!g>ddlmZddlZdZdZdZGddZy))OLSNc|jd}tt|}|j|t j |s||}t |dd|f|dd|f}|j|j}|S)acalculates the nodewise_row values for the idxth variable, used to estimate approx_inv_cov. Parameters ---------- exog : array_like The weighted design matrix for the current partition. idx : scalar Index of the current variable. alpha : scalar or array_like The penalty weight. If a scalar, the same penalty weight applies to all variables in the model. If a vector, it must have the same length as `params`, and contains a penalty weight for each coefficient. Returns ------- An array-like object of length p-1 Notes ----- nodewise_row_i = arg min 1/(2n) ||exog_i - exog_-i gamma||_2^2 + alpha ||gamma||_1 N)alpha) shapelistrangepopnpisscalarrfit_regularizedparams)exogidxrpindtmod nodewise_rows e/var/www/dash_apps/app1/venv/lib/python3.12/site-packages/statsmodels/stats/regularized_covariance.py_calc_nodewise_rowrs{6 1 A uQx.CGGCL ;;u c  tAsF|T!S&\ *D''e'4;;L c|j\}}tt|}|j|t j |s||}tj j|dd|f|dd|fj|z dz}t j||z |tj j|dzz}|S)a1calculates the nodewise_weightvalue for the idxth variable, used to estimate approx_inv_cov. Parameters ---------- exog : array_like The weighted design matrix for the current partition. nodewise_row : array_like The nodewise_row values for the current variable. idx : scalar Index of the current variable alpha : scalar or array_like The penalty weight. If a scalar, the same penalty weight applies to all variables in the model. If a vector, it must have the same length as `params`, and contains a penalty weight for each coefficient. Returns ------- A scalar Notes ----- nodewise_weight_i = sqrt(1/n ||exog,i - exog_-i nodewise_row||_2^2 + alpha ||nodewise_row||_1) Nr) rrr r r r linalgnormdotsqrt)rrrrnrrds r_calc_nodewise_weightr /s: ::DAq uQx.CGGCL ;;u c  tAsF|d1c6l&6&6|&DDEqHA A |Q ???@A Hrct|}tj| }t|D]1}t t|}|j ||||||f<3|d|dddfdzz z}|S)acalculates the approximate inverse covariance matrix Parameters ---------- nodewise_row_l : list A list of array-like object where each object corresponds to the nodewise_row values for the corresponding variable, should be length p. nodewise_weight_l : list A list of scalars where each scalar corresponds to the nodewise_weight value for the corresponding variable, should be length p. Returns ------ An array-like object, p x p matrix Notes ----- nwr = nodewise_row nww = nodewise_weight approx_inv_cov_j = - 1 / nww_j [nwr_j,1,...,1,...nwr_j,p] Nr)lenr eyer rr )nodewise_row_lnodewise_weight_lrapprox_inv_covrrs r_calc_approx_inv_covr(Ys4 AffQiZNQx758n  #1##6sCx 7b,QW5q888N rc$eZdZdZdZddZdZy)RegularizedInvCovarianceaU Class for estimating regularized inverse covariance with nodewise regression Parameters ---------- exog : array_like A weighted design matrix for covariance Attributes ---------- exog : array_like A weighted design matrix for covariance alpha : scalar Regularizing constant c||_yN)r)selfrs r__init__z!RegularizedInvCovariance.__init__s  rcz|jj\}}g}g}t|D]S}t|j||}|j |t |j|||}|j |Ut j|}t j|}t||} | |_ y)zestimates the regularized inverse covariance using nodewise regression Parameters ---------- alpha : scalar Regularizing constant N) rrr rappendr r arrayr(_approx_inv_cov) r-rrrr%r&rrnodewise_weightr's rfitzRegularizedInvCovariance.fitsyy18 6C-diieDL  ! !, /3DII|47@O  $ $_ 5  6.1HH%67-n.?A .rc|jSr,)r2)r-s rr'z'RegularizedInvCovariance.approx_inv_covs###rN)r)__name__ __module__ __qualname____doc__r.r4r'rrr*r*s".>$rr*)#statsmodels.regression.linear_modelrnumpyr rr r(r*r:rrr=s'3'T' T#L6$6$r