X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Flanguage%2Fstats%2Fregression.q;h=3706760a38ed9689b64b4d0afaedd7a05908fb0a;hb=92f198d13c9214c0d75b936f0ea0dc2684ea914b;hp=8103d914638e7abe0a3a2df48ecf913215faa1ab;hpb=7102dc8607b7e6e25bdb9806f508dce71fe76ce4;p=pspp-builds.git diff --git a/src/language/stats/regression.q b/src/language/stats/regression.q index 8103d914..3706760a 100644 --- a/src/language/stats/regression.q +++ b/src/language/stats/regression.q @@ -18,32 +18,37 @@ 02110-1301, USA. */ #include -#include + #include -#include #include +#include #include -#include +#include + +#include "regression-export.h" #include #include -#include #include -#include -#include -#include +#include #include -#include +#include +#include +#include +#include +#include +#include +#include #include -#include "gettext.h" #include +#include +#include +#include +#include +#include #include -#include -#include -#include "regression-export.h" #include -#include -#include -#include + +#include "gettext.h" #define REG_LARGE_DATA 1000 @@ -52,27 +57,27 @@ /* (specification) "REGRESSION" (regression_): *variables=custom; - statistics[st_]=r, - coeff, - anova, - outs, - zpp, - label, - sha, - ci, - bcov, - ses, - xtx, - collin, - tol, - selection, - f, - defaults, - all; + +statistics[st_]=r, + coeff, + anova, + outs, + zpp, + label, + sha, + ci, + bcov, + ses, + xtx, + collin, + tol, + selection, + f, + defaults, + all; export=custom; ^dependent=varlist; - save=residuals; - method=enter. + +save[sv_]=resid,pred; + +method=enter. */ /* (declarations) */ /* (functions) */ @@ -81,6 +86,16 @@ static struct cmd_regression cmd; /* Linear regression models. */ pspp_linreg_cache **models = NULL; +/* + Transformations for saving predicted values + and residuals, etc. + */ +struct reg_trns +{ + int n_trns; /* Number of transformations. */ + int trns_id; /* Which trns is this one? */ + pspp_linreg_cache *c; /* Linear model for this trns. */ +}; /* Variables used (both explanatory and response). */ @@ -102,7 +117,8 @@ struct file_handle *model_file; */ int pspp_reg_rc = CMD_SUCCESS; -static bool run_regression (const struct casefile *, void *); +static bool run_regression (const struct ccase *, + const struct casefile *, void *); /* STATISTICS subcommand output functions. @@ -197,7 +213,7 @@ reg_stats_coeff (pspp_linreg_cache * c) tab_text (t, 5, 0, TAB_CENTER | TAT_TITLE, _("t")); tab_text (t, 6, 0, TAB_CENTER | TAT_TITLE, _("Significance")); tab_text (t, 1, 1, TAB_LEFT | TAT_TITLE, _("(Constant)")); - coeff = c->coeff[0].estimate; + coeff = c->coeff[0]->estimate; tab_float (t, 2, 1, 0, coeff, 10, 2); std_err = sqrt (gsl_matrix_get (c->cov, 0, 0)); tab_float (t, 3, 1, 0, std_err, 10, 2); @@ -209,7 +225,7 @@ reg_stats_coeff (pspp_linreg_cache * c) tab_float (t, 6, 1, 0, pval, 10, 2); for (j = 1; j <= c->n_indeps; j++) { - v = pspp_linreg_coeff_get_var (c->coeff + j, 0); + v = pspp_coeff_get_var (c->coeff[j], 0); label = var_to_string (v); /* Do not overwrite the variable's name. */ strncpy (tmp, label, MAX_STRING); @@ -221,7 +237,7 @@ reg_stats_coeff (pspp_linreg_cache * c) for that value. */ - val = pspp_linreg_coeff_get_value (c->coeff + j, v); + val = pspp_coeff_get_value (c->coeff[j], v); val_s = value_to_string (val, v); strncat (tmp, val_s, MAX_STRING); } @@ -230,7 +246,7 @@ reg_stats_coeff (pspp_linreg_cache * c) /* Regression coefficients. */ - coeff = c->coeff[j].estimate; + coeff = c->coeff[j]->estimate; tab_float (t, 2, j + 1, 0, coeff, 10, 2); /* Standard error of the coefficients. @@ -376,7 +392,7 @@ reg_stats_bcov (pspp_linreg_cache * c) tab_text (t, 1, 1, TAB_CENTER | TAT_TITLE, _("Covariances")); for (i = 1; i < c->n_coeffs; i++) { - const struct variable *v = pspp_linreg_coeff_get_var (c->coeff + i, 0); + const struct variable *v = pspp_coeff_get_var (c->coeff[i], 0); label = var_to_string (v); tab_text (t, 2, i, TAB_CENTER, label); tab_text (t, i + 2, 0, TAB_CENTER, label); @@ -504,72 +520,185 @@ subcommand_statistics (int *keywords, pspp_linreg_cache * c) statistics_keyword_output (reg_stats_tol, keywords[tol], c); statistics_keyword_output (reg_stats_selection, keywords[selection], c); } + +/* + Free the transformation. Free its linear model if this + transformation is the last one. + */ +static bool +regression_trns_free (void *t_) +{ + bool result = true; + struct reg_trns *t = t_; + + if (t->trns_id == t->n_trns) + { + result = pspp_linreg_cache_free (t->c); + } + free (t); + + return result; +} + +/* + Gets the predicted values. + */ static int -regression_trns_proc (void *m, struct ccase *c, int case_idx UNUSED) +regression_trns_pred_proc (void *t_, struct ccase *c, int case_idx UNUSED) { size_t i; - size_t n_vars; - size_t n_vals = 0; - pspp_linreg_cache *model = m; - union value *output; + size_t n_vals; + struct reg_trns *trns = t_; + pspp_linreg_cache *model; + union value *output = NULL; + const union value **vals = NULL; + struct variable **vars = NULL; + + assert (trns != NULL); + model = trns->c; + assert (model != NULL); + assert (model->depvar != NULL); + assert (model->pred != NULL); + + vars = xnmalloc (model->n_coeffs, sizeof (*vars)); + n_vals = (*model->get_vars) (model, vars); + + vals = xnmalloc (n_vals, sizeof (*vals)); + output = case_data_rw (c, model->pred->fv); + assert (output != NULL); + + for (i = 0; i < n_vals; i++) + { + vals[i] = case_data (c, vars[i]->fv); + } + output->f = (*model->predict) ((const struct variable **) vars, + vals, model, n_vals); + free (vals); + free (vars); + return TRNS_CONTINUE; +} + +/* + Gets the residuals. + */ +static int +regression_trns_resid_proc (void *t_, struct ccase *c, int case_idx UNUSED) +{ + size_t i; + size_t n_vals; + struct reg_trns *trns = t_; + pspp_linreg_cache *model; + union value *output = NULL; const union value **vals = NULL; const union value *obs = NULL; struct variable **vars = NULL; - + + assert (trns != NULL); + model = trns->c; assert (model != NULL); assert (model->depvar != NULL); assert (model->resid != NULL); - - dict_get_vars (default_dict, &vars, &n_vars, 1u << DC_SYSTEM); - vals = xnmalloc (n_vars, sizeof (*vals)); - assert (vals != NULL); + + vars = xnmalloc (model->n_coeffs, sizeof (*vars)); + n_vals = (*model->get_vars) (model, vars); + + vals = xnmalloc (n_vals, sizeof (*vals)); output = case_data_rw (c, model->resid->fv); assert (output != NULL); - for (i = 0; i < n_vars; i++) + for (i = 0; i < n_vals; i++) { - /* Do not use the residual variable. */ - if (vars[i]->index != model->resid->index) - { - /* Do not use the dependent variable as a predictor. */ - if (vars[i]->index == model->depvar->index) - { - obs = case_data (c, i); - assert (obs != NULL); - } - else - { - vals[i] = case_data (c, i); - n_vals++; - } - } + vals[i] = case_data (c, vars[i]->fv); } - output->f = (*model->residual) ((const struct variable **) vars, + obs = case_data (c, model->depvar->fv); + output->f = (*model->residual) ((const struct variable **) vars, vals, obs, model, n_vals); free (vals); + free (vars); return TRNS_CONTINUE; } + +/* + Returns 0 if NAME is a duplicate of any existing variable name. +*/ +static int +try_name (char *name) +{ + if (dict_lookup_var (default_dict, name) != NULL) + return 0; + + return 1; +} static void -subcommand_save (int save, pspp_linreg_cache **models) +reg_get_name (char name[LONG_NAME_LEN], const char prefix[LONG_NAME_LEN]) +{ + int i = 1; + + snprintf (name, LONG_NAME_LEN, "%s%d", prefix, i); + while (!try_name (name)) + { + i++; + snprintf (name, LONG_NAME_LEN, "%s%d", prefix, i); + } +} +static void +reg_save_var (const char *prefix, trns_proc_func * f, + pspp_linreg_cache * c, struct variable **v, int n_trns) +{ + static int trns_index = 1; + char name[LONG_NAME_LEN]; + struct variable *new_var; + struct reg_trns *t = NULL; + + t = xmalloc (sizeof (*t)); + t->trns_id = trns_index; + t->n_trns = n_trns; + t->c = c; + reg_get_name (name, prefix); + new_var = dict_create_var (default_dict, name, 0); + assert (new_var != NULL); + *v = new_var; + add_transformation (f, regression_trns_free, t); + trns_index++; +} +static void +subcommand_save (int save, pspp_linreg_cache ** models) { - struct variable *residuals = NULL; pspp_linreg_cache **lc; + int n_trns = 0; + int i; assert (models != NULL); if (save) { + /* Count the number of transformations we will need. */ + for (i = 0; i < REGRESSION_SV_count; i++) + { + if (cmd.a_save[i]) + { + n_trns++; + } + } + n_trns *= cmd.n_dependent; + for (lc = models; lc < models + cmd.n_dependent; lc++) { assert (*lc != NULL); assert ((*lc)->depvar != NULL); - residuals = dict_create_var (default_dict, "residuals", 0); - assert (residuals != NULL); - (*lc)->resid = residuals; - add_transformation (regression_trns_proc, pspp_linreg_cache_free, *lc); + if (cmd.a_save[REGRESSION_SV_RESID]) + { + reg_save_var ("RES", regression_trns_resid_proc, *lc, + &(*lc)->resid, n_trns); + } + if (cmd.a_save[REGRESSION_SV_PRED]) + { + reg_save_var ("PRED", regression_trns_pred_proc, *lc, + &(*lc)->pred, n_trns); + } } } - else + else { for (lc = models; lc < models + cmd.n_dependent; lc++) { @@ -599,7 +728,7 @@ reg_print_categorical_encoding (FILE * fp, pspp_linreg_cache * c) size_t j; int n_vars = 0; struct variable **varlist; - struct pspp_linreg_coeff *coeff; + struct pspp_coeff *coeff; const struct variable *v; union value *val; @@ -608,8 +737,8 @@ reg_print_categorical_encoding (FILE * fp, pspp_linreg_cache * c) varlist = xnmalloc (c->n_indeps, sizeof (*varlist)); for (i = 1; i < c->n_indeps; i++) /* c->coeff[0] is the intercept. */ { - coeff = c->coeff + i; - v = pspp_linreg_coeff_get_var (coeff, 0); + coeff = c->coeff[i]; + v = pspp_coeff_get_var (coeff, 0); if (v->type == ALPHA) { if (!reg_inserted (v, varlist, n_vars)) @@ -632,7 +761,7 @@ reg_print_categorical_encoding (FILE * fp, pspp_linreg_cache * c) for (i = 0; i < n_vars; i++) { - coeff = c->coeff + i; + coeff = c->coeff[i]; fprintf (fp, "%s.name = \"%s\";\n\t", varlist[i]->name, varlist[i]->name); fprintf (fp, "%s.n_vals = %d;\n\t", varlist[i]->name, @@ -652,18 +781,18 @@ static void reg_print_depvars (FILE * fp, pspp_linreg_cache * c) { int i; - struct pspp_linreg_coeff *coeff; + struct pspp_coeff *coeff; const struct variable *v; fprintf (fp, "char *model_depvars[%d] = {", c->n_indeps); for (i = 1; i < c->n_indeps; i++) { - coeff = c->coeff + i; - v = pspp_linreg_coeff_get_var (coeff, 0); + coeff = c->coeff[i]; + v = pspp_coeff_get_var (coeff, 0); fprintf (fp, "\"%s\",\n\t\t", v->name); } - coeff = c->coeff + i; - v = pspp_linreg_coeff_get_var (coeff, 0); + coeff = c->coeff[i]; + v = pspp_coeff_get_var (coeff, 0); fprintf (fp, "\"%s\"};\n\t", v->name); } static void @@ -682,10 +811,10 @@ reg_has_categorical (pspp_linreg_cache * c) { int i; const struct variable *v; - + for (i = 1; i < c->n_coeffs; i++) { - v = pspp_linreg_coeff_get_var (c->coeff + i, 0); + v = pspp_coeff_get_var (c->coeff[i], 0); if (v->type == ALPHA) { return 1; @@ -703,7 +832,7 @@ subcommand_export (int export, pspp_linreg_cache * c) int n_quantiles = 100; double increment; double tmp; - struct pspp_linreg_coeff coeff; + struct pspp_coeff *coeff; if (export) { @@ -735,12 +864,12 @@ subcommand_export (int export, pspp_linreg_cache * c) for (i = 1; i < c->n_indeps; i++) { coeff = c->coeff[i]; - fprintf (fp, "%.15e,\n\t\t", coeff.estimate); + fprintf (fp, "%.15e,\n\t\t", coeff->estimate); } coeff = c->coeff[i]; - fprintf (fp, "%.15e};\n\t", coeff.estimate); + fprintf (fp, "%.15e};\n\t", coeff->estimate); coeff = c->coeff[0]; - fprintf (fp, "double estimate = %.15e;\n\t", coeff.estimate); + fprintf (fp, "double estimate = %.15e;\n\t", coeff->estimate); fprintf (fp, "int i;\n\tint j;\n\n\t"); fprintf (fp, "for (i = 0; i < %d; i++)\n\t", c->n_indeps); fprintf (fp, "%s", reg_getvar); @@ -779,7 +908,7 @@ subcommand_export (int export, pspp_linreg_cache * c) } } static int -regression_custom_export (struct cmd_regression *cmd UNUSED) +regression_custom_export (struct cmd_regression *cmd UNUSED, void *aux UNUSED) { /* 0 on failure, 1 on success, 2 on failure that should result in syntax error */ if (!lex_force_match ('(')) @@ -803,7 +932,7 @@ regression_custom_export (struct cmd_regression *cmd UNUSED) int cmd_regression (void) { - if (!parse_regression (&cmd)) + if (!parse_regression (&cmd, NULL)) return CMD_FAILURE; models = xnmalloc (cmd.n_dependent, sizeof *models); @@ -822,9 +951,9 @@ static int is_depvar (size_t k, const struct variable *v) { /* - compare_var_names returns 0 if the variable - names match. - */ + compare_var_names returns 0 if the variable + names match. + */ if (!compare_var_names (v, v_variables[k], NULL)) return 1; @@ -867,32 +996,33 @@ mark_missing_cases (const struct casefile *cf, struct variable *v, /* Parser for the variables sub command */ static int -regression_custom_variables(struct cmd_regression *cmd UNUSED) +regression_custom_variables (struct cmd_regression *cmd UNUSED, + void *aux UNUSED) { - lex_match('='); + lex_match ('='); if ((token != T_ID || dict_lookup_var (default_dict, tokid) == NULL) && token != T_ALL) return 2; - - if (!parse_variables (default_dict, &v_variables, &n_variables, - PV_NONE )) + + if (!parse_variables (default_dict, &v_variables, &n_variables, PV_NONE)) { free (v_variables); return 0; } - assert(n_variables); + assert (n_variables); return 1; } + /* Count the explanatory variables. The user may or may not have specified a response variable in the syntax. */ -static -int get_n_indep (const struct variable *v) +static int +get_n_indep (const struct variable *v) { int result; int i = 0; @@ -909,16 +1039,16 @@ int get_n_indep (const struct variable *v) } return result; } + /* Read from the active file. Identify the explanatory variables in v_variables. Encode categorical variables. Drop cases with missing values. */ -static -int prepare_data (int n_data, int is_missing_case[], - struct variable **indep_vars, - struct variable *depvar, - const struct casefile *cf) +static int +prepare_data (int n_data, int is_missing_case[], + struct variable **indep_vars, + struct variable *depvar, const struct casefile *cf) { int i; int j; @@ -926,7 +1056,7 @@ int prepare_data (int n_data, int is_missing_case[], assert (indep_vars != NULL); j = 0; for (i = 0; i < n_variables; i++) - { + { if (!is_depvar (i, depvar)) { indep_vars[j] = v_variables[i]; @@ -934,25 +1064,35 @@ int prepare_data (int n_data, int is_missing_case[], if (v_variables[i]->type == ALPHA) { /* Make a place to hold the binary vectors - corresponding to this variable's values. */ + corresponding to this variable's values. */ cat_stored_values_create (v_variables[i]); } - n_data = mark_missing_cases (cf, v_variables[i], is_missing_case, n_data); + n_data = + mark_missing_cases (cf, v_variables[i], is_missing_case, n_data); } } /* - Mark missing cases for the dependent variable. + Mark missing cases for the dependent variable. */ n_data = mark_missing_cases (cf, depvar, is_missing_case, n_data); return n_data; } +static void +coeff_init (pspp_linreg_cache *c, struct design_matrix *dm) +{ + c->coeff = xnmalloc (dm->m->size2 + 1, sizeof (*c->coeff)); + c->coeff[0] = xmalloc (sizeof (*(c->coeff[0]))); /* The first coefficient is the intercept. */ + c->coeff[0]->v_info = NULL; /* Intercept has no associated variable. */ + pspp_coeff_init (c->coeff + 1, dm); +} static bool -run_regression (const struct casefile *cf, void *cmd_ UNUSED) +run_regression (const struct ccase *first, + const struct casefile *cf, void *cmd_ UNUSED) { size_t i; - size_t n_data = 0; /* Number of valide cases. */ - size_t n_cases; /* Number of cases. */ + size_t n_data = 0; /* Number of valide cases. */ + size_t n_cases; /* Number of cases. */ size_t row; size_t case_num; int n_indep = 0; @@ -971,6 +1111,9 @@ run_regression (const struct casefile *cf, void *cmd_ UNUSED) pspp_linreg_opts lopts; assert (models != NULL); + + output_split_file_values (first); + if (!v_variables) { dict_get_vars (default_dict, &v_variables, &n_variables, @@ -997,15 +1140,15 @@ run_regression (const struct casefile *cf, void *cmd_ UNUSED) { n_indep = get_n_indep ((const struct variable *) cmd.v_dependent[k]); lopts.get_indep_mean_std = xnmalloc (n_indep, sizeof (int)); - indep_vars = xnmalloc (n_indep, sizeof *indep_vars); + indep_vars = xnmalloc (n_indep, sizeof *indep_vars); assert (indep_vars != NULL); for (i = 0; i < n_cases; i++) { is_missing_case[i] = 0; } - n_data = prepare_data (n_cases, is_missing_case, indep_vars, - cmd.v_dependent[k], + n_data = prepare_data (n_cases, is_missing_case, indep_vars, + cmd.v_dependent[k], (const struct casefile *) cf); Y = gsl_vector_alloc (n_data); @@ -1042,7 +1185,7 @@ run_regression (const struct casefile *cf, void *cmd_ UNUSED) for (i = 0; i < n_variables; ++i) /* Iterate over the variables for the current case. - */ + */ { val = case_data (&c, v_variables[i]->fv); /* @@ -1057,11 +1200,13 @@ run_regression (const struct casefile *cf, void *cmd_ UNUSED) { if (v_variables[i]->type == ALPHA) { - design_matrix_set_categorical (X, row, v_variables[i], val); + design_matrix_set_categorical (X, row, + v_variables[i], val); } else if (v_variables[i]->type == NUMERIC) { - design_matrix_set_numeric (X, row, v_variables[i], val); + design_matrix_set_numeric (X, row, v_variables[i], + val); } } } @@ -1075,8 +1220,8 @@ run_regression (const struct casefile *cf, void *cmd_ UNUSED) and store pointers to the variables that correspond to the coefficients. */ - pspp_linreg_coeff_init (models[k], X); - + coeff_init (models[k], X); + /* Find the least-squares estimates and other statistics. */