JP2002049747A - Device and method for compressing time series data - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a time series data compressor, capable of compressing time series data, while keeping a periodicity or fluctuation pattern to remain in the fluctuation of the time series data. SOLUTION: This device is provided with a pre-processing part 11 for calculating a maximum value, a minimum value and the mean value from N pieces of target data (N is an integer), while referring to N pieces of the last target data in the past including relevant target data for each of target data; a first selecting processing part 13 for performing selecting processing of the maximum value for storing the maximum value for each of target data, when the relevant target data are smaller than the mean value or for abandoning the relevant target data, when the relevant target data are larger than the mean value; and a second selecting processing part 14 for performing selecting processing of the minimum value for storing the minimum value for each of target data, when the relevant target data are larger than the mean value or for abandoning the relevant target data, when the relevant target data are smaller than the means value, and selecting processing of the maximum value and the minimum value is conducted alternately.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] The present invention relates to stock prices, foreign exchange,
The present invention relates to a time-series data compression apparatus and a time-series data compression method for compressing time-series data of precious metals, indices, futures, semiconductors, materials, and other marketable materials with price fluctuations.

[0002]

2. Description of the Related Art Conventionally, time series data, for example, stock price data has been provided to investors and the like as stock price charts such as candlestick charts on a daily or weekly basis. A database that stores this stock price data over a long period of time can display a large number of records visually when the number of records increases and the number of records exceeds the display resolution of computers, television receivers, portable electronic terminals, etc. It will be difficult. For example, when stock price data of 1000 records is displayed as a chart on a device (display area) of 600 dots in the horizontal direction, the information of a plurality of records converges to one horizontal dot, the screen is collapsed, and the display becomes difficult to see. .

[0003] Most of market data such as stock price data is composed of "open price", "high price", "low price", and "price".
This value is expressed in a format called a candle. Since the number of horizontal dots required for expression per record is at least three, the number of data that can be displayed on an image display device in a limited area is limited. It had been.

[0004] Furthermore, a general investor who does not have a dedicated line takes a long time to download an enormous amount of stock price charts and stock price data. Therefore, conventional stock price data and stock price charts are easy to use for general investors. There was a problem that was not good.

[0005] Under such circumstances, various compression methods for compressing stock price data have been proposed.

[0006]

As one of the conventional compression methods, there is a method of displaying data in units of one month in a single candle in order to prolong the display period. In the case of this method, there is a problem that an important point becomes difficult to read because the price fluctuation history in one month is simplified.

[0007] Some of the conventional display technologies include a new price bar,
There is something called the key feet, Point & Figure chart. Although these display technologies compress and display price fluctuations, they are common in that data on the time axis is lost. When the data on the time axis is lost, the following problem occurs. For example, how many months did investors take from the previous high to the next high? Or, the emphasis is on the periodicity of stock price fluctuations, such as how many months it took from the previous adjustment phase to the next adjustment phase. However, the display method of the key feet or the like cannot display the periodicity of the stock price fluctuation.

[0008] Further, generally, investors refer to price fluctuation patterns in a stock price chart to determine timing of buying and selling. As the fluctuation pattern of the stock price chart, for example, a state where the stock price is gradually rising to the bottom of the pot, a state where the vertical amplitude of the stock price is gradually reduced, and a case where the previous low point ( There is a state in which the next low point is gradually rising from the point of change (from the point of going down to going up). Such a price fluctuation pattern can be applied to the stock price of each company or the Nikkei Stock Average. Therefore, when compressing time-series data such as stock price data, it is necessary to compress a large amount of stock price data including noise while leaving a fluctuation pattern.

The present invention has been made based on the above circumstances, and has a time-series data compression apparatus and a time-series data compression apparatus capable of compressing time-series data while retaining the periodicity and fluctuation pattern of the fluctuation of the time-series data. It is an object to provide a compression method.

[0010]

In order to achieve the above object, a time-series data compression apparatus according to the present invention needs to refer to target data to be processed sequentially from time-series data. A time-series data compression apparatus for compressing data by selecting target data and discarding other target data. For each target data, a past N (including N: (Integer) number of target data, a pre-processing means for calculating a maximum value, a minimum value, and an average value from the N pieces of target data; When the value is smaller than the value, a maximum value storing process for storing the maximum value is performed, and when the target data is equal to or larger than the average value, first selection processing means for discarding the target data, When the elephant data is larger than the average value, a selection process of a minimum value for storing the minimum value is performed, and when the target data is equal to or less than the average value, a second selection processing unit for discarding the target data, Storage means for storing whether the processing of the maximum value was selected by the first selection processing means or the selection of the minimum value by the second processing selection means, It is determined whether the most recent processing was the maximum value selection processing or the minimum value selection processing with reference to the storage means. If the most recent processing was the maximum value selection processing, Switching means for causing the data to be processed by the second selection processing means, and when the most recent processing is the selection processing of the minimum value, processing the target data by the first selection processing means. It is characterized in.

[0011] In order to achieve the above object, a time-series data compression method according to the present invention comprises:
A time-series data compression method of sequentially selecting target data required by referring to target data to be processed and compressing the data by discarding other target data. Referring to the latest N (where N is an integer) past target data including the target data, a maximum value, a minimum value, and an average value are calculated from the N target data. When the target data is smaller than the average value, a selection process of a maximum value for storing the maximum value is performed.When the target data is equal to or more than the average value, the target data is processed. A first selection processing step of discarding, and for the target data, when the target data is larger than the average value, performing a selection processing of a minimum value for storing the minimum value; In the following cases, a second selection processing step of discarding the target data, and whether the most recent processing was the selection processing of the maximum value in the first selection processing step or the selection of the minimum value in the second processing selection step A storage step of storing whether the processing was a processing, and, for each target data, whether the most recent processing was the processing of selecting the maximum value or the processing of selecting the minimum value with reference to the storage unit. When the most recent processing is the maximum value selection processing, the target data is processed in the second selection processing step. When the most recent processing is the minimum value selection processing, the relevant data is And a switching step of performing the processing in the first selection processing step.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a time-series data compression device according to an embodiment of the present invention will be described with reference to the drawings. In the apparatus of the present embodiment, a case will be described in which stock price data is used as time-series data and weekly data of the stock price data is compressed. The apparatus according to the present embodiment compresses data by judging whether or not a large number of records in a database is a necessary record for each record, selecting and leaving only necessary records, and discarding other records. I do. In addition, each record of the present embodiment includes at least four fields of “open price”, “high price”, “low price”, and “close price” in addition to “time data”.

[Structure of Embodiment] FIG. 1 is a schematic block diagram of a time-series data compression apparatus according to an embodiment of the present invention. The time-series data compression device according to the present embodiment includes a database 10 storing stock price data and a database 1
Pre-processing unit 11 that performs pre-processing before performing selection processing on each record of 0, and first selection processing unit 1 that performs selection processing
A third and second selection processing unit 14, a storage unit 15 for storing whether the most recent processing was a selection processing by the first selection processing unit 13 or a selection processing by the second processing selection unit 14, and the like; A switching unit 12 for alternately performing a selection process by the first selection processing unit 13 and a selection process by the second selection processing unit 14; a display unit 16 for displaying data after selection and compression; It includes a pointer 17 for storing an address indicating the position of a record, and a control unit 20 for controlling the entire apparatus.

The preprocessing unit 11 refers to the last three past records including the record before performing the selection processing by the first selection processing unit 13 or the second selection processing unit 14 for each record of the database 10. Then, the maximum value, the minimum value, and the average value are calculated, and the results are stored in the storage unit 15 or the like. When calculating the maximum value, the highest value among the highest values in the past three records is calculated. When calculating the minimum value, the smallest low price is calculated from the low prices in the past three records. When calculating the average value of the past three records, the average value of the closing prices in the past three records is calculated. When calculating the average value, the average value may be calculated using the starting value. Alternatively, all of the maximum value, the minimum value, and the average value may be calculated using the closing price, the opening price, or the volume weighted average. Further, when processing the first record of the database, the first record is used, and when processing the second record, the first and second two records are used to calculate the maximum value, the minimum value, and the average value. .

The first selection processing unit 13 compares, for each record to be processed, the closing price of the record with the average value calculated by the preprocessing unit 11, and when the closing price of the record is smaller than the average value, A selection process (a selection process of the maximum value) for setting a flag on the record including the maximum value of the high value calculated by the preprocessing unit 11 is performed, and when the closing price of the record is equal to or more than the average value, the record is discarded.

The second selection processing unit 14 compares, for each record to be processed, the closing price of the record with the average value calculated by the preprocessing unit 11, and when the closing price of the record is larger than the average value, A process for setting a flag on a record including the minimum value of the low price calculated by the preprocessing unit (selection process of the minimum value) is performed, and when the closing price of the record is equal to or less than the average value, the record is discarded.

In the first selection processing unit 13 and the second selection processing unit 14, instead of setting a flag on a record,
The high value of the maximum value and the low value of the minimum value may be stored in a storage unit or the like.

The pointer 20 indicates the position of a record to be processed when the pre-processing unit 11, the first selection processing unit 13, the second selection processing unit 14 and the like perform processing. Is incremented each time.
Therefore, the pointer 20 is referenced not only by a main program described later but also by a subroutine.

The switching unit 12 stores the data in the storage unit 1 for each record.
5, whether the most recent processing was the maximum value selection processing by the first selection processing unit 13 or the second selection processing unit 14
It is determined whether or not the process is a minimum value selection process. If the most recent process is a maximum value selection process, the target record is processed by the second selection processing unit 14, and the most recent process is a minimum value selection process. If it is a process, the target record is processed by the first selection processing unit 13. As a result, after the selection processing is performed by the first selection processing unit 13, the selection processing is always performed by the second selection processing unit 14, and after the selection processing is performed by the second selection processing unit 14, Next, selection processing is always performed by the first selection processing unit 13. That is, in the present embodiment, the process of selecting the maximum value and the process of selecting the minimum value are performed alternately.

The control unit 20 controls the reading and writing of data in the storage unit 15 in response to requests from the preprocessing unit 11, the first selection processing unit 13, the second selection processing unit 14, and the like, and sets a pointer at the beginning of processing. Is set to 1 or the value of the pointer is incremented every time the processing of one record is completed. Further, the control unit 20 displays the compressed data on the display unit 16 by a method described later.

[Operation of Embodiment] Next, a processing procedure of the time-series data compression apparatus of the embodiment will be described with reference to FIGS. FIG. 2 is a diagram showing a processing procedure of the main program, and FIGS. 3 to 5 are diagrams showing a processing procedure of a subroutine (function) called from the main program. 2 to 5, a record is also referred to as a block.

After performing predetermined initial settings, step S1
Let's define the block length (TheBlook). In the case of stock price data, the block length (N) is set by selecting an appropriate numerical value from 3 to 10. In the present embodiment, since the pre-processing is performed by referring to the records for three weeks as described above, the block length is set to 3 here.

In step S2, since the selection processing by the first selection processing unit or the second selection processing unit has not been performed, the final detection data variable (LastLabel) indicating the result of the previous selection processing is initialized to null. I do.

In step S3, the value of the pointer 20 is set to a value "1" indicating the first record of the database.
In step S4, it is determined whether or not the record being referred to is the last record in the database 10. If it is not the last record, the process proceeds to step S5. In step S5, three subroutines shown in FIGS. 3 to 5 are executed. The positions of the records of the database referred to in the subroutines shown in FIGS. 3 to 5 are the same as the positions set in the main program (the same pointer 2).
0).

In the processing procedure of the subroutine shown in FIG. 3, after performing a predetermined initial setting, in step S21, the reference position P (the value of the pointer 20) of the current record of the database is changed from the block length (BlockSize) = 3. It is determined whether it is large. In this case, since the position P of the pointer 20 is 1, the process proceeds to step S23, and the high value from the first record to the record at the position of P (in this case, only the first record) is temporarily set as a return value. Remember. When P is 2, the highest value is calculated from the top two records, and when P is 3, the highest value is calculated from the top three records, and the maximum value is calculated. It is temporarily stored as a return value. On the other hand, when the reference position P is larger than the block length (BlockSize) = 3, the process proceeds to step S22, and the block length (BlockSize) = 3 including P in the past from the current reference position P goes back. The maximum value of the high value is calculated from the records in the range (three records), and the maximum value is temporarily stored as a return value.

In the processing procedure of the subroutine shown in FIG. 4, after performing a predetermined initial setting, in step S31, the reference position P (the value of the pointer 20) of the current record of the database is changed from the block length (BlockSize) = 3. It is determined whether it is large. In this case, since the position P of the pointer 20 is 1, the processing shifts to step S33 to temporarily store the low price from the first record to the record at the position of P (in this case, the first record) as a return value. I do. Note that P
Is 2 in the first two records, and P is 3 in the first three records.
The lowest value that is the minimum value is calculated, and the minimum value is temporarily stored as a return value. On the other hand, when the reference position P is larger than the block length (BlockSize) = 3, step S2
2, the minimum value of the low price is selected from the records (three records) in the range of the block length (BlockSize) = 3 including P in the past direction from the current reference position P in the past direction. After the calculation, the minimum value is temporarily stored as a return value.

In the processing procedure of the subroutine shown in FIG. 5, after performing a predetermined initial setting, in step S41, the reference position P (the value of the pointer 20) of the current database record is changed from the block length (BlockSize) = 3. It is determined whether it is large. In this case, since the position P of the pointer 20 is 1, the processing shifts to step S43 to temporarily store the closing value from the first record to the record at the position of P (in this case, the first record) as a return value. I do. Note that P
Is 2 in the first two records, and P is 3 in the first three records.
The average value of the closing prices is calculated, and the average value is temporarily stored as a return value. On the other hand, the reference position P is the block length (Bloc
If it is larger than (kSize) = 3, the process proceeds to step S22, in which records in the range (block size of three records) that extend backward from the current reference position P by a length of 3 including the P (BlockSize) = 3 ) Is calculated, and the average value is temporarily stored as a return value.

When the processing of the subroutine shown in FIGS. 3 to 5 ends, the processing of step S6 in FIG. 1 is performed. In step S6, it is determined whether the last detection data variable (LastLabel) indicating the result of the previous selection processing is the initial value set in step S3 or "Min" indicating that the minimum value selection processing has been performed. I do. Since the current value is the initial value, the process proceeds to step S7. Initial value or "Min"
If not, the process proceeds to step S9.

In step S7, it is determined whether or not the closing price of the record to be processed (the record indicated by the pointer 20) is smaller than the average value (dBlockAve) obtained in step S5. If smaller than the average value, step S
The process proceeds to step S8, and if it is equal to or greater than the average value, the process proceeds to step S9.

In step S8, a record having a high value equal to the maximum value (dBlockMax) obtained in step S5 is searched, and data including the time data of the record is selected as display data (a flag is set on the record). .). At this time, if there are a plurality of records having the same high value as the corresponding maximum value, a predetermined record, for example, a later record in time is selected as display data. In addition, in order to indicate that the selection process is a process of selecting the maximum value, “Max” is substituted for the last detection data variable (LastLabel) indicating the result of the previous selection process. By setting the final detection data variable to “Max” in this way, when processing the next and subsequent records,
It is possible to always execute the selection processing (steps S9 to S11) for searching for the minimum value by the second selection processing unit and search for the lowest price.

In step S9, the last detected data variable (LastLabel) indicating the result of the previous selection processing is the initial value, or "Ma" indicating that the maximum value has been selected.
x ". If the initial value or "Max",
Move to step S10. If it is not the initial value or “Max”, the process proceeds to step S12.

In step S10, the closing value of the record to be processed (the record indicated by the pointer 20) is
It is determined whether or not it is larger than the average value (dBlockAve) obtained in step S5. If the average value is larger than the average value, the process proceeds to step S11.
Move to

In step S11, a record having a low value equal to the minimum value (dBlockMin) obtained in step S5 is searched, and data including time data of the record is selected as display data (a flag is set on the record). .). At this time, if there are a plurality of records having the same low price as the corresponding minimum value, a predetermined record, for example, a later record in time, is selected as the display data. Also, “Min” is substituted for the last detection data variable (LastLabel) to indicate that the selection process is a process for selecting the minimum value. By setting the final detection data variable to “Min” in this way, when processing the next and subsequent records, the first selection processing unit always searches for the maximum value (steps S6 to S5).
8) can be executed to search for a high value.

By performing the processing from step S6 to step S11, when the closing price of each record to be processed is slightly lower than the average value of the past three records, when the closing price of the past three records is reduced. The maximum value of the high price is found from and the selection process is performed. When the closing price of the record to be processed is slightly higher than the average value of the past three records, the minimum value of the low price is found and selected from the past three records. Further, the data is compressed by alternately selecting the maximum value of the high value and the minimum value of the low value.

In step S12, the value of the pointer 20 is incremented to change the record to be processed to the next record. After this processing, step S4
Return to Hereinafter, by repeating the processing from step S4 to step S12 until the processing of the last record of the database is completed, that is, the same number of times as the number of records of the target database,
Compress the number of records.

If it is determined in step S4 that the record to be processed is the last record, step S13
Then, the closing price of the last record is selected as the display data, and the process ends.

According to the above processing procedure, for the third and subsequent records, the maximum value, the minimum value, and the average value are calculated and held in advance for each record based on the data of the past three records. The first selection processing unit 13 performs a maximum value selection process or the second selection processing unit 14 performs a minimum value selection process. The first selection processing unit 13 compares the closing price of the record to be processed with the average value, selects the maximum value when the closing price is smaller than the average value, and discards the record when the closing price is equal to or more than the average value. To process. The second selection processing unit 14 compares the closing price of the record to be processed with the average value, selects the minimum value when the closing price is larger than the average value, and discards the record when the closing price is less than the average value. To process. The switching unit 12 controls the maximum value selection processing and the minimum value selection processing to be performed alternately. By compressing records by such selection processing and discarding processing, data can be compressed while leaving periodicity and fluctuation patterns of stock prices.

FIG. 6 is a graph displayed based on the data after the actual stock price data of Konaka Co., Ltd. is compressed by the apparatus of the present embodiment. In the figure, data is compressed from February 1998 to June 2000 based on weekly data. The minimum value (low price) and the maximum value (high price) are sequentially connected with a straight line from the high price to the low price or from the low price to the high price, an arrow is attached to the tip, and the actual price is displayed at the position of each low price and the high price. Although FIG. 6 does not show the price of the closing price of the last record,
This price may be displayed.

FIG. 7 is a graph showing the stock price data of Konaka Co., Ltd. before compression by the apparatus of the present embodiment as a weekly candle. The graph in FIG. 7 shows data from February 1999 to June 2000. Comparing the corresponding periods in FIG. 6 and FIG. 7, the period from February 1999 to June 2000 displayed in FIG. 7 is compressed to approximately half or less in FIG. . While the actual total number of records of Konaka Co., Ltd. in this period is 155, the number of records after compression by the apparatus of the present embodiment is 29, and the compression ratio is about 1/3.
Further, when the actual stock price data of Sony Corporation was compressed in the same manner by the apparatus of the present embodiment, the weekly data of 280 records was compressed to 60 records. Thus, according to the device of the present embodiment, at least 1/2
The stock price data can be compressed to the order of 1/10.

Further, when looking at the graph of FIG. 6, it can be easily recognized that the fluctuation pattern of this stock price chart is a pennant. Further, it can be easily recognized from the graph of FIG. 6 that the period of the price fluctuation is gradually shortened. On the other hand, if you look closely at FIG. 7, you can see that there is a variation pattern and periodicity similar to FIG. As described above, according to the apparatus of the present embodiment, the data can be compressed while maintaining the fluctuation pattern and periodicity of the stock price chart, and the fluctuation pattern and periodicity can be easily recognized as compared with the conventional chart. Become. Therefore, the investor can easily read the date and price that should be used as a guide for the next price change from the past trend shown in FIG. In the above embodiment, the case where the stock price data of one company is compressed has been described. However, the same compression can be performed by compressing the data of the Nikkei Stock Average. Also,
According to the present embodiment, a price fluctuation pattern other than the pennant, such as a double top or double bottom, can be easily and reliably found.

Further, it is not easy to find a resistance line or an instruction line from a large amount of data by using a computer. However, according to the above-described embodiment, since the number of data is small, FIG. As shown in FIG. 7, the resistance line A and the instruction line B can be easily displayed. It is not easy to search for a fluctuation pattern or periodicity from a large number of data even by using a computer. However, finding a fluctuation pattern or periodicity using data compressed according to the present embodiment is as follows.
Because of the low computational capacity, even weak computers can be easily found. Further, it is difficult for a general investor to find a resistance line or an indicator line from the conventional complicated stock price chart shown in FIG. 7, but according to the present embodiment,
Even without drawing the resistance line A and the instruction line B, a general investor can easily recognize the instruction line and the resistance line.

In the case of stock price data, if the value of N is 2 or less, it is difficult to extract feature points, and if the value of N is 1
If it is 0 or more, it becomes too rough. Therefore, in the case of stock price data, the value of N is preferably between 3 and 10. When the data amount to be processed is large or the display device is small, the value of N can be set to 11 or more. [Effects of Embodiment] According to the above-described embodiment, when the market price is rising, a past minimum value (low price) is found, data while the market price is rising is discarded, and then data is deleted. You get it when the market price goes down. Then, while the market price continues to fall, data during that period is discarded. Thus, the data can be compressed and the data can be selected by focusing only on the past high and low prices. The average investor is paying attention to the high price when the market is raising. Also, at the time of the down price,
We are paying attention to the low price. However, investment experts are paying attention to the low prices during the uptrend. In the case of a downtrend, we are paying attention to the high price of the market. According to the present embodiment, it can be said that the market data is compressed from the viewpoint of the expert.

Further, according to the present embodiment, a large amount of stock price data can be compressed to a required size while maintaining a price fluctuation pattern and periodicity. Also on the terminal via the Internet etc.
It becomes possible to transfer a chart of stock price data according to the display ability. That is, since the transfer amount of the conventional stock price data is several tens of gigabytes, transferring this huge amount of stock price data via the Internet increases the transfer cost. However, by compressing the data with the apparatus of the present embodiment, even when stock price data is transferred to the portable terminal using the Internet, the transfer fee is low. Therefore, according to the present embodiment,
It is also possible to provide a service for transferring stock price data to mobile terminals as a business.

Further, according to the present embodiment, the disturbance data (noise) which makes it difficult to recognize the characteristic portion such as the periodicity and the fluctuation pattern of the stock price is efficiently removed, and the compression with the characteristic portion left is performed. Therefore, the characteristic portion can be easily grasped by using the compressed data, and therefore, it is possible to quickly make an accurate judgment on stock price trading. For example, according to the present embodiment, it is possible to check the price pattern of dozens of brands per minute, but if the same reading operation is performed by the conventional method shown in FIG. It is necessary, and eyes are tired because I look at the screen.

[Other Embodiments] The present invention is not limited to the above embodiments, and various modifications can be made within the scope of the invention. For example, in the above embodiment, the case where N is 3 has been described. However, by changing the value of N, the compression ratio and compression accuracy can be freely changed. That is, by setting the value of N in accordance with the amount of data to be compressed and the capability of the display device such as a portable terminal, it is possible to display the data on a small display device while leaving as much of the feature portion of the data as possible. Becomes possible.

Further, in the above embodiment, the case where the stock price data is compressed has been described. However, the present invention is not limited to this. Forex, precious metals, indices, futures,
Time-series data of semiconductors, materials, and other marketable materials with price fluctuations may be compressed.

Further, the present invention is not limited to the above embodiment, and has no open price or high price field.
This can be used when time-series data called so-called single value is compressed, or when time-series data of volume weighted average is compressed.

[0048]

As described above, according to the present invention, by setting an appropriate value of N, the periodicity and the fluctuation pattern of the time-series data can be changed according to the amount of data and the capability of the display device. It is possible to provide a time-series data compression device and a time-series data compression method that can compress time-series data while leaving it.

[Brief description of the drawings]

FIG. 1 is a schematic block diagram of a time-series data compression device according to an embodiment of the present invention.

FIG. 2 is a diagram showing a processing procedure of a main program.

FIG. 3 is a diagram showing a processing procedure of a subroutine (function) called from the main program of FIG. 2;

FIG. 4 is a diagram showing a processing procedure of a subroutine (function) called from the main program of FIG. 2;

FIG. 5 is a diagram showing a processing procedure of a subroutine (function) called from the main program of FIG. 2;

FIG. 6 is a graph displayed based on data obtained by compressing actual stock price data of Konaka Co., Ltd. using the apparatus of the present embodiment.

FIG. 7 is a graph showing stock price data of Konaka Co., Ltd. before compression by the apparatus of the present embodiment as a weekly candle.

[Explanation of symbols]

10: Database, 11: Preprocessing unit, 12:
Switching section, 13: first selection processing section, 14: second selection processing section, 15: storage section, 16: display section, 1
7: pointer, 20: control unit

Claims (7)

[Claims] 1. Time-series data for compressing data by sequentially selecting required data from time-series data by referring to target data to be processed and discarding other target data. A compression apparatus, for each target data, referring to the latest N (where N is an integer) target data including the target data in the past, and selecting a maximum from among the N target data. Preprocessing means for calculating a value, a minimum value, and an average value; and for the target data, when the target data is smaller than the average value, performing a selection process of a maximum value for storing the maximum value; A first selection processing means for discarding the target data when the target data is equal to or more than the average value; and selecting a minimum value for storing the minimum value when the target data is larger than the average value. Performing a process, when the target data is equal to or less than the average value, a second selection processing unit for discarding the target data, and whether the most recent process was the selection process of the maximum value by the first selection processing unit, or Storage means for storing whether the minimum value was selected by the second processing selection means; and for each target data, whether the most recent processing was the maximum value selection processing with reference to the storage means. Alternatively, it is determined whether or not the minimum value has been selected. If the most recent process was the maximum value selection process, the target data is set to the second value.
Switching processing for causing the first selection processing means to process the target data when the most recent processing is the selection processing of the minimum value, the processing being performed by the selection processing means. Compression device. 2. The time-series data is market data.
2. The time-series data compression device according to claim 1, wherein the target data is a record having at least a plurality of fields including an opening price, a high price, a low price, and a closing price in addition to the time data. 3. The time-series data compression device according to claim 2, wherein the maximum value is a high value, the minimum value is a low value, and the average value is an average value of closing prices. 4. The time-series data compression device according to claim 1, wherein the value of N is any one of 3 to 10. 5. A display control means for displaying, on a display means, a maximum value and a minimum value selected by the maximum value selection processing and the minimum value selection processing sequentially as a line graph connected by straight lines. The time-series data compression device according to claim 1, 2, 3, or 4, wherein: 6. Time-series data for compressing data by sequentially selecting necessary target data from time-series data by referring to target data to be processed and discarding other target data. A compression method, wherein for each target data, the latest N (where N is an integer) target data including the target data is referred to, and a maximum of the N target data is selected from among the N target data. Value, a minimum value, a pre-processing step of calculating an average value, and, for the target data, when the target data is smaller than the average value, performing a selection process of a maximum value for storing the maximum value; A first selection processing step of discarding the target data when is equal to or more than the average value; and selecting a minimum value for storing the minimum value when the target data is larger than the average value. Performing a process, when the target data is equal to or less than the average value, a second selection processing step of discarding the target data, and whether the most recent processing was the selection processing of the maximum value by the first selection processing step, or A storage step of storing whether or not the minimum value was selected in the second processing selection step; and for each target data, whether or not the most recent processing was the maximum value selection processing with reference to the storage means. Alternatively, it is determined whether or not the minimum value has been selected. If the most recent process was the maximum value selection process, the target data is set to the second value.
A switching step of causing the selected data to be processed in the selection processing step and, if the most recent processing is the selection processing of the minimum value, processing the target data in the first selection processing step. Compression method. 7. The method according to claim 1, further comprising the step of sequentially displaying the maximum value and the minimum value selected by the maximum value selection process and the minimum value selection process as a line graph connected by straight lines. 6. The time-series data compression method according to 6.