search

fabsig / GPBoost

Combining tree-boosting with Gaussian process and mixed effects models
Other
574 stars 46 forks source link

Covariance Structure #141

Closed eza494 closed 5 months ago

eza494 commented 5 months ago

Hi Dr. Fabio

I was wondering: 1) How I can add a covariance structure such as compound symmetry and AR(1) for the model? 2) Are there any ready covariance structures that can be used such as in lmer? 3) Is the covariance structure in the model below unstructured or just a matrix of 0s: gp_modelNestedSlope = gpb.GPModel(group_data=group_data_train_Actual, group_rand_coef_data=group_rand_coef_train, ind_effect_group_rand_coef=[1],likelihood="binary")

gp_modelNestedSlope.fit(y=y_train_df['Target'], X=X_train3.drop(['patientID','patientID_amalgam','group_str','sequence'],axis = 1), params={'std_dev': True, "trace":"True","optimizer_cov": "gradient_descent",'maxit' : 85})

Thank you in advance

fabsig commented 5 months ago

Please have a look at this blog post which explains how this can be done for longitudinal / panel data.