Message/Author |
|
Ahn,Taeyong posted on Monday, August 21, 2017 - 10:28 am
|
|
|
I was looking for a way to save the class probabilities for an LCA model with multiply imputed datasets. This is the syntax for the analysis with the imputation. I got the error message. What should I do? please help me. : DATA: FILE IS "C:\Users\»ç¿ëÀÚ\Desktop\XY6281Z\missimplist.dat"; TYPE = IMPUTATION; VARIABLE: names are x1 x2 x3 x4 x5 x6 x7 x8 x9 x10 x11 x12 y1 y2 y3 y4 id ; usevar = x1-x12 y1-y4 ; missing = all (999); CLASSES = c(4); idvariable = id; SAVEDATA: file is "C:\Users\»ç¿ëÀÚ\Desktop\XY6281Z" ; save = class0ut; *** WARNING in SAVEDATA command The FILE option is not available for TYPE=MONTECARLO or TYPE=IMPUTATION. The FILE option will be ignored. *** WARNING in SAVEDATA command The SAVE option is not available for TYPE=MONTECARLO or TYPE=IMPUTATION. The SAVE option will be ignored. *** ERROR Invalid symbol in data file: "*" at record #: 999, field #: 17 |
|
|
Are the output segments you show here from one run or two runs? |
|
Ahn,Taeyong posted on Monday, August 21, 2017 - 3:28 pm
|
|
|
Thank you for your response. They are from one run. |
|
|
You have to do the savedata for one imputed data set at a time. Note also that the error message at the bottom indicates that you have * as missing data in the imputed data instead of 999, so say: Missing = *; |
|
Ahn,Taeyong posted on Monday, August 21, 2017 - 10:48 pm
|
|
|
Thank you for your response. I have done the savedata for one of 10 imputed data sets. Now I am working on the other nine different datasets one by one. Then Are all 10 result files created? So how do i integrate these results? |
|
Ahn,Taeyong posted on Tuesday, August 22, 2017 - 8:25 am
|
|
|
I was told that instead of integrating the results, I had to use the user-specified start value. And I have heard that I can get value by adding something to the MODEL statement from the analysis of one of the imputed datasets. Could you show me what the code looks like? |
|
|
Answer to Aug 21 post: Take the average. Answer to Aug 22 post: I don't know what you mean. |
|
Ahn,Taeyong posted on Wednesday, August 23, 2017 - 9:46 am
|
|
|
I am doing manual 3-step of GMM with Type=Imputation (I am using 10 sets of multiplely imputed data).I've got the 3 class model in step 1. Now I'm wondering (1) how to get class assignment. As far as I understand, I need to obtain 10 sets of data file that contains class assignment information from each of the 10 sets of imputed data. And then I need to do step 3 with the 10 datasets from step 2. Now I'm also wondering (2) how to integrate the 10 sets of data. I am new at Mplus. I feel it's not easy to tell you about what I want. Do you understand what I mean? Could you help me, please. |
|
|
(1) in your Savedata command, say Save = Cprob; (2) 3-step with imputed data is not a situation that is investigated as far as I know. Just one complication is that the class ordering may shift over the different imputation runs. Therefore I would say it is not a suitable topic for a beginner. I would suggest handling missing data differently than using multiple imputation, e.g. by standard ML under MAR. You may also want to ask on SEMNET. |
|
Back to top |