KR20180116670A - MBTI-Based Recommendation Method For Resource Collaboration In IoT Environment - Google Patents

Abstract

Provided by an embodiment of the present invention is an MBTI-based IoT environment resource recommendation method for resource collaboration, which: collects the use history of a user, surrounding environment information, and the internal resource information of a user terminal, stores the collected information in CIB, and analyzes the stored use history in order to determine a personality type based on MBTI from resource information which the user used in the past; collects resource information about usable internal resources and external resources; analyzes an applied service request signal when the service request signal is applied and requests service information necessary for the requested service; acquires recommended resource information from the resource information included in the service information depending on the determined personality type, and situation information included in the service information when the service information is received; and generates a resource map by using the acquired recommended resource information. The present invention is able to recommend optimal resources according to the personality type of a user with a low performance device in real time.

Description

{MBTI-Based Recommendation Method for Resource Collaboration In IoT Environment}

The present invention relates to an IoT environment resource recommendation method.

In a smart service space such as the Internet of Things (IOT), a user desires to receive various services using his / her own lightweight personal mobile terminal. However, miniaturization of personal mobile terminals has resulted in limited resources such as small screens, limited input devices, and insufficient power. One way to solve this scarce resource problem is to configure the services that users require by collaborating with a large number of resources in the vicinity.

The resource collaboration system aims at providing users' desired services by collaborating with users' personal resources or common resources in smart space as a way to satisfy users' requirements.

However, existing resource collaboration services have some problems to solve. In particular, the existing resource collaboration based service is a resource based service rather than a user based service, and even if other users request the same service in the same place, the mobile device will recommend the same resources. That is, the service is provided based on the functions that each resource can provide, without taking into consideration individual propensity and different preferences. Therefore, there is a limitation that a personalized service can not be provided.

Korean Patent Publication No. 10-2010-0051896 (published on May 19, 2010)

An object of the present invention is to provide an MBT-based IoT environment resource recommendation method for resource collaboration in which optimal resources can be recommended in real time on a low-performance device according to a user's personality type.

According to an aspect of the present invention, a MBTI-based IOT environment resource recommendation method for resource collaboration is a resource recommendation method implemented in a mobile user terminal, wherein an internal situation information management unit manages a usage history of a user, Collecting the internal resource information and storing it in the CIB; Analyzing the usage history stored in the resource recommendation unit to identify the personality type based on MBTI from the resource information previously used by the user; Collecting resource information, which is information on internal resources and external resources available to the situation information management unit through the internal situation information management unit and the communication unit; Analyzing the applied service request signal and requesting the service information required for the requested service to the situation information management unit when the service request signal of the resource recommendation part is applied; Acquiring recommended resource information among the resource information included in the service information according to the personality type discriminated from the context information included in the service information when the resource recommendation unit receives the service information; And generating a resource map using the recommended resource information from the resource recommendation unit.

According to another aspect of the present invention, there is provided a computer program for performing MBTI-based IoT environment resource recommendation for resource collaboration recorded in a non-transitory computer readable medium including computer program instructions executable by a processor, Collecting and storing the usage history of the user, the environment information and the internal resource information of the user terminal in the CIB when the computer program instructions are executed by the processor of the computing device, analyzing the stored usage history, Determining a personality type based on MBTI from resource information previously used, collecting resource information which is information on available internal resources and external resources, analyzing the applied service request signal when a service request signal is applied , The service required for the requested service Acquiring recommended resource information of the resource information included in the service information according to the personality type discriminated from the context information included in the service information when the service information is received, And generating a resource map using the information.

The MBTI-based IoT environment resource recommendation method for resource collaboration according to the present invention identifies the personality type of the user based on the MBTI, identifies the optimal resource to the user in consideration of the determined personality type, recommends it as a recommended resource, According to the user's personality type, the inference time can be greatly reduced by excluding the resource that is not preferred by the user in advance, and the user terminal having limited performance can recommend the resource in real time. Therefore, not only can an optimal resource suitable for a user's personality type can be recommended, but an IoT environment resource recommendation system can be implemented in a low-performance user terminal, so that the user's personal information does not need to be leaked to the outside, It is possible to accurately analyze the surrounding environment of the user and maximize the resource recommendation effect.

1 illustrates the concept of a resource collaboration based smart service environment.
2 shows a concept of a resource collaboration service in a smart service environment.
FIG. 3 shows a configuration of a MBTI-based IoT environment resource recommendation system for resource collaboration according to an embodiment of the present invention.
4 shows a detailed configuration of the resource recommendation unit in Fig.
FIG. 5 shows personality types classified on the basis of MBTI.
6 shows experimental results on resource selection characteristics according to the MBTI personality type.
FIG. 7 illustrates an MBT-based IoT environment resource recommendation method for resource collaboration according to an embodiment of the present invention.
FIG. 8 to FIG. 10 show performance test results of the MBTI-based IOT environment resource recommendation system and method for resource collaboration according to the present invention.

In order to fully understand the present invention, operational advantages of the present invention, and objects achieved by the practice of the present invention, reference should be made to the accompanying drawings and the accompanying drawings which illustrate preferred embodiments of the present invention.

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, the present invention will be described in detail with reference to the preferred embodiments of the present invention with reference to the accompanying drawings. However, the present invention can be implemented in various different forms, and is not limited to the embodiments described. In order to clearly describe the present invention, parts that are not related to the description are omitted, and the same reference numerals in the drawings denote the same members.

Throughout the specification, when an element is referred to as " including " an element, it does not exclude other elements unless specifically stated to the contrary. The terms "part", "unit", "module", "block", and the like described in the specification mean units for processing at least one function or operation, And a combination of software.

FIG. 1 shows a concept of a resource collaboration based smart service environment, and FIG. 2 shows a concept of a resource collaboration service in a smart service environment.

In the smart service environment shown in FIG. 1, a user uses a small terminal possessed by a person such as a smart clock or a pendant to share a shareable resource (a public resource in FIG. 1 or an individual resource permitted by the individual) A variety of smart services can be provided through collaboration.

For example, when a user needs to perform a word processing operation, the user requests a word processing service on his smart clock, as shown in FIG. Because smart clocks are difficult to provide word processing services only with their own resources, they recommend users to search for available resources around them and to configure word processing services such as keyboard, mouse, and monitor. Once the user selects the required resources, the user can perform word processing tasks through collaboration of the selected resources.

FIG. 3 shows a configuration of a MBTI-based IoT environment resource recommendation system for resource collaboration according to an embodiment of the present invention.

Referring to FIG. 3, the IOT environment resource recommendation system includes a service mediator 100, a resource configurer 200, a collaborative resource recommender 300, and at least one UMO 400.

The service mediator 100 functions as an interface between the user terminal and the collaboration resource recommender 300 and transmits the resource request information for the selected service according to the user command to the collaboration resource recommender 300. [ The service mediator 100 transmits a service request signal rq including a service identifier for a service selected by a user command among a plurality of predetermined services to the collaboration resource recommender 300, ) To recommend collaborative resources for the service. Here, the service identifier may be a service name, for example.

When the resource map (rm) including the resource information for at least one recommended collaboration resource is transmitted from the collaboration resource recommending unit 300, the service mediator 100 transmits the resource map rm ) To the resource configuration unit (200). At this time, the service mediator 100 displays the resource map rm to the user through the interface unit (not shown) of the user terminal so that the user confirms the recommended resource information as the collaboration resource in the collaboration resource recommending unit 300 can do. When the resource map rm includes resource information for a plurality of resources capable of performing the same function, the service mediator 100 displays a plurality of resource information included in the resource map rm to the user , And may transmit the resource information on the resource selected according to the user command to the resource configuration unit 200. That is, the user can directly select a resource to be used among a plurality of resources recommended as a collaborative resource in the resource map (rm).

In addition, the service mediator 100 may integrate the information about the service requested by the user together with the resource map rm and transmit the integrated service information IB to the resource configuration unit 200.

The resource configuration unit 200 receives the integrated service information IB from the service mediator 100 and analyzes the resource map rm and the service information included in the integrated service information IB to obtain a recommended And configures the service to be provided to the user using the resource information. That is, a service that can be substantially provided to the user is configured using the recommended resources.

Meanwhile, the collaboration resource recommendation unit 300 receives the service request signal rq from the service mediator 100, analyzes the service required by the user, and recommends resources for performing the analyzed service. The collaborative resource recommending unit 300 acquires and manages information on environment conditions, surrounding resources, and resources owned by the user based on the ontology before the service request signal rq is applied. For this, the collaborative resource recommending unit 300 may communicate with the resources in advance to confirm the available resources. The available resources may be resources owned by the user or shared resources. In particular, unlike the existing resource recommendation system, the collaborative resource recommender 300 in the present invention determines the user's personality type based on MBTI (Myers-Briggs Type Indicator), identifies resources suitable for the determined personality type, It can be selected as a resource. Then, a resource map (rm) is generated using the recommended resource and transmitted to the service intermediary unit 100.

At least one Ubiquitous Mobile Object (UMO) 400 communicates with the collaborative resource recommender 300 to receive an external resource inquiry query (TQ) from the collaborative resource recommender 300, To the collaborative resource recommendation unit 300. The external resource information ER is a result of the query to the collaboration resource recommendation unit 300. [

In the present invention, a UMO refers to an object having at least one shareable resource, and capable of configuring a service through collaboration based on the resource, and the resource recommendation system of FIG. 3 is included as a component of the UMO. Each UMO can configure its collaboration services using other resources around it, and it can also provide its own resources to other UMOs. For example, in the case of a smart phone, which is a personal mobile terminal, it may provide resources (microphone, speakers, etc.) that it possesses for collaboration, in response to requests from nearby users. Each UMO can provide resources for collaborations with the permission of the owner, and the present invention assumes that all resources in the IoT environment may be used for services or for services for collaboration.

The configuration of the IOT environment resource recommendation system shown in FIG. 3 is the same as the configuration of the resource recommendation system previously proposed by the inventors of the present invention, and the MBT-based IOT environment resource recommendation system for resource collaboration of the present invention It is similar to the existing resource recommendation system.

However, while the existing resource recommendation system is configured for a resource recommendation system in a server-client environment based on a user terminal and a server, the present invention can be applied to a mobile user terminal such as a smart phone, smart clock, smart pendant, .

In the IoT environment resource recommendation system of FIG. 3, the service mediator 100 and the resource configurer 200 provide an interface between the user terminal, the collaboration resource recommender 300 and the collaboration resource recommender 300, The collaboration resource recommending unit 300 can be regarded as the IoT environment resource recommendation system. Therefore, the collaborative resource recommender 300 is included in the user terminal as the IoT environment resource recommendation system. Accordingly, the service mediator 100 may function as an interface between the user and the collaborative resource recommender 300 implemented in the user terminal, not between the user terminal and the server, or may be omitted.

The existing IoT environment resource recommendation system has a limit in recommending the optimal resource to the user by deducing the resource based only on the current environment situation and the static profile information of the user. In addition, since it takes much time to deduce recommendation resources when there are a large number of available resources or a large number of context information such as usage histories to be considered for resource inference, a user terminal having a relatively limited performance There was a limit to be included. However, the server - client - based resource recommendation system is vulnerable to personal information security due to various information about the users being transmitted to the server, and there is a need to consent to use information from other users.

On the other hand, the IoT environment resource recommendation system of the present invention identifies a user's personality type based on the MBTI, and determines an appropriate resource according to the determined personality type and deduces it as a recommended resource. Therefore, when deducing recommendation resources, it is possible to reduce the inference time more than before by eliminating the resources that the user does not like according to the personality type of the user, and it is also possible to recommend resources in real time to user terminals having limited performance. In other words, according to the present invention, by recommending recommendation resources based on the MBTI, the IOT environment resource recommendation system can be implemented in a relatively low-performance user terminal as compared with a server, so that the user's personal information need not be leaked to the outside do. In addition, it is possible to acquire very accurate information about the surrounding situation.

As shown in FIG. 3, the collaboration resource recommending unit 300 includes a resource recommender 310, a context information manager 330, 320, a Context Information Base (CIB) 330, an internal situation information management unit 340, and a communication unit 350.

The resource recommendation unit 310 first requests the user's previous resource usage history to the situation information management unit 320, receives the resource usage history, and analyzes the authorized resource usage history based on the MBTI to determine the user type.

The resource recommendation unit 310 receives and analyzes the service request signal rq from the service mediator 100 and requests service information necessary for recommending collaboration resources to the situation information management unit 320, And receives the resource information and status information from the resource manager 320 to infer the optimal resource required for constructing the service requested by the user among the currently available resources according to the user type and rule determined based on the context information, And transmits the generated map to the service mediating unit 100.

In the present invention, the resource recommendation unit 310 is configured to determine the type of the user based on the MBTI, and to recommend the optimal resource to the user according to the determined type. Since the resource recommendation unit 310 is configured to infer recommendation resources based on the MBTI, it is possible to shorten the reasoning time even when the number of usable resources is large or the amount of context information such as previous usage history is very large, Recommendation is possible.

The context information management unit 320 functions as a control unit in the collaboration resource recommendation unit 300. When the resource recommendation unit 310 transmits the query information IQ in response to the service request signal rq, IQ) into a query based on SPRQL (Simple Protocol and RDF Query Language) that can be used in the CIB 330 and transmits the query. Then, the context information management unit 320 transmits resource and context information described in the Resource Description Framework (RDF) from the CIB 330 to the resource recommendation unit 310. [

SPARQL is a semantic query language for RDF-based databases, the semantic web language. RDF is the most basic semantic web language, in which metadata about web documents such as title, author, last modified date, and copyright are described based on XML which is the basis for representing web resources.

In addition, various status information such as an environment status and a user profile can be received from the internal situation information management unit 340 and transferred to the CIB 330 and stored, or various stored status information can be read. In addition, the situation information management unit 320 can inquire about external resources available for external resources through the communication unit 350.

The situation information management unit 320 can receive or transmit information described in the resource recommendation unit 310, the CIB 330, the internal situation information management unit 340, and the RDF, and can query using the SPARQL.

The CIB 330 receives and stores the context information from the context information management unit 320 and responds to the query information SQ applied by the context information management unit 320 to send the resource and context information to the context information management unit 320 .

The internal situation information management unit 340 collects and manages various situation information such as an environmental condition and a user profile. The internal situation information management unit 340 may include a monitoring unit 341 and an internal resource management unit 342. The monitoring unit 341 acquires, stores, and manages the collected environment information, user profile information, and usage history information, such as temperature information and illumination information, through the user terminal. The internal resource manager 342 manages the internal resources, i.e., the internal resources of the user terminal.

The communication unit 350 converts the external resource verification query (SQ) applied in SPARQL form into a form suitable for the predetermined communication method in the situation information management unit 320 and transmits the converted query TQ to at least one UMO 400 And converts the external resource information ER applied according to a predetermined communication method from at least one UMO 400 to the situation information management unit 320. [

The internal situation information management unit 340 and the communication unit 350 can acquire the situation information and the external resource information periodically or non-periodically even if the service request signal rm is not applied according to the control of the situation information management unit 320 .

4 shows a detailed configuration of the resource recommendation unit in Fig.

As described above, the MBTI-based IOT environment resource recommendation system for resource collaboration according to the present invention has a similar structure to the existing IoT environment resource recommendation system. However, the collaboration resource recommendation unit 300 may be configured such that the resource recommendation unit 310 The personality type is determined based on the MBTI, and the resource is recommended according to the determined user personality type so that a user with a low performance can recommend the optimal resource to the user in real time.

4, the resource recommendation unit 310 includes a request analysis unit 311, a context information based reasoning unit 312, a user type management unit 313, a user based reasoning unit 314, and a resource map generation unit 315 .

The request analysis unit 311 receives the service request signal rq and analyzes the service requested by the user and requests the service information required for the analyzed service to the situation information management unit 320. Here, the service information may include resource information requested by the service, surrounding environment information, and user information. If the situation information management unit 320 obtains corresponding information from the CIB 330 and the internal situation information management unit 340 and transmits the information to the request analysis unit 311, Based reasoning unit 312, as shown in FIG.

The context-based reasoning unit 312 is an ontology-based reasoning unit that deduces a resource based on environmental context information among the pieces of information, which are applied through the request analysis unit 311, in order to construct a service requested by the user. At this time, the context information based reasoning unit 312 deduces the same resource regardless of which user requests the service because it inferences only the environment. The context information based reasoning unit 312 transfers the resource information as a result of the inference to the user based reasoning unit 314.

On the other hand, the user type management unit 313 requests the usage history of the user to the situation information management unit 320 and receives the usage history. Based on the usage history of the authorized user, the personality type of the user is determined based on the MBTI. The user type management unit 313 includes a usage history management unit 313a and an MBTI analysis unit 313b. The use history management unit 313a preprocesses the usage history applied from the situation information management unit 320 into a form in which the MBTI analysis unit 313b can easily analyze the user personality type. The MBTI analyzing unit 313b statistically determines the user personality type based on the MBTI based on the preprocessed usage history.

The user-based inferring unit 314 receives the inferred resource information from the context-based inferring unit 312, and determines the association between the user type and the resources determined by the user type managing unit 314 and the user's personal preference information By using the rule to filter the inferred resource information, the recommended resource information to be recommended to the user is obtained.

The user-based reasoning unit 314 includes a user type-based reasoning unit 314a and a rule-based reasoning unit 314b. The user type based reasoning unit 314a firstly filters the resource information deduced from the context information based reasoning unit 312 according to the personality type of the user determined by the user type management unit 313. [ The rule-based reasoning unit 314b performs second-order filtering based on the relationship between resources in the first-order filtered resource information and personal preference information set by each user individually. It is because the rule-based reasoning unit 314b does not filter the resource information based on the relationship between the resources, in the case of some resources, may not be mutually combined for collaboration.

The resource map generating unit 315 generates and outputs a resource map of a predetermined format, which is obtained by receiving the recommended resource information obtained from the user-based inferring unit 314.

In FIG. 4, even when the service request signal rm is not applied, the user type management unit 313 can receive the user's history in advance and determine the personality type of the user. That is, since the personality type of the user is determined in advance, the recommended resource inference time can be reduced.

Meanwhile, the MBTI-based user personality type classification performed by the MBTI analyzing unit 313b is a psychological testing tool developed for use in real life, and a unique thinking style or behavior style may be different for each person. But the way we classify people's personality through the assumption that it happens because of some basic preference tendency that we can observe.

FIG. 5 shows personality types classified on the basis of MBTI.

MBTI categorizes personality types according to the four preference trends shown in Table 1.

In each of the four trends in Table 1, two indices ((outward (E) - introverted (I)), (sense (S) - intuition (N) (J) - recognition (J))) exists, the personality type of the user can be discriminated as one of the sixteen personality types shown in FIG.

However, in the present invention, the MBTI analyzing unit 313b must determine the personality type of the user based on the previous usage history of the user, and the user type based reasoning unit 314a should filter preferred resources according to the user personality type, To do this, an analysis of the correlation between the personality type of the user and the characteristics of the resource must be preceded.

In the present invention, the MBTI test is performed on a plurality of experiment participants, the personality type of each experiment participant is first discriminated, and the resource preference tendency according to the discriminated personality type is experimented.

Table 2 shows the number of personality types of participants participating in the experiment.

And resource preference propensity of experiment participants by each personality type allowed user to select resources for five kinds of sound resource, output display resource and mouse resource for movie watching service respectively. Twelve trials were conducted twice a week and three new resources were replaced each time. In order to estimate the characteristics of the user's resource selection, the property set C is defined as Equation 1 in the present invention.

(1) in the expression (1) represents the definition of the characteristic set (C), and (2) to (4) are mathematical expressions for calculating (1). (2) is used to obtain the ratio of selecting the new resource NR (r), the popular resource PR (r), and the specific resource SR (r) among the resources r selected by the user. For example, if the number of times the user selects the mouse resource 10 times and selects the resource classified by the popular mouse 3 times among them, the number of resource selection TN (r) is 10 and the number of popular resource selection SN (r) 3, the ratio (PR (r)) of selecting the popular resource is 0.3. (2) is applied to the new resource and the specific resource used as well as the ratio (NR (r)) of selecting a new resource and the ratio (SR (r) The value can be calculated. Here, whether a new resource is defined as a resource used for the first time, or a resource whose release date is less than six months. The popularity resource is defined as a resource whose resource selection rate is more than three times the average selection rate. For example, since the average selection expectation of five mouse resources is 1/5, a resource with a selectivity of 3/5 or more is defined as a popular resource. The reference value for distinguishing a new resource from a popular resource can be flexibly defined according to an application or a service environment. (3) is an equation for obtaining the value of the average resource selection time of the user. The average resource selection time ST (r) of the user is calculated by dividing the total time TT (r) used for selecting the resources (r) necessary for constructing the service among the recommended resources into the number of resource selection times TN (r) . Experimental results show that the minimum time for selecting three resources is 6 seconds and the maximum time is 53 seconds. Therefore, the resource selection time value is calculated by dividing the unit by 5 seconds by multiplying by 0.02. Finally, (4) calculates how far the user-selected resources are far from the user in order to calculate the near resource rate (D (r)). The distance of all recommended resources in this experiment is within about 1m to 5.5m from the user. Therefore, when the average distance of selected resources is 1.5 m, it has a value of 0.8. In (3) and (4), the constant values are defined from the results of the experiment and can be changed according to the application environment or conditions.

6 shows experimental results on resource selection characteristics according to the MBTI personality type.

FIG. 6 is a graph showing the characteristics of users selecting resources according to the four preference tendencies of the MBTI personality type. FIG. 6A shows the resource selection characteristics according to the outward type (E) or the inward type (I) (T) or emotional type (F), respectively, and (c) represents the resource selection characteristic according to whether it is a tangible (S) or an intuition type (N). Finally, (d) shows the resource selection characteristic according to whether it is a judgment type (J) or a recognition type (J).

Explaining the selection characteristic of FIG. 6A as an example, the outsider E user mainly selects popular resources or new resources, and does not select a specific resource continuously. Also, the time spent by the outsourced user to select three resources is on the order of 20 seconds, and the distance selected is 3.5 meters on average. The resource selection characteristics of these users may be used by the MBTI analyzing unit 313b to deduce the MBTI personality type of the user from the resource selection history of the user.

Equation (2) is an equation that calculates which type of user's usage history is closest to the four types of preferred indicators.

는 XC와 이분되는 타입 사용자들의 특성값(C)으로로 구성된다. 예를 들어 EC는 외향형 사용자들의 집합(C)의 값이고 SC는 감각형 사용자들의 집합(C)의 값이다. UC는 현재 사용자의 사용 이력으로부터 추출된 집합(C)의 값으로 uc1에서 uc5까지 5개의 값으로 구성된다. 수학식 2에서 (5)는 집합 XC의 각 원소들과 사용자의 C값 사이에 차를 계산하고, (6)은 식(5)의 원소와 이분되는 타입의 값을 계산하기 위해서 사용된다. (7)에 의해서 결정되는 사용자의 개인 성향(UP)는 사용자의 사용 이력으로부터 획득한 값과 (5) 및 (6)로 획득되는 도5 의 각 성격 유형별 사용자 선호값 사이의 차를 이용하여 결정된다.In Equation (2), XC is a characteristic value (C) of E, S, T, P type users among the four MBTI preferred indicators,

Is the characteristic value (C) of the type users divided by XC. For example, EC is the value of the set of extrinsic users (C) and SC is the value of the set of users (C) of the users. UC is the value of the set (C) extracted from the current user's history and consists of five values from uc1 to uc5. (5) in Eq. (2) is used to calculate the difference between each element of the set XC and the C value of the user, and (6) is used to calculate the value of the type of element and dichotomous of Eq. (UP) determined by the user 7 is determined using the difference between the value obtained from the user's use history and the user preference value for each personality type in Fig. 5 obtained in (5) and (6) do.

For example, if the values of NR (r), PR (r), SR (r), ST (r), and D (r) extracted from the current user history are 0.3, 0.5, 0.9, E) The value difference with the user is 0.1 + 0.2 + 0.1 + 0.1 + 0.2 = 0.7. And the difference between the values of the inward (I) user and the inward (I) user is 0.4 + 0.4 + 0.6 + 0.5 + 0.5 = 2.4. Therefore, the current user deduces that it is more like extroversion. When calculating the remaining three preferred indicators in this way, the type of the current user can be determined as ENFP.

That is, the MBTI analyzing unit 313b can determine the personality type of the user based on the previous usage history of the user.

It is difficult to precisely classify users into one of the MBTI 16 personality types based on the above five resource selection characteristics alone. However, the two tendencies of the MBTI preference index are incompatible with each other, and if the user 's history information accumulates, the personality type can be more accurately classified. Also, if we increase the number of resource selection characteristics, more accurate classification will be possible.

In the above description, the MBTI analyzing unit 313b can determine the user's personality type based on the resource usage history of the user. Hereinafter, the user type based reasoning unit 314a determines whether the user preferences Explain that you can deduce a custom resource to be available.

To this end, in the present invention, representative attributes of resources are defined as shown in Table 3 based on a resource selection characteristic of a user.

In the case of a user with a high selection rate of resources in Table 3, a specific "model" or "manufacturer" attribute is designated as a representative attribute, and a "release date" attribute or a "registration date" attribute is set as a representative attribute . The representative attribute of the popular resource selection rate is "popularity ", the" use frequency "in which the resource is used, and the representative attribute for the user who mainly selects the proximity resource is" distance ". Finally, for users with a long selection time, it is a prudent style, so the total "usage time" used for the resource and the total "usage count" used can be representative attributes. Attributes in parentheses in Table 3 are attributes that can be used together with representative attributes

The MBTI 16 personality types have primary, secondary, tertiary and inferiority, respectively. The main function is the one that has the greatest influence on the user's decision among the four functions, and the auxiliary function is the function that is used to assist and balance the main function. The tertiary function and the inferior function are the functions which are discolored by the user due to the contradiction between the main function and the auxiliary function. For example, the main function of the ISTJ type is sensation (S), the secondary function is accident (T), the tertiary function is emotion (F), and the inferior function is intuition (N). In the present invention, the representative and auxiliary attributes of MBTI type resources are defined as shown in Table 4 based on MBTI main function and auxiliary function.

In order to classify 16 personality types more precisely, an outward-oriented (E-I) indicator is added as an additional attribute. For example, in the case of INTJ type, the representative attribute of the resource for recommendation is the 'Release Date' attribute, the secondary attribute is the 'model' attribute, and the additional attribute is 'distance'.

Table 5 is the transaction data extracted from the mouse use history of the user. In order to normalize the attribute value x such as 'distance', 'use frequency', 'use time' and the like, the equation (3) is used in the present invention.

In order to estimate the user's resource preference based on the user's usage history, the present invention defines Equation (4).

In Equation (4), (10) is used to calculate the extended selection of the resource attribute (i). The attributes of a resource are composed of various values. For example, the 'button' attribute has a limited number of values, while the 'color' attribute can have a myriad of values. Therefore, when calculating the attribute selection value of the resource selected by the user, the present invention defines v (#di) as an extended selection value reflecting the number of attribute values in the selection value using (9). In (9), #di is the number of types of values of the attribute (i) that can be obtained from the history of the user. And f is a weighting factor that affects the total selection value estimation. Here, f is defined by experience as 1 for example. However, f can be defined as a value between 0 and 1 depending on the characteristics of the application. For example, in the case of the 'button' attribute in Table 5, a button having a value of '5' among 10 times in which a user selects a mouse resource has been selected twice and five kinds of values of a button attribute have been selected. v (#di) is 2x5 / 10 = 1. (11) is used to calculate the weight of the resource attribute (i). That is, how important is the attribute i to the user in order to select the resource? The weight (w (i)) is calculated when the attribute (i) is the representative attribute, the auxiliary attribute, or the additional attribute, and separately for all the attributes. When the attribute (i) is the representative attribute (pp), the value of the weight w (i) is 1. If the representative attribute is more than one, the sum of the weights (w (i)) of the representative attribute (pp) is 1. If attribute (i) is auxiliary attribute (sp) or additional attribute (ap), it is calculated by the second expression of (11). In this case, the minimum weight (w (i)) of each property is 0.5 because each type consists of at least two attributes. If there is more than one, the value divided by the number is the weight (w (i)) and the total sum is 1. Finally, if the attribute (i) is a general attribute (gp), it is calculated by the third equation. If Table 5 is the usage history of the ENTP type user, the representative attribute (pp) is the 'Release Date' property and the weight is 1. The attribute 'model' and 'popularity', which are supplementary attributes (sp) and additional attributes (ap), are 0.5. The weight (w (i)) of the remaining attributes excluding the representative, auxiliary, and additional attributes is calculated based on Rj indicating the rank of the general attribute (gp) (j). The rank (Rj) of the general attribute has a higher priority for the attribute having the largest value of the extended selection value (v (#di)). In Table 5, the largest values of the extended selection values of 'DPI', 'Button', 'Color', 'Sensor method', 'Distance' and 'Usage time' attributes are 2.4, 1.6, 1.5, 1.2, 2.0 , 1.8. Therefore, the ranking of the attributes is 'DPI', 'Distance', 'Usage time', 'Button', 'Color', 'Sensor method' and weights are 0.27, 0.23, 0.18, 0.18, 0.09 and 0.05. Finally, the sum of the weights w (i) for all attributes (i) of the resource is 3. (12) is used to estimate a user's preference for a resource.

If an ENTP type user having a history of use as shown in Table 5 should receive a recommendation of a new mouse and five available mouse resources are searched as shown in Table 6 in an environment in which a current user requests a service, 4 can be calculated as shown in Table 7.

The numbers under the attributes in Table 7 are the weights of each attribute calculated by (11) in Equation (4). For example, the model attribute has a weight of 0.5 for ENTP type users. For the resource M1, the value of the model attribute is m40, so the extended selection value calculated by Eq. (10) is 0.7. Therefore, the preference of the model attribute of the resource M1 is 0.35. If we calculate the preferences of all resources in this way, the sum of M2 resources is 4.17, which is the highest preference value for users of ENTP type. If Equation (4) is not used, the preference for the M4 resource is calculated to be the highest.

Table 8 shows the results of estimating the preference of candidate mouse resources for users of type INTJ, ESTJ, INFP, ISFJ among the 16 types of MBTI.

Preference for other types of users can also be computed in the same way, but it seems to be representative of some types without duplicate results. As can be seen in Table 8, the values of preferences calculated for users with different personality types, even with the same candidate mouse resources, are different. Therefore, in the present invention, the user type based reasoning unit 314a can estimate the preferences of the resources based on the personality type of the user and recommend customized resources based on the preferences.

FIGS. 7 to 9 show performance test results of the MBTI-based IoT environment resource recommendation system for resource collaboration according to the present invention.

In the present invention, performance evaluation of a resource recommendation system is performed based on resource reasoning time and user satisfaction. As described above, the environment in which the resource recommendation system of the present invention is performed is the user's small mobile terminal. Therefore, it is very important to make the result of reasoning at a satisfactory time with a limited computing resource and to recommend a resource suitable for the user.

Since it is practically impossible to prepare a large amount of resources and various environmental conditions for an experiment in a real environment, a simulator for generating various resource data for experiments has been developed and experimented. The mobile terminal for performance evaluation uses smart clock type UMO with 500MHz CPU and 512MB memory.

FIG. 7 shows a simulation result for a change in reasoning time according to the number of resources with increasing number of rules. In the simulation of FIG. 7, the rules are increased from 50 to 300 in increments of 50, and resources are used in 10, 20, 30, 40, and 50 resources. Each resource has an average of seven attributes. For example, if rule inference is performed on 20 resources, a total of 140 facts are used as inputs to the reasoning engine. As can be seen from the experimental results, the reasoning time increases in proportion to the number of resources or the number of rules.

FIG. 8 compares the rule-based reasoning time after the MBTI-based preprocessing proposed in the present invention with the existing reasoning time according to the increase in the number of resources. The number of rules is 200, and the ratio of resources filtered by MBTI - based preprocessing is defined as about 30%. However, this ratio can be flexibly changed depending on the application environment.

7 and 8, it can be seen that the reasoning time increases in proportion to the number of rules, and even if the number of resources increases, a very short inference time is required compared with the conventional method.

Figure 9 shows an average of user satisfaction as an evaluation of user satisfaction between the present invention and existing technology.

The experiment performed resource recommendations for 12 movie watching services for 10 weeks. The MBTI types of the participants were ISTJ, ESTJ, ISFJ, ENTP, and INFP. Two students participated in the experiment. Recommended resources for configuring movie viewing services are display resources, output sound resources, and mouse resources. The initial resources are recommended for 15 different resources, five for each type of resource. We also replaced the three resources with other resources each time for the experiment.

Table 9 shows the user satisfaction of MBTI type-based resource reasoning results and satisfaction of rule-based resource reasoning results without using MBTI.

The score is from 1 to 5, with 5 being the highest satisfaction. In the case of ENTP type users, there was a significant difference in the satisfaction level such as the satisfaction rate was rapidly increased or decreased according to the result of the recommendation of new resource. One of the ISFJ type users was classified as ISTJ type when the MBTI type was estimated based on the history of using the resource while the experiment was continued. All INFP type users did not give 5 points at all. ISTJ and ISFJ users were generally consistently recommended and satisfied with the same resources. Figure 10 shows the average of user satisfaction between recommendation using MBTI and not recommendation. If rule - based reasoning is performed, the user 's usage history is not reflected in the reasoning, so it shows similar satisfaction regardless of the frequency of use of the resource reasoning service. However, in MBTI - based resource inference, initial satisfaction is high, but satisfaction has steadily increased as resource recommendation frequency has increased. In the case of the 4th experiment, the reason why the average satisfaction was decreased is because the users of ENTP type gave a low score because they felt the same resource inference result.

As a result, the present invention discriminates the user's personality type based on the MBTI, and selects and recommends resources according to the personality type of the discriminated user, thereby improving user satisfaction and significantly reducing the recommended resource inference time. MBT-based IoT environment resource recommendation system for resource collaboration that can be provided.

The method according to the present invention can be implemented as a computer-readable code on a computer-readable recording medium. A computer-readable recording medium includes all kinds of recording apparatuses in which data that can be read by a computer system is stored. Examples of the recording medium include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage, and the like. The computer-readable recording medium may also be distributed over a networked computer system so that computer readable code can be stored and executed in a distributed manner.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is evident that many alternatives, modifications and variations will be apparent to those skilled in the art.

Accordingly, the true scope of the present invention should be determined by the technical idea of the appended claims.

Claims (9)

A resource recommendation method implemented in a mobile user terminal,
Collecting user's use history, surrounding environment information, and internal resource information of the user terminal and storing them in the CIB;
Analyzing the stored usage history to determine a personality type based on MBTI from the resource information previously used by the user;
Collecting resource information which is information on available internal resources and external resources;
Analyzing the service request signal when the service request signal is applied, and requesting the service information required for the requested service to the situation information management unit;
Acquiring recommended resource information among the resource information included in the service information according to the personality type discriminated from the context information included in the service information when the service information is received; And
A step of generating a resource map using the obtained recommended resource information
MBTI - based IoT environment resource recommendation method for resource collaboration. The method according to claim 1,
The step of determining the personality type
Requesting and receiving the use history; And
Wherein the preference of each of the 16 types of personality types set according to the MBTI is preliminarily obtained through experiments and the preference of each of the 16 types of personality types is determined through experimentation based on the selection rate of each resource characteristic, Analyzing the preference resource characteristic and discriminating the personality type
Wherein the MBTI-based IoT environment resource recommendation method is used for resource collaboration. 3. The method of claim 2,
The step of acquiring the recommended resource information includes:
Acquiring the recommended resource information by receiving the service information, analyzing the context information included in the service information, inferring a resource for implementing the service from the resource information included in the service information based on an ontology, ; And
Calculating a resource preference according to a resource selection characteristic of each resource corresponding to the recommendation resource information obtained in the context information based reasoning unit corresponding to the personality type and obtaining the final recommendation resource information based on the calculated resource preference step
Wherein the MBTI-based IoT environment resource recommendation method is used for resource collaboration. The method of claim 3,
The step of acquiring the final recommendation resource information comprises:
Obtaining recommendation resource information corresponding to the personality type among the recommended resource information obtained from the context information based reasoning unit; And
Acquiring the final recommendation resource information based on a causal relationship between the resources and recommendation resource information corresponding to the personality type and personal preference information preset and stored by the user
Wherein the MBTI-based IoT environment resource recommendation method is used for resource collaboration. 5. The method of claim 4,
Wherein the acquiring of recommended resource information corresponding to the personality type comprises:
Two criteria for the resource preference among the four classification criteria of the sixteen personality types are obtained in advance and classified into a main function and an auxiliary function and a resource selection corresponding to each of the divided main function and auxiliary function combination Attribute is classified into three levels of a representative attribute having a high importance of the resource preference, a secondary attribute and an additional attribute, and then the representative attribute, the secondary attribute, and the additional attribute of the resource characteristic are discriminated according to the determined personality type of the user , And a recommendation resource corresponding to the user personality type is inferred by assigning weights to the representative property, the auxiliary property and the additional property of the recommended resource information to calculate the resource preference. Based IoT environment resource recommendation method. The method according to claim 1,
Further comprising the steps of: converting an external resource request (SQ) into a form suitable for a predetermined communication method and transmitting the converted query (TQ) to at least one UMO. A computer program for performing MBTI-based IoT environment resource recommendations for resource collaboration recorded in a non-transitory computer readable medium including computer program instructions executable by a processor, When executed by the < RTI ID = 0.0 >
Collecting user's use history, surrounding environment information, and internal resource information of the user terminal and storing them in the CIB;
Analyzing the stored usage history to determine a personality type based on MBTI from the resource information previously used by the user;
Collecting resource information which is information on available internal resources and external resources;
Analyzing the applied service request signal and requesting service information required for the requested service when the service request signal is applied;
Acquiring recommended resource information among the resource information included in the service information according to the personality type discriminated from the context information included in the service information when the service information is received; And
A step of generating a resource map using the obtained recommended resource information
A computer program for performing operations comprising: 8. The method of claim 7,
Wherein the collecting and storing the internal resource information of the user terminal in the CIB comprises:
Acquiring and storing the collectable environment information, the user profile information, and the use history information, and managing the internal resources of the user terminal. 8. The method of claim 7,
Further comprising the steps of: converting an external resource query (SQ) applied in SPARQL form into a form suitable for a predetermined communication method and transmitting the converted query (TQ) to at least one UMO.

Images