案件书面言语量化辅助分析系统

An Analysis System Assisting in Case-concerned Authorship Identification

书面言语鉴定对鉴定人员的经验和主观判断依赖性大,检验过程中定性的分析判断多、定量的少,特征选取的随意性较大、主观性强,缺少特征自动选取与量化分析的辅助性工具软件,凡此等等。案件书面言语量化辅助分析系统针对性地解决上述问题,实现了对别字、方音别字、实词与虚词的使用习惯、繁体字、成语、词汇套、术语、行话、被动结构、处置句、直接引语、标点符号的选用和信封三款用语等特征的半自动选取与量化分析,以及对字频、词性频率、词频、构成词的音节多寡、句长、标点符号的使用频率等特征的自动选取与量化分析,明显降低了检验人员的工作量,提高了检验效率,为更好的得到鉴定意见提供了量化数据支持。

Authorship identification deeply depends on the analyst’s experience and subjective judgment.Suchidentification is achieved more from qualitative analysis than the quantitative,and the selected features derive more out ofarbitrary and subjective choices,lacking the properness-oriented automatic selection and quantification.These hindrancesurged to develop an analysis system assisting in case-concerned authorship identification(ASAAI).The ASAAI can performsemi-automatic select and quantitative analysis on the following features of ghost word,dialect ghost word,usage habitsin content and/or function word,traditional Chinese characters,idioms,vocabulary sets,jargon,terminology,passiveconstructions,disposal sentence,direct speech,punctuations plus the words written on envelope.For the automatic select andquantitative analysis by ASAAI,the relevant features involve with the character frequency,frequency of part of speech,wordfrequency,number of the syllables constituting a word,sentence length,frequency of the used punctuation and the othersrelated.The ASAAI will significantly reduce the workload of analysts,capable of improving the test efficiency and providingthe quantitative data for expert opinion to be better drawn out.