WO2009129758A1 - Apparatus for assessing a control value - Google Patents

Apparatus for assessing a control value Download PDF

Info

Publication number
WO2009129758A1
WO2009129758A1 PCT/CZ2009/000054 CZ2009000054W WO2009129758A1 WO 2009129758 A1 WO2009129758 A1 WO 2009129758A1 CZ 2009000054 W CZ2009000054 W CZ 2009000054W WO 2009129758 A1 WO2009129758 A1 WO 2009129758A1
Authority
WO
WIPO (PCT)
Prior art keywords
block
assessing
cells
main
rows
Prior art date
Application number
PCT/CZ2009/000054
Other languages
French (fr)
Inventor
Marcel Jirina
Original Assignee
Institute Of Computer Science Of The As Cr, V.V.I.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Institute Of Computer Science Of The As Cr, V.V.I. filed Critical Institute Of Computer Science Of The As Cr, V.V.I.
Publication of WO2009129758A1 publication Critical patent/WO2009129758A1/en

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N99/00—Subject matter not provided for in other groups of this subclass

Definitions

  • This invention relates to apparatus for assessing a control value on the basis of already known facts provided by a set of data stored in a memory.
  • the apparatus for data processing solves the problem of the construction of an apparatus for assessing a control value.
  • This apparatus creates a representation of a control value that serves for controlling or operating other devices, or for displaying.
  • Processed data represents the dependence of a control value on a large number of other values (function of many variables) that are arranged in frames consisting of rows that comprise memory cells that represent the values of individual quantities.
  • the apparatus creates a representation of a control value for a query row of values of many quantities.
  • the apparatus for assessing a control value is made up of a memory, a transformation block for assessing a control value and a means for transferring or storing a control value.
  • the memory is arranged so that it consists of at least one frame, one query row including a block of cells with at least one memory cell, and a cell for control value representation.
  • the frame consists of at least two rows each of which comprises a block of cells including at least one memory cell, and a cell for representation of a response, a cell for representation of the distance, and a cell for representation of a reciprocal value of succession.
  • the transformation block for assessing a control value consists of a block for assessing a value of the distance of rows of blocks of cells from a block of, cells of a query row, a block for assessing reciprocal values of an ascending succession of rows of blocks of cells, a block for assessing the main numerator, a block for assessing the main denominator, and a block for assessing the ratio of the main numerator and the main denominator.
  • the block for assessing a value of a distance of rows of blocks of cells from a block of cells of a query row is connected to rows of blocks of cells, cells for representation of the distance of rows, and a block of cells of a query row.
  • the block for assessing reciprocal values of an ascending succession of rows of blocks of cells is connected to cells for representation of the distance, and to cells for representation of reciprocal value of succession of rows.
  • the block for assessing the main numerator consists of blocks for multiplication and a block for addition; the inputs of each of the blocks for multiplication being connected to the cell for representation of a response, and to cells for representation of a reciprocal value of succession of one row; outputs of blocks for multiplication are connected to the inputs of the block for addition the output of which is connected to a dividend's input of the block for assessing the ratio of the main numerator and the main denominator.
  • the block for assessing the main denominator consists of a block for addition the inputs of which are connected to all cells for representation of a reciprocal value of succession of all rows of all frames and the output of which is connected to a divisor's input of the block for assessing the ratio of the main numerator and the main denominator.
  • the output of the block for assessing the ratio of the main numerator and the main denominator is connected to the cell for representation of a control value.
  • i stand for succession, i.e. sequence number of a data sample or a row.
  • the succession is determined by the distance of a data sample, i.e. a block of memory cells of a row from a query sample, i.e. from the block of memory cells of a query row.
  • M stands for the total number of all samples, i.e. number of rows.
  • Rj stands for a response of a sample, represented by the cell for representation of a response, that was assigned succession value /.
  • ⁇ is a reciprocal value of succession. In this case the number of classes, i.e. number of frames, is insignificant.
  • the apparatus can also perform another transformation.
  • the block for assessing the main numerator consists of a block for addition the inputs of (which are connected to cells for representation of reciprocal value of succession of all rows of one frame and the output of which is connected to a dividend's input of the block for assessing the ratio of the main numerator and the main denominator.
  • the block for assessing the main denominator consists of a block for addition the inputs of which are connected to all cells for representation of reciprocal value of all rows of all frames and the output of which is connected to a divisor's input of the block for assessing the ratio of the main numerator and the main denominator.
  • Control value can then be interpreted as the likelihood that a query sample, i.e. query row, belongs to the same frame as the frame that the block for assessing the main numerator is related to.
  • i stands for succession, i.e. sequence number of a data sample, i.e. a row.
  • the succession is determined by the distance of a data sample, i.e. a block of cells of rows from a query sample, i.e. from the block of cells of a query row.
  • M stands for the total number of all samples, i.e. number of all rows regardless of the frame they are located in.
  • - is a reciprocal value of succession.
  • the apparatus can perform another different transformation.
  • the block for assessing the main numerator consists of a block for addition, a block for counting the number of rows, and a block for division; the inputs of the block for addition being connected to the cells for representation of a reciprocal value of succession of all rows in one frame and the inputs of the block for counting the number of rows being connected to cells for representation of a response of all rows of the same frame; the output of the block for addition is connected to a dividend's input of the block for division, the output of the block for counting the number of rows is connected to a divisor's input of the block for division the output of which is connected to a dividend's input of the block for assessing the ratio of the main numerator and the main denominator.
  • the block for assessing the main denominator consists of blocks for addition, blocks for counting the number of rows, a block for division and a block for secondary addition the output of which is connected to a divisor's input of the block for assessing the ratio of the main numerator and the main denominator and its inputs are connected to the outputs of all blocks for division; each block for division has a dividend's input connected to one block for addition the inputs of which are connected to cells of representation of reciprocal value of succession of rows in one frame, a divisor's input of the block for division is connected to output of the block for counting the number of rows the inputs of which are connected to cells for representation of a response of rows of the same frame.
  • Control value can then be interpreted as the likelihood that a query sample, i.e. query row, comes from the same class, i.e. belongs to the same frame as the class, i.e. frame, that the block for assessing the main numerator is related to.
  • N j stands for a number of samples, i.e. number of all rows of/ ,-th c technically!laturns suppliess, ; i.e réelle./ ,-th . frame
  • Figure 1 shows the overall arrangement of the apparatus according to the invention.
  • Fig. 2 shows the diagram of memory structure.
  • Fig. 3 shows the structure of transformation and links (connections) of parts of the memory and the block for assessing values of the distance, the block for assessing reciprocal values of ascending succession and the block for assessing the ratio of the main numerator and the main denominator.
  • Fig. 4 shows the diagram that illustrates connection of parts of the memory and the block for assessing the main numerator and the block for assessing the main denominator and their inner structure in cases where the number of frames is insignificant.
  • FIG. 5 shows the connection diagram of parts of the memory and the block for assessing the main numerator and the block for assessing the main denominator and their inner structure in cases where there are at least two frames of the same size.
  • Fig. 6 shows the connection diagram of parts of the memory and the block for assessing the main numerator and the block for assessing the main denominator and their inner structure in cases where there are at least two frames of different size.
  • the apparatus consists of a memory (1), a transformation block (2) for assessing a control value, and a means (3) intended to transfer or store estimated values of digitally represented dependences.
  • the memory (1) is arranged so that it is comprised of at least one frame (10), one query row (12) composed of a block (19) of cells with at least one memory cell (16), and a cell
  • Frame (10) comprises at least two rows (11) each of which is composed of a block (14) of cells (16), a cell (15) for representation of a response, a cell
  • the transformation block (2) for assessing a control value consists of a block (21) for assessing the distance of the blocks (14) of cells of rows (11) from the block (19) of cells of the query row (12), a block (22) for assessing reciprocal values of ascending succession of the blocks (14) of cells of the rows (11), a block (23) for assessing the main numerator, a block (24) for assessing the main denominator, and a block (25) for assessing the ratio of the main numerator and the main denominator; at the same time the block (21) for assessing the distance of the blocks (14) of cells of rows (11) from the block (19) of cells of the query row (12) is connected to the blocks (14) of cells of the rows (11), to cells (18) for representation of the distance of the rows (11), and to the block (19) of cells of the query row (12).
  • the structure of transformation and linkage (connection) of parts of the memory (1) and the block (21) for assessing values of the distance, the block (22) for assessing reciprocal values of ascending succession and the block (25) for assessing the ratio of the main numerator and the main denominator are shown in Fig. 3.
  • the block (21) for assessing values of the distance of the blocks (14) of cells of the rows (11) from the block (19) of cells of the query row (12) is connected to the Mocks (14) of cells of the rows (11), and to cells (18) for representation of the distance criz of the rows (11), and to the block (19) of cells of the query row (12).
  • Block (22) for assessing ' the reciprocal values of ascending succession of the rows' (11) block (14) of cells is connected to cells (18) for representation of the distance, and to cells (13) for representation of reciprocal values of succession of the rows (11).
  • Block (25) for assessing the ratio of the main numerator and the main denominator is connected through its dividend's input to the block (23) for assessing the main numerator and through its divisor's input to the block (24) for assessing the main denominator, and through its output to the cell (17) for representation of a control value.
  • the block (23) for assessing the main numerator consists of blocks (231) for multiplication and a block (232) for addition; the inputs of each of the blocks (231) for multiplication are connected to cells (15) of representation of a response and to cells (13) for representation of a reciprocal value of succession of one row (11), and the outputs of the blocks (231) for multiplication are connected to inputs of the block (232) for addition the output of which is connected to a dividend input of block (25) for assessing the ratio of the main numerator and the main denominator; block (24) for assessing the main denominator consists of a block (241) for addition the inputs of which are connected to all cells (13) for representation of a reciprocal value of succession of all rows (11) of all frames (10
  • block (23) for assessing the main numerator consists of a block (234) for addition the inputs of which are connected to cells (13) for representation of a reciprocal value of succession of all rows (11) of one frame (10), and the output of which is connected to a dividend input of block (25) for assessing the ratio of the main numerator and the main denominator.
  • Block (24) for assessing the main denominator consists of a block (245) for addition the inputs of which are connected to all cells (13) for representation of a reciprocal value of succession of all rows (11) of all frames (10) and the output of which is connected to a divisor input of block (25) for assessing the ratio of the main numerator and the main denominator.
  • Fig. 6 shows the links (connections) of parts of the memory and block (23) for assessing the main numerator and block (24) for assessing the main denominator and their inner structure in cases where there are at least two frames (10) of different size.
  • block (23) for assessing the main numerator consists of a block (235),for addition, a block (236) for counting the number of rows (11), and a block (233) for division; inputs of the block (235) for addition are connected to cells (13) for representation of a reciprocal value of succession of all rows (11) of one frame (10) and inputs of the block (236) for counting the number of rows are connected to cells (15) for representation of a response of all rows (11) of the same frame (10); output of the block (235) for addition is connected to a dividend input of block (233) for division, output of the block (236) for counting of the number of rows is connected to a divisor input of block (233) for division the output of which is connected to a dividend input of block (25) for assessing the ratio of the main numerator and the main denominator.
  • Block (24) for assessing the main denominator consists of blocks (241) for addition, blocks (242) for counting the number of rows, blocks (243) for division, and a block (244) for secondary addition the output of which is connected to a divisor input of block (25) for assessing the ratio of the main numerator and the main denominator, and its inputs are connected to outputs of all blocks (243) for division, each block (243) has a dividend input connected to one block (241) for addition the inputs of which are connected to cells (13) for representation of a reciprocal value of succession of rows (11) of one frame (10), a dividend's input of the block (243) for division is connected to the output of the block (242) for counting the number of rows the inputs of which are connected to cells (15) for representation of a response of the rows (11) of the same frame (10).
  • the apparatus in its present form can be utilized as a component in classification systems, especially personalized information systems such as, for example, warning devices watching over the driver's state of activity and distinguishing his/her ability to drive a vehicle and detecting a state when his/her ability to operate a vehicle is reduced.
  • personalized information systems such as, for example, warning devices watching over the driver's state of activity and distinguishing his/her ability to drive a vehicle and detecting a state when his/her ability to operate a vehicle is reduced.
  • Another example of possible application is the complex evaluation of the operating conditions of highly sophisticated machines, such as combustion engines, aimed to differentiate between normal operation and faulty operation requiring service.
  • the apparatus is suitable for processing digitally represented dependences on large amounts of data (dependences of high dimensions) and for dependences that other devices and approaches do not provide satisfactory results for.
  • the apparatus in its present form assesses a control value that can be further used to control, operate and supervise other machines and devices, for example to redirect unsolicited mail or display, for example, a warning signal to a tired driver.

Abstract

The apparatus for assessing a control value consists of memory (1), transformation block (2) for assessing a control value, and a means (3) intended to transfer or store a control value. Memory (1) comprises at least one frame (10), one query row (12) composed of a block (19) of cells with at least one memory cell (16), and a cell (17) for representation of a control value. Every frame (10) comprises at least two rows (11) each of which is composed of a cell (15) for representation of a response, a cell (18) for representation of the distance, and a cell (13) for representation of a reciprocal value of succession and a block (14) of cells that includes at least one memory cell (16). Transformation block (2) for assessing a control value consists of a block (21) for assessing the distance of the blocks (14) of cells rows (11) from the block (19) of cells of the query row (12), a block (22) for assessing reciprocal values of ascending succession of the blocks (14) of cells of the rows (11), a block (23) for assessing the main numerator, a block (24) for assessing the main denominator, and a block (25) for assessing the ratio of the main numerator and the main denominator. The block (21) for assessing the distance of the blocks (14) of cells of the rows (11) from the block (19) of cells of the query row (12) is connected to the blocks (14) of cells of the rows (11), to cells (18) for representation of the distance of the rows (11), and to the block (19) of cells of the query row (12). Block (22) for assessing reciprocal values of ascending succession of the block (14) of cells of the rows (11) is connected to cells (18) for representation of the distance, and to cells (13) for representation of a reciprocal value of succession of the rows (11). Block (25) for assessing the ratio of the main numerator and the main denominator is connected through its dividend's input to the block (23) for assessing the main numerator, and through its divisor's input to the block (24) for assessing the main denominator, and through its output to the cell (17) for representation of a control value.

Description

Apparatus for assessing a control value
Technical field
This invention relates to apparatus for assessing a control value on the basis of already known facts provided by a set of data stored in a memory.
Background art
There are many apparatuses that are used to assess a control value for a preset query point. They employ especially various approximation techniques such as, for example, polynomials, spline polynomials, neural networks of various types. In cases when affiliation to a specific class (classification) is substantial, other specialized apparatuses are known that make use of techniques such as k-nearest neighbour, Bayes method, kernel method. Some of these methods suffer from a so called "curse of dimensionality", i.e. demands of operations grow exponentially along with a task dimension, i.e. with a number of values that form a query point. These are the two reasons why new devices are constantly being searched for and consequently also new methods for solving this type of task. At the same time it is not expected that a universally "best" apparatus or method will be found, but rather an approach that will be optimal for a specific problem at hand.
Disclosure of the invention
The apparatus for data processing according to the invention solves the problem of the construction of an apparatus for assessing a control value. This apparatus creates a representation of a control value that serves for controlling or operating other devices, or for displaying. Processed data represents the dependence of a control value on a large number of other values (function of many variables) that are arranged in frames consisting of rows that comprise memory cells that represent the values of individual quantities. The apparatus creates a representation of a control value for a query row of values of many quantities.
The apparatus for assessing a control value is made up of a memory, a transformation block for assessing a control value and a means for transferring or storing a control value. The memory is arranged so that it consists of at least one frame, one query row including a block of cells with at least one memory cell, and a cell for control value representation. The frame consists of at least two rows each of which comprises a block of cells including at least one memory cell, and a cell for representation of a response, a cell for representation of the distance, and a cell for representation of a reciprocal value of succession. The transformation block for assessing a control value consists of a block for assessing a value of the distance of rows of blocks of cells from a block of, cells of a query row, a block for assessing reciprocal values of an ascending succession of rows of blocks of cells, a block for assessing the main numerator, a block for assessing the main denominator, and a block for assessing the ratio of the main numerator and the main denominator. The block for assessing a value of a distance of rows of blocks of cells from a block of cells of a query row is connected to rows of blocks of cells, cells for representation of the distance of rows, and a block of cells of a query row. The block for assessing reciprocal values of an ascending succession of rows of blocks of cells is connected to cells for representation of the distance, and to cells for representation of reciprocal value of succession of rows. The block for assessing the main numerator consists of blocks for multiplication and a block for addition; the inputs of each of the blocks for multiplication being connected to the cell for representation of a response, and to cells for representation of a reciprocal value of succession of one row; outputs of blocks for multiplication are connected to the inputs of the block for addition the output of which is connected to a dividend's input of the block for assessing the ratio of the main numerator and the main denominator. The block for assessing the main denominator consists of a block for addition the inputs of which are connected to all cells for representation of a reciprocal value of succession of all rows of all frames and the output of which is connected to a divisor's input of the block for assessing the ratio of the main numerator and the main denominator. The output of the block for assessing the ratio of the main numerator and the main denominator is connected to the cell for representation of a control value.
The transformation can be expressed by the following formula:
M
A = ;=1
(=1 where i stand for succession, i.e. sequence number of a data sample or a row. The succession is determined by the distance of a data sample, i.e. a block of memory cells of a row from a query sample, i.e. from the block of memory cells of a query row. M stands for the total number of all samples, i.e. number of rows. Rj stands for a response of a sample, represented by the cell for representation of a response, that was assigned succession value /. At the same time τ is a reciprocal value of succession. In this case the number of classes, i.e. number of frames, is insignificant. The apparatus can also perform another transformation. In which case its construction is modified so that the block for assessing the main numerator consists of a block for addition the inputs of( which are connected to cells for representation of reciprocal value of succession of all rows of one frame and the output of which is connected to a dividend's input of the block for assessing the ratio of the main numerator and the main denominator. Simultaneously the block for assessing the main denominator consists of a block for addition the inputs of which are connected to all cells for representation of reciprocal value of all rows of all frames and the output of which is connected to a divisor's input of the block for assessing the ratio of the main numerator and the main denominator.
This transformation is suitable when there are at least two frames and all individual frames have the same number of rows. Control value can then be interpreted as the likelihood that a query sample, i.e. query row, belongs to the same frame as the frame that the block for assessing the main numerator is related to.
The transformation can be expressed by the following formula:
Figure imgf000005_0001
where i stands for succession, i.e. sequence number of a data sample, i.e. a row. The succession is determined by the distance of a data sample, i.e. a block of cells of rows from a query sample, i.e. from the block of cells of a query row. M stands for the total number of all samples, i.e. number of all rows regardless of the frame they are located in. At the same time - is a reciprocal value of succession. Symbol i = 1(Q located under the symbol of a sum in the numerator of the present formula means that addition is performed across the rows of one frame, namely the frame that the block for assessing the main numerator is related to.
The apparatus can perform another different transformation. In which case its construction is modified so that the block for assessing the main numerator consists of a block for addition, a block for counting the number of rows, and a block for division; the inputs of the block for addition being connected to the cells for representation of a reciprocal value of succession of all rows in one frame and the inputs of the block for counting the number of rows being connected to cells for representation of a response of all rows of the same frame; the output of the block for addition is connected to a dividend's input of the block for division, the output of the block for counting the number of rows is connected to a divisor's input of the block for division the output of which is connected to a dividend's input of the block for assessing the ratio of the main numerator and the main denominator. Simultaneously the block for assessing the main denominator consists of blocks for addition, blocks for counting the number of rows, a block for division and a block for secondary addition the output of which is connected to a divisor's input of the block for assessing the ratio of the main numerator and the main denominator and its inputs are connected to the outputs of all blocks for division; each block for division has a dividend's input connected to one block for addition the inputs of which are connected to cells of representation of reciprocal value of succession of rows in one frame, a divisor's input of the block for division is connected to output of the block for counting the number of rows the inputs of which are connected to cells for representation of a response of rows of the same frame.
This transformation is suitable when there are at least two classes, i.e. frames, and individual classes, i.e. frames, don't all have the same number of rows. Control value can then be interpreted as the likelihood that a query sample, i.e. query row, comes from the same class, i.e. belongs to the same frame as the class, i.e. frame, that the block for assessing the main numerator is related to.
The transformation can also be expressed by the following formula:
Figure imgf000006_0001
where Nj stands for a number of samples, i.e. number of all rows of/ ,-th c „!la„„s„s, ; i.e „./ ,-th . frame and Nc stands for a number of samples, i.e. number of rows (11) of a certain class C, i.e. a certain frame, and a sign i= 1(C), or
Figure imgf000006_0002
means that summation is performed only for those indexes i for which the ith sample, i.e. zth row, belongs to class C, or 7.
Figure 1 shows the overall arrangement of the apparatus according to the invention. Fig. 2 shows the diagram of memory structure. Fig. 3 shows the structure of transformation and links (connections) of parts of the memory and the block for assessing values of the distance, the block for assessing reciprocal values of ascending succession and the block for assessing the ratio of the main numerator and the main denominator. Fig. 4 shows the diagram that illustrates connection of parts of the memory and the block for assessing the main numerator and the block for assessing the main denominator and their inner structure in cases where the number of frames is insignificant. Fig. 5 shows the connection diagram of parts of the memory and the block for assessing the main numerator and the block for assessing the main denominator and their inner structure in cases where there are at least two frames of the same size. Fig. 6 shows the connection diagram of parts of the memory and the block for assessing the main numerator and the block for assessing the main denominator and their inner structure in cases where there are at least two frames of different size.
Example
The apparatus consists of a memory (1), a transformation block (2) for assessing a control value, and a means (3) intended to transfer or store estimated values of digitally represented dependences. The memory (1) is arranged so that it is comprised of at least one frame (10), one query row (12) composed of a block (19) of cells with at least one memory cell (16), and a cell
(17) for representation of a control value. Frame (10) comprises at least two rows (11) each of which is composed of a block (14) of cells (16), a cell (15) for representation of a response, a cell
(18) for representation of the distance, and a cell (13) for representation of a reciprocal value of succession; at the same time the block (14) of cells (16) includes at least one memory cell (16). The transformation block (2) for assessing a control value consists of a block (21) for assessing the distance of the blocks (14) of cells of rows (11) from the block (19) of cells of the query row (12), a block (22) for assessing reciprocal values of ascending succession of the blocks (14) of cells of the rows (11), a block (23) for assessing the main numerator, a block (24) for assessing the main denominator, and a block (25) for assessing the ratio of the main numerator and the main denominator; at the same time the block (21) for assessing the distance of the blocks (14) of cells of rows (11) from the block (19) of cells of the query row (12) is connected to the blocks (14) of cells of the rows (11), to cells (18) for representation of the distance of the rows (11), and to the block (19) of cells of the query row (12); block (22) for assessing reciprocal values of ascending succession of the blocks (14) of cells of the rows (11) is connected to cells (18) for representation of the distance, and to cells (13) for representation of a reciprocal value of succession of the rows (11); and block (25) for assessing the ratio of the main numerator and the main denominator is connected through its dividend's input to the block (23) for assessing the main numerator, and through its divisor's input to the block (24) for assessing the main denominator, and through its output to the cell (17) for representation of a control value. The structure of transformation and linkage (connection) of parts of the memory (1) and the block (21) for assessing values of the distance, the block (22) for assessing reciprocal values of ascending succession and the block (25) for assessing the ratio of the main numerator and the main denominator are shown in Fig. 3. The block (21) for assessing values of the distance of the blocks (14) of cells of the rows (11) from the block (19) of cells of the query row (12) is connected to the Mocks (14) of cells of the rows (11), and to cells (18) for representation of the distance „ of the rows (11), and to the block (19) of cells of the query row (12). .Block (22) for assessing' the reciprocal values of ascending succession of the rows' (11) block (14) of cells is connected to cells (18) for representation of the distance, and to cells (13) for representation of reciprocal values of succession of the rows (11). Block (25) for assessing the ratio of the main numerator and the main denominator is connected through its dividend's input to the block (23) for assessing the main numerator and through its divisor's input to the block (24) for assessing the main denominator, and through its output to the cell (17) for representation of a control value. Fig. 4 shows the links (connections) of parts of the memory (1) and the block (23) for assessing the main numerator and the block (24) for assessing the main denominator and their inner structure in cases where the number of frames (10) is insignificant. At the same time the block (23) for assessing the main numerator consists of blocks (231) for multiplication and a block (232) for addition; the inputs of each of the blocks (231) for multiplication are connected to cells (15) of representation of a response and to cells (13) for representation of a reciprocal value of succession of one row (11), and the outputs of the blocks (231) for multiplication are connected to inputs of the block (232) for addition the output of which is connected to a dividend input of block (25) for assessing the ratio of the main numerator and the main denominator; block (24) for assessing the main denominator consists of a block (241) for addition the inputs of which are connected to all cells (13) for representation of a reciprocal value of succession of all rows (11) of all frames (10), and its output is connected to a divisor input of block (25) for assessing the ratio of the main numerator and the main denominator. The links (connections) of parts of the memory and block (23) for assessing the main numerator and block (24) for assessing the main denominator and their inner structures in cases when there are at least two frames (10) of the same size are shown in Fig. 5. At the same time block (23) for assessing the main numerator consists of a block (234) for addition the inputs of which are connected to cells (13) for representation of a reciprocal value of succession of all rows (11) of one frame (10), and the output of which is connected to a dividend input of block (25) for assessing the ratio of the main numerator and the main denominator. Block (24) for assessing the main denominator consists of a block (245) for addition the inputs of which are connected to all cells (13) for representation of a reciprocal value of succession of all rows (11) of all frames (10) and the output of which is connected to a divisor input of block (25) for assessing the ratio of the main numerator and the main denominator. Fig. 6 shows the links (connections) of parts of the memory and block (23) for assessing the main numerator and block (24) for assessing the main denominator and their inner structure in cases where there are at least two frames (10) of different size. At the same time block (23) for assessing the main numerator consists of a block (235),for addition, a block (236) for counting the number of rows (11), and a block (233) for division; inputs of the block (235) for addition are connected to cells (13) for representation of a reciprocal value of succession of all rows (11) of one frame (10) and inputs of the block (236) for counting the number of rows are connected to cells (15) for representation of a response of all rows (11) of the same frame (10); output of the block (235) for addition is connected to a dividend input of block (233) for division, output of the block (236) for counting of the number of rows is connected to a divisor input of block (233) for division the output of which is connected to a dividend input of block (25) for assessing the ratio of the main numerator and the main denominator. Block (24) for assessing the main denominator consists of blocks (241) for addition, blocks (242) for counting the number of rows, blocks (243) for division, and a block (244) for secondary addition the output of which is connected to a divisor input of block (25) for assessing the ratio of the main numerator and the main denominator, and its inputs are connected to outputs of all blocks (243) for division, each block (243) has a dividend input connected to one block (241) for addition the inputs of which are connected to cells (13) for representation of a reciprocal value of succession of rows (11) of one frame (10), a dividend's input of the block (243) for division is connected to the output of the block (242) for counting the number of rows the inputs of which are connected to cells (15) for representation of a response of the rows (11) of the same frame (10).
Industrial applicability
The apparatus in its present form can be utilized as a component in classification systems, especially personalized information systems such as, for example, warning devices watching over the driver's state of activity and distinguishing his/her ability to drive a vehicle and detecting a state when his/her ability to operate a vehicle is reduced. Another example of possible application is the complex evaluation of the operating conditions of highly sophisticated machines, such as combustion engines, aimed to differentiate between normal operation and faulty operation requiring service.
The apparatus is suitable for processing digitally represented dependences on large amounts of data (dependences of high dimensions) and for dependences that other devices and approaches do not provide satisfactory results for.
The apparatus in its present form assesses a control value that can be further used to control, operate and supervise other machines and devices, for example to redirect unsolicited mail or display, for example, a warning signal to a tired driver.

Claims

1. The apparatus for assessing a control value consisting of memory, transformation block for assessing a control value and a means intended to transfer or store a control value, characterized in that memory (1) comprises at least one frame (10), one query row (12) composed of a block (19) of cells with at least one memory cell (16), and a cell (17) for representation of a control value; at the same time every frame (10) comprises at least two rows (11) each of which is composed of a cell (15) for representation of a response, a cell (18) for representation of the distance , and a cell (13) for representation of a reciprocal value of succession, and a block (14) of cells including at least one memory cell (16); a follow-up transformation block (2) for assessing a control value consists of a block (21) for assessing the distance of the blocks (14) of cells of the rows (11) from the block (19) of cells of the query row (12), a block (22) for assessing reciprocal values of ascending succession of the blocks (14) of cells of the rows (11), a block (23) for assessing the main numerator, a block (24) for assessing the main denominator, and a block (25) for assessing the ratio of the main numerator and the main denominator; at the same time the block (21) for assessing the distance of the blocks (14) of cells of the rows (11) from the block (19) of cells of the query row (12) is connected to the blocks (14) of cells of the rows (11), to cells (18) for representation of the distance of the rows (11), and to the block (19) of cells of the query row (12); block (22) for assessing reciprocal values of ascending succession of the blocks (14) of cells of the rows (11) is connected to cells (18) for representation of the distance, and to cells (13) for representation of a reciprocal value of succession the rows (11); and block (25) for assessing the ratio of the main numerator and the main denominator is connected through its dividend's input to the block (23) for assessing the main numerator, and through its divisor's input to the block (24) for assessing the main denominator, and through its output to the cell (17) for representation of a control value.
2. The apparatus according to claim 1, characterized in that block (23) for assessing the main numerator consists of blocks (231) for multiplication and block (232) for addition; inputs of each of the blocks (231) for multiplication are connected to cells (15) of representation of a response and to cells (13) for representation of a reciprocal value of succession of one row (11), and outputs of the blocks (231) for multiplication are connected to inputs of the block (232) for addition the output of which is connected to a dividend's
' input of the block (25) for assessing the ratio of the main numerator and the main denominator; and atthe same time block (24) for assessing the main denominator consists of the block (241) for addition the output of which are connected to all cells (13) for representation of a reciprocal value of succession of all rows (11) of all frames (10), and its output is connected to a divisor's input of the block (25) for assessing the ratio of the main numerator and the main denominator.
3. The apparatus according to claim 1, characterized in that block (23) for assessing the main numerator consists of block (234) for addition the inputs of which are connected to cells (13) for representation of a reciprocal value of succession of all rows (11) of one frame (10), and the output of which is connected to a dividend's input of the block (25) for assessing the ratio of the main numerator and the main denominator, and at the same time block (24) for assessing the main denominator consists of the block (245) for addition the inputs of which are connected to all cells (13) for representation of a reciprocal value of succession of all rows (11) of all frames (10) and the output of which is connected to a divisor's input of the block (25) for assessing the ratio of the main numerator and the main denominator.
4. The apparatus according to claim 1, characterized in that the block (23) for assessing the main numerator consists of the block (235) for addition, the block (236) for counting the number of rows (11), and the block (233) for division; inputs of the block (235) for addition are connected to cells (13) for representation of a reciprocal value of succession of all rows (11) of one frame (10) and inputs of the block (236) for counting the number of rows are connected to cells (15) for representation of a response of all rows (11) of the same frame (10); at the same time output of the block (235) for addition is connected to a dividend's input of the block (233) for division, output of the block (236) for counting of the number of rows is connected to a divisor's input of the block (233) for division the output of which is connected to a dividend's input of the block (25) for assessing the ratio of the main numerator and the main denominator; and at the same time block (24) for assessing the main denominator consists of blocks (241) for addition, blocks (242) for counting the number of rows, blocks (243) for division, and the block (244) for the secondary addition the output of which is connected to a divisor's input of the block (25) for assessing the ratio of the main numerator and the main denominator, and its inputs are connected to the outputs of all blocks (243) for division, each block (243) for division has a dividend's input connected to one block (241) for addition the inputs of which are connected to cells (13) for representation of a reciprocal value of succession of rows (11) of one frame (10), a dividend's input of the block (243) for division is connected to the output of the block (242) for counting i the. i number of > rows the inputs of which are connected to cells (15) for representation of , a response of the rows (11) of the same frame (10).
PCT/CZ2009/000054 2008-04-22 2009-04-17 Apparatus for assessing a control value WO2009129758A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CZ2008-245A CZ305147B6 (en) 2008-04-22 2008-04-22 Device to determine control variable value
CZPV2008-245 2008-04-22

Publications (1)

Publication Number Publication Date
WO2009129758A1 true WO2009129758A1 (en) 2009-10-29

Family

ID=40887856

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CZ2009/000054 WO2009129758A1 (en) 2008-04-22 2009-04-17 Apparatus for assessing a control value

Country Status (2)

Country Link
CZ (1) CZ305147B6 (en)
WO (1) WO2009129758A1 (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4926385A (en) * 1987-08-05 1990-05-15 Mitsubishi Denki Kabushiki Kaisha Semiconductor memory device with cache memory addressable by block within each column
US5214610A (en) * 1989-09-22 1993-05-25 Texas Instruments Incorporated Memory with selective address transition detection for cache operation
WO1998014950A1 (en) * 1996-09-30 1998-04-09 Advanced Micro Devices, Inc. Memory block select using multiple word lines to address a single memory cell row
US6105114A (en) * 1997-01-21 2000-08-15 Sharp Kabushiki Kaisha Two-dimensional array transposition circuit reading two-dimensional array in an order different from that for writing

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4926385A (en) * 1987-08-05 1990-05-15 Mitsubishi Denki Kabushiki Kaisha Semiconductor memory device with cache memory addressable by block within each column
US5214610A (en) * 1989-09-22 1993-05-25 Texas Instruments Incorporated Memory with selective address transition detection for cache operation
WO1998014950A1 (en) * 1996-09-30 1998-04-09 Advanced Micro Devices, Inc. Memory block select using multiple word lines to address a single memory cell row
US6105114A (en) * 1997-01-21 2000-08-15 Sharp Kabushiki Kaisha Two-dimensional array transposition circuit reading two-dimensional array in an order different from that for writing

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
MARCEL JIRINA AND MARCEL JIRINA, JR.: "Classifier Based on Inverted Indexes of Neighbors II. Theory and Appendix", TECHNICAL REPORT NO. V-1041, November 2008 (2008-11-01), Institute of Computer Science, Academy of Sciences of the Czech Republic, pages 1 - 26, XP002539173, Retrieved from the Internet <URL:http://www.cs.cas.cz/cgi-bin/extftp?ftp://ftp.cs.cas.cz/pub/reports/v1041-08.pdf> [retrieved on 20090729] *
MARCEL JIRINA AND MARCEL JIRINA, JR.: "Classifier Based on Inverted Indexes of Neighbors", TECHNICAL REPORT NO. V-1034, November 2008 (2008-11-01), Institute of Computer Science, Academy of Sciences of the Czech Republic, pages 1 - 11, XP002539312, Retrieved from the Internet <URL:http://www.cs.cas.cz/cgi-bin/extftp?ftp://ftp.cs.cas.cz/pub/reports/v1034-08.pdf> [retrieved on 20090729] *
MARCEL JIRINA ET AL: "Multivariate Data Classification Using the Distribution Mapping Exponent", COMPUTATIONAL CYBERNETICS, 2006 IEEE INTERNATIONAL CONFERENCE ON, IEEE, PI, 1 August 2006 (2006-08-01), pages 1 - 6, XP031045165, ISBN: 978-1-4244-0071-3 *
MARCEL JIRINA ET AL: "Simple and Effective Probability Density Estimation and Classification", SICE-ICCAS 2006 INTERNATIONAL JOINT CONFERENCE,, 1 October 2006 (2006-10-01), pages 4479 - 4480, XP031050224, ISBN: 978-89-950038-4-8 *

Also Published As

Publication number Publication date
CZ305147B6 (en) 2015-05-20
CZ2008245A3 (en) 2009-11-04

Similar Documents

Publication Publication Date Title
Dal Pozzolo et al. Credit card fraud detection and concept-drift adaptation with delayed supervised information
CN104504400B (en) A kind of driver&#39;s anomaly detection method based on online behavior modeling
WO2019185657A1 (en) Predictive vehicle diagnostics method
CN1196984C (en) Anomaly detection system and method of teaching
CN109871862A (en) A kind of failure prediction method based on synthesis minority class over-sampling and deep learning
US20180375130A1 (en) Apparatus and method for predicting life span of fuel cell, and vehicle system having the same
CN102498445A (en) Supervised fault learning using rule-generated samples for machine condition monitoring
CN110634081A (en) Method and device for processing abnormal data of hydropower station
Aminian et al. A conformal Bayesian network for classification of Mycobacterium tuberculosis complex lineages
CN108108455A (en) Method for pushing, device, storage medium and the electronic equipment of destination
CN109964182A (en) Method and system for vehicle analysis
CN101594361A (en) Network Intrusion Detection System based on shortcut calculation of support vector machine
CN107045786A (en) The computational methods of traffic accident interval time occur for the driver based on Kaplan Meier methods
CN113643033A (en) Information processing method and server for big data wind control analysis
CN103810402A (en) Data processing method and device for genomes
WO2009129758A1 (en) Apparatus for assessing a control value
Park et al. δ-exceedance records and random adaptive walks
CN104572930A (en) Data classifying method and device
CN116933035A (en) Data anomaly detection method, device, computer equipment and storage medium
US11496379B1 (en) Network traffic analysis method and device based on multi-source network traffic data
CN111179212A (en) Method for realizing micro target detection chip integrating distillation strategy and deconvolution
Wu et al. Finding the centre: corrections for asymmetry in high-throughput sequencing datasets
TWI639908B (en) Method for detecting and diagnosing an abnormal process
FI130045B (en) Analyzing measurement results of a communications network or other target system
König et al. Advanced methods for the analysis of semiconductor manufacturing process data

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 09735709

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 09735709

Country of ref document: EP

Kind code of ref document: A1