{smcl} {* *! version 1.0 2021-08-03}{...} {vieweralsosee "mlad" "help mlad"}{...} {vieweralsosee "ml" "help ml"}{...} {vieweralsosee "python" "help python"}{...} {hline} {title:Title} {p2colset 5 13 17 2}{...} {p2col :{hi:mlad }{hline 2}}Python utility finctions for {help mlad}.{p_end} {p2colreset}{...} {pstd} {cmd: mlad} has a number of utility functions written in Python. They been accessed after importing in Python, e.g. by using {cmd:import mladutil as mu} The functions are described below. {pstd} {cmd:invlogit}({it:z}) Transforms from the log odds scale to the probability scale. {pstd} {cmd:linpred}({it:beta},{it:X},{it:eq}) - extract linear predictor. {phang2} {it:beta} - full beta matrix {phang2} {it:X} - full covariates list {phang2} {it:eq} - equation number {pstd} {cmd:rcsgen_beta}({it:x},{it:knots},{it:beta},{it:rmatrix}) Restricted cubic splines basis functions multiplied by beta matrix. {phang2} {it:x} - orginal variable {phang2} {it:knots} - knots {phang2} {it:beta} - beta matrix {phang2} {it:rmatrix} - rmatrix for orthogonalization. Use {cmd:jnp.ones((1,1))} if not orthogonalized. {pstd} {cmd:rcsgen}({it:x},{it:knots},{it:rmatrix}) Restricted cubic splines basis functions. {phang2} {it:x} - orginal variable {phang2} {it:knots} - knots {phang2} {it:rmatrix} - rmatrix for orthogonalization. Use {cmd:jnp.ones((1,1))} if not orthogonalized. {phang2} Use {cmd:vrcsgen} for a vectorized version of rcsgen. {pstd} {cmd:drcsgen}({it:x},{it:knots},{it:rmatrix}) 1st derivatove of restricted cubic splines basis functions. {phang2} {it:x} - orginal variable {phang2} {it:knots} - knots {phang2} {it:rmatrix} - rmatrix for orthogonalization. Use {cmd:jnp.ones((1,1))} if not orthogonalized. {phang2} Use {cmd:vdrcsgen} for a vectorized version of drcsgen. {pstd} {cmd:sumoverid}({it:id},{it:x},{it:Nid}) - sum X over id {phang2} {it:id} - id indicator {phang2} {it:x} - variables to be summed {phang2} {it:Nid} - Number of unique individuals {pstd} {cmd:vecquad_gl}({it:fn},{it:a},{it:b},{it:Nnodes},{it:arglist}) Gauss-Legendre quadrature {phang2} {it:fn} - function to be integrated {phang2} {it:a} - lower limit of integral {phang2} {it:b} - upper limit of integral {phang2} {it:Nnodes} - Number of nodes {phang2} {it:arglist} - List of additional arguments for function {it:fn} {pstd} {cmd:vecquad_gh}({it:fn},{it:Nnodes},{it:arglist}) Gauss-Hermite quadrature {phang2} {it:fn} - function to be integrated {phang2} {it:Nnodes} - Number of nodes {phang2} {it:arglist} - List of additional arguments for function {it:fn} {pstd} {cmd:weibsurv}({it:t},{it:lam},{it:gam}) - Weibull survival function. {phang2} {it:t} - time {phang2} {it:lam} - lambda {phang2} {it:gam} - gamma {pstd} {cmd:weibdens}({it:t},{it:lam},{it:gam}) - Weibull density function. {phang2} {it:t} - time {phang2} {it:lam} - lambda {phang2} {it:gam} - gamma {pstd} {cmd:mlvecsum}({it:Z},{it:X},{it:eq}) - equivalent of mlvecsum in Stata {phang2} {it:Z} - expression {phang2} {it:X} - full list of covariates {phang2} {it:eq} - equation number {pstd} {cmd:mlmatsum}({it:Z},{it:X},{it:eq}) - equivalent of mlvecsum in Stata {phang2} {it:Z} - expression {phang2} {it:X} - full list of covariates {phang2} {it:eq} - equation number {pstd} {cmd:tanhsinh_quad}({it:Nnodes},{it:N}) - tanh-sinh quadrature {phang2} {it:Nnodes} - Number of nodes {phang2} {it:N} - defines range (-N,N) to evaluate integral {phang2} Returns 2 Nx1 vectors: nodes and weights. {pstd} {cmd:bs}({it:x},{it:degree},{it:knots},{it:bknots},{it:calcintercept},{it:calcderiv}) - B-splines {phang2} {it:x} - X variables {phang2} {it:degree} - degree of B-spline {phang2} {it:knots} - internal knots {phang2} {it:bknots} - boundary knots {phang2} {it:calcintecept} - True/False (default) {phang2} {it:calcderiv} - True/False (default) {phang2} Returns array of splines or 2 arrays with {it:calcderiv} option {phang2} USe {cmd:vbs}() for vmap version (and jitted) {pstd} {cmd:ns}({it:x},{it:knots},{it:bknots},{it:calcintercept},{it:calcderiv}) - Natural splines {phang2} {it:x} - X variables {phang2} {it:knots} - internal knots {phang2} {it:bknots} - boundary knots {phang2} {it:calcintecept} - True/False (default) {phang2} {it:calcderiv} - True/False (default) {phang2} Returns array of splines or 2 arrays with {it:calcderiv} option {phang2} USe {cmd:vns}() for vmap version (and jitted)