JP4584010B2 - Median value calculation apparatus, median value calculation method, and median value calculation program - Google Patents

Description

  The present invention relates to a median value calculation device, a median value calculation method, and a median value calculation program for calculating a median value from a plurality of data.

  Conventionally, data normalization using an average value of all data has been performed for the purpose of obtaining a practical data distribution. For example, in MPEG (moving picture experts group), which is a compression encoding method for image data, a practical data distribution is calculated by normalizing the image data using an average value (for example, patent literature). 1).

  On the other hand, data normalization using an average value may be an obstacle to data normalization because it is easily influenced by the magnitude of the data value. For this reason, data normalization using a median value (a value at the center when data samples are sorted) instead of an average value is often performed.

JP 2003-37843 A

  However, in the prior art using the median value, it takes a lot of time to calculate the median value because it requires data sorting to calculate the median value. There is a problem that can not be. Note that when the median value is calculated on hardware, there is a problem that it is necessary to construct a complicated sort circuit for data sort processing.

  The present invention has been made in view of the above, and provides a median value calculation device, a median value calculation method, and a median value calculation program for simply calculating a median value.

In order to solve the above-described problems and achieve the object, the present invention is a median value calculation device for calculating a median value from a plurality of data, and each of the plurality of data is set to a predetermined value according to each data value. Range distribution means for distributing any of the plurality of ranges divided into ranges, and data located at the center of the plurality of data from the plurality of ranges in which each of the plurality of data is distributed by the range distribution means approximation to approximately calculate the value of the center of the center range as the median value based the central range detecting means for detecting the center range range allocated to the value of the detected center range by said central range detecting means characterized in that a median value calculating means.

In order to solve the above-described problems and achieve the object, the present invention is a median value calculation device for calculating a median value from a plurality of data, and each of the plurality of data is set to a predetermined value according to each data value. Range distribution means for distributing any of the plurality of ranges divided into ranges, and data located at the center of the plurality of data from the plurality of ranges in which each of the plurality of data is distributed by the range distribution means A center range detecting means for detecting a range to which the values are assigned as a center range, and an approximate median value calculation for approximately calculating a median value by multiplying a value of the center of the center range detected by the center range detecting means by a coefficient Means.

Further, the present invention is characterized in that, in the above invention, the range distribution means distributes the plurality of data to any one of the plurality of ranges divided at equal intervals.

Also, the present invention is characterized in that, in the above invention, the range distribution means distributes the plurality of data to any one of the plurality of ranges divided into power-of-two ranges.

Further, in the present invention according to the above invention, the range distribution unit distributes the plurality of data to any of the plurality of ranges divided according to a distribution density of each data value in the plurality of data. It is characterized by.

The present invention is also a median value calculation method applied to a median value calculation device for calculating a median value from a plurality of data, wherein each of the plurality of data is divided into a predetermined value range according to each data value. A range distribution step of allocating to any of the plurality of ranges, and data located at the center of the plurality of data are distributed from the plurality of ranges in which each of the plurality of data is distributed by the range distribution step a central range detection step of detecting a range around range, approximate median value calculation step of approximately calculating the value of the center of the center range based on the value of the center range detected by the central range detecting step as a median value wherein the containing and.

The present invention is also a median value calculation program for causing a computer to execute a method for calculating a median value from a plurality of data, wherein the computer sets each of the plurality of data to a predetermined value according to each data value. A range distribution procedure for allocating one of a plurality of ranges divided into ranges, and data located at the center of the plurality of data from the plurality of ranges in which each of the plurality of data is distributed by the range distribution procedure a central range detection procedure for detecting a distribution obtained range centered range, approximate median of approximately calculating the value of the center of the center range based on the value of the detected center range by said central range detection procedure as the median value A value calculation procedure is executed.

According to the present invention , each data in a plurality of data is assigned to one of a plurality of ranges divided into predetermined values according to the value, and a data value located at the center of the plurality of data from the plurality of ranges is determined. After the distributed central range is detected, the median value is approximately calculated from the central range, so that the median value can be easily calculated. In addition, since the sort processing of data that is conventionally performed when calculating the median value is not required, it is possible to easily calculate the median value without constructing a complicated sort circuit on hardware. .

In addition, according to the present invention , since a plurality of ranges to which a plurality of data are distributed are divided at equal intervals, a plurality of data including distributed data values can be efficiently distributed.

Further, according to the present invention , since a plurality of ranges into which a plurality of data are distributed are divided into power-of-two ranges, a plurality of data including data having a wide range of data values can be efficiently distributed.

In addition, according to the present invention , since a plurality of ranges for distributing a plurality of data are divided into ranges according to the distribution density of a plurality of data values, a plurality of data can be distributed more efficiently according to the distribution of data values. Is possible.

In addition, according to the present invention , the median value is approximately calculated as the average value of the upper and lower limits of the center range, so that the median value can be easily calculated.

In addition, according to the present invention , the median value is approximately calculated by multiplying the average value of the upper and lower limits of the central range by a coefficient, so that adjustment of the median value according to the use after normalization of a plurality of data Is possible.

  Hereinafter, preferred embodiments of a median value calculation apparatus, a median value calculation method, and a median value calculation program according to the present invention will be described in detail with reference to the accompanying drawings.

[Outline and Features (Example 1)]
In the following first embodiment, the outline and features of the median value calculation apparatus according to the first embodiment, the configuration and processing of the median value calculation apparatus will be described, and finally the effects of the first embodiment will be described.

  First, the outline and features of the median value calculation apparatus will be described with reference to FIG. FIG. 1 is a diagram showing an outline of a median value calculation apparatus.

  An outline of the median value calculation device is to calculate a median value from a plurality of data. Then, the median value calculation device distributes each data in the plurality of data to one of a plurality of ranges divided by a predetermined value according to the value, and is located at the center of the plurality of data from the plurality of ranges. It is characterized in that after detecting a central range to which data is distributed, a median value is approximately calculated from the central range.

  For example, as shown in FIG. 1, a case will be described in which a total of 101 data is distributed into 11 ranges at equal intervals of 10 steps. First, each of the 101 pieces of data is distributed into 11 ranges divided at equal intervals according to the value of each data. Next, the number of each data allocated to the range of 11 is sequentially cumulatively counted, and the range 5 including the 51st count number data located at the center of the total number of data is detected as the central range. Subsequently, 45 which is an average value of the upper limit value 50 and the lower limit value 40 in the range 5 detected as the central range is approximately calculated as a median value. Therefore, it is possible to easily calculate the median value from a plurality of data.

[Configuration of Median Value Calculation Device (Example 1)]
Next, the configuration of the median value calculation apparatus according to the first embodiment will be described with reference to FIG. 2 and FIG. 3 in addition to FIG. FIG. 2 is a block diagram illustrating a configuration of the median value calculation apparatus according to the first embodiment, and FIG. 3 is a diagram illustrating a configuration of a range distribution table of the storage unit.

  As shown in FIG. 2, the median value calculation device 10 is configured by connecting an input unit 11, an output unit 12, an input / output control IF unit 13, a storage unit 14, and a control unit 15 via a predetermined bus. The processing of these units will be described below.

  The input unit 11 is a unit that receives input of various types of information, and includes an operation panel, a switch, a keyboard, buttons, and the like. For example, as shown in FIG. 4, on the range distribution instruction screen, a setting for setting the distribution range as an equal interval of an initial value 0 and an interval value 10 is received from the operator.

  The output unit 12 is an output unit that outputs various types of information, and includes a monitor (or display or operation panel), a speaker, a lamp, and the like. For example, the processing result (median value calculation processing result, data normalization processing result) stored in the data storage unit 14a described later is output.

  The input / output control IF unit 13 is a unit that controls data transfer. For example, the input / output control IF unit 13 reads out data processed by the control unit 15 (to be described later) from a storage unit 14 (to be described later) and transmits the data to the output unit 12. .

  The storage unit 14 is a memory that stores input / output data and various types of processing data by the control unit 15, and includes a data storage unit 14a and a sorting table 14b as shown in FIG. In the storage unit 14, the data storage unit 14 a first stores a plurality of data input from the input unit 11. Further, the median value calculated by the approximate median value calculation unit 15c described later is stored. The distribution table 14b stores data processed by the range distribution processing unit 15a described later. For example, as shown in FIG. 3, the initial value is 0 and the number of data distributed by the range distribution processing unit 15a is stored in association with the range 1 to the range 11 divided at equal intervals of 10 increments. The total number of data obtained by accumulating the number of data in the range is stored.

  The control unit 15 is a processor that has an internal memory for storing programs and control data that define various processing procedures, and executes various processes using these programs, and is particularly closely related to the present invention. As shown in FIG. 2, a range distribution processing unit 15a, a center range detection unit 15b, and an approximate median value calculation unit 15c are provided. The range distribution processing unit 15a corresponds to “range distribution unit” described in the claims, the center range detection unit 15b also corresponds to “center range detection unit”, and the approximate median value calculation unit 15c includes “approximate median”. Corresponds to “value calculation means”.

  Among such control units, the range distribution processing unit 15a is a processing unit that reads a plurality of data stored in the data storage unit 14a and distributes the data to a plurality of ranges divided into predetermined values according to the values of each data. is there. Specifically, the range distribution processing unit 15a reads a plurality of data from the data storage unit 14a, distributes each of the plurality of data into a plurality of ranges divided into predetermined values according to each data value, and the distribution is performed. Count the number of data. For example, as shown in FIG. 1 and FIG. 3, a plurality of data is allocated to a range 1 to a range 11 divided into equal intervals of 10 at an initial value of 0, and the number of data allocated to each range is counted. .

  The center range detection unit 15b is a processing unit that reads the processing data of the range distribution processing unit 15a from the distribution table 14b and detects the center range. Specifically, the center range detection unit 15b reads the processing result of the range distribution processing unit 15a from the distribution table 14b, sequentially counts the number of data distributed to each range, and counts at the center of the total number of data. A distribution range corresponding to the number is detected as a central range. For example, as shown in FIG. 1, the number of data distributed from the range 1 to the range 11 by the range distribution processing unit 15a is sequentially accumulated, and the 51st count located at the center of a plurality of data whose total data is 101. The range 5 to which the data is distributed is detected as the central range.

  The approximate median value calculation unit 15c is a processing unit that calculates an approximate median value from the center range detected by the center range detection unit 15b. Specifically, the approximate median value calculation unit 15c reads the detection range by the center range detection unit 15b from the data storage unit 14a, and calculates the approximate median value from the detection range. For example, as shown in FIG. 1, the median value is approximately calculated by obtaining 45 which is an average value of the upper limit value 50 and the lower limit value 40 of the center range. The data normalization unit 15d is a processing unit that normalizes data from the median value calculated by the approximate median value calculation unit 15c.

[Processing Procedure of Median Value Calculation Device (Example 1)]
Next, FIG. 4 and FIG. 5 in addition to FIG. 1 show the flow of the approximate median value calculation process. FIG. 4 is a diagram showing a screen for instructing a data range distribution range, and FIG. 5 is a flowchart showing a flow of approximate median value calculation processing.

  As illustrated in FIG. 5, the range distribution processing unit 15a receives a range distribution instruction from the input unit 11 via the input / output processing IF unit 13 (Yes in step S501). For example, as shown in FIG. 4, the range distribution instruction is executed by setting the distribution range to the initial value 0 and the interval value 10 on the range distribution instruction screen displayed on the input unit 11 and pressing the start button. Is called. Following this, upon receiving the range distribution instruction, the range distribution processing unit 15a reads all data stored in the data storage unit 14a via the input / output control IF unit 13 (step S502).

  Subsequently, the range distribution processing unit 15a performs a range distribution process for all the read data (step S503). Specifically, as shown in FIG. 1, the range distribution processing unit 15 a sequentially distributes each data into ranges 1, 2, etc., which are equally spaced in 10 increments according to the value of the data, and is distributed to each range. Count the number of data and the total number of data. As a result, as shown in FIG. 3, the initial value is 0 and the number of data distributed by the range distribution processing unit 15a in association with the range 1 to the range 11 divided at equal intervals of 10 increments, The total number of data obtained by accumulating the number of data is stored in the distribution table 14b of the storage unit 14.

  Subsequently, the center range detection unit 15b reads the processing data from the sorting table 14b and performs center range detection processing (step S504). Specifically, as shown in FIG. 1, the center range detection unit 15 b sequentially counts the number of count data in each range from the range 1 and accumulates at the center of a plurality of data whose total data is 101. The range 5 (40 or more and less than 50) including the 51st count data is detected as the central range.

  Subsequently, the approximate median value calculation unit performs an approximate median value calculation process with reference to the detection processing data of the center range detection unit 15b stored in the data storage unit 14a (step S505). Specifically, as shown in FIG. 1, the approximate median value calculation unit 15 c calculates 45 as an average value of the upper limit value 50 and the lower limit value 40 of the range 5 (40 to less than 50) detected as the central range. Approximately calculated as median value. Then, the approximate median value calculation unit 15c outputs the processing result from the output unit 12 via the input / output control IF unit, and ends the approximate median value calculation process.

[Effect of Example 1]
As described above, according to the first embodiment, each data in a plurality of data is distributed to any one of a plurality of ranges divided into predetermined values according to the values, and a plurality of data is distributed from the plurality of ranges. Since the median value is approximately calculated from the central range after detecting the central range to which the data value located at the center is distributed, the median value can be easily calculated. Further, since the sort processing of data that is performed when calculating the median value is not required, it is possible to easily calculate the median value without constructing a complicated sort circuit on hardware.

  Further, according to the first embodiment described above, since a plurality of ranges in which a plurality of data are distributed are equally spaced, a plurality of data including distributed data values can be efficiently distributed.

  Furthermore, according to the first embodiment, since the median value is approximately calculated as an average value of the upper limit value and the lower limit value in the center range, the median value can be easily calculated.

  Although the median value calculation apparatus according to the first embodiment has been described so far, the present invention may be implemented in various different forms other than the above-described embodiments. Therefore, in the following, as the median value calculation apparatus according to the second embodiment, various different modes are classified into (1) to (4) and described.

(1) Change of data distribution range In the first embodiment described above, a case has been described in which the data distribution range is set at equal intervals of 10 (see FIG. 1), but the present invention is not limited to this. The equally spaced width may be set to a value such as 1, 100, 1000, or the like.

  In the first embodiment, the range in which the data is distributed is set at an equal interval. However, the present invention is not limited to this. For example, as shown in FIG. 6, the range in which each data is distributed is a power of two. You may make it be the range of. In this way, if a plurality of ranges divided by a predetermined value for distributing a plurality of data are set to a power-of-two range, a plurality of data composed of data having a wide range of data values can be efficiently distributed. Become.

  In this case, the approximate median value calculation unit 15c can also calculate by the method described in the first embodiment after converting the bit string in the center range to the decimal system, but simply calculates the bit string as it is. You may make it do. Specifically, as shown in FIG. 6, it is possible to easily calculate, as an approximate median value, a bit with 1 in the leftmost portion where 1 appears in the bit string.

  In the first embodiment, the range in which the data is distributed is equally spaced. However, the present invention is not limited to this, and the range of the distribution range is determined according to the distribution density of each data value in the plurality of data. Each width may be changed. For example, as shown in FIG. 7, when the distribution density of each data value in a plurality of data is high between 150 and 250, the range between 150 and 250 is divided into 25 increments such as range 3 and range 9, A plurality of ranges divided into 10 increments such as 4 to 8 may be used. And it is good also as one range of the range 1 or the range 11 without distinguishing between 1-100 and 300-400 with low distribution density of a data value. In this way, if a plurality of ranges in which each data is distributed is divided into a plurality of ranges divided according to the distribution density of each data value in the plurality of data, the plurality of data can be more efficiently used according to the distribution density of the data values. It becomes possible to distribute to.

(2) Calculation of median value In the above-described first embodiment, the median value is obtained by the average value of the upper limit value and the lower limit value of the central range. However, a value near the average value may be used as the median value. Alternatively, a value calculated by multiplying an average value calculated from the center range by an appropriate coefficient may be approximately calculated as a median value. As a result, the median value can be adjusted according to the use after the data normalization. For example, when compressing image data by the MPEG method, the median value is adjusted by multiplying an appropriate coefficient, thereby increasing the compression rate of the image data and reducing the bit amount during encoding. Safety and failure prevention can be achieved.

(3) Device Configuration Each component of the median value calculation device 10 shown in FIG. 2 is functionally divided and does not necessarily need to be physically configured as illustrated. That is, the specific form of dispersion and integration of the median value calculation device 10 is not limited to the illustrated one. For example, the range distribution processing unit 15a and the central range detection unit 15b are integrated, or approximate to the central range detection unit 15b. Integration of the median value calculation unit 15c or distribution of the approximate median value calculation unit 15c for each calculation item, such as all or part of each component in arbitrary units according to various loads and usage conditions It can be configured functionally or physically distributed and integrated. Furthermore, all or a part of each processing function performed in each device may be realized by a CPU and a program that is analyzed and executed by the CPU, or may be realized as hardware by wired logic.
(4) Median Value Calculation Program Various processes described in the first embodiment can be realized by executing a program prepared in advance on a computer. In the following, an example of a computer that executes a median value calculation program having the same function as that of the first embodiment will be described with reference to FIG. FIG. 8 is a diagram illustrating a computer that executes a median value calculation program.

  As shown in the figure, a computer 20 as a median value calculation device is configured by connecting an operation panel 21, a display 22, an HDD 23, a RAM 24, a ROM 25, and a CPU 26 via a bus 27. Here, the operation panel 21 corresponds to the input unit 11 shown in FIG. 2, and the display 22 corresponds to the output unit 12.

  The ROM 25 has a median value calculation program that exhibits the same function as in the first embodiment, that is, as shown in FIG. 8, a range distribution program 25a, a center range detection program 25b, an approximate median value calculation program 25c, A data normalization program 25d is stored in advance. In addition, about 25a-25d, you may integrate or disperse | distribute suitably similarly to each component of 10 of the median value calculation apparatus shown in FIG.

  Then, the CPU 26 reads out these programs 25a to 25d from the ROM 25 and executes them. As shown in FIG. 8, each program 25a to 25d has a range distribution process 26a, a center range detection process 26b, and an approximate median value calculation. Functions as a process 26c and a data normalization process 26d. Each of the processes 26a to 26d corresponds to the range distribution processing unit 15a, the center range detection unit 15b, the approximate median value calculation unit 15c, and the data normalization unit 15d illustrated in FIG.

  Further, as shown in FIG. 8, the HDD 23 is provided with a data table 23a and a sorting table 23b. The data table 23a and the distribution table 23b correspond to the data storage unit 14a and the distribution table 14b shown in FIG. 2, respectively. The CPU 26 registers each processing result in the data table 14 a and the distribution table 23 b, reads the data group 24 a and the distribution table data 24 b from the data table 23 a and the distribution table 23 b, stores them in the RAM 24, and stores them in the RAM 24. Each processing is executed based on the data group 24a and the sorting table data 24b.

  By the way, the above-described programs 25a to 25d are not necessarily stored in the ROM 25 from the beginning. For example, a flexible disk (FD), a CD-ROM, an MO disk, a DVD disk, an optical disk inserted into the computer 20 can be used. “Portable physical media” such as magnetic disks and IC cards, or “fixed physical media” such as hard disk drives (HDD) provided inside and outside the computer 20, public lines, the Internet, LAN, WAN Alternatively, each program may be stored in “another computer (or server)” connected to the computer 20 via the computer 20 and the like, and the computer 20 may read and execute each program therefrom.

  As described above, the median value calculation apparatus, the median value calculation method, and the median value calculation program according to the present invention are useful when simply calculating the median value from a plurality of data, and in particular, when calculating the median value. This is suitable for the case where the median value is simply calculated on the hardware without constructing a complicated sort circuit for the sort process.

It is a figure which shows the outline | summary and characteristic of the median value calculation apparatus which concern on Example 1. FIG. It is a block diagram which shows the structure of the median value calculation apparatus which concerns on Example 1. FIG. It is a figure which shows the structure of the distribution table of the memory | storage part which concerns on Example 1. FIG. FIG. 10 is a diagram illustrating a screen for instructing data range distribution according to the first embodiment. 6 is a flowchart illustrating a flow of approximate median value calculation processing according to the first embodiment. It is a figure which shows the different form of a range distribution means. It is a figure which shows the different form of a range distribution means. It is a figure which shows the computer which performs a median value calculation program.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Median value calculation apparatus 11 Input part 12 Output part 13 Input / output IF control part 14 Storage part 14a Data storage part 14b Distribution table 15 Control part 15a Range distribution process part 15b Center range detection part 15c Approximate median value calculation part 15d Data normalization Part 20 Computer 21 Operation Panel 22 Display 23 HDD (Hard Disk Drive)
24 RAM (Random Access Memory)
25 ROM (Read Only Memory)
26 CPU (Central Processing Unit)

Claims (7)

A median value calculation device for calculating a median value from a plurality of data,
Range distribution means for distributing each of the plurality of data according to each data value to any of a plurality of ranges divided into predetermined value ranges;
Center range detection means for detecting, as a central range, a range in which data located at the center of the plurality of data is distributed from the plurality of ranges in which each of the plurality of data is distributed by the range distribution means;
An approximate median value calculating means for approximately calculating the value of the center of the center range based on the value of the center range detected by said central range detecting means as a median value
The median value calculation apparatus characterized by comprising a. A median value calculation device for calculating a median value from a plurality of data,
Range distribution means for distributing each of the plurality of data according to each data value to any of a plurality of ranges divided into predetermined value ranges;
Center range detection means for detecting, as a central range, a range in which data located at the center of the plurality of data is distributed from the plurality of ranges in which each of the plurality of data is distributed by the range distribution means;
Approximate median value calculating means for approximately calculating a median value by multiplying a coefficient of the center value of the central range detected by the central range detecting means ;
The median value calculation apparatus characterized by comprising a. The range distribution means, median value calculating apparatus according to claim 1 or 2, characterized in that for distributing the plurality of data to one of the plurality of ranges that are divided at equal intervals. 3. The median value calculation apparatus according to claim 1, wherein the range distribution unit distributes the plurality of data to any of the plurality of ranges divided into power-of-two ranges. 3. The range distribution unit according to claim 1 or 2 , wherein the range distribution unit distributes the plurality of data to any of the plurality of ranges divided according to a distribution density of each data value in the plurality of data. Median value calculation device. A median value calculation method applied to a median value calculation device that calculates a median value from a plurality of data,
A range distribution step of distributing each of the plurality of data according to each data value to any of a plurality of ranges divided into a predetermined value range;
A center range detection step of detecting, as a central range, a range in which data located at the center of the plurality of data is distributed from the plurality of ranges in which each of the plurality of data is distributed by the range distribution step;
An approximate median value calculation step of approximately calculating the value of the center of the center range based on the value of the center range detected by the central range detecting step as a median value
The median value calculation method characterized by including. A median value calculation program for causing a computer to execute a method of calculating a median value from a plurality of data,
In the computer,
A range distribution procedure for distributing each of the plurality of data according to each data value to any of a plurality of ranges divided into a predetermined value range;
A center range detection procedure for detecting, as a central range, a range in which data located at the center of the plurality of data is distributed from the plurality of ranges in which each of the plurality of data is distributed by the range distribution procedure;
An approximate median value calculating procedure for approximately calculating the value of the center of the center range based on the value of the detected center range by said central range detection procedure as the median value
The median value calculation program characterized by performing this.

Images