M/Ph+dZddlmZddlmZddlmZmZddlZ ddl Z ddl m Z mZddlmcmZdd Zdd Zdd Zd ZdZddZdZddZdZdZdZddZdS)z/ Miscellaneous utility code for VAR estimation ) frequencies)asbytes) array_likeint_likeN)statslinalgcskipct}tjfdt|D}|dkrt j|d||}|S)u Make predictor matrix for VAR(p) process Z := (Z_0, ..., Z_T).T (T x Kp) Z_t = [1 y_t y_{t-1} ... y_{t - p + 1}] (Kp x 1) Ref: Lütkepohl p.70 (transposed) has_constant can be 'raise', 'add', or 'skip'. See add_constant. c`g|]*}|z |ddd+S)N)ravel).0tlagsys ^/var/www/html/test/jupyter/venv/lib/python3.11/site-packages/statsmodels/tsa/vector_ar/util.py z!get_var_endog..s<III!!AdFQJ-"%++--IIInT)prependtrend has_constant)lennparrayrangetsa add_trend)rrrrnobsZs`` r get_var_endogr"ss q66D IIIIIuT47H7HIIIJJA || M!T'3 5 5 5 Hrcn|dkrd}n+|dvrd}n$|dkrd}n|dkrd}ntd ||S) Nr )rncrctcttzUnkown trend: ) ValueError)r trendorders rget_trendorderr,(sa || +   $ % 1%11222 rr$cg}t|tr|g}td|dzD]V}|D]Q}t|tst|}|dt|zdz|zRW|dkr|dd|dkr|dd|dkr|dd|t|t jrt j|}n$t|d stj |}|j dkr |d d d f}t|j dD]b}t|t jrt|j |}nd t|z}|||z|c|S) z Produce list of lag-variable names. Constant / trends go at the beginning Examples -------- >>> make_lag_names(['foo', 'bar'], 2, 1) ['const', 'L1.foo', 'L1.bar', 'L2.foo', 'L2.bar'] r$L.rconstrr'ztrend**2Nndimexog) isinstancestrrappendinsertpdSeries DataFramehasattrrasarrayr1shapecolumns)names lag_orderr+r2 lag_namesiname exog_names rmake_lag_namesrD7sI%1i!m $ $22 2 2DdC(( !4yy   SQZ^D0 1 1 1 1 2 QG$$$A~~G$$$A~~J'''  dBI & & $<%%DDv&& $:d##D 9>>4=Dtz!}%% 8 8A$ -- , Q00 "SVVO   Z!^Y 7 7 7 7 rc|j\}}}||krtd||z}tj||f}tj|d|d|<|dkr/d|tj||tj||z f<|S)z Return compansion matrix for the VAR(1) representation for a VAR(p) process (companion form) A = [A_1 A_2 ... A_p-1 A_p I_K 0 0 0 0 I_K ... 0 0 0 ... I_K 0] z'coefs must be 3-d with shape (p, k, k).r$)axisN)r<r*rzeros concatenatearange)coefspk1k2kpresults r comp_matrixrPbs IAr2 RxxBCCC aB Xr2h  F.Q///F3B3K 1uu67ryR  ")BrE"2"223 Mrc"ddlm}ddlm}ddl}|t d}t |d5}||}dddn #1swxYwYd}t d|vr(|dz }t d|v( |dz }|}||} | r| \} } } nJtj |d |dz  d } t| }t| } t| } t dtjt dtjt dtji}|| }|| dz z}||| dd|z}|| }tj|||}| |fS)u` Parse data files from Lütkepohl (2005) book Source for data files: www.jmulti.de r)deque)datetimeNz<(.*) (\w)([\d]+)>.*rbz*/r$Tz\s+) delimiterheaderF)indexQMA)startfreqperiods) collectionsrRrSrecompileropenpopleftmatchgroupsr7read_csv to_recordsrintr BQuarterEnd BMonthEndBYearEnd rollforward date_range)pathrRrSr_regexflinesto_skiplinemyearr\ start_pointdataroffsetsoffsetinc start_daterls rparse_lutkepohl_datar{sQ"!!!!!!!!!!! III JJw677 8 8E dD  QaG $--u}} . .1  $--u}} . .1 }} KK    &'hhjj #D$   Kwqy A A A ZeZ $ $  D Ak""K t99D  k-// k+-- k*,,GT]F K!O $C##HHT1a$8$899C?J T]FZfaHHHJ  s AA"A皙?cLtjd|dz z S)Nr$r')rnormppf)alphas rnorm_signif_levelrs :>>!eai- ( ((rctj|d}|tjtj||z S)Nr)rdiagsqrtouter)acfrs r acf_to_acorrrs4 73q6??D $--.. ..rdc ttj|}|j}|j\} } } t |dd}t |tr|dkrtd||| f} d}n||| f} |tj | }|tj t||||z ||| } tj ||| f} |mtj |dkr-t|| jdkstd | |z } | dd| dfxx| dd| dfz cc<n| dd| df| dd| df<t|d dd }|3|j| | fks|j| fkstd || ddd| f<t| |D]X}| dd|f}t| D]:}|tj||| dd||z dz fjjz };Y| | S)a  Simulate VAR(p) process, given coefficients and assuming Gaussian noise Parameters ---------- coefs : ndarray Coefficients for the VAR lags of endog. intercept : None or ndarray 1-D (neqs,) or (steps, neqs) This can be either the intercept for each equation or an offset. If None, then the VAR process has a zero intercept. If intercept is 1-D, then the same (endog specific) intercept is added to all observations. If intercept is 2-D, then it is treated as an offset and is added as an observation specific intercept to the autoregression. In this case, the intercept/offset should have same number of rows as steps, and the same number of columns as endogenous variables (neqs). sig_u : ndarray Covariance matrix of the residuals or innovations. If sig_u is None, then an identity matrix is used. steps : {None, int} number of observations to simulate, this includes the initial observations to start the autoregressive process. If offset is not None, then exog of the model are used if they were provided in the model initial_values : array_like, optional Initial values for use in the simulation. Shape should be (nlags, neqs) or (neqs,). Values should be ordered from less to most recent. Note that this values will be returned by the simulation as the first values of `endog_simulated` and they will count for the total number of steps. seed : {None, int} If seed is not None, then it will be used with for the random variables generated by numpy.random. nsimulations : {None, int} Number of simulations to perform. If `nsimulations` is None it will perform one simulation and return value will have shape (steps, neqs). Returns ------- endog_simulated : nd_array Endog of the simulated VAR process. Shape will be (nsimulations, steps, neqs) or (steps, neqs) if `nsimulations` is None. )seed nsimulationsT)optionalrz3nsimulations must be a positive integer if providedNr$z*2-D intercept needs to have length `steps`initial_valuesr')rmaxdimzoinitial_values should have shape (p, k) or (k,) where p is the number of lags and k is the number of equations.)rrandom RandomStatemultivariate_normalr<rr3rgr*eyerGrreshaper1rrdotT)rJ interceptsig_ustepsrrrrsrmvnormrKk result_shapeugenrOrygenjs rvarsimrsX   D  ) )B$GkGAq!<$GGGL,$$P):):NOOOqz  $eQ/  }q  728CJJ''l0B C C K KLZ_ab c cD X|UA. / /F 79   ! !y>>TZ]22 !MNNN)qqqt QQQqrrT " AAAabbDzqqqt 0@4XYZZZN!$A...2F1$2N2NOPP P%qqq!t 1e__::aaac{q : :A BF58VAAAac!eG_%6779 9DD : >>, ' ''rc ||}n(#t$rt|ts|}YnwxYw|S)N)rW Exceptionr3rg)lstrBrOs r get_indexrsX4 $$$   Ms "==chtj|dd\}}tj|}|||fS)zt Returns ------- W: array of eigenvectors eigva: list of eigenvalues k: largest eigenvector TF)leftright)reigrargmax) sym_arrayeigvaWrs r eigval_decompr s9z)$e<<= 0 Number of seasons (e.g. 12 for monthly data and 4 for quarterly data). len_endog : int >= 0 Total number of observations. first_period : int, default: 0 Season of the first observation. As an example, suppose we have monthly data and the first observation is in March (third month of the year). In this case we pass 2 as first_period. (0 for the first season, 1 for the second, ..., n_seasons-1 for the last season). An integer greater than n_seasons-1 are treated in the same way as the integer modulo n_seasons. centered : bool, default: False If True, center (demean) the dummy variables. That is useful in order to get seasonal dummies that are orthogonal to the vector of constant dummy variables (a vector of ones). Returns ------- seasonal_dummies : ndarray (len_endog x n_seasons-1) rr$N)remptyrGr) n_seasons len_endog first_periodcentered season_exogrAs rseasonal_dummiesrAs2A~~xA'''1}}h 9q=9:: y1}%% H HAFGK<94?i?B C C  ) 1y= (K}r)r r )r )r$N)r|)rNNN)rF)__doc__statsmodels.compat.pandasrstatsmodels.compat.pythonrstatsmodels.tools.validationrrnumpyrpandasr7scipyrrstatsmodels.tsa.tsatoolsrtsatoolsr"r,rDrPr{rrrrrrrrrrs{211111------========&&&&&&&&&     .    ((((V:444n))))/// R(R(R(R(j   &""""""r