Method for a sequential prediction of binary element's state in a binary process and the system for the method implementation

• US 6,581,047 B1
• Filed: 04/07/1999
• Issued: 06/17/2003
• Est. Priority Date: 04/07/1999
• Status: Expired due to Fees

• Current Assignee: Inesa Incorporated
• Original Assignee: Inesa Incorporated
• Inventors: Vinarskiy, Ilya, Raykhman, Alexander M.
• Primary Examiner(s): STARKS, WILBERT L
• Application Number:
  US09/287,848
• Time in Patent Office: 1,532 Days
• Field of Search: 375/240, 375/240.12, 375/253, 358/261.2, 382/243, 706/22
• US Class Current: 706/21
• CPC Class Codes: G06N 3/126 Evolutionary algorithms, e.g. genetic algorithms or genetic programming

Abstract

The method for a sequential prediction of a binary element'"'"'s state in a binary process and the system for the method implementation are based on the binary process representation by a unique sequence of binary elementary behavioral functions. These elementary behavioral functions are used for building a genotype of the binary process'"'"' behaviors. The behavioral genotype is periodically updated through the analysis of the binary process history. The behavioral genotype set defines a set of predicting functions with respect of an evaluation of the binary process'"'"' current instability level. The natural selection algorithm that operates upon the set of predicting functions is used in predicting procedures for a non-iterative determination of a state of the forthcoming binary process'"'"' element. If the binary process is non-stationary, then a cardinal number of the set of predicting functions is equal or less than a cardinal number of the behavioral genotype set.

The method allows performing real time predictions of binary processes, including non-stationary processes, regardless of the possible speed and memory limitations of computing means.

The predicting system for the method implementation increases the efficiency of functioning of any entity that includes this system as a subsystem.

Claims

1. A method for predicting a state of a binary element in a binary process and method being performed with the aid of computing means, said method comprising the steps of:

providing a computer-readable medium including instructions for;

forming a binary string for said binary process to obtain a history storage so that each new element of said binary process is an element of said binary string;

decomposing said binary string into structures of a minimal length to uniquely represent said binary string, said structures are built from opposed to each other adjacent sequences of identical states of said binary process element, such that a propagation of every said structure in both directions of said binary process domain returns a regular order;

selecting elements that are similar to said binary structures and may includes said structures;

combining said elements into an initial set of elements;

evaluating said binary string instability level if said instability level is not pre evaluated;

generating a resulting set of elements which is built from said initial set of elements depending on the evaluated level of said binary string instability;

sequentially predicting the state of forthcoming elements accumulating in a group in said binary string by applying an element vise predicting procedure acting upon said resulting set of elements;

updating said initial set of elements and reevaluating of said binary string instability level;

returning to the generating step of said method if additional predictions are needed;

whereby said sequential prediction of said binary element'"'"'s state in stationary, quasi-stationary, and non stationary binary process is generated.

2. A method as claimed in claim 1, wherein:

said binary process may be a direct output of any entity, which is defined on a space of discrete binary processes; and

said binary process may be a mapping of any other than said binary process kind entity'"'"'s output onto the space of discrete binary processes.

3. A method as claimed in claim 1, further including the step of:

generating said initial set of elements only if it is permitted by setting up an updating mode for said initial set to at least one of two binary components of a first permitting vector, said first permitting vector having a first component and a second component, said first component indicates a fact of said initial set predetermination, and said second component indicates a fact of beginning of creation of said initial set based on some schedule.

4. A method as claimed in claim 3, when at least one component of said first permitting vector is put to the updating mode for said initial set, further including the steps of:

determining a statistically stable said decomposing structures with a maximal period;

generating a first set containing all possible said decomposing structures including said maximal period structures, so that each element of said first set possesses the property of orthogonality; and

generating said initial set of elements via the union of said first set and a second set of two opposite to each other elements, whose composition is a countable sequence of the identical binary elements of said binary process.

5. A method as claimed in claim 1, wherein;

said element vise predicting procedure is a retrospective predicting procedure if said element vise predicting procedure is defined on an interval of said binary string domain which is antecedent to the latest observed element of said binary string; and

said element vise predicting procedure is an actual predicting procedure if said element vise predicting procedure is defined on an interval of said binary string domain which includes the latest observed element of said binary string; and

said predicting procedure is cyclic and each cycle is composed of two stages, and may contain several predictions.

6. A method as claimed in claim 5, further including the steps of:

beginning each cycle of said predicting procedure at a starting point when an error of prediction occurs; and

performing the first stage of each said predicting cycle through generation of a resulting set of elements by transforming said initial set into said resulting set, executed at the starting point of each said predicting cycle, so that a composition of said resulting set of elements depends on said evaluation of said current instability level of said binary string;

performing the second stage of each said predicting cycle through a creation of a functional model of each forthcoming element of said binary process belonging to said current predicting cycle, and through a generation of a prediction permitting indicator, so that said prediction is an output of the Boolean conjunction of said model and said permitting indicator.

7. A method as claimed in claim 6, further including the steps of:

determining the component of said estimating predicting efficiency vector function, that has an extreme value corresponding to the degree of the binary string stability; and

selecting said resulting set of elements based on a composition of the set that corresponds to the extreme component of said estimating predicting efficiency vector function, such that each member of said resulting set possesses certain common and specific properties that depend on the object of prediction.

8. A method as claimed in claim 7 wherein:

said resulting set'"'"'s common properties are

first, said resulting set is a mapping of said initial set;

second, the cardinal number of said resulting set is inversely related to the evaluated said binary string instability level; and

third, the state of each element of said resulting set is identical to the state of said binary string'"'"'s element which is observed at the beginning of said predicting cycle.

9. A method as claimed in claim 7 wherein:

said resulting set'"'"'s specific property of any non stationary binary process is that said resulting set is a subset of said initial set.

10. A method as claimed in claim 7 wherein:

said resulting set'"'"'s specific property of some of non stationary binary process is that a resulting set'"'"'s member is such a member of said initial set that can be observed on the interval closest to the beginning of the predicting cycle, and whose period is the maximal period among all members of said initial set observed on said interval; and

said resulting set'"'"'s specific property of some of non stationary binary process is that a resulting set'"'"'s member is such a member of said initial set that can be observed on the closest to the beginning of the predicting cycle interval, and can be distinguished as a periodic function containing its full period; and

said resulting set'"'"'s specific property of some of non stationary binary process is that there is a relationship between at least one known member of said resulting set of elements and some other members of said resulting set related to said known member, such that for any binary element'"'"'s state denoted by s, it is true that (1s2s)(2s2s), (2s2s)(2s1s), (1s3s)(2s3s) v (3s3s), (2s3s)(1s3s) v (3s3s),(3s3s)(1s3s)v(2s3s), where the left part of each statement defines known member of said resulting set, and the right part of each statement defines all other members of said resulting set that are related to said known member; and

said resulting set'"'"'s specific property of some of non stationary binary process is that there is a relationship between a sequence of two or three identical binary process elements found on the binary process decomposition, which are adjoined to the beginning moment of a predicting cycle and some members of said resulting set related to said sequence of binary process elements, such that for any binary element'"'"'s state denoted by s, it is true that 2s(2s2s) v (2s1s), 3s(3s3s) v (3s2s) v (3s1s), where the left part of each statement defines said sequence of binary process elements, and the right part of each statement defines all members of said resulting set that are related to said sequence of binary process elements; and

said resulting set'"'"'s specific property of some of non stationary binary process is that the first member of said initial set is a member of said resulting set, and both members of said second set are the members of said resulting set.

11. A method as claimed in claim 7 wherein:

said resulting set'"'"'s specific property of another non stationary binary process is that a resulting set'"'"'s member is such a member of said initial set that can be observed on the interval closest to the beginning of the predicting cycle, and can be distinguished as a periodic function containing its full period; and

said resulting set'"'"'s specific property of another non stationary binary process is that there is a relationship between at least one known member of said resulting set of elements and some other members of said resulting set related to said known member, such that for any binary element'"'"'s state denoted by s, it is true that (1s2s)(2s2s), (2s2s)(2s1s), (1s3s)(2s3s) v (3s3s), (2s3s)(1s3s) v (3s3s), (3s3s)(1s3s) v (2s3s), where the left part of each statement defines known member of said resulting set, and the right part of each statement defines all other members of said resulting set that are related to said known member; and

said resulting set'"'"'s specific property of another non stationary binary process is that there is a relationship between a sequence of two or three identical binary process elements found on the binary process decomposition, which are adjoined to the beginning moment of a predicting cycle and some members of said resulting set related to said sequence of binary process elements, such that for any binary element'"'"'s state denoted by S, it is true that 2s(2s2s) v (2s1s), 3s(3s3s) v (3s2s) v (3s1s), where the left part of each statement defines said sequence of binary process elements, and the right part of each statement defines all members of said resulting set that are related to said sequence of binary process elements; and

said resulting set'"'"'s specific property of another non stationary binary process is that both members of said second set are the members of said resulting set.

12. A method as claimed in claim 6 wherein:

said functional model of a forthcoming element of said binary process is a nonlinear transformation of a base model;

said nonlinear transformation of the base model includes an inversion of said model'"'"'s output based on a result of a conjunction of two binary components of a third permitting vector;

said first component of the third permitting vector is an indicator of the inversion mode initiation, and said second component of the third permitting vector indicates the beginning of said inversion when an output of a current predicting efficiency estimating function is within a specified distance of a predetermined threshold value, and the second component of said third permitting vector also indicates the end of said inversion if a new error of prediction is observed;

said base model is a union of those members of said resulting set that are allowed to participate in the base model;

the allowance of participation in said base model includes that each member of said resulting set participates in the current cycle of predictions since the beginning of said cycle until this member produces an incorrect prediction, and this member may again participate in the prediction only in the next predicting cycle;

said prediction permitting indicator is a conjunction of three binary components of a fourth permitting vector;

said first component of the fourth permitting vector is an indicator of the prediction blocking mode initiation, and said second component of the fourth permitting vector reflects the fact of beginning of said prediction blocking mode when a value of said current predicting efficiency estimating function is within a specified distance of another predetermined threshold value, and the second component of said fourth permitting vector also indicates the end of said blocking if a new error of prediction is observed, and said third component of the fourth permitting vector is a conjunction of all members of said resulting set that participate in said base model.

13. A method as claimed in claim 1, further including the step of:

evaluating said instability of said binary string only if it is permitted by setting up an evaluation updating mode to at least one of two binary components of a second permitting vector, whose first component indicates a fact of said evaluation predetermination, and whose second component indicates a fact of beginning of said evaluation based on some schedule.

14. A method as claimed in claim 7, when at least one component of a second permitting vector is put to the evaluation updating mode further including the step of:

performing said binary string instability evaluation through a calculation of an estimating vector function of predicting efficiency involving but not limited to a retrospective predicting procedure, which is specific for each component of said estimating vector function and is defined on said binary string domain.

15. A method as claimed in claim 14, wherein:

said estimating vector function of predicting efficiency is a vector whose each component is an aggregation function of a sequence of outputs of said retrospective predicting procedure, which acts upon a specific for this component pre selected set of elements, which is a mapping of said initial set of elements;

one of said pre selected sets can be said initial set itself; and

any component of said estimating vector function of predicting efficiency can be predetermined to control the composition of said resulting set of elements.

16. A method as claimed in claim 1 being embedded in a method of time series analysis further including the steps of:

mapping a non binary time series into said vectorial binary process, such that each said component of vectorial binary process component is a product of said time series'"'"' quantification, such quantification is based on values of corresponding components of a vector of quantizing levels;

selecting a first group of components of said vectorial binary process that are suitable for a prediction of the event that a forthcoming element of each member of said first group of components will be located higher than some predetermined position on said time series range;

selecting a second group of components of said vectorial binary process that are suitable for a prediction of the event that a forthcoming element of each member of said second group of components will be located lower than some predetermined position on said time series range;

predicting the forthcoming element'"'"'s state of each binary process in both groups; and

finding a pair of the most proximate, in terms of a location on said time series range, predictions such that the first element of said pair belongs to a binary process from said first group and the second element of said pair belongs to a binary process from said second group;

whereby the position of said time series value in its range is predicted to be in a narrowed corridor, and said corridor is bound by functions of said time series discrete argument.

17. A method as claimed in claim 16 wherein:

selecting said groups of components of said vectorial binary process may involve evaluating predicting efficiency of each component by applying said retrospective predicting procedure;

predicting the forthcoming element'"'"'s state of each binary process from both of the groups of components involves said actual predicting procedure.

18. A method as claimed in claim 1 being embedded in an algorithm for predicting the quality of schedule based communication further including the steps of:

sending a schedule based test signal to each communicator;

receiving a response to said test signal from each communicator;

classifying and then distributing said communicators into two groups by a binary criterion of quality of communication having said responses as values for an independent variable, such that members of one group are classified as having a good quality of communication, and members of the other group are classified as having a bad quality of communication; and

predicting a forthcoming element'"'"'s state in a binary process representing said quality of communication of each communicator;

generating a notifying message for any communicator attempting to contact a member of the group with predicted bad quality of communication;

generating a notifying message if said bad quality of communication is predicted for any of participants during communication.

19. A predicting algorithm as claimed in claim 1 being embedded in a computer based game wherein:

said algorithm is a part of said game'"'"'s background algorithm;

in said game the human intelligence and the artificial intelligence compete so that human tries to challenge a computer program by creating binary puzzles the computer program must solve;

in said game'"'"'s background algorithm said binary puzzle is represented by a binary string of a finite length;

during the play, each element of said binary string is inserted sequentially into a computing device, whose artificial cognitive algorithm must predict the state of said binary element before said element has been inserted;

the number of correct and incorrect predictions is evaluated on at least one binary string length so that human wins and respectively artificial mind loses if some predetermined threshold is exceeded by said evaluation;

in said background algorithm different said evaluating thresholds and said algorithms internal parameters can be setup by a player to get a game with different levels of complexity;

whereby depending on how skilled the human player is, a different result of said competition is expected.

29. A system as claimed in claim 16 having computing and controlling means for receiving, processing, transferring, and exchange of information between the elements of said system and between said system and an environment thereof wherein including:

means for signal to data conversion and means for data to signal conversion;

means for mapping of any non binary process on the space of binary processes;

means for inserting, writing, reading, and modifying data represented by scalars, vectors, and matrices;

means for storing a region of said binary process, forming a binary string so that each new element of said binary process is an element of said binary string;

means for numerical processing and for quantitative analysis of said binary string including means for decomposition, means for data sampling of said decomposition data, means for statistical analysis of said decomposition data, means for creating data sets having functional forms as members of said data sets, means for aggregating strings of values;

whereby said means aid accurate and stable functioning of said predicting system.

30. A system as claimed in claim 16 being embedded in a measuring system further including:

means for an input signal filtering;

means for performing measuring block procedures;

means for sequential prediction of a binary elements'"'"' state in a binary process;

means for calculating a resulting measurement; and

means for controlling a system'"'"'s working modes.

31. A system as claimed in claim 30 wherein:

said input signal and a filtered input signal are used together for generating a signal to noise ratio signal;

said signal to noise ratio variable is considered a non binary process and converted to a binary process representing current level of disturbance evaluation in said system;

the signal to noise evaluation is predicted element by element through the utilization of said predicting subsystem;

each measuring block procedure is executed at a pace related to said signal to noise ratio evaluation, and only that measuring block procedure output is accounted in said resulting measurement, whose predicted signal to noise ratio evaluation indicates a low level of disturbance in said measuring system;

said resulting measurement is an aggregation of a number of permitted for said aggregation outputs of said measuring block procedure so that the part of disturbance in said system output becomes minimized;

said measuring system'"'"'s input is switched to a noiseless testing signal when the high level of disturbance is predicted, so that said measuring system may provide compensation for a multiplicative disturbance;

whereby the accuracy of any measuring system can be significantly improved in those cases when further filtration of any kind is not effective.

20. A computing means implemented system for a sequential prediction of a binary element'"'"'s state in a binary process, comprising:

a computer-readable medium residing in a computing means implemented system further comprising;

a first subsystem including means for determination of said initial set of elements;

a second subsystem including means for prediction of a binary element'"'"'s state in a binary process;

computing and controlling means, which receive, process, transfer, and exchange information between said system'"'"'s elements and between said system and the environment thereof;

whereby said system implements a sequential prediction of any binary process and can be embedded in any entity that permits mapping of at least one of the entity'"'"'s output onto the space of binary processes.

21. A system as claimed in claim 20 wherein:

said first subsystem is a plurality of channels, and each of said channels serves for the purpose of determination of said initial set of a corresponding single binary process;

said second subsystem is a plurality of channels, and each of said channels serves for the purpose of prediction of a binary element'"'"'s state in a corresponding single binary process;

each single channel of said first subsystem transfers information about said initial set composition to each corresponding single channel of said second subsystem.

22. A system as claimed in claim 20 having said first subsystem further including:

means for mapping an output of an object of prediction into at least one binary sequence, means for binary process formation hereafter;

each single channel of said first subsystem comprising means for binary process history storage, means for creation of said initial set, means for data sampling, and means for binary process shortest history storage.

23. A system as claimed in claim 20 having said single channel of said first subsystem wherein:

said means for binary process history storage memorize each sequential element of said binary process coming to a first input thereof from said means for binary process formation;

said means for binary process history storage, being controlled by an output of said means for data sampling, form a binary string at an output of said means for binary process history storage;

said binary string goes from said means for binary process history storage to a first input of said means for creation of said initial set, and said binary string also goes to a first input of said means for binary process shortest history storage, and a second input thereof receives information from the first output of said means for creation of said initial set, and said first output of said initial set is also connected to an input of said means for data sampling;

a second input of said means for creation of said initial set is a control input, that is external to the predicting system for a first component assignment of said first permitting vector;

a third input of said means for creation of said initial set receives information from said corresponding single channel of said second subsystem about the status of said error of prediction indicator to be used for the determination of a second component of said first permitting vector; and

said second output of said means for creation of said initial set is a second output of said single channel of said first subsystem, and the output of means for binary process shortest history storage is a first output of said single channel of said first subsystem.

24. A system as claimed in claim 20 having said second subsystem further including:

each single channel of said second subsystem comprising a main predicting contour including means for selection of main resulting set and means for actual prediction; and

each single channel of said second subsystem comprising an auxiliary instability evaluating contour, which is a plurality of tracts connected to means for evaluating of the binary process current instability.

25. A system as claimed in claim 24 having said main predicting contour of said single channel of said second subsystem wherein:

a first input of said means for selection of main resulting set is utilized for determination of members of main resulting set, in case of external to the predicting system control;

the second input of said means for selection of main resulting set receives data from said means for binary process shortest history storage;

a third input of said means for selection of main resulting set receives data from said means for evaluating of said binary process current instability, so that said composition of the main resulting set depends on a value of said third input;

an output of said means for selection of main resulting set is connected to a third input of said means for actual prediction;

a first input of said means for actual prediction is a control input for assignment of said first component of said third permitting vector in case of external to the predicting system control;

a second input of said means for actual prediction is a control input for assignment of said first component of said fourth permitting vector in case of external to the predicting system control;

a first output of said means for actual prediction is intrinsically a prediction of a binary element'"'"'s state in the corresponding binary process;

a second output of said means for actual prediction is a binary indicator of said error of prediction that comes to said third input of said means for creation of said initial set of said corresponding channel in said first subsystem to be used for determination of said second component of said first permitting vector; and

said second output of said means for actual prediction also goes to each tract of said auxiliary instability evaluating contour.

26. A system as claimed in claim 24 having said auxiliary instability evaluating contour of said single channel of said second subsystem further including:

each said single tract comprising consecutively connected means for assigning an auxiliary resulting set, means for retrospective prediction, and means for analyzing of said binary process retrospective instability.

27. A system as claimed in claim 26 having said single tract of said auxiliary instability evaluating contour of said single channel of said second subsystem wherein:

a first input of said means for assigning an auxiliary resulting set receives information about the composition of said initial set of elements from said second output of said corresponding channel of said first subsystem;

a second input of said means for assigning an auxiliary resulting set is a control input for assignment of said first component of said second permitting vector in case of external to the predicting system control;

a third input of said means for assigning an auxiliary resulting set receives information about the status of said error of prediction indicator to be used for determination of said second component of said second permitting vector;

an output of said means for assigning an auxiliary resulting set contains information about a composition of said auxiliary resulting set, which is tract specific; and

said means for retrospective prediction performs said retrospective predicting procedure acting upon elements of said auxiliary resulting set, so that information about each correct and each incorrect prediction is aggregated by said means for analyzing of said binary process retrospective instability, whose output is a quantitative evaluation of said corresponding binary string instability, which is also an input of said means for evaluating of said binary process current instability.

28. A system as claimed in claim 24 having said auxiliary instability evaluating contour of said single channel of said second subsystem wherein:

said means for evaluating of said binary process current instability select a value among the values represented by the outputs of each of said tracts, where said value is the best indicator of said binary process current instability, and said means for evaluating of said binary process current instability reveal at an output thereof an ordering number of that tract at the output of which the best indicating value was found.

32. A system as claimed in claim 20 being embedded in an another system with essential non linearity further including:

said system having essential nonlinear element;

means for said nonlinear element'"'"'s input measurement;

means for a sequential prediction of a binary elements'"'"' state in a binary process; and

means for performing any actions preventing harmful affection of said system'"'"'s non linearity.

33. A system as claimed in claim 32 wherein:

said nonlinear element'"'"'s input is measured and evaluated by comparison to a predetermined threshold value that is a parameter of said nonlinear element'"'"'s snap condition; and

means for binary process formation of said predicting means generates said vectorial binary process whose each component characterizes certain distance to said threshold;

said predicting means provide a prediction of a forthcoming element of each component of said vectorial binary process; and

means for performing any actions preventing harmful affection of said system'"'"'s non linearity receive information about said forthcoming element'"'"'s state of said vectorial binary process, analyses said forecast, and generates appropriate control actions.

34. A system as claimed in claim 20 being embedded in a multi channel data acquisition system providing a selection of a most informative channel further including:

several sources of data;

means for transmitting, receiving, and processing data over multiple informative channels;

means for generating an indicator, value of which is related to a signal to noise ratio appearing at an input of said receiving means of each said informative channel;

means for a sequential prediction of a binary elements'"'"' state in a vectorial binary process; and

means for selecting an informative channel of said multi channel data acquisition system.

35. A system as claimed in claim 34 wherein:

every said indicator, that is related to signal to noise ratio, is evaluated by at least one predetermined signal to noise ratio level, and a corresponding binary string is generated and stored in said means for a sequential prediction of a binary elements'"'"' state in a vectorial binary process;

a forthcoming element'"'"'s state of each said binary string is predicted; and

the channel whose signal to noise ratio is predicted to be the highest among all said channels, is selected as a preferable channel for the next data transmissions

whereby with each moment of a transaction, said data acquisition system collects data with a lowest disturbance magnitude.

36. A method for predicting a state of a binary element in a binary process, said method being implemented within computer-readable medium of computing means, said method comprising the steps of:

converting a sequence of values of a variable related to a functioning entity into a plurality of binary processes by using a computing means;

forming a binary string of a finite length for each said binary process to obtain a history storage so that each new element of said binary process is an element of said binary string;

generating for each said binary string an initial set of predicting solutions;

evaluating each said binary string'"'"'s transients effect on said binary string'"'"'s predictability if said transients effect is not pre evaluated;

generating for each binary string a resulting set of predicting solutions which is built from said initial set of predicting solutions depending on said evaluated binary string'"'"'s transients effect;

sequentially predicting the state of each element in a temporal group of forthcoming elements in each said binary string through the application of an element vise predicting procedure acting upon each said resulting set of predicting solutions;

updating each said initial set of predicting solutions and reevaluating each said binary string'"'"'s transients effect, if necessary; and

returning to generating each said resulting set of predicting solutions;

whereby said sequential prediction of said binary element'"'"'s state in each stationary, quasi-stationary, and non stationary binary process is generated for said functioning entitle by said computing means.

37. A method as claimed in claim 36, wherein:

said binary process is a mapping of a process of said functioning entity on the space of discrete binary processes providing the-process of said entity functioning internal representation in said computing system.

38. A method as claimed in claim 36, further including the steps of:

generating said initial set of elements only if it is permitted by setting up an updating mode of said initial set to at least one of two binary components of a first permitting vector, said permitting vector having a first component and a second component, said first component indicates that said initial set was predetermined, and said second component indicates whether to begin creation of said initial set based on some schedule.

39. A method as claimed in claim 38, when at least one component of said first permitting vector is put to the updating mode of said initial set, further including the steps of:

expanding each said binary string into uniquely representing said binary string functional structures of a minimal length, which are built from opposed to each other adjacent sequences of identical states of said binary process element, such that a propagation of every said structure in both directions of said binary process domain returns a binary periodical function, which is a prototype of said predicting solution;

determining a statistically robust said functional structures with a maximal period;

generating a first set of predicting solutions containing all possible said structures including said maximal period functional structures, so that each element of said first set possesses the property of orthogonality; and

generating said initial set of predicting solutions via the union of said first set and a second set of two opposite to each other predicting solutions, whose composition is a countable sequence of the identical binary elements of said binary process.

40. A method as claimed in claim 36, wherein:

said element vise predicting procedure is a retrospective predicting procedure if said element vise predicting procedure is defined on an interval of said binary string domain which is antecedent to the latest observed element of said binary string; and

said element vise predicting procedure is an actual predicting procedure if said element vise predicting procedure operates upon the latest observed element of said binary string; and

said predicting procedure is non iterative, is cyclic, and each cycle is composed of two stages, and may contain several predictions.

41. A method as claimed in claim 40, further including the steps of:

beginning each cycle of said predicting procedure at a starting point when an error of prediction occurs;

performing the first stage of each said predicting cycle through generation of a resulting set of predicting solutions by mapping said initial set on said resulting set, executed at the starting point of each said predicting cycle, so that a composition of said resulting set of predicting solutions depends on said evaluation of said current transients effect of said binary string; and

performing the second stage of each said predicting cycle through generation of a functional model of each forthcoming element from a temporal group of said binary string'"'"'s elements belonging to said current predicting cycle, and through a generation of a prediction permitting indicator, so that said element vise prediction is an output of the Boolean conjunction of said model and said permitting indicator.

42. A method as claimed in claim 41, further including the steps of:

determining a component of said estimating predicting efficiency vector function, that has an extreme value; and

selecting said resulting set of predicting solutions based on a composition of said specific set that corresponds to the extreme component of said estimating predicting efficiency vector function, such that each member of said resulting set possesses certain common and specific properties that depend on said entity'"'"'s specificity.

43. A method as claimed in claim 42, wherein:

said resulting set'"'"'s common properties are

first, said resulting set is a mapping of said initial set;

second, the cardinal number of said resulting set is inversely relates to the evaluated said binary string'"'"'s transients effect on predictability; and

third, if being observed at the beginning of said predicting cycle, the binary state of each member of said resulting set of predicting solution is identical to the state of said binary string'"'"'s element having observed at the beginning of the same predicting cycle.

44. A method as claimed in claim 42, wherein:

said resulting set'"'"'s specific property of any non stationary binary process is that said resulting set is a subset of said initial set of predicting solutions.

45. A method as claimed in claim 42, wherein:

said resulting set'"'"'s specific property of some of non stationary binary process is that a resulting set'"'"'s member is such a member of said initial set of predicting solutions that can be observed on the interval next to the beginning of the predicting cycle, and whose period is the maximal period among all observed on said interval members of said initial set;

said resulting set'"'"'s specific property of some of non stationary binary process is that a resulting set'"'"'s member is such a member of said initial set that can be observed on the interval next to the beginning of the predicting cycle, and can be distinguished as a periodic function containing its full period;

said resulting set'"'"'s specific property of some of non stationary binary process is that there is a relationship between at least one known member of said resulting set of predicting solutions and some other members of said resulting set related to said known member, such that for any binary element'"'"'s state denoted by s, it is true that (1s2s)(2s2s), (2s2s)(2s1s), (1s3s) (2s3s) v (3s3s), (2s3s)(1s3,s) v (3s3s), (3s3s)(1s3s) v (2s3s), where the left part of each statement defines known member of said resulting set, and the right part of each statement defines all other members of said resulting set that are related to said known member;

said resulting set'"'"'s specific property of some of non stationary binary process is that there is a relationship between a sequence of two or three identical binary process elements found on the binary process decomposition, which are adjoined to the beginning moment of a predicting cycle and some members of said resulting set related to said sequence of binary process elements, such that for any binary element'"'"'s state denoted by s, it is true that 2s (2s2s) v (2s1s), 3s(3s3s) v (3s2s) v (3s1s), where the left part of each statement defines said sequence of binary process elements, and the right part of each statement defines all members of said resulting set that are related to said sequence of a binary process elements; and

said resulting set'"'"'s specific property of some of non stationary binary process is that the first member of said initial set of predicting solutions is a member of said resulting set, and both members of said second set are the members of said resulting set.

46. A method as claimed in claim 42, wherein:

said resulting set'"'"'s specific property of another non stationary binary process is that a resulting set'"'"'s member is such a member of said initial set that can be observed on the interval next to the beginning of the predicting cycle, and can be distinguished as a periodical function containing its full period;

said resulting set'"'"'s specific property of another non stationary binary process is that there is a relationship between at least one known member of said resulting set of elements and some other members of said resulting set related to said known member, such that for any binary element'"'"'s state denoted by s, it is true that (1s2s)(2s2s), (2s2s)(2s1s), (1s3s)(2s3s) v (3s3s), (2s3s) (1s3s) v (3s3s), (3s3s)(1s3s) v (2s3s), where the left part of each statement defines known member of said resulting set, and the right part of each statement defines all other members of said resulting set that are related to said known member;

said resulting set'"'"'s specific property of another non stationary binary process is that there is a relationship between a sequence of two or three identical binary process elements found on the binary process decomposition, which are adjoined to the beginning moment of a predicting cycle and some members of said resulting set related to said sequence of binary process elements, such that for any binary element'"'"'s state denoted by s, it is true that 2s (2s2s) v (2s1s), 3s (3s3s) v (3s2s) v (3s1s), where the left part of each statement defines said sequence of binary process elements, and the right part of each statement defines all members of said resulting set that are related to said sequence of binary process elements; and

said resulting set'"'"'s specific property of another non stationary binary process is that both members of said second set are the members of said resulting set.

47. A method as claimed in claim 41, wherein:

said functional model of a forthcoming element of said binary process is a nonlinear transformation of a base model;

said nonlinear transformation of the base model includes an inversion of said model'"'"'s output based on a result of a conjunction of two binary components of a third permitting vector;

said first component of the third permitting vector is an indicator of the inversion mode initiation, and said second component of the third permitting vector indicates the beginning of said inversion when an output of a current predicting efficiency estimating function is within a specified distance of a predetermined threshold value, and the second component of said third permitting vector also indicates the end of said inversion if a new error of prediction is observed;

said base model is a union of those members of said resulting set of predicting solutions that are allowed to participate in the base model;

the allowance of participation in said base model includes that each member of said resulting set participates in the current cycle of predictions since the beginning of said cycle until this member produces an incorrect prediction, and this member may again participate in the prediction only in the next predicting cycle;

said prediction permitting indicator is a conjunction of three binary components of a fourth permitting vector; and

said first component of the fourth permitting vector is an indicator of the prediction blocking mode initiation, and said second component of the fourth permitting vector reflects the fact of beginning of said prediction blocking mode when a value of said current predicting efficiency estimating function is within a specified distance of another predetermined threshold value, and the second component of said fourth permitting vector also indicates the end of said blocking if a new error of prediction is observed, and said third component of the fourth permitting vector is a conjunction of all members of said resulting set that participate in said base model.

48. A method as claimed in claim 36, further including the step of:

evaluating said transients effect of said binary string only if it is permitted by setting up an evaluation updating mode to at least one of two binary components of a second permitting vector, whose first component indicates a fact of said evaluation predetermination, and whose second component indicates a fact of beginning of said evaluation based on some schedule.

49. A method as claimed in claim 48, when at least one component of said second permitting vector is set up to the evaluation updating mode further including the step of:

performing said binary string'"'"'s transients effect evaluation through a calculation of an estimating vector function of predicting efficiency involving said retrospective predicting procedure, which is specific for each component of said estimating vector function and is defined on said binary string domain.

50. A method as claimed in claim 49, wherein:

said estimating vector function of predicting efficiency is a vector whose each component is an aggregation function of a sequence of outputs of said retrospective predicting procedure, which acts upon a specific for this component pre selected set of predicting solutions, which is a mapping of said initial set of predicting solutions;

one of said pre selected sets can be said initial set itself; and

any component of said estimating vector function of predicting efficiency can be predetermined controlling the composition of said resulting set of predicting solutions.

51. A computer program product for predicting of a state of a binary element in a binary process having a functioning entity or an improvement of said functioning entity, said product executable in a computer readable environment, said product having instructions for:

converting an observable sequence of values of a variable related to said functioning entity into a plurality of binary processes providing for said sequence of values internal representation in said product;

forming a binary string of a finite length for each said binary process history storage so that each new element of said binary process is an element of said binary string;

generating for each said binary string an initial set of predicting solutions;

evaluating each said binary string'"'"'s transients effect on said binary string'"'"'s predictability if said transients effect is not pre evaluated;

generating for each binary string a resulting set of predicting solutions which is built from said initial set of predicting solutions depending on said evaluated binary string'"'"'s transients effect;

sequentially predicting the state of each element in a temporal group of forthcoming elements in each said binary string through the application of an element vise predicting procedure acting upon each said resulting set of predicting solutions;

updating each said initial set of predicting solutions and reevaluating each said binary string'"'"'s transients effect; and

returning to generating step for each said resulting set of predicting solutions if additional predictions are needed;

whereby said sequential prediction of said binary element'"'"'s state in each stationary, quasi-stationary, and non stationary binary process providing said internal representation of said sequence of values for said functioning entity or for said improvement of said functioning entity by said product is generated.

52. A product as claimed in claim 51, wherein:

said observable sequence of values of a variable denoted u(t) and related to said entity'"'"'s functioning is converted into said plurality of binary processes by an application of a quantizing procedure denoted B(λ

,u⁰), whose one possible realization for continuous u(t) is described by formulas;

53. A product as claimed in claim 52, wherein:

formula (2) defines a vectorial time sampling procedure; and

formulas (3), (4) define an amplitude quantification procedure.

54. A product as claimed in claim 52, further including:

performing a two stage quantification procedure according to the formula (1) such that said continuous variable u(t) is mapped onto the space of discrete processes, and then said variable experiences said amplitude quantification;

whereby, said continuous variable u(t) becomes converted into said plurality of binary processes x(τ

) suitable for processing by said product.