src/tools/analysis-principal-components.c

   1 /*
   2  * analysis-principal-components.c:
   3  *
   4  * Author:
   5  *   Andreas J. Guelzow  <aguelzow@pyrshep.ca>
   6  *
   7  * (C) Copyright 2009 by Andreas J. Guelzow  <aguelzow@pyrshep.ca>
   8  *
   9  *
  10  * This program is free software; you can redistribute it and/or modify
  11  * it under the terms of the GNU General Public License as published by
  12  * the Free Software Foundation; either version 2 of the License, or
  13  * (at your option) any later version.
  14  *
  15  * This program is distributed in the hope that it will be useful,
  16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18  * GNU General Public License for more details.
  19  *
  20  * You should have received a copy of the GNU General Public License
  21  * along with this program; if not, see <https://www.gnu.org/licenses/>.
  22  */
  23
  24 #include <gnumeric-config.h>
  25 #include <glib/gi18n-lib.h>
  26 #include <gnumeric.h>
  27 #include <tools/analysis-principal-components.h>
  28 #include <tools/analysis-tools.h>
  29 #include <value.h>
  30 #include <ranges.h>
  31 #include <expr.h>
  32 #include <func.h>
  33 #include <numbers.h>
  34
  35 static gboolean
  36 analysis_tool_principal_components_engine_run (data_analysis_output_t *dao,
  37                                                analysis_tools_data_generic_t *info)
  38 {
  39         int l = g_slist_length (info->input), i;
  40         GSList *inputdata;
  41
  42         GnmFunc *fd_mean;
  43         GnmFunc *fd_var;
  44         GnmFunc *fd_eigen;
  45         GnmFunc *fd_mmult;
  46         GnmFunc *fd_munit;
  47         GnmFunc *fd_sqrt;
  48         GnmFunc *fd_count;
  49         GnmFunc *fd_sum;
  50         GnmFunc *fd_and;
  51         GnmFunc *fd_if;
  52
  53         GnmExpr const *expr;
  54         GnmExpr const *expr_count;
  55         GnmExpr const *expr_munit;
  56         GnmExpr const *expr_and;
  57
  58         int data_points;
  59         GnmExprList *and_args = NULL;
  60         GnmEvalPos ep;
  61
  62         if (!dao_cell_is_visible (dao, l, 9 + 3 * l)) {
  63                 dao_set_bold (dao, 0, 0, 0, 0);
  64                 dao_set_italic (dao, 0, 0, 0, 0);
  65                 dao_set_cell (dao, 0, 0,
  66                               _("Principal components analysis has "
  67                                 "insufficient space."));
  68                 return 0;
  69         }
  70
  71         fd_mean = gnm_func_lookup_or_add_placeholder ("AVERAGE");
  72         gnm_func_inc_usage (fd_mean);
  73         fd_var = gnm_func_lookup_or_add_placeholder ("VAR");
  74         gnm_func_inc_usage (fd_var);
  75         fd_eigen = gnm_func_lookup_or_add_placeholder ("EIGEN");
  76         gnm_func_inc_usage (fd_eigen);
  77         fd_mmult = gnm_func_lookup_or_add_placeholder ("MMULT");
  78         gnm_func_inc_usage (fd_mmult);
  79         fd_munit = gnm_func_lookup_or_add_placeholder ("MUNIT");
  80         gnm_func_inc_usage (fd_munit);
  81         fd_sqrt = gnm_func_lookup_or_add_placeholder ("SQRT");
  82         gnm_func_inc_usage (fd_sqrt);
  83         fd_count = gnm_func_lookup_or_add_placeholder ("COUNT");
  84         gnm_func_inc_usage (fd_count);
  85         fd_sum = gnm_func_lookup_or_add_placeholder ("SUM");
  86         gnm_func_inc_usage (fd_sum);
  87         fd_and = gnm_func_lookup_or_add_placeholder ("AND");
  88         gnm_func_inc_usage (fd_and);
  89         fd_if = gnm_func_lookup_or_add_placeholder ("IF");
  90         gnm_func_inc_usage (fd_if);
  91
  92         dao_set_bold (dao, 0, 0, 0, 0);
  93         dao_set_italic (dao, 0, 0, 0, 11 + 3 * l);
  94         dao_set_format (dao, 0, 0, 0, 0,
  95         /* translator info: The quotation marks in the next strings need to */
  96         /* remain since these are Excel-style format strings */
  97                         _("\"Principal Components Analysis\";"
  98                           "[Red]\"Principal Components Analysis is invalid.\""));
  99         dao_set_align (dao, 0, 0, 0, 0,
 100                        GNM_HALIGN_LEFT, GNM_VALIGN_BOTTOM);
 101
 102         dao->offset_row++;
 103         analysis_tool_table (dao, info, _("Covariances"), "COVAR", TRUE);
 104         dao->offset_row--;
 105
 106         for (i = 1, inputdata = info->input; inputdata != NULL; i++, inputdata = inputdata->next)
 107                 analysis_tools_write_label (inputdata->data, dao, info, 0, 9 + 2 * l + i, i);
 108
 109         eval_pos_init_sheet (&ep,
 110                              ((GnmValue *)(info->input->data))->v_range.cell.a.sheet);
 111         data_points = value_area_get_width (info->input->data, &ep) *
 112                 value_area_get_height (info->input->data, &ep);
 113
 114         for (i = 0; i < l; i++)
 115                 and_args = gnm_expr_list_prepend
 116                         (and_args,
 117                          gnm_expr_new_binary
 118                          (gnm_expr_new_constant (value_new_int (data_points)),
 119                           GNM_EXPR_OP_EQUAL,
 120                           make_cellref (1 + i, 3 + l)));
 121         expr_and = gnm_expr_new_funcall (fd_and, and_args);
 122         dao_set_cell_expr (dao, 0, 0,
 123                            gnm_expr_new_funcall3
 124                            (fd_if,
 125                             expr_and,
 126                             gnm_expr_new_constant (value_new_int (1)),
 127                             gnm_expr_new_constant (value_new_int (-1))));
 128         dao_set_merge (dao,0,0,2,0);
 129         set_cell_text_col (dao, 0, 3 + l,
 130                            _("/Count"
 131                              "/Mean"
 132                              "/Variance"
 133                              "//Eigenvalues"
 134                              "/Eigenvectors"));
 135         dao_set_cell (dao, 0, 11 + 3 * l, _("Percent of Trace"));
 136         dao_set_italic (dao, 0, 9 + 2 * l, 1 + l, 9 + 2 * l);
 137         dao_set_percent (dao, 1, 11 + 3 * l, 1 + l, 11 + 3 * l);
 138
 139         for (i = 1, inputdata = info->input; inputdata != NULL; i++, inputdata = inputdata->next) {
 140                 expr = gnm_expr_new_constant (value_dup (inputdata->data));
 141
 142                 dao_set_cell_expr (dao, i, 3 + l,
 143                                    gnm_expr_new_funcall1 (fd_count, gnm_expr_copy (expr)));
 144                 dao_set_cell_expr (dao, i, 4 + l,
 145                                    gnm_expr_new_funcall1 (fd_mean, gnm_expr_copy (expr)));
 146                 dao_set_cell_expr (dao, i, 5 + l,
 147                                    gnm_expr_new_funcall1 (fd_var, expr));
 148         }
 149
 150         expr_count = gnm_expr_new_binary (make_cellref (0,-4), GNM_EXPR_OP_DIV,
 151                                           gnm_expr_new_binary (make_cellref (0,-4), GNM_EXPR_OP_SUB,
 152                                                                gnm_expr_new_constant (value_new_int (1))));
 153         expr = gnm_expr_new_funcall1
 154                 (fd_eigen, gnm_expr_new_binary
 155                  (expr_count, GNM_EXPR_OP_MULT, make_rangeref (0, - (5 + l), l - 1, - 6)));
 156         dao_set_array_expr (dao, 1, 7 + l, l, l + 1, expr);
 157
 158         for (i = 1; i <= l; i++) {
 159                 dao_set_align (dao, i, 9 + 2 * l, i, 9 + 2 * l,
 160                                GNM_HALIGN_CENTER, GNM_VALIGN_BOTTOM);
 161                 dao_set_cell_printf (dao, i, 9 + 2 * l, "\xce\xbe%i", i);
 162                 dao_set_cell_expr (dao, i, 11 + 3 * l,
 163                                    gnm_expr_new_binary (make_cellref (0,- 4 - 2 * l),
 164                                                         GNM_EXPR_OP_DIV,
 165                                                         gnm_expr_new_funcall1
 166                                                         (fd_sum,
 167                                                          dao_get_rangeref (dao, 1, 7 + l, l, 7 + l))));
 168         }
 169
 170         expr_munit =  gnm_expr_new_funcall1 (fd_munit, gnm_expr_new_constant (value_new_int (l)));
 171         expr = gnm_expr_new_funcall2 (fd_mmult,
 172                                       gnm_expr_new_binary
 173                                       (gnm_expr_new_funcall1
 174                                        (fd_sqrt, gnm_expr_new_binary
 175                                         (gnm_expr_new_constant (value_new_int (1)),
 176                                          GNM_EXPR_OP_DIV,
 177                                          make_rangeref (0, - 5 - l, l - 1, - 5 - l))),
 178                                        GNM_EXPR_OP_MULT,
 179                                        gnm_expr_copy (expr_munit)),
 180                                       make_rangeref (0, - 2 - l, l - 1, - 3));
 181         expr = gnm_expr_new_funcall2 (fd_mmult, expr,
 182                                       gnm_expr_new_binary
 183                                       (gnm_expr_new_funcall1
 184                                        (fd_sqrt, make_rangeref (0, - 3 - l, l - 1, - 3 - l)),
 185                                        GNM_EXPR_OP_MULT,
 186                                        expr_munit));
 187         dao_set_array_expr (dao, 1, 10 + 2 * l, l, l, expr);
 188
 189         gnm_func_dec_usage (fd_mean);
 190         gnm_func_dec_usage (fd_var);
 191         gnm_func_dec_usage (fd_eigen);
 192         gnm_func_dec_usage (fd_mmult);
 193         gnm_func_dec_usage (fd_munit);
 194         gnm_func_dec_usage (fd_sqrt);
 195         gnm_func_dec_usage (fd_count);
 196         gnm_func_dec_usage (fd_sum);
 197         gnm_func_dec_usage (fd_and);
 198         gnm_func_dec_usage (fd_if);
 199
 200         dao_redraw_respan (dao);
 201         return 0;
 202 }
 203
 204 gboolean
 205 analysis_tool_principal_components_engine (G_GNUC_UNUSED GOCmdContext *gcc, data_analysis_output_t *dao, gpointer specs,
 206                                    analysis_tool_engine_t selector, gpointer result)
 207 {
 208         analysis_tools_data_generic_t *info = specs;
 209
 210         switch (selector) {
 211         case TOOL_ENGINE_UPDATE_DESCRIPTOR:
 212                 return (dao_command_descriptor
 213                         (dao, _("Principal Components Analysis (%s)"), result)
 214                         == NULL);
 215         case TOOL_ENGINE_UPDATE_DAO:
 216                 prepare_input_range (&info->input, info->group_by);
 217                 dao_adjust (dao, 1 + g_slist_length (info->input),
 218                             12 + 3 * g_slist_length (info->input));
 219                 return FALSE;
 220         case TOOL_ENGINE_CLEAN_UP:
 221                 return analysis_tool_generic_clean (specs);
 222         case TOOL_ENGINE_LAST_VALIDITY_CHECK:
 223                 return FALSE;
 224         case TOOL_ENGINE_PREPARE_OUTPUT_RANGE:
 225                 dao_prepare_output (NULL, dao, _("Principal Components Analysis"));
 226                 return FALSE;
 227         case TOOL_ENGINE_FORMAT_OUTPUT_RANGE:
 228                 return dao_format_output (dao, _("Principal Components Analysis"));
 229         case TOOL_ENGINE_PERFORM_CALC:
 230         default:
 231                 return analysis_tool_principal_components_engine_run (dao, specs);
 232         }
 233         return TRUE;  /* We shouldn't get here */
 234 }