GWR - significance column in attribute table

[MirelaTurk]

Hi.

I am running GWR in my research and I don't know what do mean column 'Significance' and 'Pseudo-T' in attribute table after running GWR. These are new column in layer's attribute table after running GWR. What does it mean this 0 and 1? Can anyone help me. 

Thank you. 

[DanPatterson]

Geographically Weighted Regression (GWR) (Spatial Statistics)—ArcGIS Pro | Documentation

for reference and see

How Geographically Weighted Regression (GWR) works—ArcGIS Pro | Documentation

under Interpreting messages and diagnostics

[Omar_A · Accepted solution]

Hi Mirela,

The "Significance" column in the attribute table indicates whether the parameter estimate for each variable is statistically significant or not. A value of 1 indicates that the parameter estimate is significant at the specified level (usually 95% confidence interval) and a value of 0 indicates that the parameter estimate is not significant.

The "Pseudo-T" column in the attribute table represents the t-statistic for each parameter estimate in GWR. The t-statistic measures how many standard errors the estimate is away from zero. The larger the absolute value of the t-statistic, the more significant the estimate. The significance level of the estimate can be determined from the associated p-value.

More about the GWR here:

How Geographically Weighted Regression (GWR) works—ArcGIS Pro | Documentation

Best,

Omar

[MirelaTurk]

Thank you very much. Yes, it helps. 

[LeonKincy]

Hello. What column is the associated p-value?  I have the following fields:

Std. Error, Pseudo-T, Predicted , Residual, Std. Residual, Influence, Cook's D, Condition Number, Local R-Squared. Can you explain each of these? I can't find any documentation on them.

[MirelaTurk]

Hi. The associated p-value can be determed according to the Pseudo-T, Student's table and degrees of freedom. 

In my case it was 0.01or 99% significance. I don't know if it vould be determinated before running the GWR.

For other statistical indicator you wrote look at some book about statistic. I also can't find any documentation on them. Local R-Squared is R-Squared for each entity in the attribute table (rows in attribute table). 

I hope it helps.

Mirela

[JamalNUMAN]

I got confused regarding the equation used to calculate the Cook’s D.

• Do we divide on MSE or σ2 (variance)
• How the Cook’s D is integrated to GWR

https://www.machinelearningplus.com/machine-learning/cooks-distance/

https://www.statisticshowto.com/cooks-distance/

https://www.youtube.com/watch?v=PVJYdMjfrAA&t=580s