X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Fmath%2Flinreg.h;h=39cda89f295806bc46cf9930aac99dddaad94c17;hb=edd5c738dfef01c90d02e06a33b93fc9d38320b8;hp=834b0922469c4f711bf035e5c6607c1d99c50fdd;hpb=b5b474193e450bba97610065df0518c08074a7fb;p=pspp diff --git a/src/math/linreg.h b/src/math/linreg.h index 834b092246..39cda89f29 100644 --- a/src/math/linreg.h +++ b/src/math/linreg.h @@ -1,5 +1,5 @@ /* PSPP - a program for statistical analysis. - Copyright (C) 2005 Free Software Foundation, Inc. Written by Jason H. Stover. + Copyright (C) 2005, 2011 Free Software Foundation, Inc. Written by Jason H. Stover. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -16,20 +16,17 @@ #ifndef LINREG_H #define LINREG_H -#include + #include -#include #include - -struct variable; -struct pspp_coeff; -union value; +#include +#include enum { - PSPP_LINREG_CONDITIONAL_INVERSE, - PSPP_LINREG_QR, - PSPP_LINREG_SWEEP, + LINREG_CONDITIONAL_INVERSE, + LINREG_QR, + LINREG_SWEEP, }; @@ -91,21 +88,20 @@ typedef struct pspp_linreg_opts_struct pspp_linreg_opts; */ -struct pspp_linreg_cache_struct +struct linreg_struct { - int n_obs; /* Number of observations. */ + double n_obs; /* Number of observations. */ int n_indeps; /* Number of independent variables. */ int n_coeffs; /* The intercept is not considered a coefficient here. */ /* - The variable struct is ignored during estimation. It is here so - the calling procedure can find the variable used in the model. + Pointers to the variables. */ const struct variable *depvar; + const struct variable **indep_vars; - gsl_vector *residuals; - struct pspp_coeff **coeff; + double *coeff; double intercept; int method; /* Method to use to estimate parameters. */ /* @@ -118,7 +114,6 @@ struct pspp_linreg_cache_struct column of the design matrix. */ double depvar_mean; - double depvar_std; gsl_vector *indep_means; gsl_vector *indep_std; @@ -126,19 +121,12 @@ struct pspp_linreg_cache_struct Sums of squares. */ double ssm; /* Sums of squares for the overall model. */ - gsl_vector *ss_indeps; /* Sums of squares from each - independent variable. */ double sst; /* Sum of squares total. */ double sse; /* Sum of squares error. */ double mse; /* Mean squared error. This is just sse / dfe, but since it is the best unbiased estimate of the population variance, it has its own entry here. */ - gsl_vector *ssx; /* Centered sums of squares for independent - variables, i.e. \sum (x[i] - mean(x))^2. */ - double ssy; /* Centered sums of squares for dependent - variable. - */ /* Covariance matrix of the parameter estimates. */ @@ -150,62 +138,52 @@ struct pspp_linreg_cache_struct double dfe; double dfm; - /* - 'Hat' or Hessian matrix, i.e. (X'X)^{-1}, where X is our - design matrix. - */ - gsl_matrix *hat; - - double (*predict) (const struct variable **, const union value **, - const void *, int); - double (*residual) (const struct variable **, - const union value **, - const union value *, const void *, int); - /* - Returns pointers to the variables used in the model. - */ - int (*get_vars) (const void *, const struct variable **); - struct variable *resid; - struct variable *pred; + int dependent_column; /* Column containing the dependent variable. Defaults to last column. */ + int refcnt; + bool origin; }; -typedef struct pspp_linreg_cache_struct pspp_linreg_cache; +typedef struct linreg_struct linreg; -/* - Allocate a pspp_linreg_cache and return a pointer - to it. n is the number of cases, p is the number of - independent variables. - */ -pspp_linreg_cache *pspp_linreg_cache_alloc (size_t n, size_t p); +linreg *linreg_alloc (const struct variable *, const struct variable **, + double, size_t, bool); -bool pspp_linreg_cache_free (void *); +void linreg_unref (linreg *); +void linreg_ref (linreg *); /* Fit the linear model via least squares. All pointers passed to pspp_linreg are assumed to be allocated to the correct size and initialized to the values as indicated by opts. */ -int -pspp_linreg (const gsl_vector * Y, const gsl_matrix * X, - const pspp_linreg_opts * opts, pspp_linreg_cache * cache); - -double -pspp_linreg_predict (const struct variable **, const union value **, - const void *, int); -double -pspp_linreg_residual (const struct variable **, const union value **, - const union value *, const void *, int); +void linreg_fit (const gsl_matrix *, linreg *); + +double linreg_predict (const linreg *, const double *, size_t); +double linreg_residual (const linreg *, double, const double *, size_t); +const struct variable ** linreg_get_vars (const linreg *); + /* - All variables used in the model. + Mean of the independent variable. */ -int pspp_linreg_get_vars (const void *, const struct variable **); - -const struct pspp_coeff *pspp_linreg_get_coeff (const pspp_linreg_cache - *, - const struct variable - *, - const union value *); +double linreg_get_indep_variable_mean (const linreg *, size_t); +void linreg_set_indep_variable_mean (linreg *, size_t, double); + +double linreg_mse (const linreg *); + +double linreg_intercept (const linreg *); + +const gsl_matrix * linreg_cov (const linreg *); +double linreg_coeff (const linreg *, size_t); +const struct variable * linreg_indep_var (const linreg *, size_t); +size_t linreg_n_coeffs (const linreg *); +double linreg_n_obs (const linreg *); +double linreg_sse (const linreg *); +double linreg_ssreg (const linreg *); +double linreg_dfmodel (const linreg *); +double linreg_sst (const linreg *); +void linreg_set_depvar_mean (linreg *, double); +double linreg_get_depvar_mean (const linreg *); #endif