mssm Documentation

This is the documentation of the public API of mssm. The entire source code is available for inspection on GitHub.

mssm is a Python toolbox to estimate Generalized Additive Mixed Models (GAMMs), Generalized Additive Mixed Models of Location Scale and Shape (GAMMLSS), and more general smooth (mixed) models (GSMMs) in the sense defined by Wood, Pya, & Säfken (2016). mssm is an excellent choice for the modeling of multi-level time-series data, often estimating additive models with separate smooths for thousands of levels in a couple of minutes.

You can either use the side-bar on the left to navigate through the document tree or click on the links below. To get started, we suggest that you first work through the Getting Started section, which includes installation instructions. Then you should complete the Tutorial, in which you will familiarize yourself with the syntax of mssm and the mssmViz package. mssmViz offers functions to visualize & validate models estimated via the mssm toolbox. If you are interested in adding your own custom smooth basis and penalty matrices, head to the Programmer’s Guide section. If you just want to take a look at the api, check out the api section.

Getting Started

• Getting Started
  • Installation

Tutorial

• mssm Tutorial
  • Data structure
  • The Formula class
  • Practicing with Formulas & Estimating Additive Models
  • Beyond Gaussian Models: Estimating GAMMs and GAMMLSS in mssm
  • Advanced topics
  • References

Programmer’s Guide

• Programmer’s Guide
  • Adding a New Family
  • Adding a New GAMLSSFamily
  • Adding a New GSMMFamily
  • Adding a New Marginal Smooth Basis
  • Adding a New Penalty Matrix

api

• api
  • mssm.models module
    • GAMM
    • GAMMLSS
    • GSMM
  • mssm.src.python.compact_rep module
    • computeH()
    • computeHSR1()
    • computeV()
    • computeVSR1()
  • mssm.src.python.compare module
    • compare_CDL()
  • mssm.src.python.custom_types module
    • ConstType
    • Constraint
    • Fit_info
    • LambdaTerm
    • PenType
    • Reparameterization
    • TermType
    • VarType
  • mssm.src.python.exp_fam module
    • Binomial
    • Family
    • GAMLSSFamily
    • GAMMALS
    • GAUMLSS
    • GSMMFamily
    • Gamma
    • Gaussian
    • Identity
    • InvGauss
    • LOG
    • LOGb
    • Link
    • Logit
    • MULNOMLSS
    • Poisson
    • PropHaz
    • est_scale()
  • mssm.src.python.file_loading module
    • clear_cache()
    • read_cor_cov_single()
    • read_cov()
    • read_cov_no_cor()
    • read_dtype()
    • read_no_cor_cov_single()
    • read_unique()
    • read_unique_single()
    • setup_cache()
  • mssm.src.python.formula module
    • Formula
    • build_model_matrix()
    • build_penalties()
    • build_sparse_matrix_from_formula()
    • lhs
  • mssm.src.python.gamm_solvers module
    • PIRLS_newton_weights()
    • PIRLS_pdat_weights()
    • apply_eigen_perm()
    • back_track_alpha()
    • calculate_edf()
    • calculate_term_edf()
    • check_drop_valid_gammlss()
    • check_drop_valid_gensmooth()
    • compute_S_emb_pinv_det()
    • compute_eigen_perm()
    • compute_lgdetD_bsb()
    • computetrVS3()
    • correct_coef_step()
    • correct_coef_step_gammlss()
    • correct_coef_step_gen_smooth()
    • correct_lambda_step()
    • correct_lambda_step_gamlss()
    • correct_lambda_step_gen_smooth()
    • deriv_transform_eta_beta()
    • deriv_transform_mu_eta()
    • drop_terms_S()
    • drop_terms_X()
    • extend_lambda_step()
    • form_cross_prod_mp()
    • form_eta_mp()
    • gd_coef_smooth()
    • grad_lambda()
    • handle_drop_gammlss()
    • handle_drop_gsmm()
    • identify_drop()
    • init_step_gam()
    • initialize_extension()
    • keep_XTX()
    • keep_eta()
    • newton_coef_smooth()
    • read_XTX()
    • read_eta()
    • read_mmat()
    • restart_coef()
    • restart_coef_gammlss()
    • solve_gamm_sparse()
    • solve_gamm_sparse2()
    • solve_gammlss_sparse()
    • solve_generalSmooth_sparse()
    • step_fellner_schall_sparse()
    • test_SR1()
    • undo_extension_lambda_step()
    • update_PIRLS()
    • update_coef()
    • update_coef_and_scale()
    • update_coef_gammlss()
    • update_coef_gen_smooth()
    • update_scale_edf()
  • mssm.src.python.matrix_solvers module
    • compute_B()
    • compute_Linv()
    • compute_block_B_shared()
    • compute_block_B_shared_cluster()
    • compute_block_linv_shared()
    • cpp_backsolve_tr()
    • cpp_chol()
    • cpp_cholP()
    • cpp_dqrr()
    • cpp_qr()
    • cpp_qrr()
    • cpp_solve_L()
    • cpp_solve_LXX()
    • cpp_solve_am()
    • cpp_solve_coef()
    • cpp_solve_coefXX()
    • cpp_solve_coef_pqr()
    • cpp_solve_qr()
    • cpp_solve_tr()
    • cpp_symqr()
    • est_condition()
    • map_csc_to_eigen()
    • map_csr_to_eigen()
    • translate_sparse()
  • mssm.src.python.penalties module
    • DifferencePenalty
    • IdentityPenalty
    • Penalty
    • TP_pen()
    • adjust_pen_drop()
    • embed_in_S_sparse()
    • embed_in_Sj_sparse()
    • embed_shared_penalties()
  • mssm.src.python.repara module
    • reparam()
    • reparam_model()
  • mssm.src.python.smooths module
    • B_spline_basis()
    • TP_basis_calc()
    • bbase()
    • convolve_event()
    • tpower()
  • mssm.src.python.terms module
    • GammTerm
    • build_ir_smooth_series()
    • build_linear_term()
    • f
    • fs
    • get_linear_coef_info()
    • i
    • irf
    • l
    • li()
    • ri
    • rs
  • mssm.src.python.utils module
    • DummyRhoPrior
    • GAMLSSGSMMFamily
    • REML()
    • RhoPrior
    • adjust_CI()
    • approx_smooth_p_values()
    • computeAr1Chol()
    • compute_REML_candidate_GSMM()
    • compute_Vb_corr_WPS()
    • compute_Vp_WPS()
    • compute_bias_corrected_edf()
    • compute_reml_candidate_GAMM()
    • correct_VB()
    • estimateVp()
    • print_parametric_terms()
    • print_smooth_terms()
    • sample_MVN()
    • updateVp()