in case of  multiple variables(+100 variables),the following statement is OK?   proc fmm data=XXX; class XXX;←think separately about model? model A=,B=C=,・・・・・・ = / k=X; run; ... View more

①please tell me the case of following ERROR   ERROR: A problem occured during model estimation.  Diagnostic information is as follows:           Matrix not positive definite         OCCURRED IN: cholesky_in_place in MOD logit_model         Logistic regression failed at iteration 1         -Hessian not positive definite.         OCCURRED IN: run_logit_model_threads in MOD logit_model         Could not fit logit model.  The problem occurred in:         Group:  1.         OCCURRED IN: mstep_beta in MOD estimation         M-step for BETA failed at iteration 1.         Could not fit logit model. We suggest         using the BETA PRIOR statement.         OCCURRED IN: run_em_modelfit in MOD estimation   ②for example, when I want to incorporate all of the following variables,proc lca program is right?   House number Customer number employer number sex height(cm) weight(kg)      product 　　year    period 112　　　　　　　　　 1 　　　　　　LO2113 　　1 　　161 　　58 　　　      A0001         2018    3 112　　　　　　　　    2 　　　　　　 LO2223 　　2 　　140 　　60　　　      A0002         2017    4 113 　　　　　　　　　3 　　　　　　LO2113 　　1 　　131 　　73 　　　      A0001         2013    2 113 　　　　　　　　　4 　　　　　　LO2113          2          170         121            A0003         2014   6 221 　　　　　　　　　5 　　　　　　LO2004          2          121         48              A0002         2015   7 222 　　　　　　　　　6 　　　　　　LO2113          2          135          50              A0002         2017   8 223 　　　　　　　　　7 　　　　　　LO2113          2          140         .                   A0002         2016   1   PROC LCA DATA=out OUTPARAM=param1 OUTPOST=post1 OUTEST=est1 OUTSTDERR=stderr1; NCLASS 5; id employer number;  item sex;   CATEGORIES 2; COVARIATES height weight year period; clusters house number ;   NSTARTS 50;   RHO PRIOR=1; SEED 1000; RUN;    id：Is multile variables is available?       2.  when I want to standarize, would I better to use not covariates statement,but weight statement?       3.  what statement do I use to incorporate product variables? ... View more

I am trying to find  clustering method in case of  large data and features.no teacher learning. until now, I found k-means k-means++, spectral clustering ,minibatch k-means,DBSCAN,x-means,Laten Class Analysis、mixed normal distribution. ①if there are other methods,please let me know. ②I used SAS Enterprise guide, can I use  sas procedure of  k-means++, spectral clustering ,minibatch k-means,DBSCAN ? ... View more