/**
* 数字+数字合并,zheng
*
* @param terms
*/
public static void recognition(Term[] terms) {
int length = terms.length - 1;
Term from = null;
Term to = null;
Term temp = null;
for (int i = 0; i < length; i++) {
if (terms[i] == null) {
continue;
} else if (".".equals(terms[i].getName())) {
// 如果是.前后都为数字进行特殊处理
to = terms[i].getTo();
from = terms[i].getFrom();
if (from.getTermNatures().numAttr.flag && to.getTermNatures().numAttr.flag) {
from.setName(from.getName() + "." + to.getName());
TermUtil.termLink(from, to.getTo());
terms[to.getOffe()] = null;
terms[i] = null;
i = from.getOffe() - 1;
}
continue;
} else if (!terms[i].getTermNatures().numAttr.flag) {
continue;
}
temp = terms[i];
// 将所有的数字合并
while ((temp = temp.getTo()).getTermNatures().numAttr.flag) {
terms[i].setName(terms[i].getName() + temp.getName());
}
// 如果是数字结尾
if (temp.getTermNatures().numAttr.numEndFreq > 0) {
terms[i].setName(terms[i].getName() + temp.getName());
temp = temp.getTo();
}
// 如果不等,说明terms[i]发生了改变
if (terms[i].getTo() != temp) {
TermUtil.termLink(terms[i], temp);
// 将中间无用元素设置为null
for (int j = i + 1; j < temp.getOffe(); j++) {
terms[j] = null;
}
i = temp.getOffe() - 1;
}
}
}
/**
* 新词熵及其左右熵
*
* @param all
*/
public static double leftRightEntropy(List<Term> all) {
// TODO Auto-generated method stub
double score = 0;
NewWordNatureAttr newWordAttr = null;
Term first = all.get(0);
// 查看左右链接
int twoWordFreq = TwoWordLibrary.getTwoWordFreq(first.getFrom(), first);
score -= twoWordFreq;
// 查看右连接
int length = all.size() - 1;
Term end = all.get(all.size() - 1);
twoWordFreq = TwoWordLibrary.getTwoWordFreq(end, end.getTo());
score -= twoWordFreq;
// 查看内部链接
for (int i = 0; i < length; i++) {
score -= TwoWordLibrary.getTwoWordFreq(all.get(i), all.get(i + 1));
}
if (score < -3) {
return 0;
}
// 首字分数
newWordAttr = first.getTermNatures().newWordAttr;
score += getTermScore(newWordAttr, newWordAttr.getB());
// 末字分数
newWordAttr = end.getTermNatures().newWordAttr;
score += getTermScore(newWordAttr, newWordAttr.getE());
// 中词分数
double midelScore = 0;
Term term = null;
for (int i = 1; i < length; i++) {
term = all.get(i);
newWordAttr = term.getTermNatures().newWordAttr;
midelScore += getTermScore(newWordAttr, newWordAttr.getM());
}
score += midelScore / (length);
return score;
}
public void recognition() {
if (branch == null) {
return;
}
int length = terms.length - 1;
Term term = null;
for (int i = 0; i < length; i++) {
if (terms[i] == null) {
continue;
} else {
from = terms[i].getFrom();
terms[i].score = 0;
terms[i].selfScore = 0;
}
branch = branch.getBranch(terms[i].getName());
if (branch == null || branch.getStatus() == 3) {
reset();
continue;
}
offe = i;
// 循环查找添加
term = terms[i];
sb.append(term.getName());
if (branch.getStatus() == 2) {
term.selfScore = branch.getParam().getScore();
}
boolean flag = true;
while (flag) {
term = term.getTo();
branch = branch.getBranch(term.getName());
// 如果没有找到跳出
if (branch == null) {
break;
}
switch (branch.getStatus()) {
case 1:
sb.append(term.getName());
continue;
case 2:
sb.append(term.getName());
score = branch.getParam().getScore();
tempNature = branch.getParam().getNature();
to = term.getTo();
makeNewTerm();
continue;
case 3:
sb.append(term.getName());
score = branch.getParam().getScore();
tempNature = branch.getParam().getNature();
to = term.getTo();
makeNewTerm();
flag = false;
break;
default:
System.out.println("怎么能出现0呢?");
break;
}
}
reset();
}
}