Developing of the law-case automatic categorizing system based on law lexicons
GUAN Li-he~1,YANGGang~2,LI Yong-li~2 ( 1.Department of Computera and Information, Chongqing Jiaotong University, Chongqing 400074,China; 2.School of Information Science and Engineering, Lanzhou University, Gansu Lanzhou 730000,China)
The paper discussed and developed a Law-case automatic categorizing system, which makeis a subsystem of the "Law-case Analyzing System".Firstly,the characteristic-word weight tables of each category are gained out of a mass of law-case training documents which have been categorized already.Secondly,the weight summation is conducted for each category based on the weight tables of the characteristic Words.Finally,the related law case falls under the category which is at the leafage of the Category tree and gets the biggest weiht sum.The paper also presents and andyzes two important formaula. the characteristic word weight formula and the weight-summing formula,puts forward a new word-parting algorithm based on law lexicons,which is the core module of the system. The experiment shows the excellent generality,expansibility and satisfactory categorizing nicety of the system.