Memory overflow in the Moran's I space statistics

In the fields of economics, resource management, biological geography, Political Geography, and population statistics, there are often the following research demands:

• What is the clearest boundary between the wealthy and poor areas in the study area?
• Can I find the location of the abnormal consumption mode in the study area?
• What is the high incidence of diabetes in the study area?

. Common analysis methods include clustering and outlier analysis anselin local monran's I and Ripley's K. For more information, see

Http://resources.arcgis.com/zh-cn/help/main/10.1/index.html#//005p0000000m000000

Http://resources.arcgis.com/zh-cn/help/main/10.1/index.html#//005p0000000z000000

However, when the analysis model in the ArcGIS Spatial Statistics tool is used for analysis, when the data volume exceeds a certain amount, the memory overflow computing may be interrupted, Which is puzzling. Memory overflow occurs when I use a layer with 30 thousand points for analysis.

Next, we will analyze the cause of the problem. First, read the help document of the analysis tool. The parameters are as follows:

Input featured class: input data, that is, the generated vertex data (test2010.shp)

Input filed: fields to be analyzed (such as F12)

Generate Report: whether to generate a report

Conceptualization of spatial relationships: Concept of spatial relationships

Distance Method: Distance Method

Standardization: Standardization

Distance band or threshold distance: distance threshold

Weights matrix file: Weight Matrix File

Only the first and second parameters are required. Other parameters have default values.

Check the help and find a prompt:

Note: When running this tool, there may be insufficient memory. This is usually the case when you choose to create a conceptual and/or distance range or threshold distance between elements that have many adjacent spatial relationships. Generally, you do not want to define the spatial relationship that will give elements thousands of adjacent elements. You want all elements to have at least one adjacent point, and almost all elements have at least eight adjacent points.

It seems that the distance band or threshold distance parameter uses the default value. Analysis normal!