the AddClassBreak method. This method is used when classes are meaningful. | ||

**K-Means Spatial**

By default, `KMeansSpatial`

classification method works with 2-dimensional data - using objects' centroids (X, Y).

To perform an algorithm in 3D, set `Field`

property with Z-data. You can use one of the built-in fields, such as `GIS_COORD_Z`

, `GIS_CENTER_Z`

, or any numerical existing field.

2022/11/16 01:14