CodePudding user response:
The first module, is the result of the lag fixed effects model, R - squared log likelihood is model fitting of some indicators, such as numerical point of view is very good, important come,,, z - aim-listed probability shows that under the 1% level significantly below, shows that the results significantly W * dep. Var coefficient indicates that overflow effect is obvious, or this area to overflow in other areas of the obvious (model decision ha ha)
The second module, is the result of the lag model random effects, most of the above explanation, specific coefficient of the above is not ok, of course still have analysis of the value of the individual indicators,
The third module spatial error model of the results of the fixed effect, can see the pn is not significant, but the spillover effect is significantly, it is in other parts of the region's flooded
The fourth module error drifts random effect regardless of the fixed effect is good,
Generally should choose the rest of the SLM SEM of the fixed effect model is compared this two models, the building should be the right solution is to use the building Lord hope