KR101340464B1 - Intelligent Scheduling Management Agent - Google Patents

Abstract

The present invention provides a user information collecting means for collecting user information through a mobile device, a context preprocessing means for constructing a context using the user information, and a schedule for recommending or feeding back a schedule by inferring a user's behavior using the context. And guide means, wherein the context is a hierarchical location context, a hierarchical time context, a weather context and a schedule context tree.

Description

Intelligent Scheduling Management Agent {Intelligent Scheduling Management Agent}

The present invention provides a user information collecting means for collecting user information through a mobile device, a context preprocessing means for constructing a context using the user information, and a schedule for recommending or feeding back a schedule by inferring a user's behavior using the context. And guide means, wherein the context is a hierarchical location context, a hierarchical time context, a weather context and a schedule context tree.

Recently, as the use of mobile devices such as smart phones and smart pads is popularized, interest in personalization services such as efficient scheduling is increasing.

However, in order to provide a personalization service, it is very important to predict the user's next situation by analyzing the context information of the user individual obtained in real time.

However, currently used schedule management applications remain at the level of providing an interface for a user to record a schedule and providing an alarm when the time comes.

Therefore, there is an urgent need for an intelligent schedule management agent that uses the user's contextual information to infer the current behavior of the user to confirm whether the scheduled schedule has been performed and to help the user perform the schedule properly.

In addition, there is an urgent need for intelligent schedule management agents that can recommend optimal schedules to users using contextual information.

In addition, there is an urgent need for an intelligent schedule management agent that has a feedback function of encouraging the user to modify, delete, or perform a schedule by using the user's context information.

On the other hand, Korean Patent Laid-Open Publication No. 10-2011-0039733 as a related art related to the present invention discloses an individual daily life management apparatus and a management method.

In general, the apparatus for managing daily life includes a data collection unit for collecting context information of a user through a mobile device, a behavior pattern inference unit for inferring user pattern information by applying a probability model to the context information, and the context information. And a tagging UI unit configured to form the behavior pattern information as a visualized interface, and receiving the tagging information from the user, and analyzing the context information, the behavior pattern information, and the tagging information, and generating metadata as meaningful information. It is configured to include a data generator.

It infers behavior patterns by using contextual information, but it does not proceed to the function of assisting user's schedule execution or recommending new schedule after inferring behavior.

In addition, in order to infer a user's behavior and manage schedules, it is necessary to process context information about time or location more hierarchically, but it is not described therein.

An object of the present invention is to provide an intelligent schedule management agent that checks whether a scheduled schedule is performed by using the user's context information and infers the current behavior of the user and helps the user to perform the schedule properly.

Another object of the present invention is to provide an intelligent schedule management agent that can recommend an optimal schedule to a user by using context information of the user.

Another object of the present invention is to provide an intelligent schedule management agent that has a feedback function of encouraging a user to modify, delete, or perform a schedule by using context information of the user.

In addition, the present invention provides an intelligent schedule management agent that hierarchically processes context information about a time or location, which is particularly important for user inference and schedule management.

It also provides an intelligent schedule management agent that infers user's behavior and manages schedules using hierarchical time and location context information.

The present invention provides a user information collecting means for collecting user information through a mobile device, a context preprocessing means for constructing a context using the user information, and a schedule for recommending or feeding back a schedule by inferring a user's behavior using the context. Including the guide means, wherein the context is a hierarchical location context, a hierarchical time context, a weather context and a schedule context tree, by providing an intelligent schedule management agent, to solve the above technical problem.

In the present invention, the hierarchical location context is leveled in order of distant order from the main activity base of the user as a reference position, by providing an intelligent schedule management agent, to solve the above technical problem.

In the present invention, the schedule context tree comprises a basic schedule, a detailed schedule, a hierarchical location context for each detail schedule and a hierarchical time context for each location, by providing an intelligent schedule management agent, the technical problem To solve.

In the present invention, the technical problem is solved by providing an intelligent scheduling agent, which recommends or feeds back the schedule, and uses the similarity between the hierarchical location context and the hierarchical time context and the schedule context tree. I will.

In the present invention, by recommending or feeding back the schedule, by using the frequency included in the schedule context tree, by providing an intelligent schedule management agent, to solve the technical problem.

The intelligent scheduling agent of the present invention has a remarkable effect of inferring the user's current behavior by using the user's context information and confirming whether the scheduled schedule is performed and helping the user perform the schedule properly.

In addition, it has a remarkable effect of recommending an optimal schedule to the user by using the user's context information.

In addition, it has a remarkable effect of using the user's context information to perform feedback such as modifying, deleting, or encouraging the user to perform a schedule.

In addition, it has the remarkable effect of hierarchically processing context information about time or location which is especially important for user's behavior inference and schedule management.

It also has the remarkable effect of inferring user's behavior and managing schedule using hierarchical time and location context information.

1 is a diagram schematically showing a main configuration of an intelligent schedule management agent according to an embodiment of the present invention.
2 is a diagram illustrating an example of a hierarchical location context preprocessed using user information in an intelligent scheduling agent according to an embodiment of the present invention.
3 is a diagram illustrating an example of a hierarchical time context preprocessed using user information in an intelligent scheduling agent according to an embodiment of the present invention.
4 is a diagram illustrating an example of a Bayesian network for behavior inference in an intelligent scheduling agent according to an embodiment of the present invention.
5 is a diagram illustrating an example of a schedule context tree using hierarchical location and time context in an intelligent scheduling agent according to an embodiment of the present invention.
6 is a diagram illustrating an example of a flow of managing a schedule in an intelligent schedule management agent according to an embodiment of the present invention.

The terms and words used in the present specification and claims should not be construed as limited to ordinary or dictionary terms and the inventor may properly define the concept of the term in order to best describe its invention It should be construed as meaning and concept consistent with the technical idea of the present invention.

Therefore, the embodiments described in the present specification and the configurations shown in the drawings are merely the most preferred embodiments of the present invention and are not intended to represent all of the technical ideas of the present invention. Therefore, various equivalents It should be understood that water and variations may be present.

Before describing the present invention with reference to the accompanying drawings, it should be noted that the present invention is not described or specifically described with respect to a known configuration that can be easily added by a person skilled in the art, Let the sound be revealed.

1 is a diagram schematically showing a main configuration of an intelligent schedule management agent according to an embodiment of the present invention.

First, the intelligent scheduling agent according to an embodiment of the present invention is configured to utilize information collected by other programs or devices included in the mobile device by being installed in the user's mobile device. It may be configured to be included in (not shown).

The intelligent schedule management agent according to an embodiment of the present invention includes a schedule management unit 100, a schedule management setting unit 200 and the context DB, the schedule management unit 100 is configured to include a schedule guide unit 110. .

Depending on the design conditions, some means, module, or context DB of the schedule management unit 100 and the schedule management setting unit 200 may be included in a server (not shown) interworking with the agent.

The schedule management setting unit 200 is configured to include a schedule management interface and schedule program interworking means.

The schedule management interface is an interface provided by the agent according to the present invention, an interface for inputting, modifying, and deleting a schedule, an interface showing a result of whether the schedule is being performed properly after inferring an action, an interface for recommending a schedule, Interface for feeding back schedule.

The schedule program interlocking means is a means for interworking with other schedule programs included in a device on which an agent according to the present invention is installed, and configured to extract and use schedule information input from another schedule program.

Depending on the design conditions, it can be configured to continuously update the schedule input from another schedule program through synchronization with other schedule programs.

The schedule management unit 100 includes a user information collecting means, a context preprocessing means, a schedule guide unit 110, and the schedule guide unit 110 is configured to include a behavior inference module, a schedule recommendation module, and a schedule feedback module.

The user information collecting means collects information processed by a GPS, a timer, a weather information program, etc. included in the user's mobile device, and delivers the information to the context preprocessing means.

Preferably, the GPS coordinate information, time information, weather information is collected in real time to deliver to the context preprocessing means.

Here, the collected user information is a kind of low-level context information, and it is important to preprocess it and convert it into a context for use in the agent of the present invention.

This is because in order to intelligently manage a user's schedule, information about time or location must be broken down and processed hierarchically.

The context preprocessing means performs a function of preprocessing user information obtained through the user information collecting means and converting the user information into a hierarchical position context and a hierarchical time context for use in the schedule guide unit 110.

It also preprocesses user information and converts it into a weather context.

It also converts the user's basic and detailed schedules into a schedule context tree.

Also, among the obtained user information, information not required by the agent is removed.

2 is a diagram illustrating an example of a hierarchical location context preprocessed using user information in an intelligent scheduling agent according to an embodiment of the present invention.

The context preprocessing means converts the GPS received coordinate information transmitted in real time from the user information collecting means into a location context, and converts it into a hierarchical location context composed of multiple levels.

In other words, it organizes the schedule context tree that is used to determine whether schedules are performed through action inference and when performing schedule recommendations and feedback.

In the case of Figure 2 is an example consisting of six levels, in order to verify the agent according to an embodiment of the present invention is divided based on the Gongju University Cheonan campus.

In other words, it can be said that the reference position is the user's main activity base, and the levels are sequentially leveled away from it.

Depending on the design conditions, the reference position and the number of levels separating the positions can be changed and can be automatically processed by the statistical processing or set by the user.

location level Average distance (m) Individual similarity diagram District One 4000 0.1 Neighborhood 2 700 0.3 Street 3 100 0.7 Building 4 60 0.85 Floor 5 30 0.95 Room 6 10 1.0

[Table 1] shows that the representative locations are divided into six levels with the maximum distance (4000m) from Cheongan Campus (dormitory or lecture room) to Gongju University Cheonan Bus Terminal.

Here, the individual similarity is given based on the accuracy of the location information and is a relative meaning and a value arbitrarily set.

3 is a diagram illustrating an example of a hierarchical time context preprocessed using user information in an intelligent scheduling agent according to an embodiment of the present invention.

The time information transmitted in real time from the user information collecting means is converted into a time context, but is converted into a hierarchical time context composed of multiple levels.

Hierarchical temporal contexts are also used to determine whether schedules are performed through behavioral inference and to construct schedule context trees for schedule recommendations and feedback.

In the case of Fig. 3, the week is divided into seven equal parts by week day, the day is divided into four quarters by six hours, and the quarter is further divided into six equal parts by one hour.

In this case, the conversion to the three-level hierarchical time context as described above makes it possible to index 168 (7x4x6) branches per week due to such processing.

The weather context may or may not be processed hierarchically, and when not processed hierarchically, the weather context is processed by converting the weather information collected by the user information collecting means into a context.

The schedule context tree is configured to include detailed schedules in subordinate nodes of the basic schedule, and to include hierarchical location contexts and hierarchical time contexts for each detailed schedule.

When converted into a hierarchical location context, a hierarchical time context, a weather context, and a schedule context tree, it is stored in the context DB in real time, so that the schedule guide unit 110 continuously refers to the context DB.

4 is a diagram illustrating an example of a Bayesian network for behavior inference in an intelligent scheduling agent according to an embodiment of the present invention.

The schedule guide 110 monitors whether a schedule is performed by inferring a user's behavior using a hierarchical location context and a hierarchical time context, and recommends or recommends a schedule.

The behavior inference module performs a function of inferring a user's behavior using a hierarchical location context, a hierarchical time context, a weather context, and a schedule context tree.

Preferably, the behavioral inference module is designed to use a Bayesian network, which is a directional acyclic network of arcs expressing dependencies between nodes and numbers representing numbers, and differs according to design conditions. Of course, reasoning techniques can be used.

Basic schedule Detailed Schedule study Certificate, TOEIC, lecture Exercise Swimming, fitness, jogging, climbing religion Buddhism, Protestantism, Catholicism Appointment Meal, meeting, gathering Leisure Movies, Shopping, Watching, Travel health Internal medicine, surgery and dentistry

In Table 2, examples of basic schedules and detailed schedules are described. The type and number of basic schedules and detailed schedules may vary depending on design conditions.

In addition, the schedule context tree is determined by using schedule information of the schedule management interface and / or schedule program linkage means of the schedule management setting unit 200.

When using the Bayesian network for behavior inference, use a schedule context tree, but configure it to include additional day and weather information.

That is, it inserts the basic schedule, detailed schedule, and hierarchical location context and hierarchical time context information for each schedule, and also inserts weather context, and the day of the week information is inserted separately from the time context. do.

After inducing the Bayesian network, the behavior inference module calculates the probability of the basic schedule and the detailed schedule using the current user information by using the following Equation 1 and Equation 2 to infer the behavior of the user. Do this.

Where C is the context and BP is the default schedule.

For example, in order to calculate the probability that the basic schedule of the current user is 'study', Equation 1 is used to calculate the probability that the basic schedule for each of four kinds of contexts, that is, time, day, location, and weather, is studied.

After that, Equation 2 is used again to multiply the probabilities for each context to find the probability that the basic schedule is 'study'.

Based on the results of all six basic schedules, the highest probability of behavior is inferred.

In order to obtain a detailed schedule, the above equations [1] and [2] are used in the same manner as the probability for each basic schedule, but the detailed schedule DP is substituted instead of the basic schedule BP.

Overall, after performing context preprocessing using the user information collected in real time, the Bayesian network is used to perform probability calculation.

5 is a diagram illustrating an example of a schedule context tree using hierarchical location and time context in an intelligent scheduling agent according to an embodiment of the present invention.

The schedule recommendation module included in the schedule guide unit 110 performs a function of recommending a schedule by using a schedule context tree to measure similarity based on a location and a time.

Here, the method of measuring similarity based on the location and time includes a method of using the number of contexts that match the method of using frequency.

First, measure similarity based on location and time, but use frequency as follows.

Referring to FIG. 5, a subordinate node of a basic schedule includes a detailed schedule, includes a hierarchical location context for each detailed schedule, and a schedule context tree including a hierarchical time context for each location.

Here, it is assumed that the frequency included for each node has already been determined through a statistical process, and the frequency of the upper node is the sum of the frequencies of its own lower nodes.

The schedule recommendation module calculates the score for each node by multiplying the weights according to the frequency and the level for each node of the schedule context tree using Equation 3 below.

Here, the weight according to the level can be arbitrarily or automatically set statistically according to the design conditions.

After obtaining the score for each node, the probability for each schedule is calculated according to [Equation 4]. And if the probability value obtained according to [Equation 5] is the maximum value, the schedule of the highest probability at a specific time or location is obtained.

Here, S means a schedule, V means a location and time context, and Context means a schedule context.

Second, measure similarity based on location and time, but use the number of contexts that match.

Table 3 below compares the locations included in the hierarchical location context with the locations included in the schedule context tree to obtain the number of matching contexts, and shows the individual similarity level of the level corresponding to the number (Table 1). (Ref)).

In other words, if the positions included in the hierarchical position context are included in the positions included in the context tree or the two are the same, the satisfaction level becomes 'Y' and the similarity is 1, the maximum value. .

vaginal goal Number of matches satisfaction Similarity 1131 112162 2 N 0.3 1131 113161 4 Y 1.0 112322 112222 3 N 0.7 112322 112322 6 Y 1.0

When using a hierarchical time context, the number of matching contexts and / or a numerical comparison is performed. [Table 4] below compares the time included in the hierarchical time context and the schedule context tree sequentially to calculate satisfaction. The case of obtaining is described.

Query (A) Goal B Common number Numerical Comparison (= A-B) satisfaction 113 116 2 -3 N 113 123 One -10 N 721 720 2 One N 321 321 3 0 Y

Depending on the design conditions, the similarity and satisfaction scores used in [Table 3] and [Table 4] can be statistically processed and automatically set or adjusted according to the agent's situation.

After the schedule recommendation module calculates the most probable schedule at a specific time and / or location, the schedule recommendation module performs a function of recommending the optimal schedule when the user inputs or modifies the schedule.

The schedule feedback module operates in conjunction with the action inference module and the schedule recommendation module, and performs a function of feeding back a schedule. For example, when the schedule feedback module is located at a time or position that is significantly lower than the time or position necessary for executing the schedule, the schedule is automatically generated. It performs the function of encouraging modification and deletion or schedule.

6 is a diagram illustrating an example of a flow of managing a schedule in an intelligent schedule management agent according to an embodiment of the present invention.

First, the user information collecting means collects user information such as location, time and weather and transmits the information to the context preprocessing means. (S100)

The context preprocessing means forms a hierarchical location context, a hierarchical time context, a weather context, and a schedule context tree by using the user information and the schedule information acquired by the schedule management setting unit 200. (S200)

The behavior inference module uses the Bayesian network to infer the user's current behavior. (S300)

After the behavior is inferred, it is compared with the user's planned schedule to see if the schedule is being followed. (S400)

Here, the schedule feedback module may notify the user that the schedule is not being performed through the schedule management interface and notify the schedule to be followed.

The schedule recommendation module recommends a schedule using a schedule of the maximum probability using location similarity and / or time satisfaction and / or frequency when the user enters or modifies the schedule through a schedule management interface. (S500)

In addition, the schedule feedback module feeds back to modify, delete or encourage related schedules using location similarity and / or time satisfaction. (S500)

The context DB is updated in real time and the part for scheduling is also updated. (S600)

1 to 6 described above, only the main matters of the present invention are described, and the present invention is limited to the configurations and functions of FIGS. 1 to 6 as long as various designs are possible within the technical scope. It is not obvious.

100: schedule management
200: schedule management section

Claims (5)

User information collecting means for collecting user information through a mobile device;
Context preprocessing means for constructing a context using the user information;
Schedule guide means for inferring a user's behavior using the context and recommending or feeding back a schedule;
The context is a hierarchical location context, a hierarchical time context, a weather context, and a schedule context tree,
Recommend or feedback the above schedule,
Using similarity of the hierarchical location context and the hierarchical time context with a schedule context tree, or
Intelligent calendar agent, characterized in that to use the frequency included in the schedule context tree.
The method of claim 1,
And the hierarchical location context is leveled in order of distance away from the user's main activity base as a reference location.
The method of claim 1,
The schedule context tree is characterized in that it comprises a basic schedule, a detailed schedule, a hierarchical location context for each detail schedule and a hierarchical time context for each location.
delete delete

Images