Java版的BosonNLP分词

package com.test; import org.junit.Test; import com.boso.Bosonnlp; import com.enumBean.TypeBean; import com.mashape.unirest.http.exceptions.UnirestException; import com.util.Util; public class BosoNnlpTest { @Test public void sentiment() throws UnirestException { String [] data = new String[]{"中国共产党", "美好的世界"}; System.out.println(Bosonnlp.determine(TypeBean.SENTIMENT.getDesc(),data)); //第一个值为非负面概率，第二个值为负面概率，两个值相加和为 1 //[[0.6519134382562579,0.34808656174374203],[0.92706110187413,0.07293889812586994]] } /** * 命名实体识别（NER）是指识别文本中具有特定意义的实体，主要包括人名、地名、机构名、 * 专有名词等。命名实体识别是信息提取、问答系统、句法分析、机器翻译等应用领域的重要基础工具 * @author wangbo */ @Test public void ner() throws UnirestException { String [] data = new String[]{"对于该小孩是不是郑尚金的孩子，目前已做亲子鉴定，结果还没出来，纪检部门仍在调查之中。成都商报记者 姚永忠"}; System.out.println(Bosonnlp.ner(data)); } /** * 依存文法分析核心思想为将一个线性描写的句子表述为成员之间的搭配与驱动关系 * @author wangbo */ @Test public void depparser() throws UnirestException { String [] data = new String[]{"我以最快的速度吃了午饭"}; System.out.println(Bosonnlp.determine(TypeBean.DEPPARSER.getDesc(),data)); //[{"head":[6,6,3,4,5,1,-1,6,6], //"role":["SBJ","MNR","VMOD","DEC","NMOD","POBJ","ROOT","VMOD","OBJ"], //"tag":["PN","P","AD","VA","DEC","NN","VV","AS","NN"],"word":["我","以","最","快","的","速度","吃","了","午饭"]}] } /** * 以将文本自动进行关键词分析，给出每个词语相应的权重 * @author wangbo */ @Test public void keywords() throws UnirestException { String [] data = new String[]{"病毒式媒体网站：让新闻迅速蔓延"}; System.out.println(Bosonnlp.determine(TypeBean.KEYWORDS.getDesc(),data)); //[[[0.5686631749811326,"蔓延"], //[0.5671956747680966,"病毒"], //[0.33993439662135194,"迅速"], //[0.30642011458238383,"网站"], //[0.26392731183346607,"新闻"],[0.23807884315568398,"媒体"],[0.12876489756725826,"式"],[0.0504282185215189,"让"]]] } /** * 将新闻文本归类到预设的 14 个分类当中 0体育1教育2财经3社会4娱乐5军事6国内7 科技8 互联网9房产10国际11女人12 汽车13游戏 * @author wangbo */ @Test public void classify() throws UnirestException { String [] data = new String[]{ "俄否决安理会谴责叙军战机空袭阿勒颇平民", "邓紫棋谈男友林宥嘉：我觉得我比他唱得好", "Facebook收购印度初创公司"}; String str = Bosonnlp.determine(TypeBean.CLASSIFY.getDesc(),data); String classify[] = str.substring(str.indexOf("[")+1, str.indexOf("]")).split(","); for (String string : classify) { System.out.println(Util.classify(Integer.parseInt(string))); } //[5,4,8] } @Test public void suggest() throws UnirestException { String data = "\"傻逼\""; System.out.println(Bosonnlp.suggest(TypeBean.SUGGEST.getDesc(),data)); } /** * 将中文时间描述转换为三种标准的时间格式的时间字符串：1) 时间点（timestamp，表示某一具体时间时间描述）; 2) 时间量（timedelta，表示时间的增量的时间描述）; 3)时间区间（timespan，大于一天的有具体起始和结束时间点的时间描述），以方便处理。在转换时以尽量保留原始时间描述中所含信息为原则。 * @author wangbo */ @Test public void time() throws UnirestException { String [] data = new String[]{}; System.out.println(Bosonnlp.time(TypeBean.TIME.getDesc(),data,"2013年二月二十八日下午四点三十分二十九秒")); //{"type":"timestamp","timestamp":"2013-02-28 16:30:29"} } }