/**
* 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;
}
/**
* 从一个词的词性到另一个词的词的分数
*
* @param form 前面的词
* @param to 后面的词
* @return 分数
*/
public static double compuScore(Term from, Term to) {
double frequency = from.getTermNatures().allFreq + 1;
if (frequency < 0) {
return from.getScore() + MAX_FREQUENCE;
}
int nTwoWordsFreq = TwoWordLibrary.getTwoWordFreq(from, to);
double value =
-Math.log(
dSmoothingPara * frequency / (MAX_FREQUENCE + 80000)
+ (1 - dSmoothingPara) * ((1 - dTemp) * nTwoWordsFreq / frequency + dTemp));
if (value < 0) value += frequency;
if (value < 0) {
value += frequency;
}
return from.getScore() + value;
}
