JPH0696105A - Statistical analyzer - Google Patents

Abstract

PURPOSE:To enable a user to carry out statistical analysis processing with a small memory space and computer power and to reduce the burden on or error of a user. CONSTITUTION:When non-display data arranged in a table form are inputted, processing to store these input data in a magnetic disk 13 is executed (step S1). Next, processing to convert the non-display data to code data and processing to prepare a list is executed (steps S2-S8). Afterwards, processing to correct the table of conversion to the code data and the list is executed (steps S9-S12). Further, statistical analysis processing is executed to the code data based on non-quantative classification (step S13). When this processing is completed, processing to convert the code data contained in the analyzed result to the non-display data is executed (step S14). When this processing is completed, processing to output the analyzed result is executed (step S15).

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a statistical analysis device for performing statistical analysis processing on a plurality of non-quantitative data arranged in a table format based on a quantification classification method.

[0002]

2. Description of the Related Art In recent years, as the performance of personal computers and the like has improved, various spreadsheet software suitable for statistical analysis processing has been developed.

One of the statistical analysis methods used in this spreadsheet software is the quantity classification method.

This quantification classification method is intended for non-quantitative data such as questionnaire results, weather, and days of the week.
General statistical analysis methods for quantitative data (for example,
Multiple regression analysis methods, discriminant analysis methods, factor / principal component analysis methods, etc.) are obtained. There are four analysis methods from Class 1 to Class 4 in this quantification classification method.

When data is transferred from spreadsheet software to a program that executes such a quantified classification method, category data must be transferred.

Conventionally, as this transfer method, a method of transferring non-quantity data as it is and a method of converting the non-quantity data into code data and transferring the code data have been considered.

[0007]

However, in the former method, since non-quantitative data is represented by a long character string, a large memory space is required, and a large computer is required to perform data collation at high speed. There was a problem that it needed power.

On the other hand, in the latter method, since the input data is represented by code data such as A, B, C, ..., The above problem does not occur.

However, this method has a problem that the user has to prepare the code data prior to the statistical analysis process and a problem that the data expression is difficult for a human to understand, and thus an error is likely to occur. It was

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a statistical analysis device which can perform statistical analysis processing with a small memory space and computer power, and which can reduce a user's burden and mistakes. .

[0011]

In order to achieve the above object, the present invention provides means for converting input non-quantitative data into code data, and a quantification classification method for the conversion output of this means. A means for performing statistical analysis processing on the basis of this and a means for returning the code data contained in the processing result of this means to non-quantity data are provided.

[0012]

According to the above configuration, the category data is input to the statistical analysis device in the form of non-quantitative data. This input data is automatically converted into code data inside the device and then subjected to statistical analysis processing. This processing result is output after the code data is converted back to non-quantity data.

As a result, the user can handle the category data as non-quantitative data, so that the user's burden and mistakes can be reduced.

Since the statistical analysis processing can be executed based on the code data, the memory space and computer power can be saved.

[0015]

Embodiments of the present invention will now be described in detail with reference to the drawings.

FIG. 1 is a flow chart showing the operation of the main part of one embodiment of the present invention, and FIG. 2 is a block diagram showing the configuration of one embodiment.

First, the configuration of one embodiment will be described with reference to FIG.

In the figure, reference numeral 11 denotes a central processing unit (hereinafter referred to as "CPU") which executes various processes such as a statistical analysis process. Reference numeral 12 denotes a keyboard for inputting various data into the apparatus by a user's external operation. Thirteen
Is the data input from the keyboard 12 or the CPU 11
It is a magnetic disk for holding the processing result of the above.

Reference numeral 14 is a read-only memory (hereinafter referred to as ROM) that stores a control program for the CPU 11. Reference numeral 15 is a random access memory (hereinafter referred to as “RAM”) used as a work memory when the CPU 11 executes various processes.

Reference numeral 16 denotes a display device for displaying the processing result of the CPU 11 and the like as an image. Similarly, 17 is a printer for printing out the processing result on paper and outputting it. A bus 18 connects the CPU 11 to the keyboard 12 and the like.

The operation of the above configuration will be described with reference to FIG.

In FIG. 1, (1) input data storage processing, (2) data conversion processing and list creation processing,
(3) Conversion table and list correction processing, (4) statistical analysis processing, (5) data inverse conversion processing, (6) analysis result output processing are sequentially executed. Hereinafter, these processes will be described in order.

(1) Input Data Storing Process In this input data storing process, a process of storing the input data in the magnetic disk 13 is executed (step S
1). The data input process is performed by the user using the keyboard 12, for example. This input data has a structure in which category data is arranged in a table format.

FIG. 3 is a diagram showing an example of input data. The input data shown in the figure is the data of the quantification class 1,
For example, it is data for predicting the distribution amount of newspapers to station shops and the like. This data includes 4 items: "weather", "day of the week", "game of baseball of specific team the day before" (hereinafter referred to as "game of baseball of the day before"), "presence or absence of horse race on the previous day".
It consists of one item. The category data for each item is represented by non-quantitative data. For example, the category data of “weather” is represented by three non-quantity data of “fine”, “cloudy”, and “rainy”.

Such input data is transferred to the magnetic disk 13
In the case of storing in, each row is stored as one record. Therefore, in the case of FIG. 3, the number of records is R
There will be 5 from 1 to R5.

When the storage process of the input data is completed, the process of converting the non-quantity data into the code data and the process of creating the list used for correcting the conversion result are executed. These processes are executed in the form of simultaneous parallel progress.

(2) Data conversion process and list creation process In these processes, first, a process of transferring one record of data from the magnetic disk 13 to the RAM 15 is executed (step S2). This transfer processing is performed from the record on the head side. Therefore, in this case, the record R1 is stored in the RAM 15 as shown in FIG.

When this transfer processing is completed, the RAM 15 is updated to 1
A process of reading the non-quantity data for the item and determining whether or not this data is the data that first appears (step S3).

This determination processing is performed from the head item. Therefore, in this case, with respect to the non-quantity data “fine” of the item “weather”, a process of determining whether or not this is the data that first appears. In the case of the present example, this non-quantitative data “fine” is the data that first appeared, so it is determined that the data first appeared.

When such a determination result is obtained, a process of assigning code data to the non-quantity data "fine" which is the determination target is executed (step S4). This allocation process is performed, for example, from the magnetic disk 13 to the RAM 1
5 is carried out using the list previously transferred.

FIG. 5 is a diagram showing an example of the structure of this list. As shown in the figure, this list consists of a column for code data and a column for each item. For each item, the correspondence relationship between code data and non-quantitative data and the appearance frequency of this non-quantitative data are shown. Has become. In the column of code data, a plurality of code data, for example, “A”,
“B”, “C”, “D”, “E”, ... Are written.

In such a configuration, the above-mentioned allocation processing is performed by writing the non-quantity data "fine" in the column of the non-quantity data "weather". For example, when the code data “A” is assigned to the non-quantity data “fine”, the non-quantity data “fine” is written in the same line as the code data “A”. This state is shown in FIG.

By executing such allocation processing, the non-quantity data "fine" is stored in a state corresponding to the code data "A".

When the code data allocation process is completed,
A process of incrementing the appearance frequency of the non-quantity data “fine” by 1 is executed (step S5). This process also
This is done using the list above. As a result, the appearance frequency of the non-quantity data “fine” becomes “1” as shown in FIG.

When the increment process is completed, a process of converting the non-quantity data "fine" into the code data A is executed (step S6). As a result, the non-quantity data “fine” of the item “weather” of the record R1 stored in the RAM 15 is converted into code data “A” as shown in FIG. 4B.

When this conversion process is completed, a process for determining whether or not the above-described process has been repeated for the number of items is executed (step S7). In this case, the process is executed only for the first item "weather", so it is determined that the process is not repeated.

When such a determination result is obtained, the process returns to step S3, and the above-described processing is performed on the non-quantity data of the next item. So in this case,
The above-described processing is executed for the non-quantity data “month” of the item “day of the week”.

Similarly, the above-mentioned processing is executed for all the items of the record R1. As a result, the non-quantitative data of all the items of the record R1 are shown in FIG.
As shown in (c), both are converted into code data “A”. In addition, the appearance frequency of each non-quantity data is shown in FIG.
It is set to "1" as shown in (b).

After this, a process for determining whether or not the above-described process has been executed for all the records R1 to R5 is executed (step S8). In this case, since the process is executed only for the first record R1, it is determined that the process is not repeated.

When such a determination result is obtained, the process returns to step S2, and the above-described processing is executed for the next record. Therefore, in this case, the above-described processing is performed on the record R2.

The non-quantity data of the items "weather" and "presence or absence of horse race on the previous day" of the record R2 are the same "clear" and "present" as in the record R1. Therefore, in this case, the determination result of step S3 is not the first appearing data.

When such a determination result is obtained, it is not necessary to execute the code data allocation process (step S5). Therefore, in this case, only the appearance frequency increment process (step S5) and the data conversion process (step S6) are executed.

On the other hand, the non-quantity data of the items "day of the week" and "game of baseball day before" is different from the record R1. As a result, in this case, the code data allocation process (step S5) is executed.

From the above, when the processing of the record R2 is completed, the contents converted into the code data are as shown in FIG. 4 (d), and the list is as shown in FIG. 6 (c).

When the processing of the record 2 is completed, the above-described processing is sequentially executed on the records R3, R4 and R5. As a result, a conversion table of input data as shown in FIG. 4 (e) is obtained. Moreover, a list as shown in FIG. 7A is obtained.

As described above, the data conversion process and the list creation process (correctly, the entry process) are completed. When these processes are completed, the conversion table and list modification processing is started.

(3) Conversion Table and List Correction Process In this correction process, first, a list output process is executed (step S9). This output processing is performed, for example, by supplying the list data stored in the RAM 15 to the display device 16 and the printer 17. As a result, the list is provided to the user in the form of images and paper.

When the list is provided to the user, the user performs the same data grouping work and the low appearance frequency data grouping work, for example. Here, the content and necessity of these operations will be described.

First, although the same data grouping work of (1) can be regarded as having the same meaning for each item, a plurality of non-quantitative data represented by different character strings (hereinafter referred to as “identical data”). This is the work of detecting "data") and selecting one non-quantitative data (hereinafter referred to as "specific data") from these.

If the same data as described above exists,
The analysis accuracy of the statistical analysis processing described below may decrease. In order to solve this problem, it is possible to replace the same data with specific data, for example. Therefore,
In this embodiment, the user is made to perform the same data detection work and specific data selection work, and based on this work result,
The same data is automatically replaced with specific data inside the device.

In the list of FIG. 7 (a), non-display data "none" and "none" of the item "baseball game of the previous day"
Can be considered to have the same meaning but different strings. Therefore, in this case, the non-quantity data “none” and “none” are detected as the same data. Then, one of the two non-quantity data “none” and “none”, for example, the non-quantity data “none” is selected as the specific data. As a result, non-quantity data "none"
Are automatically replaced with non-quantity data “none” inside the device.

Next, the low-frequency data grouping work is to detect non-quantitative data (hereinafter referred to as "low-frequency data") having extremely low occurrence frequency for each item, and perform similar non-quantitative data. This is an operation of detecting data (hereinafter referred to as “similar data”).

The presence of such low-frequency data makes it easy to include an error in the analysis result in the case of quantification classification. In order to solve this problem, it is possible to replace low-frequency data with similar data, for example. Therefore, in this embodiment, the user is made to perform the detection work of low-frequency data and the detection of similar data, and based on the result of this work, the low-frequency data is automatically replaced with similar data. There is.

In the list of FIG. 7E, the low frequency data as described above does not exist. Therefore, suppose now that the non-quantitative data of the item "weather" is "clear", "cloudy",
The appearance frequency of "rain" is "100", "200",
The above work will be described as "1".

In this case, the appearance frequency of the non-quantity data "rain" is extremely lower than the appearance frequencies of the other non-quantity data "clear" and "cloudy". Therefore, in this case, the non-quantity data “rain” is detected as the low frequency data. The non-quantity data "rain" is more similar to "cloudy" than "clear". Therefore, in this case, the non-quantity data “cloudy” is detected as similar data. As a result, the non-quantity data “rain” is automatically replaced with “cloudy” inside the device.

The above operations may be automatically performed inside the apparatus. However, in such a case, the load on the device becomes extremely large, and it is not always possible to perform highly accurate work. Further, if it is performed inside the device, it becomes difficult to give the working accuracy a degree of freedom. Therefore, in this embodiment, the above-mentioned work is executed by the user.

When the above-mentioned work is completed, the work of inputting a correction command for correcting the conversion table and the list is executed based on the work result. This input work is performed by the user using the keyboard 12.

At this time, the CPU 11 is executing a process of determining whether or not a correction command is input (step S10). As a result, when a correction command is input,
The process designated by this modification command is executed (step S11).

In this case, if the modification instruction is related to the same data grouping work, a process of replacing the same data with specific data is executed on the conversion table. Also,
On the list, a process of deleting low-frequency data, a process of clearing the appearance frequency of low-frequency data, and a process of incrementing the appearance frequency of similar data are executed.

In the above example, the conversion table is converted from the state shown in FIG. 4E to the state shown in FIG. That is, the code data D of the item "game of baseball on the day before" is rewritten to the code data C.

Further, the list is converted from the state shown in FIG. 7 (e) to the state shown in FIG. 7 (b). That is, the non-quantity data “none” in the item “baseball game of the previous day” is deleted. Further, the appearance frequency of the non-quantity data “none” is cleared, and conversely, the appearance frequency of the non-quantity data is incremented from “1” to “2”.

Although details are omitted, similarly, when the correction instruction is related to the low-frequency data grouping work, the code data is rewritten on the conversion table, and the low-frequency data is displayed on the list. Is deleted and the appearance frequency is rewritten.

When this correction process is completed, a process of transferring the corrected list to the magnetic disk 13 is executed (step S12). When this transfer process is completed, a statistical analysis process is performed (step S13). If there is no correction instruction, the correction process of step S11 is not executed, but the transfer process of step S12 is executed.

(4) Statistical Analysis Processing This statistical analysis processing is executed using the conversion table left in the RAM 15. Therefore, this statistical analysis processing is performed using the code data. This allows this process to be executed with a small memory space and computer power.

When this statistical analysis process is completed, a data inverse conversion process for returning the code data included in this analysis result to the non-quantity data is executed (step S14).

(5) Data Reverse Conversion Processing This data reverse conversion processing is executed with reference to the list stored in the magnetic disk 13. When this inverse conversion process ends, a process of outputting the analysis result is executed (step S15).

(6) Analysis Result Output Processing This analysis result output processing is performed by supplying the analysis result to the display device 16 and the printer 17.
Thereby, the user can obtain the analysis result by the image and the paper. In this case, various methods such as tables and graphs are used as the method of displaying the analysis result.

When this output process is completed, the analysis result is converted to R
A process of transferring from the AM 15 to the magnetic disk 13 is executed (step S16). With the above, the processing is completed.

According to this embodiment described in detail above, the following effects can be obtained.

(1) First, since a function for converting non-quantitative data into code data and a function for converting code data into non-quantitative data are provided inside the apparatus, statistical analysis processing is executed based on the code data. be able to. Further, data input work and analysis result analysis work can be performed based on non-quantity data.

As a result, the statistical analysis processing can be executed with a small memory space and computer power. Further, it is possible to reduce mistakes of the user in the data input work and the analysis result analysis work.

(2) Since the function of providing the list in which the predetermined items are described and the function of correcting the conversion table and the list are provided, the accuracy of the processing result of the statistical analysis processing can be improved.

(3) Further, by providing the list to the user, the user can perform the same data grouping work and the infrequent data grouping work. Therefore, the conversion table and the list can be corrected. This can be performed accurately with a simple configuration, and the degree of freedom of this correction can be increased.

(4) Further, since the blank table is prepared in advance and the data is written in this table in parallel with the data conversion processing, the data conversion processing is completed immediately. In addition, it is possible to provide the user with a list in which predetermined items are entered.

Although one embodiment of the present invention has been described in detail above, the present invention is not limited to such an embodiment.

(1) For example, in the above embodiment, the case where the list is modified based on the modification command has been described. However, if this list is used only for the data reverse conversion process, the data reverse conversion process can be performed without modification. Therefore, in this case, it is not necessary to modify the list.

(2) Further, in the above embodiment, the case where the blank table is formed in advance has been described. However, in the present invention, the list may be formed from the beginning after the data conversion process is completed.

(3) Further, in the above embodiment, the case where the user is made to perform various works for correcting the conversion result has been described. However, in the present invention, this work may be automatically performed inside the device.

(4) In the above embodiment, the list is provided to the user prior to the statistical analysis processing.
The case of modifying the conversion table has been described. However, when the correction function of the conversion table is omitted, the provision function of the list may be omitted.

(5) In addition to this, it goes without saying that the present invention can be variously modified and implemented without departing from the scope of the invention.

[0081]

As described above in detail, according to the present invention, statistical analysis processing can be performed with a small memory space and computer power, and a user's burden and mistakes can be reduced. The purpose is to provide a device.

[Brief description of drawings]

FIG. 1 is a flow chart for explaining the operation of an embodiment of the present invention.

FIG. 2 is a block diagram showing the configuration of an embodiment of the present invention.

FIG. 3 is a diagram showing an example of an input data table.

FIG. 4 is a diagram for explaining a conversion table generation operation.

FIG. 5 is a diagram showing an example of a list.

FIG. 6 is a diagram illustrating a list generation operation.

FIG. 7 is a diagram for explaining a list generation operation.

[Explanation of symbols]

11 ... CPU, 12 ... Keyboard, 13 ... Magnetic disk, 14 ... ROM, 15 ... RAM, 16 ... Display device, 17 ... Printer, 18 ... Bus.

Claims (2)

[Claims] 1. A statistical analysis device for performing statistical analysis processing on a plurality of non-quantitative data arranged in a tabular form based on a quantification classification method, and data input means for inputting the non-quantitative data, Data conversion means for converting the non-quantitative data input by the data input means into code data, and statistical analysis processing based on the quantified classification method for the code data converted by the data conversion means. Statistical analysis means, data inverse conversion means for returning the code data included in the processing result of the statistical analysis means to the non-quantity data, and data for outputting the processing result subjected to the conversion processing of the data inverse conversion means A statistical analysis device comprising: an output unit. 2. Appearance frequency detection means for detecting the appearance frequency of the non-quantity data input by the data input means, the non-quantity data input by the data input means, and conversion by the data conversion means. List output means for outputting a list showing the correspondence between the code data and the appearance frequencies detected by the appearance frequency detection means, and a correction for correcting the code data converted by the data conversion means A correction command input unit for inputting a command; and a data correction unit for correcting the code data converted by the data conversion unit based on the correction command input by the correction command input unit. The statistical analysis device according to claim 1.