I used proc spatialreg to analyze spatial econometrics model.
However, error messages kept showing up depending on model or data. "Opimization failed. Interpret the estimates with care" or
"Optimization cannot improve the function value". In addition, although the message 'Algorithm converged' comes out, standard error, t value, and p value don`t appear.
I specified quasi-Newton method and I specified maximum number of iterations as 1000.
Algorithm converged. |
1 | 12675 | 148.614569 | 85.29 | <.0001 |
1 | 32.448024 | 0.170667 | 190.12 | <.0001 |
1 | 505.773196 | 139.973396 | 3.61 | 0.0003 |
1 | -585.435224 | 12.982785 | -45.09 | <.0001 |
1 | 214.707052 | 12.250951 | 17.53 | <.0001 |
1 | -499.383547 | 12.682656 | -39.38 | <.0001 |
1 | 214.837050 | 1.364687 | 157.43 | <.0001 |
1 | -240.843630 | 0.983443 | -244.90 | <.0001 |
0 | -199.938586 | . | . | . |
0 | 176.465831 | . | . | . |
1 | 98.594504 | 3.551837 | 27.76 | <.0001 |
1 | 1629.535839 | 152.198891 | 10.71 | <.0001 |
0 | 1.000000 | . | . | . |
1 | 11625708 | 0.255397 | 4.552E7 | <.0001 |
-1 | 94905077 | . | . | . * |
ERROR: Optimization failed. Interpret the estimates with care. |
1 | 4877.620429 | 5802.227026 | 0.84 | 0.4005 |
1 | 27.446622 | 1.054131 | 26.04 | <.0001 |
1 | -738.228100 | 161.714703 | -4.57 | <.0001 |
1 | -460.278202 | 49.892925 | -9.23 | <.0001 |
1 | 317.332993 | 23.506006 | 13.50 | <.0001 |
1 | -514.798669 | 37.320073 | -13.79 | <.0001 |
1 | 134.349259 | 93.345199 | 1.44 | 0.1501 |
1 | -192.431771 | 19.548520 | -9.84 | <.0001 |
1 | -172.167729 | 18.373370 | -9.37 | <.0001 |
1 | 221.460165 | 26.588296 | 8.33 | <.0001 |
1 | 383.475678 | 26.552452 | 14.44 | <.0001 |
1 | 1420.679084 | 299.021935 | 4.75 | <.0001 |
0 | 0.377923 | . | . | . |
0 | -3.805885 | . | . | . |
1 | 660740 | 74.384999 | 8882.71 | <.0001 |
WARNING: Optimization cannot improve the function value. |
0 | 5033.903709 | . | . | . |
1 | 31.303281 | 0.028436 | 1100.83 | <.0001 |
1 | 569.039293 | 43.805973 | 12.99 | <.0001 |
0 | -527.967647 | . | . | . |
0 | 438.671229 | . | . | . |
0 | -363.941741 | . | . | . |
0 | 84.402772 | . | . | . |
0 | -110.553241 | . | . | . |
0 | -86.368103 | . | . | . |
1 | 223.582940 | 2.153441 | 103.83 | <.0001 |
0 | 381.665965 | . | . | . |
0 | 790.001959 | . | . | . |
0 | 1.000000 | . | . | . |
0 | 3031879 | . | . | . |
-1 | 81288778 | . | . | . * |
Try to use </> to load the output, rather than cutting and pasting. The former will preserve column id's, etc. Then please also include your code and some information about the dataset (response variable, number of records, etc.). All I can say is that it might (stressing might) be that your model is over-parameterized for the amount of data. Another possibility is multi-collinearity.
SteveDenham
Available on demand!
Missed SAS Innovate Las Vegas? Watch all the action for free! View the keynotes, general sessions and 22 breakouts on demand.
ANOVA, or Analysis Of Variance, is used to compare the averages or means of two or more populations to better understand how they differ. Watch this tutorial for more.
Find more tutorials on the SAS Users YouTube channel.