/**
* 对给定的文本进行分类
* @param text 给定的文本
* @return 分类结果 1 2 3 4 0
*/
public static Int16 Classify(String text)
{
var terms = BaseNLP.CutWord(text); //中文分词处理(分词后结果可能还包含有停用词）
List <String> term = DropStopWords(terms); //去掉停用词，以免影响分类
// String Classes = tdm.getTraningClassifications();//分类
List <ClassifyResult> crs = new List <ClassifyResult>(); //分类结果
Int32 ValidWordNum = 0;
//MI=new double[terms.length][Classes.length];
for (Int16 i = 0; i < 4; i++)
{
List <float> probility = new List <float>();
// String Ci = Classes[i];//第i个分类
int Ci = i; //第i个分类
probility = CalcProd(term, Ci); //计算给定的文本属性向量terms在给定的分类Ci中的分类条件概率
ClassifyResult cr = new ClassifyResult(probility, i);
crs.Add(cr);
ValidWordNum = probility.Count - 1;
}
if (!(Convert.ToDouble(ValidWordNum) / term.Count > 0.8 && ValidWordNum > 5))
{
foreach (var item in term)
{
if (!IsKeyWord(item))
{
return(0);
}
}
}
//对最后概率结果进行排序
crs.Sort(ClassifyResult.Compare);
//返回概率最大的分类
return(++crs[3].classification);//先加再return
}
private Dictionary <String, int> CalcuCaseWordsFrequen(int casetpye)
{
Dictionary <String, int> WordsFrequency = new Dictionary <String, int>();
foreach (var item in trainingSet[casetpye])
{
Dictionary <String, int> SentanceWordsFrequency = BaseNLP.SentanceTF(item);
WordsFrequency = MergeDict(WordsFrequency, SentanceWordsFrequency);
}
return(WordsFrequency);
}
protected Dictionary <String, int> NormalTF(List <Pair> cutWordResult)
{
return(BaseNLP.TF(cutWordResult));
}
/**
* 去掉停用词
* @param text 给定的文本
* @return 去停用词后结果
*/
static List <String> DropStopWords(List <Pair> oldWords)
{
return(BaseNLP.DropStopWords(oldWords));
}
protected List <Pair> CutWord(String file)
{
return(BaseNLP.CutWord(file));
}
