1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
|
/*
* 文節の構造metawordをソートする
*
* 文節に対する複数の構造の候補をソートする
*
* Copyright (C) 2000-2007 TABATA Yusuke
*
*/
/*
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*/
#include <stdlib.h>
#include <math.h>
#include <anthy/segment.h>
#include <anthy/ordering.h>
#include <anthy/feature_set.h>
#include <anthy/splitter.h>
#include <anthy/diclib.h>
#include "sorter.h"
static void *cand_info_array;
static double
calc_probability(struct feature_list *fl)
{
struct feature_freq *res, arg;
res = anthy_find_feature_freq(cand_info_array,
fl, &arg);
if (res) {
double pos = (double)res->f[15];
double neg = (double)res->f[14];
double prob = pos / (pos + neg);
prob = prob * prob;
/**/
return prob;
}
return 0;
}
static void
mw_eval(struct seg_ent *prev_seg, struct seg_ent *seg,
struct meta_word *mw)
{
int pc;
struct feature_list fl;
double prob;
(void)seg;
anthy_feature_list_init(&fl);
/**/
anthy_feature_list_set_cur_class(&fl, mw->seg_class);
anthy_feature_list_set_dep_word(&fl, mw->dep_word_hash);
anthy_feature_list_set_dep_class(&fl, mw->dep_class);
anthy_feature_list_set_mw_features(&fl, mw->mw_features);
/* 前の文節の素性 */
if (prev_seg) {
pc = prev_seg->best_seg_class;
} else {
pc = SEG_HEAD;
}
anthy_feature_list_set_class_trans(&fl, pc, mw->seg_class);
anthy_feature_list_sort(&fl);
/* 計算する */
prob = 0.1 + calc_probability(&fl);
if (prob < 0) {
prob = (double)1 / (double)1000;
}
anthy_feature_list_free(&fl);
mw->struct_score = RATIO_BASE * RATIO_BASE;
mw->struct_score *= prob;
/*
anthy_feature_list_print(&fl);
printf(" prob=%f, struct_score=%d\n", prob, mw->struct_score);
*/
/**/
if (mw->mw_features & MW_FEATURE_SUFFIX) {
mw->struct_score /= 2;
}
if (mw->mw_features & MW_FEATURE_WEAK_CONN) {
mw->struct_score /= 10;
}
}
static void
seg_eval(struct seg_ent *prev_seg,
struct seg_ent *seg)
{
int i;
for (i = 0; i < seg->nr_metaword; i++) {
mw_eval(prev_seg, seg, seg->mw_array[i]);
}
}
static void
sl_eval(struct segment_list *seg_list)
{
int i;
struct seg_ent *prev_seg = NULL;
for (i = 0; i < seg_list->nr_segments; i++) {
struct seg_ent *seg;
seg = anthy_get_nth_segment(seg_list, i);
seg_eval(prev_seg, seg);
prev_seg = seg;
}
}
static int
metaword_compare_func(const void *p1, const void *p2)
{
const struct meta_word * const *s1 = p1;
const struct meta_word * const *s2 = p2;
return (*s2)->struct_score - (*s1)->struct_score;
}
void
anthy_sort_metaword(struct segment_list *seg_list)
{
int i;
/**/
sl_eval(seg_list);
/**/
for (i = 0; i < seg_list->nr_segments; i++) {
struct seg_ent *seg = anthy_get_nth_segment(seg_list, i);
if (seg->mw_array) { /* 不正なメモリアクセスを行うバグの修正 */
qsort(seg->mw_array, seg->nr_metaword, sizeof(struct meta_word *),
metaword_compare_func);
}
}
}
void
anthy_infosort_init(void)
{
cand_info_array = anthy_file_dic_get_section("cand_info");
}
|
