KR102250035B1 - Appointment information Calling System During Call Reflecting Personal Preference and Method - Google Patents

Abstract

The present invention relates to an appointment information calling system during a call reflecting personal preference to guide appointment information with a specified hour in voice. According to the present invention, the appointment information calling system comprises: a recording module for recording, storing, and managing conversations during a call; an appointment term sensing module extracting preset appointment terms from the conversation, and determining that the conversation includes an appointment when the sum of weights of the extracted appointment terms exceeds a reference value, thereby confirming the appointment terms; an appointment information extraction module indexing the conversations including the confirmed appointment terms and m conversations stored in a call recording database, extracting a hour matching a preset appointment discovery rule from the indexed conversations during the call, and extracting a place satisfying a constraint through named-entity recognition (NER) with respect to the place to generate appointment information; an appointment hour search module determining whether an appointment rejection, an appointment rejection recommendation, or an appointment execution according to a preset personal preference, and recalculating a moving time according to constraints including an appointment subject, place, and hour to arrange the hours matched with the appointment information in a time series according to an overlapping hour calculation rule; and an appointment output module converting the appointment information into voice to outputs the same according to the sorted hours.

Description

Appointment information Calling System During Call Reflecting Personal Preference and Method

The present invention relates to a system and method for calling appointment information during a call reflecting personal preferences. More specifically, when a preset appointment term is extracted from the voice of a sender or receiver, and the sum of weights of the extracted appointment terms exceeds a reference value. It is determined that the voice contains an appointment, and a rule is extracted from the voice containing the appointment to generate specific appointment information for the appointment time, but when real-time confirmation of the schedule is required during a call, a constraint reflecting the personal preference (departure It relates to a technology that guides appointment information by voice in consideration of location, travel time, traffic conditions at the time, distance, preferred place, importance of meetings before/before/after the time.

It is common to make calls for conversations and communication between people or for inquiring and exchanging information. For example, to call the other party and find out by asking for a third party's phone number, email, etc., to determine a meeting place between you and the other party by asking where the other party's current location is, or to schedule a meeting with the other party. In everyday life, such as checking the other party's schedule, calls are one of the main means of communication.

In the case of receiving a conversation from the other party asking for a third party's contact information while on a call, the user stops the call with the other party and exits the call app by pressing the home button on the smartphone and opens the contacts app. It searches for a contact and sends it to the other party as a text message using the SMS app. Meanwhile, a wired terminal user also performs actions such as searching for a contact information and sending a text message after having a conversation asking the other party to wait for a while.

As another example, in deciding where to meet with the other party during a call, the user runs the Map app while the call with the other party is stopped, and sets the current location of the other party on the map and the current location of the other party found out by asking the other party. After determining the meeting place, the location information of the intermediate point is communicated to the other party through conversation or sent by text.

As another example, even when scheduling a meeting with the other party while on a call, both the user and the other party run their own schedule app while the call is stopped, check the free time, and ask each other for the schedule through the conversation again. You will have a schedule. Of course, after you schedule a meeting by phone and finish the call, you also register the meeting schedule in your own schedule app.

However, as described above, it is a hassle to take a series of actions in handling tasks such as providing third-party contact information during a call, providing callers intermediate point location information, and scheduling callers meetings. In addition, there is a problem in that it is inconvenient to inform the found information through conversation again, in particular, the conversation is cut off between the callers.

In addition, it is difficult to immediately process the above tasks during a call, so after disconnecting the call with the other party, you can find a third-party contact in Contacts, perform a map search to find out the location information of the intermediate point, or search the schedule app. After figuring out when it is best to schedule a meeting with the other party, it is cumbersome to call the other person again and inform them of these details through conversation. Particularly, when the above task request call is received from the counterpart while driving the vehicle, there is a problem of causing inconvenience to the user, such as having to stop driving the vehicle and perform a task required for task processing in a parked state.

Therefore, in the process of handling specific tasks such as providing third-party contact information during a call, providing intermediate point location information for callers, scheduling meetings for callers, etc., while ensuring a seamless call, the result of task processing is shared between callers and the contents of the conversation. There is an urgent need to develop technologies that enable better communication, such as this development.

Republic of Korea Patent Publication No. 10-2017-0084551 is to solve the above-described problem, during a call, a specific task requested from one party (user side) or the other party (the other party) is based on voice recognition. A technology for processing a specific task during a call by processing it with an intelligent agent and providing it to a caller is disclosed.

In the case of receiving a command related to a specific task processing request from the terminal A or the terminal B, the prior patent obtains the call connection information and the caller information from the call call between the terminal A and the terminal B through the call system and responds to the received command. The user command is interpreted from the result of speech recognition or natural language processing, and information necessary for command execution is obtained from terminal A or terminal B or an external DB based on the analyzed user command, and generated by using the acquired command execution information. It is characterized in that the result of executing one command is provided to the terminal A or the terminal B or an external DB.

That is, the prior patent interprets a user command during a call and performs a function corresponding thereto. For example, if a command is given to "send a text" during a call, a text is transmitted, or if a command is given to "tell me this week's schedule", the saved schedule details, etc. The result of executing the command of is provided to the terminal.

However, the prior patent is a structure in which the caller has to give a voice command to the terminal in order to process a specific task during a call. Therefore, there is a limitation that specific task processing is limited to commands recognized by the caller.

Accordingly, the present applicant extracts a preset appointment term from the voice of the sender or receiver, determines that the voice contains the appointment if the sum of the weights of the extracted appointment terms exceeds the reference value, and extracts the rule from the voice containing the appointment. When a specific appointment information is created with a time, but it is necessary to check the schedule in real time during a call, the constraints reflecting the personal preference (departure location, travel time, traffic conditions at the time, distance, preferred place, corresponding/before/after) Considering the importance of meetings, etc.), I would like to propose a system that guides appointment information by voice.

Korean Patent Publication No. 10-2017-0084551 (2017.07.20)

An object of the present invention is to determine that the voice contains an appointment when the sum of the weights of the appointment terms extracted from the voice of the sender or the receiver exceeds a reference value, and by extracting the rule from the voice containing the appointment, a specific time zone is voiced. It is to guide you.

It is an object of the present invention to provide constraints (departure location, travel time, traffic conditions at that time, distance, preferred place, corresponding/previous/ Since then, it is to inform the appointment information by voice considering the importance of the meeting.

The objects of the present invention are not limited to the above-mentioned objects, and other objects and advantages of the present invention that are not mentioned may be understood by the following description and will be clearly understood by those skilled in the art from the following description. In addition, it will be easily understood that the objects and advantages of the present invention can be realized by the means shown in the claims and combinations thereof.

In order to achieve this technical problem, the system for calling appointment information during a call reflecting the preferences of individuals of the present invention includes: a recording module for recording, storing and managing conversations during a call; An appointment term sensing module for extracting a preset appointment term from the conversation, and determining that the appointment term is determined by determining that an appointment is included in the conversation when the sum of the weights of the extracted appointment terms exceeds a reference value; Conversations containing the confirmed appointment terms and m conversations previously stored in the call recording DB are indexed, the time zone corresponding to the preset appointment discovery rule is extracted from the conversation during a call and the indexed conversation, and object name recognition for a place ( An appointment information extraction module for generating appointment information by extracting a place meeting the constraint through (NER); Determine whether to reject an appointment, recommend refusal of an appointment, or fulfill an appointment according to a preset individual's preference, recalculate the travel time according to constraints including the person, place, and time of the appointment, and calculate the time zone matched with the appointment information as an overlapping time zone. An appointment time zone search module that sorts in time series according to a rule; And an appointment output module that converts the appointment information into voice and outputs it according to the sorted time zone.

Preferably, the appointment term sensing module converts the conversation during a call into text, extracts the text previously stored in the appointment term DB among the converted texts as appointment terms, and the sum of the extracted appointment terms and matched weights is greater than or equal to the reference value. , It is characterized in that the appointment term is determined by determining that the extracted appointment term contains the appointment content.

The appointment information extraction module indexes the m conversations previously stored in the call recording DB and the conversations during a call containing the determined appointment terms, converts the conversations during the call and the indexed conversations into text through natural language processing, and converts the converted text into text. Among them, the text containing the time zone (date + time) corresponding to the appointment discovery rule is extracted, the sender is designated as the appointment target, and the place is extracted from the converted text through object name recognition (NER), and the appointment target and time zone And generating appointment information including a place.

The appointment time zone search module determines whether to reject an appointment, recommend rejection of an appointment, or fulfill an appointment according to the individual's preference indexed from the personal preference/constraints DB, but if the individual's preference value is the fulfillment of the appointment, the appointment target, place, and time zone Recalculate travel time to the place by indexing constraints including, and sorting in a time series to be overlapped according to the overlapping time zone calculation rule by indexing existing appointment information that may overlap from the recipient appointment DB. do.

In accordance with an embodiment of the present invention based on the above-described system, a method for calling appointment information during a call reflecting a personal preference includes the steps of: (a) extracting a preset appointment term from a conversation by an appointment term sensing module; (B) determining whether the sum of weights of the appointment terms extracted by the appointment term sensing module exceeds a reference value; (c) determining, by determining, by the appointment information extraction module, that an appointment is included in the conversation, when the sum of the weights exceeds the reference value as a result of the determination in step (b); (D) indexing, by the appointment information extraction module, a conversation including the confirmed appointment term and m conversations previously stored in the call recording DB; Appointment information extraction module extracts the time zone corresponding to the preset appointment discovery rule from the conversation during the call and the conversation indexed, and extracts the place that meets the constraint condition through object name recognition (NER) for the place to generate appointment information. Step (e); The appointment time zone search module determines whether an appointment rejection, appointment rejection recommendation, or fulfillment of an appointment is determined according to a preset individual's preference, and the travel time is recalculated according to constraints including the appointment target, place, and time zone, and matched with the appointment information. (F) arranging time zones in a time series according to an overlapping time zone calculation rule; And (h) converting, by the appointment output module, the appointment information into voice and outputting it according to the sorted time zone.

Step (f) includes the step (f-1) of determining whether to reject the appointment, recommend rejection of the appointment, or fulfill the appointment according to the preference value of the individual indexed from the personal preference/constraint DB by the appointment time zone search module; As a result of the determination in step (f-1), if the personal preference value is the fulfillment of the appointment, the appointment time zone search module indexes the constraints including the appointment target, place, and time zone, and recalculates the travel time to the place (f -2) step; And (f-3), by the appointment time zone search module, indexing existing appointment information that may overlap from the recipient's appointment DB, and sorting the appointment information in a time series according to the overlapping time zone calculation rule.

According to the present invention as described above, if the sum of the weights of the promise terms extracted from the voice of the sender or the receiver exceeds the reference value, it is determined that the voice contains the promise, and the time zone of the voice containing the appointment is specified according to a preset rule. There is an effect of guiding appointment information by voice.

According to the present invention as described above, when it is necessary to check the schedule in real time during a call, the constraints reflecting the personal preference from the conversation during the call (departure location, travel time, traffic conditions, distance, preferred place, corresponding/ Considering the importance of meetings before/after), it is effective to guide appointment information by voice.

1 is a block diagram showing a system for calling appointment information during a call reflecting personal preferences according to an embodiment of the present invention.
2 is a block diagram showing an appointment term sensing module of an appointment information calling system during a call that reflects an individual's preference according to an embodiment of the present invention.
3 is a block diagram showing an appointment information extraction module of an appointment information calling system during a call that reflects an individual's preference according to an embodiment of the present invention.
4 is a block diagram showing an appointment time zone search module of an appointment information calling system during a call that reflects an individual's preference according to an embodiment of the present invention.
FIG. 5 is an exemplary view showing components included in constraints according to personal preferences of an appointment time zone search module according to an appointment information calling system during a call that reflects an individual's preference according to an embodiment of the present invention.
6 is a flow chart illustrating a method for calling appointment information during a call reflecting personal preferences according to an embodiment of the present invention.
7 is a flow chart showing a detailed process of step S612 of the system for calling appointment information during a call reflecting personal preferences according to an embodiment of the present invention.

Specific features and advantages of the present invention will become more apparent from the following detailed description based on the accompanying drawings. Prior to this, terms or words used in the present specification and claims are based on the principle that the inventor can appropriately define the concept of the term in order to describe his or her invention in the best way. It should be interpreted as a corresponding meaning and concept. In addition, when it is determined that a detailed description of known functions and configurations thereof related to the present invention may unnecessarily obscure the subject matter of the present invention, it should be noted that the detailed description thereof has been omitted.

In the following, a specific reference will be omitted, but the system for calling appointment information during a call reflecting the preference of an individual according to an embodiment of the present invention is one of a tablet, a laptop, or a smartphone capable of communicating with a server connected through an information communication network. It is mounted on the smart device of the company, and is driven by an application that is distributed and installed online.

As shown in Fig. 1, the system 100 for calling appointment information during a call reflecting the preference of an individual according to an embodiment of the present invention includes a recording module 102 for recording, storing and managing a call voice, and a call voice ( Hereinafter, an appointment term sensing module 104 that extracts a preset appointment term from'conversation'), but determines that the appointment term is determined by determining that an appointment is included in the dialog when the sum of the weights of the extracted appointment terms exceeds the reference value. ), and the conversations containing the confirmed appointment terms and m conversations previously stored in the call recording DB, extracting the time zone corresponding to the preset appointment discovery rule from the conversations during a call and the indexed conversations, and objects for places An appointment information extraction module 106 that generates appointment information by extracting a place that meets the constraints through Named-Entity Recognition (NER), and an appointment rejection, appointment rejection recommendation, or appointment according to a preset individual's preference. Appointment time zone search module 108 that determines whether to fulfill, recalculates travel time according to constraints including appointment target, place, and time zone, and sorts the time zones matched with appointment information in a time series according to the overlapping time zone calculation rule , And an appointment output module 110 that converts the appointment information into voice and outputs it according to the sorted time zone.

At this time, conversations stored in the call recording DB are stored as the number of call recordings t (most recently recorded conversations) to call recordings t-i (conversations recorded before the i-th) according to the set number i value. In addition, when there is appointment information, a voice including time zone and place information is output, and if there is no appointment information, a voice in the form of "no appointment details" is output.

In addition, m, which is the number of conversations indexed from the call recording DB, is set to four by default, and is configured to include the current conversation and the previous three conversations, and the number of previous conversations can be added or changed through setting change. In addition, a personal preference/constraint DB for storing and managing set values and constraints for personal preference and a recipient appointment DB for storing and managing existing appointment information will be described below.

Referring to FIG. 2, the appointment term sensing module 104 converts a conversation during a call into text, extracts a text previously stored in an appointment term DB among the converted texts as an appointment term, and includes weights matched with the extracted appointment term. If the sum is greater than or equal to the reference value, it is determined that the extracted appointment term contains the appointment content, and the appointment term is determined.

In addition, the appointment term sensing module 104 switches the appointment output mode to the On state when the appointment term is confirmed while the appointment output mode is Off when an initial telephone call is received.

인 경우, 약속이 포함된 것으로 판단하며, 약속 용어 DB에는 [표 1]에 나타낸 바와 같이 약속 용어별로 매칭된 약속 가능성이 수치화되어 저장된다.In addition, the probability that the appointment content was included in the appointment term is calculated by [Equation 1],

In the case of, it is determined that an appointment is included, and the likelihood of an appointment matched for each appointment term is numerically stored and stored in the appointment term DB as shown in [Table 1].

[Equation 1]

는 약속 용어이고,

는 약속 가능성이며,

는 기준값이다.here,

Is an appointment term,

Is the promise possibility,

Is the reference value.

[Table 1]

Referring to FIG. 3, the appointment information extraction module 106 indexes m conversations pre-stored in the call recording DB and the conversation during a call including the determined appointment term, and processes the conversation during a call and the indexed conversation through natural language processing. Converts to text, extracts the text containing the time zone (date + time) corresponding to the appointment discovery rule from among the converted texts, designates the sender as an appointment target, and from the text converted through object name recognition (NER) By extracting the place, appointment information including the person, time zone, and place of the appointment is generated.

The time zone included in the appointment information may be expressed as'November 15th 13:00 ~ 15:00', and the time zone may include any one of a date, a day of the week, or a time.

Specifically, according to the appointment discovery rule preset in the appointment discovery rule DB, the appointment information extraction module 106 displays the dialog included in the call tone, that is, the converted text "Yes! Let's meet at 13:00 ~ 15:00 on November 15th." ." In the case of, the time zone "November 15th 13:00~15:00" is set to be used as it is.

In addition, the appointment discovery rule determines that the word is a time-related word when the converted text contains time-related words (today, tomorrow, this week, next week, morning, morning, afternoon, etc.).

In addition, if the word included in the converted text is "Yes! Is it tomorrow time?" as shown in [Table 2], "Tomorrow" is a word related to time (date), so change it to time (date) and convert A sender included in a word included in one text is designated as an appointment target, and the word "Gangnam Station" is extracted as place information through object name recognition (NER). In [Table 2], the extracted appointment subjects, time zones, and places are indicated in red boxes.

[Table 2]

As such, the appointment information extraction module 106 is configured to convert words included in the converted text according to the appointment discovery rule, converts'today is today() -> November 15th', and'tomorrow is today() + 1 -> November 16th','sand is today() +2 -> November 17th', and'next Monday is week1stday(today()) + 3 -> November 18th ', converts to'morning -> 07:00 ~ 10:00', and converts'Tomorrow -> 07:00 ~ 10:00 on November 16, etc. to include time zone Create

In addition, the appointment information extraction module 106 sets all '24 hours' of the date to the time zone when there is only a date in the converted text, and when there is only the date and the meeting start time, the'corresponding date + meeting start time-meeting start Time + 2 hours' is set as the time zone, and if the date and time section (morning, morning, evening, etc.) are included, the time zone can be set as'corresponding date + time section'. At this time, the detailed time interval is shown in [Table 3] below.

[Table 3]

Referring to FIG. 4, the appointment time search module 108 determines whether to reject an appointment, recommend refusal of appointment, or fulfill the appointment according to the preference of the individual indexed from the personal preference/constraint DB, but the personal preference value is the person who fulfills the appointment. In this case, the travel time to the place is recalculated by indexing the constraints including the appointment target, place, and time zone, and the existing appointment information that may be overlapped from the recipient's appointment DB is indexed and overlapped according to the overlapping time zone calculation rule. The appointment information with high probability is arranged in order, that is, in a time series.

On the other hand, if the personal preference value indexed from the personal preference/constraints DB is appointment rejection, the appointment output mode is switched to Off, and the individual preference value indexed from the personal preference/constraint DB is the appointment rejection advisor. In this case, a preset appointment rejection advisory guide sound or message is output and the appointment output mode is switched to the Off state according to the feedback value, or the appointment information sorting process according to the overlapping time zone calculation rule is performed.

Referring to FIG. 5, components included in the constraints of the appointment time zone search module 108 for recalculating travel time include weights assigned to each of the appointment target, time/weather, and place, and the weight is the user It may be set in advance or automatically changed to a weight applied to an existing appointment information selection case. In addition, weather-related values can be replaced with values automatically collected through an information and communication network.

Specifically, among the components included in the constraints of the appointment time zone search module 108 1) The appointment target is the importance of the appointment target (high, low), the family's preference for the appointment (high, low), and the possibility of delaying the end time of the appointment. (30 minutes, 0 minutes), and friend status (1, 0), 2) Time/weather is likelihood of another appointment (high, low), possibility of rain (0 ~ 100%), cold (high, low) Low), heat (high, low), snowiness (high, medium, small, no), and my mental state (expected) (bored, tired), and 3) the location is blocked at the time (1.0 ~ 2.0), place preference, car accessibility, parking availability, and public transportation accessibility.

In addition, among the components included in the personal preference of the appointment time zone search module 108, my mental state (S) is set to [tired] if the amount of previous appointments on the day is higher than the predetermined number (A), and the previous 1 If the number of appointments with the appointment target whose monthly friend value is '1' is lower than the set number (B), it can be set to [bored], and this setting value can be changed through user manipulation. In addition, it is desirable to understand the probability of another appointment occurring as the probability that the appointment was made at the corresponding time on the same day of the previous days.

In addition, in the calculation of overlapping possible time zones based on constraints based on personal preferences, the preferences of individuals indexed from the personal preferences/constraints DB when the appointment time zone search module 108 connects appointments in a time series (provision target, time zone, and place). ), it is determined whether the appointment information is fulfilled or rejected, but if the probability of rejecting the appointment is more than a preset value, it is determined that the requested appointment is rejected.

)에 영향을 미치는 요소는 약속 이행 요소(Positive Elements:

), 및 약속 거절 요소(Negative Elements:

)로 구성된다.At this time, the possibility of rejecting the appointment (

The factors that influence the positive elements (Positive Elements:

), and Negative Elements:

).

)는 약속 대상자 중요도(

), 가족의 약속 대상자 선호도(

) 및 장소 선호도(

)를 포함한다. 즉, 사용자가 기 설정한 약속 대상자의 중요도와, 가족이 특정한 약속 대상자와의 만남을 싫어하는지 여부와, 약속 장소가 술집인 경우를 배제시킬 수 있다.First, the Positive Elements:

) Is the importance of the person to be promised (

), family preferences (

) And place preferences (

). That is, it is possible to exclude the importance of the appointment target set by the user, whether the family dislikes meeting with a specific appointment target, and the case where the appointment place is a bar.

)는 다른 약속이 생길 가능성(

), 우천 가능성(

), 눈오는 정도(

), 나의 심리상태(

), 추위 여부(

), 더위 여부(

), 및 자가용 접근성(

)을 포함하며, 약속을 거절할 가능성은 아래의 [수학식 2]에 의해 계산된다.Also, the negative elements (negative elements:

) Is the likelihood that another appointment will occur (

), possibility of rain (

), snowing degree (

), my mental state (

), whether it is cold (

), whether it is hot (

), and car accessibility (

), and the probability of rejecting the appointment is calculated by the following [Equation 2].

[Equation 2]

) 여부 판단은,

≥ 70 이면 거절, 70 ≥

≥ 40 이면 거절 권고,

< 40 이면 약속 이행으로 판단할 수 있다. 여기서 각 수치는 %이다.At this time, the possibility of rejecting the appointment of the appointment time zone search module 108 (

) Whether or not,

Refuse if ≥ 70, 70 ≥

If ≥ 40, recommend rejection,

If <40, it can be judged as fulfilling the promise. Here, each figure is a percentage.

)을 다음 기준(

)에 의해 재계산을 수행하는데, 이전, 추가될 약속의 시간 중첩에 영향을 미치는 요소는 눈(비)오는 정도, (이전)약속의 지연 가능성, 이동수단, 및 해당일시 막히는 정도가 포함되며, 아래의 [수학식 3]에 의해 계산된다.In addition, the appointment time zone search module 108 is the moving time of the previous/added appointment (

) To the next criterion (

), and the factors that affect the time overlap of the appointments to be added before and after include the degree of snow (rain), the possibility of delay of the (previous) appointment, the means of transportation, and the degree of blockage at the time, It is calculated by the following [Equation 3].

[Equation 3]

이고, 눈(비) 오는 정도의 가중치는

이며, 해당일시 막히는 정도의 가중치는

이고, (이전)약속의 지연 가능성에 대한 가중치는

이다. 또한, 이동수단이 지하철인 경우 이동시간은 적설량과 관계가 없기 때문에

는 "무"로 설정되고, 이동수단이 지하철인 경우 막히는 정도와 무관함으로

은 "원활"로 설정된다.At this time, the weight of the means of transportation is

And the weight of the degree of snow (rain)

And the weight of the degree of blockage at the time is

And the weight for the (previous) promise delay probability is

to be. In addition, when the means of transportation is the subway, the travel time is not related to the amount of snowfall.

Is set to "no", and if the means of transportation is a subway,

Is set to "seamless".

In addition, the order of the overlapped appointment information is arranged in the order of the existing appointment information that overlaps the requested time zone at least a little, the existing appointment information in the overlapping time zone while preceding the request time zone, and the existing appointment information in the overlapping time zone while behind the request time zone. , The rules stored in the overlapping time zone calculation rule DB can be changed.

In addition, the rules for calculating overlapping time zones are when only the start time is specified (case 1), when only the time zone is specified (case 2), when only the date is specified (case 3), and the date and time interval (morning, morning, evening, etc.) It can be classified into a case specified as (case 4), a case where only the day of the week is specified (case 5), and a weekend (case 6).

The time zones for case 1 to case 6 and overlapping time zones according to the time zones for case 1 to case 6 are calculated as follows with reference to [Table 4] below.

[Table 4]

First, in case 1 (when only the start time is specified), if the time zone is 2:00 pm on October 11, sort the overlapping time from -1 hours before the start time to +3 hours after the start time, and change the time zone to'October 11 Sun 1:00 pm ~ 4:00 pm' is set as the overlapping time.

In case 2 (when the time zone is specified), if the time zone is from 2:00 pm to 3:00 pm on October 11, the time zone is '10 by sorting as overlapping time from -1 hours before the start time to +3 hours after the start time. 1:00 pm to 4:00 pm on the 11th of the month' is set as the overlapping time.

In case 3 (when only the date is specified), if the time zone is October 11, the date + 00:00 ~ 24:00 is sorted as the overlapping time, and'October 11 00:00 ~ 24:00' is overlapped. Set by time.

In case 4 (if specified by date and time section), if the time zone is the morning of October 11 by sorting the corresponding date + time zone as overlapping time,'October 11 06:00 ~ 11:00 'Is set as the overlapping time, and when the time zone is the morning of October 11th,'October 11th 06:00 ~ 13:00' is set as the overlapping time.

In case 5 (when only the day of the week is specified), if the time zone is Thursday, the days of the week + 00:00 ~ 24:00 for 4 weeks after the present are sorted as overlapping time, and the overlapping time is'October 24th 00 :00 ~ 24:00','October 31 00:00 ~ 24:00','November 7 00:00 ~ 24:00', and'November 14 00:00 ~ 24:00' Set as follows.

In case 6 (weekend), this Saturday + 00:00 ~ 24:00, Sunday + 00:00 ~ 24:00 are sorted as overlapping time, and'October 19 00:00 ~ 24:00, October 20th 00:00 ~ 24:00' is set as the overlapping time.

Then, the appointment output module 110 converts the appointment information received from the appointment time zone search module 108 into voice, outputs it through the speaker according to the sorted time zone, and switches the appointment output mode to the Off state at the same time as output.

Hereinafter, a method of calling appointment information during a call reflecting personal preference according to an embodiment of the present invention will be described with reference to FIG. 6.

First, the appointment term sensing module extracts a preset appointment term from the conversation (S602). At this time, the appointment term sensing module converts the conversation during a call into text, and extracts the text previously stored in the appointment term DB among the converted texts as the appointment term.

Subsequently, it is determined whether the sum of the weights of the appointment terms extracted by the appointment term sensing module exceeds a reference value (S604).

As a result of the determination in step S604, if the sum of the weights exceeds the reference value, the appointment term sensing module determines that the appointment is included in the conversation and determines the appointment term (S606).

Subsequently, the appointment information extraction module indexes the conversations including the confirmed appointment terms and m conversations previously stored in the call recording DB (S608).

Then, the appointment information extraction module extracts the time zone corresponding to the preset appointment discovery rule from the conversation during the call and the conversation indexed, and extracts a place that meets the constraint condition through object name recognition (NER) for the place. Generates (S610).

Subsequently, the appointment time zone search module determines whether the appointment rejection, appointment rejection recommendation, or fulfillment of the appointment is determined according to the preset individual's preference, and the travel time is recalculated according to the constraints including the target person, place, and time of the appointment, and the appointment information and The matched time zones are arranged in time series according to the overlapping time zone calculation rule (S612).

Then, the appointment output module converts the appointment information into voice and outputs it according to the sorted time zone (S614).

Hereinafter, a detailed process of step S612 of the method for calling appointment information during a call reflecting personal preference according to an embodiment of the present invention will be described with reference to FIG. 7.

After step S610, the appointment time search module determines whether to reject the appointment, recommend rejection of the appointment, or fulfill the appointment according to the preference value of the individual indexed from the personal preference/constraint DB (S702).

As a result of the determination in step S702, if the personal preference value is the fulfillment of the appointment, the appointment time zone search module indexes the constraints including the appointment subject, the place, and the time zone, and recalculates the travel time to the place (S704).

Subsequently, the appointment time zone search module indexes existing appointment information that may be overlapped from the recipient's appointment DB, and arranges the appointment information in a time series according to the overlapping time zone calculation rule (S706).

On the other hand, as a result of the determination in step S702, when the personal preference value is the appointment rejection, the appointment time zone search module switches the appointment output mode to the Off state (S708).

And, as a result of the determination in step S702, when the personal preference value is the appointment rejection recommendation, the appointment time search module outputs a preset appointment rejection recommendation guide sound or message, and turns off the appointment fulfillment or appointment output mode according to the feedback value. To the state (S710).

The appointment information extraction module converts the conversation during a call and the indexed conversation into text through natural language processing (S802).

Subsequently, among the texts converted by the appointment information extraction module, a text including a time zone (date + time) corresponding to the appointment discovery rule is extracted (S804).

Subsequently, the sender is designated as an appointment target from the text extracted by the appointment information extraction module, and a place is extracted from the converted text through object name recognition (NER) (S806).

Then, the appointment information extraction module generates appointment information including an appointment subject, a time zone, and a place (S808).

In this case, the appointment discovery rule is preferably understood as a rule for converting the converted text to conform to a preset format when time-related words or date-related words are included.

As described above, according to an embodiment of the present invention as described above, if the sum of the weights of the promise terms extracted from the voice of the sender or the receiver exceeds the reference value, it is determined that the voice contains the promise, and the rule is extracted from the voice containing the promise. Thus, there is an effect of guiding appointment information specific to the time zone by voice.

Although described and illustrated in connection with a preferred embodiment for illustrating the technical idea of the present invention as described above, the present invention is not limited to the configuration and operation as illustrated and described as described above, and deviates from the scope of the technical idea. It will be well understood by those skilled in the art that many changes and modifications are possible to the present invention without. Accordingly, all such appropriate changes and modifications and equivalents should be considered to be within the scope of the present invention.

100: Appointment information calling system during a call reflecting personal preferences
102: recording module
104: appointment term sensing module
106: appointment information extraction module
108: appointment time zone search module
110: promise output module

Claims (6)

A recording module for recording, storing, and managing conversations during a call;
A preset appointment term is extracted from the conversation, but if the sum of the weights of the extracted appointment terms exceeds the reference value, it is determined that the conversation contains the appointment, and the appointment is made.
An appointment term sensing module for determining a term;
Conversations containing the confirmed appointment terms and m conversations previously stored in the call recording DB are indexed, the time zone corresponding to the preset appointment discovery rule is extracted from the conversation during a call and the indexed conversation, and object name recognition for a place ( An appointment information extraction module for generating appointment information by extracting a place meeting the constraint through (NER);
Calculate the time zone matched with the appointment information by recalculating the travel time according to the constraints including the appointment target, place, and time zone, judging whether or not to reject the appointment, recommend rejection of the appointment, or fulfill the appointment according to the preset individual's preference. An appointment time zone search module that sorts in time series according to a rule; And
It includes an appointment output module for converting the appointment information into voice and outputting it according to the sorted time zone,
The appointment time zone search module determines whether to reject an appointment, recommend rejection of an appointment, or fulfill an appointment according to the preference of the individual indexed from the personal preference/constraints DB. The travel time is recalculated according to a specific criterion, and the amount of snow (rain), the possibility of delay in the (previous) appointment, and the means of transportation and the degree of temporary blockage are weighted and promised as factors that affect the overlapping time of the appointment. Recalculate the travel time of
The appointment time zone search module sets an overlapping time zone in which the time zone matched with the appointment information may overlap with a request time zone specified by any one of a start time, time zone, date, date_time zone, day of the week, and weekend, and A call appointment information calling system reflecting personal preference, characterized in that the appointment information is arranged in a time series in order of high probability appointment information. The method of claim 1,
The promise term sensing module,
Converts the conversation during a call into text, extracts the text previously stored in the appointment term DB among the converted texts as appointment terms, and if the sum of the extracted appointment terms and the matched weights is greater than or equal to the reference value, the extracted appointment terms include the appointment content. An appointment information calling system in a call reflecting an individual's preference, characterized in that the appointment term is determined by determining that it has been made. The method of claim 1,
The appointment information extraction module,
In-call conversations containing confirmed appointment terms and m conversations previously stored in the call recording DB are indexed, the busy conversations and the indexed conversations are converted to text through natural language processing, and appointment discovery rules and An appointment including the target person, time zone, and place by extracting the text containing the corresponding time zone (date + time), designating the sender as the appointment target, and extracting the place from the converted text through object name recognition (NER) Appointment information calling system during a call reflecting personal preferences, characterized in that generating information. The method of claim 1,
The appointment time zone search module,
Personal preferences/constraints Determine whether to reject an appointment, recommend a rejection of an appointment, or fulfill an appointment according to the individual's preference indexed from the DB, but if the individual's preference value is the fulfillment of an appointment, constraints including the person, place, and time of the appointment It reflects the preferences of individuals, characterized by recalculating the travel time to the place by indexing, indexing existing appointment information that may overlap from the recipient appointment DB, and sorting in a time series to be overlapped according to the overlapping time zone calculation rule. Appointment information calling system during a call. (a) extracting, by the appointment term sensing module, a preset appointment term from the conversation;
(b) determining whether the sum of weights of the appointment terms extracted by the appointment term sensing module exceeds a reference value;
(c) if the sum of the weights exceeds the reference value as a result of the determination in step (b), the appointment information extraction module determines that the appointment is included in the conversation, and thus the appointment term is determined;
(d) the appointment information extraction module indexing the conversations including the confirmed appointment terms and m conversations previously stored in the call recording DB;
(e) The appointment information extraction module extracts the time zone corresponding to the preset appointment discovery rule from the conversation during the call and the conversation indexed, and extracts the place that meets the constraints through object name recognition (NER) for the place to make an appointment. Generating information;
(f) Appointment time zone search module determines whether to reject appointment, recommend rejection of appointment, or fulfill appointment according to preset personal preferences, recalculate travel time according to constraints including appointment target, place, and time zone, and appointment information Arranging the time zones matched with chronologically according to an overlapping time zone calculation rule; And
(h) the appointment output module converts the appointment information into voice and outputs it according to the sorted time zone,
In step (f), the appointment rejection, appointment rejection recommendation, or appointment fulfillment is determined according to the personal preference indexed from the personal preference/constraint DB in the appointment time search module, but if the personal preference value is the appointment fulfillment, The transfer time of the appointment to be transferred/added is recalculated according to a specific criterion. The degree of snow (rain), the possibility of delay in the (previous) appointment, the means of transportation and the temporary blockage as factors affecting the time overlap of the appointment. Recalculate the travel time of the appointment by weighting the degree
In the appointment time zone search module, the time zone matched with the appointment information sets an overlapping time zone that is likely to overlap with the requested time zone specified by any one of start time, time zone, date, date_time zone, day of the week, and weekend. A method for calling appointment information during a call reflecting personal preference, characterized in that the appointment information is arranged in a time series in order of high probability appointment information. The method of claim 5,
The step (f),
(f-1) determining, by the appointment time zone search module, whether the appointment rejection, appointment rejection recommendation, or appointment fulfillment according to the individual preference value indexed from the personal preference/constraints DB;
(f-2) As a result of the determination in step (f-1) above, when the personal preference value is the fulfillment of the appointment, the appointment time zone search module indexes the constraints including the appointment target person, place, and time zone, and the travel time to the place. Recalculating; And
(f-3) The appointment time zone search module indexes the existing appointment information that may be overlapped from the recipient's appointment DB, and arranges the appointment information in a time series according to the overlapping time zone calculation rule.
Calling appointment information during a call reflecting personal preference, characterized in that it comprises.

Images