KR102354725B1 - Apparatus and method for classifying data for process mining - Google Patents

Abstract

The present invention relates to an apparatus and a method for classifying data. The apparatus receives an unclassified data table and a mapping schema data table, extracts a predetermined number of unclassified sample tables from the unclassified data table, converts each of the predetermined number of unclassified sample tables into an image to generate a predetermined number of unclassified images, extracts a preset number of mapping sample tables from the mapping schema data table, generates a preset number of mapping images by converting each of the preset number of mapping sample tables into an image, calculates the mapping accuracy between the preset number of unclassified images and the preset number of mapping images using a convolutional neural network, and classifies the data of the unclassified data table based on the mapping accuracy to classify the data for checking a relation schema.

Description

Apparatus and method for classifying data for process mining

The following embodiments relate to a technique for classifying data in an unclassified data table to check a relation schema.

Recently, interest in the introduction of a process-aware information system (PAIS) is increasing in the management and operation management areas of various industrial fields. Accordingly, the importance of large-capacity business, process, and execution logs and process and mining to discover useful knowledge from them is also being emphasized.

Process mining is a study that extracts useful knowledge from event logs provided by information systems. It provides new techniques for discovering, monitoring, and improving processes. can be applied to

By using process 　 mining, useful information can be discovered based on the business processing records that occur in business 　 processes, and the information discovered through 　 process 　 mining can be used for business, process, and innovation of the company. With the development of internet and computing technology and the increase of data, it is expected that the field and market size to which process and mining are applied will gradually expand.

Existing process mining methodologies start with the assumption that there is an event log. The event log is created by converting the event data extracted from the information system into an MXML or XES file through the event mapping process, and domain knowledge is required in this process.

Therefore, in the process of mapping event data to the event log, an expert who knows the domain of the corresponding event data checks the relation schema of the event data based on this and maps the event data to the event log.

In the face of increasing complexity and data volume, a method for mapping event data into event logs without the help of experts with domain knowledge is required.

An object of the present invention is to classify data included in an unclassified data table so that a relation schema of data can be checked.

A data classification apparatus according to an embodiment of the present invention includes: a receiver configured to receive an unclassified data table and a mapping schema data table; an unclassified sample extraction unit for extracting a preset number of unclassified sample tables from the unclassified data table; an unclassified image generator configured to convert each of the preset number of unclassified sample tables into images to generate a preset number of unclassified images; a mapping sample extraction unit for extracting a preset number of mapping sample tables from the mapping schema data table; a mapping image generator for generating a preset number of mapping images by converting each of the preset number of mapping sample tables into an image; a mapping calculator configured to calculate mapping accuracy between the preset number of unclassified images and the preset number of mapped images using a convolutional neural network; and a classifier configured to classify data of the unclassified data table based on the mapping accuracy to check a relation schema.

In this case, the unclassified image generating unit converts the predetermined number of unclassified sample tables into a relational frequency matrix to generate an unclassified relational frequency matrix, and an image using the value of the unclassified relational frequency matrix as a color value of a pixel can be converted to an unclassified image.

In this case, the non-classification relationship frequency matrix may be a matrix counting the number of non-redundant data related to data included in a specific field included in the non-classification sample table.

At this time, the mapping image generation unit converts each of the predetermined number of mapping sample tables into a relation frequency matrix to generate a mapping relation frequency matrix, and converts the value of the mapping relation frequency matrix into an image using the color value of the pixel. You can create a mapping image.

In this case, the mapping relationship frequency matrix may be a matrix counting the number of non-overlapping data related to data included in a specific field included in the mapping sample table.

In this case, if the mapping accuracy indicating a matching probability between the unclassified image and the mapping image is equal to or greater than a preset mapping value, the classification unit is configured to use the relation schema information of the mapping sample table corresponding to the mapping image to determine the ratio You can check the relation schema by classifying the data in the classification data table.

In this case, the relation schema information of the mapping sample table is information for classifying each field of the mapping sample table, and may be one of a case ID, a timestamp, a resource, and an attribute.

A method for classifying data in a data classification apparatus according to an embodiment of the present invention includes: receiving an unclassified data table and a mapping schema data table; extracting a predetermined number of unclassified sample tables from the unclassified data table; generating a preset number of unclassified images by converting each of the preset number of unclassified sample tables into images; extracting a preset number of mapping sample tables from the mapping schema data table; generating a preset number of mapping images by converting each of the preset number of mapping sample tables into an image; calculating mapping accuracy between the preset number of unclassified images and the preset number of mapped images using a convolutional neural network; and classifying data of the unclassified data table based on the mapping accuracy to check a relation schema.

In this case, the step of generating the preset number of unclassified images by converting each of the preset number of unclassified sample tables into images includes converting the preset number of unclassified sample tables into a relational frequency matrix to form a non-classified relationship generating a frequency matrix; and generating an unclassified image by converting a value of the unclassified relational frequency matrix into an image having a color value of a pixel.

In this case, the step of generating a predetermined number of mapping images by converting each of the predetermined number of mapping sample tables into images includes converting each of the predetermined number of mapping sample tables into a relation frequency matrix to generate a mapping relation frequency matrix to do; and converting a value of the mapping relation frequency matrix into an image using a color value of a pixel to generate a mapping image.

In this case, the step of classifying the data of the unclassified data table based on the mapping accuracy and checking the relation schema may include, if the mapping accuracy indicating a matching probability between the unclassified image and the mapping image is greater than or equal to a preset mapping value , the data of the unclassified data table may be classified using relation schema information of the mapping sample table corresponding to the mapping image.

In this case, the relation schema information of the mapping sample table is information for classifying each field of the mapping sample table, and may be one of a case ID, a timestamp, a resource, and an attribute.

In the present invention, the relation schema of the unclassified data table can be identified by extracting a sample table from each of the unclassified data table and the mapping schema data table, and classifying the data by comparing the image of the sample table.

1 is a diagram illustrating a configuration of a data classification apparatus for classifying data of an unclassified data table according to an embodiment of the present invention.
2 is a flowchart illustrating a process of classifying data in an unclassified data table in a data classification apparatus according to an embodiment of the present invention.
3 is a diagram illustrating an example of a mapping schema data table according to an embodiment of the present invention.
4 is a diagram illustrating an example of converting a mapping sample table into a mapping relation frequency matrix according to an embodiment of the present invention.
5 is a diagram illustrating an example of converting a mapping relation frequency matrix into a mapping image according to an embodiment of the present invention.

Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. However, since various changes may be made to the embodiments, the scope of the patent application is not limited or limited by these embodiments. It should be understood that all modifications, equivalents and substitutes for the embodiments are included in the scope of the rights.

The terms used in the examples are used for the purpose of description only, and should not be construed as limiting. The singular expression includes the plural expression unless the context clearly dictates otherwise. In the present specification, terms such as "comprise" or "have" are intended to designate that a feature, number, step, operation, component, part, or combination thereof described in the specification exists, but one or more other features It should be understood that this does not preclude the existence or addition of numbers, steps, operations, components, parts, or combinations thereof.

Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which the embodiment belongs. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and should not be interpreted in an ideal or excessively formal meaning unless explicitly defined in the present application. does not

In addition, in the description with reference to the accompanying drawings, the same components are given the same reference numerals regardless of the reference numerals, and the overlapping description thereof will be omitted. In describing the embodiment, if it is determined that a detailed description of a related known technology may unnecessarily obscure the gist of the embodiment, the detailed description thereof will be omitted.

In addition, in describing the components of the embodiment, terms such as first, second, A, B, (a), (b), etc. may be used. These terms are only for distinguishing the elements from other elements, and the essence, order, or order of the elements are not limited by the terms. When it is described that a component is “connected”, “coupled” or “connected” to another component, the component may be directly connected or connected to the other component, but another component is between each component. It will be understood that may also be "connected", "coupled" or "connected".

Components included in one embodiment and components having a common function will be described using the same names in other embodiments. Unless otherwise stated, descriptions described in one embodiment may be applied to other embodiments as well, and detailed descriptions within the overlapping range will be omitted.

Hereinafter, an apparatus and method for classifying data for process mining according to an embodiment of the present invention will be described in detail with reference to FIGS. 1 to 5 .

1 is a diagram illustrating a configuration of a data classification apparatus for classifying data of an unclassified data table according to an embodiment of the present invention.

Referring to FIG. 1 , the data classification apparatus 100 includes a receiver 101 , an unclassified sample extractor 102 , a mapping sample extractor 103 , an unclassified image generator 104 , and a mapping image generator 105 . ), a mapping calculation unit 106 and a classification unit 107 may be included.

The receiver 101 receives the unclassified data table 110 and the mapping schema data table 120 .

Here, the unclassified data table 110 is configured in the form of a table of event data for which a relation schema is not defined.

And, the mapping schema data table 120 is configured in the form of a table of event logs in which the relation schema is already defined, and mapping from event data to event logs is completed in advance.

The unclassified data table 110 and the mapping schema data table 120 may be data generated in the same business process process, but may be data generated in a similar business process process although not the same.

The mapping sample extraction unit 103 extracts a preset number of mapping sample tables from the mapping schema data table 120 .

3 is a diagram illustrating an example of a mapping schema data table according to an embodiment of the present invention.

Referring to FIG. 3 , the mapping sample extractor 103 selects a preset number of fields 310 , 320 , and 330 from the mapping schema data table 120 , and selects a preset number of fields 310 , 320 , and 330 from the selected fields 310 , 320 , 330 . A preset number of mapping sample tables 340 of a set size are extracted.

In this case, the preset number of mapping sample tables extracted from the selected field may be set to, for example, 1000 to 2000.

The mapping image generator 105 generates a preset number of mapping images by converting each of the preset number of mapping sample tables into an image.

The mapping image generation unit 105 converts each of the predetermined number of mapping sample tables into a relation frequency matrix to generate a mapping relation frequency matrix, and converts the value of the mapping relation frequency matrix into an image using the color value of the pixel to convert the mapping image can create

In this case, the mapping relationship frequency matrix is a matrix counting the number of non-redundant data related to data included in a specific field included in the mapping sample table.

4 is a diagram illustrating an example of converting a mapping sample table into a mapping relation frequency matrix according to an embodiment of the present invention.

Referring to FIG. 4 , the mapping image generator 105 counts the number of non-overlapping data included in other fields related to data included in a specific field (the first field) included in the mapping sample table 410 , A mapping relation frequency matrix 420 that is the generated matrix is generated.

5 is a diagram illustrating an example of converting a mapping relation frequency matrix into a mapping image according to an embodiment of the present invention.

Referring to FIG. 5 , the mapping image generator 105 may generate the mapping image 520 by converting the value of the mapping relation frequency matrix 510 into an image using the color value of the pixel.

More specifically, the mapping image generating unit 105 may bind three values of the mapping relation frequency matrix 510 and convert it into a color of one pixel corresponding to the RGB color value, and the mapping relation frequency matrix 510 . A mapping image 520 may be generated by converting all values of α into pixel colors.

The unclassified sample extraction unit 102 extracts a preset number of unclassified sample tables from the unclassified data table 110 . In this case, the extraction of the unclassified sample table is performed in the same way as the extraction of the matching sample table.

The unclassified image generator 104 generates a preset number of unclassified images by converting each of the preset number of unclassified sample tables into images.

The unclassified image generating unit 104 converts each of a preset number of unclassified sample data into a relational frequency matrix to generate an unclassified relational frequency matrix, and an image in which the value of the unclassified relational frequency matrix is the color value of a pixel. It can be converted to create an unclassified image.

In this case, the non-classification relationship frequency matrix is a matrix counting the number of non-redundant data related to data included in a specific field included in the non-classification sample table.

In addition, the non-classification relation frequency matrix is generated in the same manner as the mapping relation frequency matrix, and the non-classification image is generated in the same manner as the matching image.

The mapping calculator 106 calculates mapping accuracy between a preset number of unclassified images and a preset number of mapping images using a convolutional neural network (CNN).

More specifically, the mapping calculation unit 106 inputs a preset number of unclassified images and a preset number of mapping images to the convolutional neural network to check the mapping image having a similarity with the unclassified image equal to or greater than a preset probability, and non-classification By counting the mapping images having a similarity with the image equal to or greater than a preset probability, the retrieved probability is calculated as the mapping accuracy between the image and the preset number of mapping images.

The classification unit 107 may classify the data of the unclassified data table based on the mapping accuracy to check the relation schema. In this case, the classifier 107 may be included in the convolutional neural network.

More specifically, if the mapping accuracy indicating the matching probability between the unclassified image and the mapped image is greater than or equal to a preset mapping value, the classification unit 107 uses the relation schema information of the mapping sample table corresponding to the mapping image to the unclassified data table. You can check the relation schema by classifying the data of

In this case, the relation schema information of the mapping sample table is information for classifying each field of the mapping sample table, and may be one of a case ID, a timestamp, a resource, and an attribute.

2 is a flowchart illustrating a process of classifying data in an unclassified data table in a data classification apparatus according to an embodiment of the present invention.

Referring to FIG. 2 , the data classification apparatus 100 receives an unclassified data table and a mapping schema data table ( 210 ).

Then, the data classification apparatus 100 extracts a preset number of unclassified sample tables from the unclassified data table ( 212 ).

Then, the data classification apparatus 100 converts a predetermined number of unclassified sample tables into relational frequency metrics and generates them as nonclassified relational frequency metrics ( 214 ).

In this case, the non-classification relationship frequency matrix is a matrix counting the number of non-redundant data related to data included in a specific field included in the non-classification sample table.

Then, the data classification apparatus 100 generates an unclassified image by converting the value of the unclassified relational frequency matrix into an image having a pixel color value ( 216 ).

Meanwhile, the data classification apparatus 100 extracts a preset number of mapping sample tables from the mapping schema data table ( 218 ).

Then, the data classification apparatus 100 converts each of the predetermined number of mapping sample tables into a relational frequency matrix to generate a mapping relational frequency matrix ( 220 ).

In this case, the mapping relationship frequency matrix is a matrix counting the number of non-overlapping data related to data included in a specific field included in the mapping sample table.

Then, the data classification apparatus 100 generates a mapping image by converting the value of the mapping relation frequency matrix into an image using the color value of the pixel ( 222 ).

Here, steps 212 to 216 and 218 to 220 may be simultaneously processed in parallel, but steps 212 to 216 or 218 to 220 may be performed first.

Then, the data classification apparatus 100 calculates the mapping accuracy between the preset number of unclassified images and the preset number of mapped images using the convolutional neural network ( 224 ).

And, if the mapping accuracy indicating the matching probability between the unclassified image and the mapped image is equal to or greater than a preset mapping value, the data classification apparatus 100 uses the relation schema information of the mapping sample table corresponding to the mapping image to the unclassified data table. Classify the data of (226).

In this case, the relation schema information of the mapping sample table is information for classifying each field of the mapping sample table, and may be one of a case ID, a timestamp, a resource, and an attribute.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the embodiment, or may be known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic such as floppy disks. - includes magneto-optical media, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

Software may comprise a computer program, code, instructions, or a combination of one or more thereof, which configures a processing device to operate as desired or is independently or collectively processed You can command the device. The software and/or data may be any kind of machine, component, physical device, virtual equipment, computer storage medium or apparatus, to be interpreted by or to provide instructions or data to the processing device. , or may be permanently or temporarily embody in a transmitted signal wave. The software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored in one or more computer-readable recording media.

As described above, although the embodiments have been described with reference to the limited drawings, those skilled in the art may apply various technical modifications and variations based on the above. For example, the described techniques are performed in an order different from the described method, and/or the described components of the system, structure, apparatus, circuit, etc. are combined or combined in a different form than the described method, or other components Or substituted or substituted by equivalents may achieve an appropriate result.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

100: data classification device
101: receiver
102: unclassified sample extraction unit
103: mapping sample extraction unit
104: unclassified image generation unit
105: mapping image generation unit
106: mapping calculator
107: classification unit
110: unclassified data table
120: mapping schema data table

Claims (13)

a receiver for receiving the unclassified data table and the mapping schema data table;
an unclassified sample extraction unit for extracting a preset number of unclassified sample tables from the unclassified data table;
an unclassified image generator configured to convert each of the preset number of unclassified sample tables into images to generate a preset number of unclassified images;
a mapping sample extraction unit for extracting a preset number of mapping sample tables from the mapping schema data table;
a mapping image generator for generating a preset number of mapping images by converting each of the preset number of mapping sample tables into an image;
a mapping calculator configured to calculate mapping accuracy between the preset number of unclassified images and the preset number of mapped images using a convolutional neural network; and
A classification unit that classifies data in the unclassified data table based on the mapping accuracy to check a relation schema
A data classification device comprising a.
According to claim 1,
The unclassified image generation unit,
An unclassified relationship frequency matrix is generated, which is a matrix generated by counting the number of non-duplicate data related to data included in a specific field included in the predetermined number of unclassified sample tables, and the value of the unclassified relationship frequency metric Converts to an image with pixel color values to create an unclassified image.
data classification device.
delete According to claim 1,
The mapping image generation unit,
Counting the number of non-overlapping data related to data included in a specific field included in each of the predetermined number of mapping sample tables to generate a mapping relationship frequency matrix, which is a matrix generated respectively, and the value of the mapping relationship frequency matrix Converting to an image with pixel color values to create a mapping image
data classification device.
delete According to claim 1,
The classification unit,
When the mapping accuracy indicating the matching probability between the unclassified image and the mapping image is equal to or greater than a preset mapping value, the data of the unclassified data table is obtained using relation schema information of the mapping sample table corresponding to the mapping image. Classifying and checking the relation schema
data classification device.
7. The method of claim 6,
The relation schema information of the mapping sample table is,
As information for classifying each field of the mapping sample table, one of a case ID, a timestamp, a resource, and an attribute
data classification device.
receiving an unclassified data table and a mapping schema data table;
extracting a predetermined number of unclassified sample tables from the unclassified data table;
generating a preset number of unclassified images by converting each of the preset number of unclassified sample tables into images;
extracting a preset number of mapping sample tables from the mapping schema data table;
generating a preset number of mapping images by converting each of the preset number of mapping sample tables into an image;
calculating mapping accuracy between the preset number of unclassified images and the preset number of mapped images using a convolutional neural network; and
Classifying the data of the unclassified data table based on the mapping accuracy to check the relation schema
A method of classifying data in a data classification device comprising a.
9. The method of claim 8,
The step of generating the predetermined number of unclassified images by converting each of the predetermined number of unclassified sample tables into images,
generating an unclassified relationship frequency matrix, which is a matrix generated by counting the number of non-redundant data related to data included in a specific field included in the preset number of unclassified sample tables; and
generating an unclassified image by converting the value of the unclassified relational frequency matrix into an image having a color value of a pixel
A method of classifying data in a data classification device comprising a.
9. The method of claim 8,
The step of generating a preset number of mapping images by converting each of the preset number of mapping sample tables into an image,
generating a mapping relationship frequency matrix, which is a matrix generated by counting the number of non-overlapping data related to data included in a specific field included in each of the preset number of mapping sample tables; and
generating a mapping image by converting the value of the mapping relation frequency matrix into an image having a color value of a pixel
A method of classifying data in a data classification device comprising a.
9. The method of claim 8,
The step of classifying the data of the unclassified data table based on the mapping accuracy to check the relation schema,
When the mapping accuracy indicating the matching probability between the unclassified image and the mapping image is equal to or greater than a preset mapping value, the data of the unclassified data table is obtained using relation schema information of the mapping sample table corresponding to the mapping image. to classify
A method of classifying data in a data classification device.
12. The method of claim 11,
The relation schema information of the mapping sample table is,
As information for classifying each field of the mapping sample table, one of a case ID, a timestamp, a resource, and an attribute
A method of classifying data in a data classification device.
13. A computer-readable recording medium in which a program for executing the method of any one of claims 8 to 12 is recorded.

Images