To successfully control something, one has to possess knowledge of critical points, weak links, high-risk groups, right moments, advantageous conditions and hazards. The system CLANST will find them with the help of innovative algorithms and give You a powerful tool of control.
CLANST is the only system with a unique technology of automatic detection of critical situations in an unlimited stream of information.
CLANST is a universal analytic system based on analyzing interdependences of relative frequencies in homogeneous groups (or "clans") of objects, events, which they undergo, and concomitant parameters. The CLANST performs an analysis of interrelations of parameters and events and at the same time assorts them into homogeneous groups or "clans". The term "clan", which is one of the basic concepts of the system, refers to an aggregation of objects, which have a similar behavior. A "behavior" of an object is understood as an aggregation of frequencies of occurrence of different types of events. A "similar behavior" is defined by similarity of frequencies sets for different objects. Studying regularities in changes of relative frequencies of parameters and events for different clans and analyzing principles of automatic formation of the clans in most cases allows to extract new knowledge of processes occurring in a field under study and to find more appropriate decisions from results of analysis.
Performance capabilities of the system can be efficiently realized in an analysis in various spheres:
- disease, injury and death rate;
- crimes and delinquencies;
- catastrophes, accidents and malfunctions;
- stock, foreign exchange and goods markets;
- employment, unemployment, education rate;
- deliveries, suppliers, sales and contracts;
- events and regions;
- websites and internet resources activity;
The principle features of the system are:
- capability of simultaneous analysis of data of any kind - symbolic, numeric, mixed, - preliminary transformation and distribution law hypotheses not being required;
- absence of any restrictions for initial data volume;
- automatic formation of homogeneous groups (clans) of objects, events, parameters;
- identification of all possible interrelations and interdependences of formed homogeneous groups, even though the groups are of different nature and units of measure;
- simultaneous identifying all regularities in a system under studies and analyzing it as a whole.
The capabilities of the system are best demonstrated by a concrete example, which can illustrate the way of presenting initial data and results of an analysis for any of the fields listed above or any other sphere of research interest.
The demonstration of the system is based on an analysis of medical statistics of many years for patients working in extreme natural conditions.
The initial data (N - a patient number, S - record of service in extreme conditions, M - a disease type according to the International Classification of Diseases, P - gender, W - age, D - a post category, R - a region of work in extreme conditions; size of an initial data file is unlimited):
In this example objects are patients (N), events are values of subclasses of diseases according to the International Classification of Diseases (M), concomitant parameters are S, P, W, D, R.
The first stage of work of the system CLANST consists in classifying types of events (diseases) into classes, that correspond to the first level of the International Classification of Diseases, and identifies homogeneous groups under every parameter. The below displays one of the set of diagrams derived in the result of this operation. This diagram helps to identify percentage of frequencies of diseases of different classes for different age groups. Borders of age groups are automatically defined by the system. Similar diagrams are created for each of the rest of the parameters.
For every disease class a set of diagrams, representing correlation of different parameters, is plotted. The picture below shows one of these diagrams - the diagram of frequencies of occurrence of class C diseases for different age groups under different values of record of service in extreme conditions.
The final operation of the first stage is to identify risk groups for every type of events (a disease class). The pictures below represent the diagrams of risk groups only for two disease classes - C and I.
At its second stage of work the system CLANST proceeds to more detailed examination of subclasses of events (diseases) and revelation of homogeneous groups of the diseases of the 2nd level according to the International Classification of Diseases, corresponding homogeneous in parameter values groups of objects being formed. The picture below representing one of the set of the diagrams derived at this stage, illustrates the dynamics of correlation of frequencies of homogeneous groups of 2nd level diseases for different age groups.
For every group of homogeneous diseases, diagrams of interrelated frequency change under every possible combination of different parameters are plotted. The picture below illustrates the dynamics of frequency change for I11 group disease with consideration of the combination of post categories and age groups.
The 2nd stage of work ends by forming risk groups for every disease group of the 2nd level and identifying relative frequencies for every group.
The 3rd stage of work of the system enables to define interrelations of event types (disease classes) of the 1st level for every group of parameters.
The diagram below shows frequencies of concomitant diseases of other classes for male patients with diagnosis C.
Similar diagrams are created for all groups of parameters and all disease classes.
The 4th stage identifies interrelations of different event types (disease classes) of the 2nd level for each group of homogeneous parameters. The diagram below represents frequencies of concomitant groups of diseases for patients with diagnosis I10 for one of the post categories.
Diagrams are plotted for all homogeneous groups of diseases and parameters.
The 5th stage defines interrelations of groups of disease classes of the 1st level for each group of parameters. The diagram below shows frequencies of concomitant classes of diseases for patients, who have diseases of joint class C-I. Similar diagrams are plotted for all homogeneous groups of classes and parameters.
This analysis can be applied to any structured aggregation of information data, which allows transforming it into a certain standard form, containing information about objects, occurring events and concomitant parameters.
Such representation enables to use special processing methods, realized in the system CLANST, and to obtain deeper knowledge of a field of studies and regularities in it, which approximates us to the profounder understanding of the reality.
![[]](/img/k/klimenko_b_f/clanstechnology1/tabl.jpg)
![[]](/img/k/klimenko_b_f/clanstechnology1/ris01.jpg)
![[]](/img/k/klimenko_b_f/clanstechnology1/ris02.jpg)
![[]](/img/k/klimenko_b_f/clanstechnology1/ris03-04.jpg)
![[]](/img/k/klimenko_b_f/clanstechnology1/ris05.jpg)
![[]](/img/k/klimenko_b_f/clanstechnology1/ris06.jpg)
![[]](/img/k/klimenko_b_f/clanstechnology1/ris07.jpg)
![[]](/img/k/klimenko_b_f/clanstechnology1/ris08.jpg)
![[]](/img/k/klimenko_b_f/clanstechnology1/ris09.jpg)
![[]](/img/k/klimenko_b_f/clanstechnology1/ris10.jpg)
![[]](/img/k/klimenko_b_f/clanstechnology1/log.jpg)