Trying to investigate whether the Autoregressive behaviour is as a result of Lags (SER) or errors (SEM).
Diagnistic check on the weights matrix
The output is as below
spatdiag, weights(W)
Diagnostic tests for spatial dependence in OLS regression
Fitted model
------------------------------------------------------------
logHPRICE = SPOOL + BQUARTER + WFENCE + PSIZE + NROOMS
------------------------------------------------------------
Weights matrix
------------------------------------------------------------
Name: W
Type: Distance-based (inverse distance)
Distance band: 0.0 < d
The spatial error model is significant and the lag is not. This means you should run a spatial error model
Based on the diagnoses obtained, it must be evaluated whether the spatial influence on the model is due to: i) error between the observations and the predictions; ii) the spatial lag or average of values of the variable obtained from the neighbors.
Therefore, when evaluating the statistics of the Lagrange multiplier:
i) If significant for spatial error, a spatial error model (SEM) should be run.
ii) If significant for spatial lag, a spatial lag model (SLM) should be run.
- Badges
- Science topic
- Similar topics
- Methodology
- Research Methodology