WO2017104863A1

WO2017104863A1 - System and method for conversion between heterogeneous data stored through data collection api

Info

Publication number: WO2017104863A1
Application number: PCT/KR2015/013842
Authority: WO
Inventors: 조용상; 이재호
Original assignee: 한국교육학술정보원
Priority date: 2015-12-16
Filing date: 2015-12-17
Publication date: 2017-06-22
Also published as: KR101719926B1

Abstract

The present invention relates to a system and method for conducting exchange between heterogeneous data collected through a data collection API, and is capable of carrying out semantic matching and mapping with classes and properties of learning activity data by using a first data collection API, a first storage area, and data mapping and matching instances which convert first learning activity data into learning activity data according to a second data profile that is different from a first data profile.

Description

System and method for converting between heterogeneous data stored through data collection API

The present invention relates to a system and method for converting heterogeneous data collected through a data collection API. More specifically, the present invention relates to structural and semantic design of transformation rules between heterogeneous data generated from various platforms and software for learning analysis. A system and method for converting.

Terminals such as personal computers, laptops, mobile phones, etc. may be configured to perform various functions. Examples of such features include a variety of application-driven features, web browsing capabilities, data and voice communications capabilities, the ability to take photos or videos with the camera, voice recording capabilities, playback of music and audio files through the speaker system, images and videos Display function. Some terminals include additional functionality to play games, while others are implemented as multimedia devices. Moreover, recent terminals can receive a broadcast or multicast signal to watch a video or television program. All these functions can be used as a function for learning in the field of education.

Meanwhile, as smartphones and social network services (SNS) are becoming more common, the types and amounts of data generated by billions of users around the world are growing at an explosive rate. The focus is on data utilization. The government is also in line with global trends by presenting a new government operating paradigm to open up public data and foster new industries based on data through government 3.0 policy.

In particular, big data is more like a picture that expresses individual needs and actions, rather than data representing a group, so it is expected to be very useful in all fields for personalized customized service. In the field of education, learning analytics technology, which collects and analyzes data related to learning, intentionally or unintentionally, has begun to be applied to big data applications.

However, the reality is that it is difficult to collect data systematically and accurately because the learning environment of users to collect data, that is, the platform and the software, is very diverse. To solve this problem, some standardization organizations have established a systematic data collection system by standardizing data profiles and data collection APIs for data collection.

On the other hand, because standardized data collection systems are composed of heterogeneous structures and APIs, in some situations, an educational institution or service provider may face a situation where heterogeneous data collection standards are used.

In this specification, we propose a rule, procedure, and method for securing interoperability of heterogeneous data collection systems and data collected by APIs.

An object of the present invention is to provide interoperability between learning data collected and stored from heterogeneous data collection systems and APIs.

Specifically, the present invention defines a structural syntactic mapping method through systematic identification steps and procedures for a data profile of a heterogeneous data collection system, and also provides a system of homogeneous and similar meaning data through semantic mapping between collected data. The goal is to provide interoperability between data so that it can be recognized.

Technical problems to be achieved in the present invention are not limited to the technical problems mentioned above, and other technical problems not mentioned above will be clearly understood by those skilled in the art from the following description. Could be.

A data mapping and matching system according to an embodiment of the present invention for realizing the above object includes a first data collection API for collecting first learning activity data performed in a first learning environment according to a first data profile; A first storage storing first learning activity data collected by the first data collection API and first metadata which is metadata of the first learning activity data; A data mapping and matching instance for converting the first learning activity data into learning activity data according to a second data profile that is different from the first data profile, wherein the data mapping and matching instance is assigned to the first data profile. Meaning of the structure and syntactic mapping instance table corresponding to the class and attribute of the learning activity data according to the second data profile and the learning activity data according to the first data profile and the second The meaning of the learning activity data according to the data profile includes a semantic instance table corresponding to the ontology rule, grasps the data profile of the first learning activity data from the first metadata, and the structural and syntactic mapping instance. The first using the table Converts a class and an attribute of one learning activity data into a class and an attribute according to the second data profile, and uses the semantic instance table to convert the meaning of the first learning activity data according to the ontology rule. Can be converted to

In addition, a unique identification number may be assigned to each class and attribute of the structural and syntactic mapping instance table.

In addition, the data mapping and matching instance may add the unclassified class to the structural and syntactic mapping instance table when the class of the first learning activity data is not included in the structural and syntactic mapping instance table. Can be.

In addition, the data mapping and matching instance may add the non-included attribute to the structural and syntactic mapping instance table when the attribute of the first learning activity data is not included in the structural and syntactic mapping instance table. Can be.

Also, the first metadata may include at least one of a name of a data profile of first learning activity data stored in the first repository, a standard name and version of the data profile, and a location of the first repository.

In addition, the data mapping and matching instance, when the class or attribute of the first learning activity data does not correspond to the mapping instance table or the meaning of the first learning activity data does not correspond to the semantic instance table The exception may be made to the first learning activity data.

The apparatus may further include a second storage configured to store second learning activity data collected according to the second data profile, wherein the data mapping and matching instance is converted into learning activity data according to the second data profile. One learning activity data may be transmitted to the second repository.

In addition, the second store performs a suitability check to determine whether the received first transformed learning activity data has a class, an attribute, and a meaning according to the second data profile. The first learning activity data may be stored, and an exception may be thrown if not suitable.

On the other hand, the data mapping and matching method according to an embodiment of the present invention for realizing the above-described problem is structural and syntactic mapping by collecting the learning activity data using the first data collection API, the first storage and the data mapping and matching instance And a semantic matching method comprising: a first step of collecting, by a first data collection API, first learning activity data performed in a first learning environment according to a first data profile; A second step of storing, by the first repository, first metadata that is metadata of the first learning activity data and the first learning activity data; A third step of the data mapping and matching instance querying first metadata of the first learning activity data stored in the first repository to determine a data profile of the first learning activity data; A fourth step of the data mapping and matching instance receiving the first learning activity data from the first repository; A fifth step of the data mapping and matching instance converting the classes and attributes of the first learning activity data into classes and attributes according to a second data profile different from the first data profile; And a sixth step of the data mapping and matching instance converting the meaning of the first learning activity data into a meaning according to the second data profile, wherein the data mapping and matching instance is assigned to the first data profile. Meaning of the structure and syntactic mapping instance table corresponding to the class and attribute of the learning activity data according to the second data profile and the learning activity data according to the first data profile and the second Meaning of learning activity data according to a data profile includes a semantic instance table corresponding to an ontology rule, and in the fifth step, the data mapping and matching instance is transformed using the structural and syntactic mapping instance table. In the sixth step, the data mapping and Ching instance may be converted in accordance with said ontology rule using the semantic instance table.

Further, after the fourth step, if the class of the first learning activity data received by the data mapping and matching instance is not included in the structural and syntactic mapping instance table, the data mapping and matching instance is not included. Class can be added to the structural and syntactic mapping instance table.

Further, after the fourth step, if the attribute of the first learning activity data received by the data mapping and matching instance is not included in the structural and syntactic mapping instance table, the data mapping and matching instance is not included. May add attributes to the structural and syntactic mapping instance table.

The class or attribute of the first learning activity data does not correspond to the mapping instance table in the fifth step, or the meaning of the first learning activity data corresponds to the semantic instance table in the sixth step. If not, after the sixth step, the data mapping and matching instance may exception the first learning activity data.

In addition, after the sixth step, the data mapping and matching instance receives the second learning activity data collected according to the second data profile from the first learning activity data converted into learning activity data according to the second data profile. A seventh step of transmitting to the second storage to store; may further include.

In addition, after the seventh step, a suitability check is performed to determine whether the converted first learning activity data received by the second repository has a class, an attribute, and a meaning according to the second data profile. The method may further include storing the converted first learning activity data and, if not appropriate, exception processing.

The present invention is to systematically collect data generated by learning activities to provide interoperability between heterogeneous data utilized for accurate learning analysis.

Specifically, the process of mapping the structure and syntax of heterogeneous data and the semantic matching process for comparing and analyzing the meaning of each data are used to find, transform, and exchange the agreement or significance of the data to provide an accurate understanding of the data for learning analysis. Can provide.

Meanwhile, the effects obtainable in the present invention are not limited to the above-mentioned effects, and other effects not mentioned above will be clearly understood by those skilled in the art from the following description. Could be.

The following drawings, which are attached to this specification, illustrate one preferred embodiment of the present invention, and together with the detailed description thereof, serve to further understand the technical spirit of the present invention. It should not be construed as limited.

1 conceptually illustrates a binding structure in which heterogeneous data is commonly followed.

2 illustrates one embodiment of a heterogeneous data exchange system that may be implemented in accordance with the present invention.

3 is a flowchart illustrating an operation of a transformation system through learning data mapping and matching according to an embodiment of the present invention.

4 illustrates that heterogeneous data profiles are structurally and syntactically mapped according to an embodiment of the present invention.

FIG. 5 conceptually shows a data mapping and matching sequence composed of two steps when there are two data profiles for collecting data in an educational field according to an embodiment of the present invention.

6 illustrates the concept of converting data while managing entities and attributes of the data profile by the ontology mapping instance according to a predetermined rule when there are multiple data profiles for collecting data in the education field according to an embodiment of the present invention. Indicates.

First, data used in the system of the present invention is generally represented by a triple binding structure as shown in FIG.

1 conceptually illustrates the binding structure of data used in the system of the present invention.

For example, as shown in FIG. 1, the bound data includes "who, what, did," in addition to the direct data, "in what circumstances, what tool did you use, what data you created, If so, the number of points earned ”is also bound.

Hereinafter, with reference to the drawings will be described a preferred embodiment of the present invention. In addition, one Example described below does not unduly limit the content of this invention described in the Claim, and the whole structure demonstrated by this Embodiment is not necessarily required as a solution of this invention.

Hereinafter, with reference to the drawings will be described in detail the learning data collection and conversion system proposed by the present invention. 2 illustrates one embodiment of a process and system for generating, transmitting and storing training data that may be implemented in accordance with the present invention.

Referring to FIG. 2, the transformation system 100 through the training data mapping and matching according to the present invention includes a learning environment 10, a data profile 20, a data collection API 30, a data store 40, and a data store profile. 50, and data mapping and matching process 60, and the like.

However, since the components illustrated in FIG. 2 are not essential, the data mapping and matching system 100 having more components or fewer components may be implemented.

In addition, the components illustrated in FIG. 2 are connected to each other interdependently, and each component may be implemented separately or in combination with each other, as illustrated in FIG. 2.

Hereinafter, the components will be described in order.

The learning environment 10 may be composed of

various environments

11 and 12, and a user may generate learning data while participating in a learning activity by utilizing contents, services, web links, and software provided by the learning environment 10. do.

This training data is generated according to the standardized data profile 20. Examples of the data profile 20 include IMS Caliper Metric Profile 21, xAPIs Recipes 22, and the like. The data profile 20 may be used in various ways by educational institution, region, and country.

Data generated according to the data profile 20 is acquired, stored and transmitted by the data collection API 30, which also collects and stores

heterogeneous forms

31 and 32 according to the diversity of the data profile 20. Can be operated as. That is, data generated according to IMS Caliper Metric Profile 21 may be operated by IMS Caliper Sensor APIs 31, and data generated according to xAPIs Recipes 22 may be operated by xAPIs 32.

Data sent by the data collection API 30 is preserved in the data collection store 40, where data exchange and matching instances 60 need to be exchanged between the

stores

41 and 42 that preserve the heterogeneous data. A conversion request can be made to receive the converted data. The data mapping and matching instance 60 looks up the

metadata

51 and 52 to obtain a profile for the data profile and location applied to the requested repositories and transforms the data according to the conversion rules for the data profile to be converted. Afterwards, the data interoperability providing function is returned to the data storage 60.

A representative example of the system is a learning analysis service. In learning analytics services, service providers typically provide a learning environment (10) that operates on a variety of devices, and IMS Caliper (21) or xAPI (data collection schemes provided by standardization organizations such as the IMS Global Learning Consortium and Advanced Distributed Learning). Data is generated and stored using

data collection APIs

31 and 32 that partially or fully comply with standards such as 22).

The data collected may include session information, quiz and test results, reading activity history, assignment performance history, media utilization, etc., and these data are stored in the learning

data stores

41 and 42 and analyzed and then displayed in a dashboard form. The visualization information is processed and delivered to the user.

In the meantime, the data profile 20 uses different data profiles and vocabularies, and in particular, the xAPI standard encourages users to create and use their own profiles under the name xAPIs Recipes 22. Therefore, heterogeneous data is generated by the data collection API.

First, the procedure before heterogeneous data exchange is performed is described.

The structural and syntactic information of the data profile to be converted is registered in the data mapping and matching instance 60 (S110).

In other words, it corresponds to an identification scheme process that assigns an identification number, such as a URI, to classes and attributes of a data profile.

Next, structural and syntactic mapping between heterogeneous data profiles is performed (S120).

This is the process of structural and syntactic mapping for data transformation in the following steps.

FIG. 4 illustrates that heterogeneous data profiles are structurally and syntactically mapped according to an embodiment of the present invention. As illustrated in FIG. 4, the classes and attributes used in each data profile have unique identification values as URIs. Has an identification system that can be assigned and mapped to an N: M relationship.

Specifically, FIG. 5 conceptually illustrates a data mapping and matching sequence composed of two steps when there are two data profiles for collecting data in the education field according to an embodiment of the present invention. As described above, this relationship mapping maps the identification value 111 assigned to the class and the identification values 112 for the class-specific attributes into a hierarchical structure.

Next, semantic matching is performed (S130).

That is, in the following steps, the meanings of the vocabularies and sentences used in each data profile for data conversion are analyzed to match the degree of agreement and significance.

FIG. 6 illustrates a concept of converting data while managing classes and attributes of the data profile according to a predetermined rule when an ontology mapping instance has a plurality of data profiles for collecting data in an educational field according to an embodiment of the present invention. As shown in FIG. 5 and FIG. 6, for the semantic mapping, the instance table 140 using the mapping rule for the vocabulary used in each data profile is used.

Next, a procedure for collecting and storing data will be described.

First, the data collection API 30 is notified of the occurrence of the learning activity (S210).

When the user participates in learning, the learning environment 10 is used. When a learning activity is performed in the learning environment 10, the data collection API 30 is notified that data is generated in a synchronous or asynchronous manner.

Next, the data collection API 30 recognizing that the data is generated is bound to the generated data according to the data profile and stored in the temporary storage module (end-point) of the API (S220).

That is, the generated data is temporarily stored by binding the data according to the data profile.

Next, the data collection API 30 requests authentication to transmit the temporarily stored data to the data storage 40 (S310).

This is a procedure for authenticating between the data collection API 30 and the data store 40, and the data store 40 may approve the data collection API 30 when it is authenticated.

Thereafter, the data temporarily stored by the data collection API 30 is transmitted to the data storage 40 (S320).

Next, the data storage 40 may perform a conformity check to test whether the received data is created according to the data profile (S330). However, step S330 may be selectively performed.

Hereinafter, a process of converting data stored in a data store will be described.

When mutual data exchange is required in the storage 40 where heterogeneous data is stored, when transferring data from one storage 40 to another storage 40, or from several data storages 40 to one storage 40 When collecting data in a batch, first, the data store 40 makes a data conversion request to the data mapping and matching instance 60 (S410).

Next, the data mapping and matching instance 60 queries the profile metadata of the data store in order to grasp the data profile applied to each data store, transmission and reception location information, and the like before receiving the data (S420).

In other words, according to which data profile the data stored in the data store is generated, and where the data is stored.

Next, the data store 40 transmits the data requesting the transformation to the data mapping and matching instance 60 (S430).

Next, data received by the data mapping and matching instance 60 is converted into classes and attributes to be converted (S510).

The data mapping and matching instance 60 converts the received classes and attributes of the received data into classes and attributes to be converted using the structural and syntactic mapping instance tables.

Thereafter, the data mapping and matching instance 60 matches the meaning of the received data (S520).

The data mapping and matching instance 60 extracts a vocabulary and a sentence used in the received data and the transformed data, and matches the same meaning or meaning with the ontology rules of the instance table.

The data mapping and matching instance 60 may exception the received data when the mapping in step S510 or the matching in step S520 is not performed.

Next, the data mapping and matching instance 60 transmits the converted data to the data store 40 to be transmitted (S440).

Thereafter, the data store 40 checks the suitability of the received data, and stores the nonconforming data by exception processing (S450).

The data store 40 selectively performs suitability check on the received converted data, and stores an exception if it is not correctly mapped or semantically matched.

Meanwhile, new structures, attributes, and vocabularies may be received when data is transmitted from the data store 40 to the data mapping and matching instance 60 to perform the structural, syntactic mapping (S510) and semantic matching (S520) processes. In this case, the data mapping and matching instance 60 may update the semantic filter and mapper 113 illustrated in FIG. 6 using ontology rules.

Ontology rules can be updated in an automated manner using the structure, syntax, and vocabulary arrangement of initially constructed data and newly received data.

On the other hand, the present invention can also be embodied as computer readable codes on a computer readable recording medium. Computer-readable recording media include all kinds of recording devices that store data that can be read by a computer system. Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disks, optical data storage devices, and the like, which are also implemented in the form of carrier waves (for example, transmission over the Internet). Include. The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion. In addition, functional programs, codes, and code segments for implementing the present invention can be easily inferred by programmers in the art to which the present invention belongs.

The transformation system and method through the learning data mapping and matching described above may not be limitedly applied to the configuration and method of the above-described embodiments, and the embodiments may be modified in various ways so that various modifications may be made. Or some may be selectively combined.

Claims

In the system that collects learning activity data to perform structural and syntactic mapping and semantic matching,

A first data collection API that collects first learning activity data performed in a first learning environment according to a first data profile;

A first storage storing first learning activity data collected by the first data collection API and first metadata which is metadata of the first learning activity data;

And a data mapping and matching instance for converting the first learning activity data into learning activity data according to a second data profile different from the first data profile.

The data mapping and matching instance,

A structural and syntactic mapping instance table corresponding to classes and attributes of learning activity data according to the first data profile and classes and attributes of learning activity data according to the second data profile, and learning activity data according to the first data profile Meaning of the learning activity data according to the second data profile includes a semantic instance table corresponding to the ontology rule,

Grasp the data profile of the first learning activity data from the first metadata,

Converting the classes and attributes of the first learning activity data into classes and attributes according to the second data profile using the structural and syntactic mapping instance tables;

And a meaning of the first learning activity data is converted into a meaning according to the ontology rule according to the ontology rule using the semantic instance table.
The method of claim 1,

And a unique identification number is assigned to each of the classes and attributes of the structural and syntactic mapping instance table.
The method of claim 1,

The data mapping and matching instance,

If the class of the first learning activity data is not included in the structural and syntactic mapping instance table,

Adding the non-class to the structural and syntactic mapping instance table.
The method of claim 1,

The data mapping and matching instance,

If the attribute of the first learning activity data is not included in the structural and syntactic mapping instance table,

Adding the non-included attribute to the structural and syntactic mapping instance table.
The method of claim 1,

The first metadata includes at least one of a name of a data profile of first learning activity data stored in the first repository, a standard name and version of the data profile, and a location of the first repository. system.
The method of claim 1,

The data mapping and matching instance,

Exceptions to the first learning activity data, if the class or attribute of the first learning activity data does not correspond according to the mapping instance table or the meaning of the first learning activity data does not correspond according to the semantic instance table. Processing and data mapping and matching system.
The method of claim 1,

A second repository for storing second learning activity data collected according to the second data profile;

And the data mapping and matching instance transmits the first learning activity data converted into learning activity data according to the second data profile to the second repository.
The method of claim 7, wherein

The second reservoir,

A suitability check for determining whether the converted first learning activity data received has a class, an attribute, and a meaning according to the second data profile,

If appropriate, store the converted first learning activity data,

Data mapping and matching system, if appropriate.
1. A method of collecting learning activity data using a first data collection API, a first repository and a data mapping and matching instance to perform structural and syntactic mapping and semantic matching,

A first step of the first data collection API collecting first learning activity data performed in a first learning environment according to a first data profile;

A second step of storing, by the first repository, first metadata that is metadata of the first learning activity data and the first learning activity data;

A third step of the data mapping and matching instance querying first metadata of the first learning activity data stored in the first repository to determine a data profile of the first learning activity data;

A fourth step of the data mapping and matching instance receiving the first learning activity data from the first repository;

A fifth step of the data mapping and matching instance converting the classes and attributes of the first learning activity data into classes and attributes according to a second data profile different from the first data profile; And

And a sixth step of the data mapping and matching instance converting the meaning of the first learning activity data into a meaning according to the second data profile.

The data mapping and matching instance,

A structural and syntactic mapping instance table corresponding to classes and attributes of learning activity data according to the first data profile and classes and attributes of learning activity data according to the second data profile, and learning activity data according to the first data profile Meaning of the learning activity data according to the second data profile includes a semantic instance table corresponding to the ontology rule,

In the fifth step,

The data mapping and matching instance is converted using the structural and syntactic mapping instance table,

In the sixth step,

The data mapping and matching instance is converted according to the ontology rule using the semantic instance table.
The method of claim 9,

And a unique identification number is assigned to each of the classes and attributes of the structural and syntactic mapping instance table.
The method of claim 9,

After the fourth step,

If the class of the first learning activity data received by the data mapping and matching instance is not included in the structural and syntactic mapping instance table,

And said data mapping and matching instance adds said non-class to the structural and syntactic mapping instance table.
The method of claim 9,

After the fourth step,

If the attribute of the first learning activity data received by the data mapping and matching instance is not included in the structural and syntactic mapping instance table,

The data mapping and matching instance

And adding the not included attribute to the structural and syntactic mapping instance table.
The method of claim 9,

The first metadata includes at least one of a name of a data profile of first learning activity data stored in the first repository, a standard name and version of the data profile, and a location of the first repository. Way.
The method of claim 9,

The class or attribute of the first learning activity data does not correspond to the mapping instance table in the fifth step, or the meaning of the first learning activity data does not correspond to the semantic instance table in the sixth step. Occation,

And after the sixth step, the data mapping and matching instance performs an exception on the first learning activity data.
The method of claim 9,

After the sixth step,

The data mapping and matching instance transmits the first learning activity data converted into learning activity data according to the second data profile to a second repository that stores second learning activity data collected according to the second data profile. The seventh step; data mapping and matching method further comprising.
The method of claim 15,

After the seventh step,

A suitability check for determining whether the transformed first learning activity data received by the second repository has a class, attribute, and meaning according to the second data profile,

If appropriate, store the converted first learning activity data,

If not appropriate, exception handling; data mapping and matching method further comprising.
Instructions that may be executed by the digital processing device to perform a method of collecting learning activity data using a first data collection API, a first repository and data mapping and matching instances to perform structural and syntactic mapping and semantic matching In the recording medium which the program of the present invention is tangibly implemented and can be read by the digital processing apparatus

The method for performing semantic matching with the structural and syntactic mapping,

A first step of the first data collection API collecting first learning activity data performed in a first learning environment according to a first data profile;

A second step of storing, by the first repository, first metadata that is metadata of the first learning activity data and the first learning activity data;

A third step of the data mapping and matching instance querying first metadata of the first learning activity data stored in the first repository to determine a data profile of the first learning activity data;

A fourth step of the data mapping and matching instance receiving the first learning activity data from the first repository;

A fifth step of the data mapping and matching instance converting the classes and attributes of the first learning activity data into classes and attributes according to a second data profile different from the first data profile; And

And a sixth step of the data mapping and matching instance converting the meaning of the first learning activity data into a meaning according to the second data profile.

The data mapping and matching instance,

A structural and syntactic mapping instance table corresponding to classes and attributes of learning activity data according to the first data profile and classes and attributes of learning activity data according to the second data profile, and learning activity data according to the first data profile Meaning of the learning activity data according to the second data profile includes a semantic instance table corresponding to the ontology rule,

In the fifth step,

The data mapping and matching instance is converted using the structural and syntactic mapping instance table,

In the sixth step,

The data mapping and matching instance is converted according to the ontology rule using the semantic instance table.