about K nearest neighbors and Contiguity edges corners in Cluster and Outlier Analysis

[KAMATAYOKO]

Hello,

I use spatial data of 10x10 square polygons with side length of 100m and perform LISA(Cluster and Outlier Analysis) analysis in various ways. Various ways mean spatial weighting matrices constructed with different spatial relationships.

Then the results of the LISA analysis differed when K nearest neighbors was set to eight neighbors and when Contiguity edges corners was used. I would like to know why the results are different.

Thank you,

YOKO