Goal of research
- Make a description of the data of the judicial system on a scale from one court to the aggregate of all courts.
- Analyze patterns of each court and the relationship of the similarity of states between courts.
- Create methodology of formation and analysis of the data on the basis of the modern research in the field of data science (Data Science, including Data Analysis, Computer Vision, Cluster analysis, Pattern recognition).
- Approbate methods, approaches and algorithms by computational experiments on the real data of the commercial courts.
Mathematical methods include:
- сomparison of probability distributions;
- Cluster analysis;
- OLAP (On-Line Analytical Processing) cubes.
Interpretation and verification of the collected data is based on the modern legal theory, including legal positivism (G. Hart and his followers), R. Dworkin’s theory, Critical Legal Studies.
Also traditional legal methods are used:
- classical legal dogmatic methods;
- comparative legal method.
Empirical verification of the hypothesis is based on the contemporary sociological methodology (in the aggregate with big data analysis).
Empirical base of research
Research is based on the judicial decisions that are published in Commercial Case Files (http://kad.arbitr.ru/). As the data is full, the analysis results in comprehensive assessment of judicial practice characteristics.
Results of research
First of all, the study proved that the developed methodology, based on the Data-driven paradigm, is effective and promising for the study of big data of complex systems in general. It especially concerns with litigation.
The collection, preliminary processing and analysis of data of the Russian commercial proceedings have been conducted.
As a result, a network structure is constructed. The major role is devoted to the time lines (representations) of the states of the courts (individual). Clusters are identified on each time line: periods of bifurcations, periods of trend dynamic stability and periods of stochastic dynamic stability (without a pronounced “smooth trend”). The pattern analysis of dynamically stable clusters for the repeatability of periods of systems evolution is carried out. In addition, secondary connections (measures of cluster similarity among themselves) are also defined on the described network structure. Finally, the construction of a generalized practice and practice, that is general to the courts, has been made. The two named virtual objects have significantly different characteristics.
There are some methodological gaps due to the lack of scientifically developed methods for determination of the direction of evolution in terms of optimal transformation, transformation of probability measures. The theoretical task of constructing a method for determining this direction may be part of further research. In the case of the aforementioned task, the results of the analysis of the legal proceedings should be clarified, since the new methods can potentially determine the “hidden” periods of bifurcation, additional to those that are already discovered.
Also the analysis of multifactorial and complex states of the network structure of the Russian commercial proceedings have to be conducted. It is necessary to classify the various bifurcations in the states of the courts, to clarify the results of clustering and pattern analysis, to make a basis for the interpretation stage. At present, the resulting data structure is so complex and large that it is extremely difficult to carry out interpretative analysis.
The introduction of additional factors clarifying the state, the connection of bifurcations with external influences, changes in the legal field, comments of the Supreme Court, etc. can be useful in describing the evolution of Arbitrazh proceedings as complex system and individual commercial courts of a particular interest.
Thus, there are prospects for the future research. However, due to the fact that the methodology is based on the techniques of the scientific frontline and the research itself is a breakthrough, there is incompleteness in some aspects of the results. Nevertheless, it does not reduce the value of the methodology of complex systems analysis.
Level of implementation, recommendations on implementation or outcomes of the implementation of the results
Results can be used in law-making and law-enforcement practice of government agencies, especially in commercial courts activity.
 Porter F. Testing Consistency of Two Histograms, California Institute of Technology Lauritsen Laboratory for High Energy Physics, Pasadena, California, 2008.
 Stryukov R.K., Shashkin A.I. “O modificacii metoda blizhayshih sosedey. Vestnik VGU, Seriya Sistemniy analiz I infomatsionniye tehnologii. № 1. 2015, Filipovych, Roman; Resnick, Susan M.; Davatzikos, Christos (2011). "Semi-supervised Cluster Analysis of Imaging Data". NeuroImage. 54 (3): 2185–2197. PMC 3008313 . PMID 20933091. doi:10.1016/j.neuroimage.2010.09.074.
 Dworkin R. Taking Rights Seriously. 1978. Dworkin R. Empire of La. Cambridge, Mass., 1986.