Noumenon based criminal justice case inference method by means of linear programming

Abstract

The invention discloses a noumenon based criminal justice case inference method by means of linear programming. The method comprises the following steps that 1) data of criminal cases is expressed and stored in the vector form scientifically and effectively; S2) a model established on the basis of GA-KNN is used for inference; S3) prior knowledge is used effectively via linear programming, namely, an output result of an algorithm is evaluated according to subjective experiences of law experts, and the execution precision of a case inference system can be improved effectively; and S4) an addition, deletion, modification and checking module of a mass of criminal case data is used to provide an operational interface for data access by a user. According to technical schemes of the invention, the prior knowledge of the law experts can be used to adjust parameters learned by machine learning in the data screening phase, cases more correlated to a case to be judged serve as a training set and are provided for a training neural network, and the decision precision of the whole system is improved.

Description

technical field [0001] The invention relates to an ontology-based criminal trial case reasoning method with the aid of linear programming, in particular to a ontology-based criminal trial case retrieval method with the aid of GA-KNN and linear programming. Background technique [0002] The main tasks of the case-based decision-making system based on case reasoning are the reasoning and decision-making of cases and the establishment and maintenance of evidence and case databases. [0003] Ontology is a concept derived from philosophy. It originally meant the theory of existence, its essence and laws, and was later introduced by the field of computer science. It specifically refers to the clarification, formalization, and standardization of shared conceptual models. Essential concepts, and also emphasize the relationship between these essential concepts. The ontology of a certain field can explicitly and formally express the various concepts and the relationship between conce...

AI Technical Summary

Problems solved by technology

[0005] However, none of these methods can be directly applied to the characteristics of my country's legal system.

In 2012, the article "Criminal Sentencing Decision-Making System Based on Neural Network" proposed a criminal case sentencing decision-making system based on artificial neural network, which input various sentencing plots of the case into the neural network, and used the learning and reasoning of the neural network Ability to obtain the final sentencing result, but this method does not have the step of screening the case characteristics closely related to the type of criminal sentence, but directly inputs all the data into the neural network for measurement, so the accuracy still needs to be improved

But neither of these two methods takes into account the combination of subjective experience for reasoning

In the second scheme, the prior knowledge of legal experts is not used to participate in adjusting the parameters learned by machine learning to improve the accuracy of the system.

Images