Atomic Operation Classification for Effective Legal Supervision Model Construction

With the rapid development of big data technology, more and more complex legal supervision models are established to achieve different supervision goals. How to reduce the constructive difficulty and improve the usability is an essential challenge to be solved. In this paper, we propose a framework to classify the operations into different groups and named them as atomic operations. First, we divide the atomic operations into four different categories, including the basic operations, basic machine learning operations, specific functions and some specific operations for special areas. Secondly, for different categories, we simple give the implementation of different operations, and select specific operation as an example to issullate the detailed implementation process. Based on the atomic operations proposed in this paper, we can not only reduce the construction difficulty of the legal supervision model, but also can improve the interpretability of different constructed models, as they are formed by different atomic operations.