KR20120111855A - Apparatus and method for generating story according to user information - Google Patents

Abstract

PURPOSE: An apparatus and a method for generating a story according to user information collected from at least one electronic device are provided to easily recognize a life pattern and life information by configuring user life information according to user information collected from at least one electronic device. CONSTITUTION: An information collecting unit(110) collects log information including life information of a user from an electronic device. An analyzing unit(120) analyzes collected log information. The analyzing unit determines at least one topic. A story generating unit(130) generates one sentence which indicates the life information by using one topic. [Reference numerals] (100) Controlling unit; (110) Information collecting unit; (120) Analyzing unit; (130) Story generating unit; (140) Storing unit; (150) Displaying unit; (160) Input unit; (170) Communication module

Description

Apparatus and method for generating story according to user information {APPARATUS AND METHOD FOR GENERATING STORY ACCORDING TO USER INFORMATION}

The present invention relates to an apparatus and method for generating a story according to user information, and more particularly, to an apparatus and method for providing a user by summarizing a story of user's living information according to user information collected from at least one electronic device. It is about.

With the development of wired and wireless technology, it is possible to continuously collect information related to users' daily lives. For example, users always carry a mobile device with the development of mobile communication technology. Accordingly, the mobile device may collect various information such as a call log, a photo, a music file, and location information of the user.

As described above, the mobile device which is always carried by the user may collect daily information of the user. In this case, the mobile device may provide various services to the user based on the collected daily information of the user.

However, the current mobile device does not provide a service considering the user's convenience based on the user's daily information. Accordingly, there is a need for a technology in which a mobile device can provide the user with the user's life information by summarizing the story based on the user's daily information.

Accordingly, an object of the present invention is to provide an apparatus and method for generating a story according to user information collected by at least one electronic device.

Another object of the present invention is to provide an apparatus and method for generating and providing a story to a user according to user information collected by at least one electronic device.

Another object of the present invention is to provide an apparatus and method for generating an hourly story based on user information collected by at least one electronic device.

Another object of the present invention is to provide an apparatus and method for generating a story by theme based on user information collected by at least one electronic device.

According to a first aspect of the present invention for achieving the objects of the present invention, an apparatus for summarizing daily information of a user comprises: an information collecting unit for collecting log information including daily information of the user from at least one electronic device; An analysis unit determining at least one topic representing daily information of a user by analyzing log information collected by the information collecting unit, and at least one representing daily information of a user using at least one topic determined by the analysis unit Characterized in that it comprises a story generating unit for generating a sentence.

According to a second aspect of the present invention, a method for summarizing daily information of a user includes: collecting log information including daily information of a user from at least one electronic device, and log information collected by the information collecting unit; Determining at least one topic representing the user's daily information by analyzing the information, and generating at least one sentence representing the user's daily information using the at least one topic determined by the analyzer. It is done.

As described above, the user's life information is composed as a story and provided to the user according to the user information collected through the at least one electronic device, so that the user can easily recognize his / her life information and life pattern.

1 is a view showing the configuration of a story generating apparatus according to user information according to an embodiment of the present invention;
2 is a diagram illustrating a detailed configuration of an analysis unit of a story generating device according to an embodiment of the present invention;
3 is a diagram illustrating a detailed configuration of a feature extraction unit of a story generating device according to an embodiment of the present invention;
4 is a diagram illustrating a story generation procedure according to user information according to an embodiment of the present invention;
5 is a diagram illustrating a procedure for configuring a topic segment according to an embodiment of the present invention;
6 is a diagram illustrating an object recognition procedure according to an embodiment of the present invention;
7 is a diagram illustrating a procedure for generating a story by time according to an embodiment of the present invention;
8 is a diagram illustrating a procedure for generating a story for each topic according to an embodiment of the present invention; and
9 is a diagram illustrating a display screen of a story according to an embodiment of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear. The following terms are defined in consideration of the functions of the present invention, and may be changed according to the intentions or customs of the user, the operator, and the like. Therefore, the definition should be based on the contents throughout this specification.

Hereinafter, a description will be given of a technology for providing a user by summarizing a user's life information according to user information collected from at least one electronic device into a story.

In the following description, the story generating device collects daily information of the user and generates a story summarizing the living information of the user based on the collected daily information.

In the following description, it is assumed that the story generating device is composed of a mobile phone. However, the story generating device may be any one of a computer, a digital camera, a digital home appliance, a digital telephone, a digital project, a home server, a digital video recorder, a digital satellite broadcast receiver, a set-top box, and a digital TV broadcast receiver capable of processing data. Can be configured.

1 is a block diagram of a story generating apparatus based on user information according to an exemplary embodiment.

As shown in FIG. 1, the story generating apparatus includes a controller 100, an information collector 110, an analyzer 120, a story generator 130, a storage 140, a display 150, and an input unit. And 160.

The controller 100 controls the overall operation of the story generating device. The controller 100 controls the display unit 150 to display the story provided from the story generation unit 130.

The information collecting unit 110 collects daily information of a user by using log data of at least one electronic device. For example, the information collecting unit 110 collects daily information of a user using data received from the outside, data generated inside, and data stored in an external storage. Here, the data received from the outside includes weather, latitude and longitude information, blog information, social data, and the like. The internally generated data includes call history, short message service (SMS), multi-media message service (MMS), and latitude and longitude information measured by a global positioning system (GPS) and a schedule. And address book and the like. Data stored in the external storage includes an image photographed image, a list of played music files, TV viewing information, and the like.

The analyzer 120 estimates an episode for representing the user's daily information by analyzing the information collected by the government collector 110. For example, the analyzer 120 is configured as shown in FIG. 2.

The story generator 130 generates a story summarizing and expressing daily information of the user by using the episode estimated by the analyzer 120. For example, the story generator 130 generates a story by time using the episode estimated by the analyzer 120. For another example, the story generator 130 may generate a story for each topic by using an episode estimated by the analyzer 120. In this case, the story generator 130 may generate a story having a format different from that of the feature extracted using the significant verb. In addition, the story generator 130 may generate a story in consideration of the age and region of the user.

The storage unit 140 may include a program storage unit for storing a program for controlling the operation of the story generating device and a data storage unit for storing data generated during program execution. For example, the storage unit 140 stores the user schedule information collected by the government collecting unit 110 in the form as shown in Table 1 below.

Data1 Data2 Type Type_detail Data1 Data2 Data3 Data4 Data5 Data6 1318813200000 1318816800000 A 0 OS class Changwon National University Building 55 Listen to OS lessons from room 55110 to 11:00 1318825080000 1318825080000 T 0 Lotteria in front of the school. Waiting for a friend .... Hamburger is delicious ~ 0.0 0.0 Salim-dong, Changwon-si, Korea 1318824360000 1318824360000 S 2 01051141415 If you're bored, stay inside Lotteria. Wait inside. 1289809757249 1289809757249 W 0 Sunny Changwon-si 1289780978000 1289780978000 P 0 35.25032446 128.60073712 Guam 2-dong, Masan, South Korea 1289831348000 1289831348000 G 0 35.2397167 128.69623801 Yong-dong, Changwon-si, Korea / sdcard / DCIM / Camera / 2010-11-17 22.59.30.jpg ... ... ... ... ... ... ... ... ... ...

The display unit 150 displays the state information, the menu screen, and the story information of the story generating apparatus under the control of the controller 100. For example, the display unit 150 displays the story generated by the story generator 130 under the control of the controller 100 so that the user can check the story.

The input unit 160 provides input data generated by a user's selection to the controller 100.

The communication module 170 processes signals transmitted and received through an antenna for voice and data communication.

In the above configuration, the control unit 100 may perform the functions of the information collecting unit 110, the analyzing unit 120, and the story generating unit 130. In the present invention, it is separately configured to describe each function separately. Therefore, in actual implementation, the controller 100 may be configured to process all or some functions of the information collecting unit 110, the analyzing unit 120, and the story generating unit 130.

2 illustrates a detailed configuration of an analysis unit of a story generating device according to an embodiment of the present invention.

As illustrated in FIG. 2, the analysis unit 120 includes a topic segment configuration unit 201, a feature extraction unit 203, a topic determination unit 205, and an episode determination unit 207.

The topic segment configuration unit 201 configures a topic segment that is a set of candidate topics for composing a story using user schedule information collected by the information collection unit 110. For example, the topic segment configuration unit 201 divides the scan guns into time intervals and configures the topic segment with the scan guns included in each time interval. If the schedule information is a scan gun, the topic segment configuration unit 201 configures each schedule information into each topic segment. Subsequently, the topic segment configuration unit 201 maps an auxiliary event whose occurrence time of each auxiliary event is included in a time interval of the topic segment to the corresponding topic segment. If there is no topic segment to which an auxiliary event is mapped, the topic segment constructer 201 configures a dummy segment using an auxiliary event not mapped to the topic segment. In this case, a time interval of the dummy segment is preset. For example, the dummy segments are configured at one hour intervals. In this case, the scan gun represents schedule information that is mainly used to construct a story among user schedule information collected by the information collection unit 110. The incident event represents information excluding the scan gun from user schedule information collected by the information collecting unit 110.

The feature extracting unit 203 extracts a feature from scan guns and auxiliary events included in the topic segment configured in the topic segment constructing unit 201. That is, the feature extraction unit 203 extracts features from each topic segment. For example, the feature extraction unit 203 is configured as shown in Figure 3 below to extract the feature. Here, the qualities represent data that may be sentence components. For example, the qualities include verbs, entity names, sentence association information, topic information associated with entity names, and additional information that may be sentence elements. In this case, the entity name may vary depending on at least one of a classification, an entity name recognizer, and a sentence template. However, only the entity name defined in the environment setting may be used as shown in Table 2 below.

Object name name Tag information Details Additional features EXT_NONE Additional feature recognition tag, not used as actual object name but used internally in system Person PERSON Group name ORG date DAY time TIME nomination LOCATION stuff OBJECT Event name, thing Photo file name JPEG-FILENAME Picture file name in JPEG format

The topic determiner 205 determines a topic by using the features extracted by the feature extractor 203 in each topic segment. For example, the topic determination unit 205 determines a topic in consideration of the occurrence probability of the topic as shown in Equation 1 below. In this case, the topic determiner 205 determines a topic in each topic segment.

Here, T ^* denotes an optimal topic in a topic segment, T denotes a set of all topics, and S _t denotes a topic score for topic t. Here, S _t may be represented by Equation 2 below.

Here, S _t denotes a topic score for topic t, L denotes a type of logs included in a topic segment, Wx denotes a weight for the corresponding log type, and P (word) corresponds to a corresponding topic. The probability of occurrence of a related word is represented, and P (exts) represents a characteristic of an additional feature.

The episode determiner 207 selects an optimal topic segment for each time period. In this case, the episode determination unit 207 recognizes the topic of the optimal topic segment as an epid of the corresponding time. When there are a plurality of topic segments in the same time interval, the episode determiner 207 compares the topic scores of the topics for each topic segment determined by the topic determiner 205 and selects an optimal topic segment in the corresponding time interval. Choose. Then, when the topics of the consecutive topic segments are the same, the episode determiner 207 merges the consecutive topic segments into one topic segment.

In addition, the episode determination unit 207 determines the main topic that can represent the most important feature of the user schedule information in consideration of the episodes of each time interval. For example, the episode determination unit 207 determines the main topic as shown in Equation 3 below.

Where T ^** represents the main topic, T represents a set of all topics, EP represents an episode of each time interval, d represents the length of the episode, and S _t represents a topic t Indicates a topic score.

In addition, the episode determiner 207 may estimate the user's emotion in consideration of the occurrence probability of the emotion word identified in the process of determining the episode of each time interval.

3 illustrates a detailed configuration of a feature extraction unit of a story generating device according to an embodiment of the present invention.

As shown in FIG. 3, the feature extracting unit 203 includes a social information extracting unit 301, a location information extracting unit 303, a text extracting unit 305, and an object name setting unit 307. do.

The social information extracting unit 301 extracts social information such as mobile phone information, group information, etc. from the scan and auxiliary events included in each topic segment configured in the topic segment monk 201.

The location information extractor 303 extracts location information, such as latitude, longitude, address, and place name, from scans and auxiliary events included in each topic segment configured in the topic segment monk 201. For example, the location information extraction unit 303 sets the extracted address information to an entity name of LOCATION.

In addition, the location information extractor 303 extracts distance information. Here, the distance information includes a moving distance of the story generating device, a distance to a destination for moving, and the like.

The text extractor 305 extracts a feature from text information of scan and auxiliary events included in each topic segment configured in the topic segment 201. For example, the text extractor 305 extracts a verb-containing text, extracts a title text, extracts a formatted text, extracts a file name, extracts sentence information, and emotions using an entity name recognizer.

In the case of extracting the verb-containing text, the text extracting unit 305 checks whether a word included in the text is included in the noun dictionary. In this case, the text extracting unit 305 removes the survey included in the word and checks whether it is included in the noun dictionary.

When the word is included in the noun dictionary, the text extracting unit 305 checks the entity name of the word. For example, the text extraction unit 305 checks the entity name of the recognized word using the entity name dictionary as shown in Table 3 below.

noun Object name Topic bicycle OBJECT festival OBJECT Visit, festival station LOCATION Travel, work friend PERSON Life, picnic blackboard class House LOCATION ... ... ...

The text extractor 305 may separately configure an entity name pattern dictionary including nouns having a predetermined pattern to reduce the size of the noun dictionary as shown in Table 4 below.

noun Object name Topic ~ Library LOCATION Study Exam / Certificate ~ Reading room LOCATION Study Exam / Certificate School LOCATION Study Exam / Certificate Museum of Art LOCATION Spectator painting ~ Exhibition Hall LOCATION Spectator painting ~ School LOCATION Class festival presentation University LOCATION Class festival presentation University LOCATION Class festival presentation ~ Match OBJECT Athletic Competition Arboretum LOCATION Travel Movie theater LOCATION Preview Fire LOCATION Drawing ~ Restaurant LOCATION Anniversary Cake OBJECT Anniversary Cake OBJECT Anniversary Insurance OBJECT finance Savings OBJECT finance Securities OBJECT finance ~ Bank LOCATION finance Insurer LOCATION finance Baked ~ OBJECT Cooking Fried ~ OBJECT Cooking Clause ~ OBJECT Cooking Roasted ~ OBJECT Cooking Chorin ~ OBJECT Cooking Seasoning ~ OBJECT Cooking Minced ~ OBJECT Cooking Bibin ~ OBJECT Cooking Chinchin ~ OBJECT Cooking Chasun ~ OBJECT Cooking Sliced ~ OBJECT Cooking Toss ~ OBJECT Cooking Pot ~ OBJECT Cooking ~ Bowl OBJECT Cooking Cutting OBJECT Cooking Rice bowls OBJECT Cooking Cooking OBJECT Cooking To stew OBJECT Cooking Meat OBJECT Cooking Ribs OBJECT Cooking Recipe OBJECT Cooking Studio LOCATION Picture Photographer OBJECT Picture ... lens OBJECT Picture Animals LOCATION Pets Dog ~ LOCATION Ewan Finger OBJECT pain Ligament OBJECT pain

When using the entity name pattern dictionary as shown in Table 4, the text extracting unit 305 checks whether the word is included in the entity name pattern dictionary if the word is not included in the noun dictionary.

When the word is not included in the noun dictionary and the entity name pattern dictionary, the text extractor 305 checks whether the word is included in the verb dictionary or the verb utilization pattern dictionary. For example, the verb dictionary is configured as shown in Table 5 below.

topic verb synonyms tmplKey study Convey Counting One study Counting Convey 2 Entrance examination Gathering Convene 3 Travel Gathering Convene 3 Athletic Gathering Convene 3 Entrance examination Through Stuck, passing 4 ... ... ... ...

For another example, the verb utilization pattern dictionary is configured as shown in Table 6 below.

Conjugation Pattern Circular pattern To ~ | ~ That I went To Did you Shiki To cry ~ Woo To Oh Pulled To ... ...

Here, "~ |" indicates that "~" can be used in multiple patterns.

In the case of extracting the title text, the text extracting unit 305 checks whether a word included in the title text is included in the noun dictionary. When the word is included in the noun dictionary, the text extracting unit 305 checks the entity name of the word.

As described above, when extracting the verb-containing text and the title text, the text extraction unit 305 checks whether or not the verb is included in the recognized entity name. If the recognized entity name does not include a verb, the text extractor 305 adds a basic verb to the entity name.

In the case of the sentence linkage information extraction, the text extractor 305 extracts the sentence linkage information from the scan cases and auxiliary events included in each topic segment configured in the topic segment 201. In this case, the text extracting unit 305 extracts sentence linking information using a sentence linking ending dictionary as shown in Table 7 below.

key Connection name ... high, ..., old book, ... Conjoint Or, or, or, me, ... Detachable I, I, I, I, I, ... Opposition So, it's ... Cause If, if, ... Condition I'm going to ... Purpose / Intention Dependency ... high, low, ... Method / degree dependence

For example, if the sentence "Cheol and Young-hee went home, and Yeong-min went home," the text extractor 305, "Cheol and Young-hee went home" and "Young-min" Also went home together. " At this time, the text extraction unit 305 is inserted into the information "joint (circular" ga ")" and "go (circular" ga ")" is a joint type.

In the case of extracting the emotion information, the text extractor 305 extracts the emotion information from the scan cases and the auxiliary events included in each topic segment configured in the topic segment 201. In this case, the text extractor 305 extracts emotion information using an emotion dictionary as shown in Table 8 below.

Emotional nouns / verbs Emotion information Glad PLEASURE Happy PLEASURE mad ANGER Split ANGER ㅠㅠ SADNESS fence SADNESS ... ...

When extracting the standardized text, the text extraction unit 305 extracts a feature from data of a standardized form included in an injection case and an auxiliary event included in each topic segment configured in the topic segment 201. do. In this case, the text extracting unit 305 extracts a feature from the data in the standardized format using a standardized pattern dictionary as shown in Table 9 below.

Card company Basic verb pattern KB Card Pay 1234 | Mdd | H: m :: \\ [KB (?: Check | card) \\] (?: [^] +?) 'S card (?:. *?) ? [0-9] month [0-9]? [0-9] day) ([0-9]? [0-9] \\: [0-9]? [0-9]) ([0 -9,] +) using circle ([^] +) BC / NH card Pay 3124 | M / d | h: m :: \\ [(?:. *?) Approval \\] ([0-9,] +) 원 [^] + (?: [^] +?) [0-9]? [0-9] \\ / [0-9]? [0-9]) ([0-9]? [0-9] \\: [0-9]? [0- 9]) ([^] +) $ Shinhan Card Pay 1243 | Md d | h: m :: Shinhan Card (?: [^] +?) ([0-9]? [0-9] Month [0-9]? [0-9])) ( [0-9]? [0-9] \\: [0-9]? [0-9]) ([^] +) ([0-9,] +) Using circle (?:. *?) Samsung Card Pay 1243 | M / d | h: m :: \\ [Use Samsung Card \\] ([0-9]? [0-9] \\ / [0-9]? [0-9]) ([0 -9]? [0-9] \\: [0-9]? [0-9]) ([^] +) ([0-9,] +) ([^] +) $ Lotte Card Pay 3124 | M / d | h: m :: Lotte Card (?: [^] +?) Won ([0-9,] +) won ([^] +) ([0-9]? [0-9 ] \\ / [0-9]? [0-9]) ([0-9]? [0-9] \\: [0-9]? [0-9]) ([^] +) $ Hana SK Card Pay 1234 | M / d | h: m :: \\ [Hana SK Card \\] \\ ([0-9 *] + \\) (?: [^] +?) [0-9] \\ / [0-9]? [0-9]) ([0-9]? [0-9] \\: [0-9]? [0-9]) ([^ /] +) \\ / ([0-9,] +) won \\ / approved \\ / ([^] +) $ Hyundai Card Pay 234 || h: m :: \\ [Hyundai Card [^]] * \\] \\-Approval (?: [^] +?) ([0-9]? [0-9] \\: [0-9]? [0-9]) ([0-9,] +) one \\ ([^)] + \\) ([^] +) $ ... ... ...

When extracting the photo file name, the text extracting unit 305 extracts the path information and tag information of the photo-related file included in the scan gun and the sub-event included in each topic segment configured in the topic segment monk 201. Extract the entity name by using

The entity name setting unit 307 sets input data to a predetermined entity name. For example, set the data of the file name column in the photo log to the object name OBJECT.

The feature extracting unit 203 checks occurrence frequency of each topic by using extracted feature information. For example, the entity name among the features extracted by the feature extraction unit 203 includes corresponding topic information as shown in Table 2 above. Accordingly, the feature extraction unit 203 updates the frequency of occurrence of the topic according to the extracted topic information. For example, when the address extracted by the location information extractor 301 is "Seogwipo-si, Jeju-do, Korea," and "Jeju-do" and "Seogwipo-si" are set to the topic of travel in the noun dictionary, the feature extractor 203 may travel. Set the frequency of occurrences of the topic to 2.

The feature extraction unit 203 configured as described above may have a different structure depending on the format of log data. Accordingly, the configuration of the feature extraction unit 203 shown in FIG. 3 may be changed according to the format of log data.

The features extracted by the feature extracting unit 203 configured as described above may be managed in the form as shown in Table 10 below.

Structure name Cebu Government Main quality management Manage verbs and entity names that can be sentence components Additional feature management A feature that cannot be a sentence component but affects topic decisions Sentence linking ending management Verbs: linkage information between verbs Topic frequency information by log Manage frequency information / data required for topic determination

As shown in Table 3, among the features extracted by the feature extraction unit 203, other feature information except for the frequency information of the sentence connection ending and the log-specific topic are classified and managed as the main feature and the additional feature. For example, the social information extracting unit 301, the location information extracting unit 303, the text extracting unit 305, and the entity name setting unit 307 respectively classify the extracted features into main or additional features. do. In another example, the feature extracting unit 203 may extract features extracted from the social information extracting unit 301, the location information extracting unit 303, the text extracting unit 305, and the entity name setting unit 307. It may be configured to include a feature extraction post-processor to classify the main feature or additional features.

As described above, the story generating device determines the episode according to the user's daily information by analyzing log information collected from at least one electronic device. At this time, the story generating device generates a story according to the user's daily information, as shown in FIG.

4 illustrates a story generation procedure based on user information according to an exemplary embodiment of the present invention.

Referring to FIG. 4, in operation 401, the story generating device collects log information from at least one electronic device. For example, the story generating device collects daily information of the user using data received from the outside, data generated inside, and data stored in an external storage. Here, the data received from the outside includes weather, latitude and longitude information, blog information, social data, and the like. The internally generated data includes call logs, short messages (SMS), multimedia messages (MMS), latitude and longitude information measured by satellite navigation equipment (GPS), schedules and address books. Data stored in the external storage includes an image photographed image, a list of played music files, TV viewing information, and the like.

In operation 403, the story generating apparatus constructs a topic segment that is a set of candidate topics for composing a story using the collected log information. For example, the story generating device configures a topic segment as shown in FIG. 5.

After configuring the topic segment in step 403, the story generating apparatus proceeds to step 405 to extract the features from the scan gun and the auxiliary event of each topic segment. For example, the story generating apparatus extracts a feature from each topic segment as shown in FIG. 6 using the feature extracting unit 203 configured as shown in FIG. 3. Here, the qualities represent data that may be sentence components. For example, the qualities include verbs, entity names, sentence association information, topic information associated with entity names, and additional information which may be sentence elements.

After extracting the feature from each topic segment, the story generating apparatus proceeds to step 407 to determine a topic for each topic segment in consideration of the frequency of occurrence of the topic in the logs. For example, the story generating apparatus determines a topic of each topic segment by using Equation 1.

After determining the topic for each topic segment, the story generating apparatus proceeds to step 409 to determine the optimal topic segment in each time interval. For example, when there are a plurality of topic segments in the same time interval, the story generating apparatus selects an optimal topic segment of the corresponding time interval by comparing topic scores of topics included in each topic segment. In this case, the story generating apparatus recognizes the topic of the optimal topic segment as an epid of the corresponding time.

In operation 411, the story generating apparatus merges the topic segments having the same topic into one topic segment. For example, if the topics of consecutive topic segments are the same, the story generating device merges the consecutive topic segments into one topic segment. In another example, when the topics of the same segment of the same time interval are the same, the story generating apparatus merges the topic segments into one topic segment.

After merging the topic segments having the same topic, the story generating apparatus proceeds to step 413 and determines whether to generate a story over time.

If it is determined to generate the hourly story, the story generating apparatus proceeds to step 415 to generate each story by time using the episode determined in step 411. For example, the story generating device generates a story by time as shown in FIG. 7.

In operation 417, the story generating apparatus displays the story generated by the time on the display unit. For example, the story generating apparatus displays a story generated by time as shown in FIG. 9A on the display unit. In this case, the story generating device may output not only a text-type sentence but also multimedia information related to the sentence.

On the other hand, if it is decided not to generate a story by time, the story generating device recognizes that the story by theme is generated. Accordingly, the story generating apparatus proceeds to step 419 to determine the main topic. That is, the story generation device determines the main topic that can represent the most important feature of the user schedule information in consideration of the episodes of each time section. For example, the story generating apparatus determines a main topic by using Equation 3.

After determining the main topic, the story generating apparatus proceeds to step 421 to merge the qualities in the episode related to the main topic to generate a story related to one main topic. In this case, the story generating device may generate a story related to the weather, the user's emotion as well as the main topic. For example, the story generating device generates a story by theme as shown in FIG. 8.

In operation 417, the story generating apparatus displays the story generated for each subject on the display unit. For example, the story generating apparatus displays a story generated for each subject on the display unit as illustrated in FIG. 9B. In this case, the story generating device may output not only a text-type sentence but also multimedia information related to the sentence.

Thereafter, the story generating device ends the present algorithm.

5 illustrates a procedure for configuring a topic segment according to an embodiment of the present invention.

Referring to FIG. 5, after collecting log information in step 401 of FIG. 4, the story generating apparatus checks whether a main event exists in step 501. In this case, the scan gun represents schedule information that is mainly used to construct a story among the collected user schedule information.

If there is no scan gun, the story generating apparatus proceeds to step 507 and determines whether to generate a dummy segment.

On the other hand, if there is a main event, the story generating device proceeds to step 503 to form a topic segment with the main event. For example, when the schedule information is a main event, the story generating device lists the schedule information in chronological order and organizes each schedule information into topic segments. If the occurrence time of the main event overlaps, the story generating device divides the topic segment based on the main event of the short time interval. For example, if there is a schedule A from 13:00 to 14:00 and a schedule B from 13:00 to 15:00, the story generating device constructs a topic segment for each of schedules A and B from 13:00 to 14:00. do. In addition, the story generating device configures a topic segment for schedule B from 14:00 to 15:00.

After configuring the topic segment, the story generating apparatus proceeds to step 505 and, when the occurrence time of the auxiliary event is included in the time segment of the topic segment, maps the corresponding auxiliary event to the topic segment. Here, the auxiliary event represents information excluding the scan gun from the collected user schedule information.

In operation 507, the story generating apparatus determines whether to generate a dummy segment. For example, the story generating device generates a dummy segment using an auxiliary event not mapped to a topic segment. Accordingly, the story generating apparatus checks whether there is an auxiliary event not mapped to the topic segment.

If there is no auxiliary event not mapped to the topic segment, the story generating device recognizes that it does not generate a dummy segment. Accordingly, the story generating apparatus proceeds to step 405 to extract the features from each topic segment.

On the other hand, if there is an auxiliary event not mapped to the topic segment, the story generating device proceeds to step 509 to generate a dummy segment using the auxiliary event not mapped to the topic segment. In this case, the dummy segment is set to a predetermined time interval.

In operation 405, the story generating apparatus extracts a feature from each topic segment.

Hereinafter, a description will be given of a method for extracting a feature from text data in the text extractor 305.

6 illustrates an object recognition procedure according to an embodiment of the present invention.

Referring to FIG. 6, after configuring a topic segment in step 403 of FIG. 4, the story generating apparatus determines in step 601 whether to perform object name recognition using an object name recognizer. For example, the story generating apparatus checks whether there is a candidate word for entity name recognition through the individual name recognizer.

If the entity name recognition is not performed using the entity name recognizer, the story generating apparatus recognizes that there is no candidate terminal for entity name recognition. Accordingly, the story generating device ends the present algorithm. In other words, the story generating apparatus proceeds to step 407 to determine the topic of each topic segment.

Meanwhile, when performing entity name recognition, the story generating device recognizes that there is a candidate terminal requiring entity name recognition. In step 603, the story generating apparatus checks a word for object name recognition. For example, when extracting the qualities from the text of "Cheol and Alice went home," the story generating apparatus checks "Cheolwa" for object name recognition.

In operation 605, the story generating apparatus checks the entity name of the word identified in operation 603. For example, in the case of confirming the entity name of "heolhwa", the story generating apparatus checks whether "heol" is included in the noun dictionary after removing "he" which is a survey from "heolhwa". If the "withdrawal" is not included in the noun dictionary, the story generation device may check whether the "withdrawal" is included in the entity name pattern dictionary configured as shown in Table 4 below.

When the entity name of the word checked in step 603 is confirmed, the story generating apparatus proceeds to step 611 to update a frequency of occurrence of the topic of the recognized entity name. For example, when the PERSON entity name of the word "friend" is confirmed, the story generating device adds one occurrence frequency of "life" and "excursion" which are topics of PERSON.

On the other hand, if the entity name of the word cannot be confirmed, the story generating apparatus proceeds to step 607 to check whether the word is included in the verb dictionary or verb utilization pattern dictionary.

If the word is not included in the verb dictionary or the verb utilization pattern dictionary, the story generating device recognizes that the feature cannot be extracted from the word. Accordingly, the story generating apparatus proceeds to step 609 to check whether the next word exists.

When the word is included in the verb dictionary or the verb utilization pattern dictionary in step 607, the story generating device recognizes that the verb estimation is successful. Accordingly, the story generating apparatus proceeds to step 611 to update the frequency of occurrence of the topic according to the verb information.

In operation 609, the story generating apparatus determines whether a next word exists.

If the next word does not exist, the story generating device recognizes that all the qualities are extracted. Accordingly, the story generating apparatus proceeds to step 407 to determine the topic of each topic segment.

On the other hand, if there is a next word, the story generating apparatus proceeds to step 603 to identify the next word to extract the feature.

Although not shown, the story generating apparatus checks whether the verb is included in the recognized entity name. For example, in the case of extracting the nouns of "Abstract", "Younghee", "House", and the verb "Go" from the text of "Cheon and Young-Hee went home," the story generating device is "Hole", "Young-Hee". The noun of "house" is understood to include the verb "go". However, when no verb is included in the recognized entity name, the story generating device estimates a verb to be added to the entity name in the basic verb dictionary.

7 illustrates a procedure for generating a story by time according to an embodiment of the present invention.

Referring to FIG. 7, if it is determined in step 419 that the story is generated by time, the story generating device generates an intermediate sentence for each episode using the main features of each episode in step 701. For example, the story generating device identifies topics and verbs in the main qualities of the episode. Subsequently, the story generating device selects a sentence template corresponding to the topic and a verb in Table 11 below. Here, the sentence template indicates a sentence format constituted by a specific topic and a verb.

Topic verb A significant verb Heraldry template Required study Convey Counting You gave <OBJECT> <> from <LOCATION> on <DAY> <TIME> OBJECT study Counting Convey Passed <OBJECT> <> from <LOCATION> on <DAY> <TIME> OBJECT Admission, Travel, Sports Gathering Convene <ORG> to <LOCATION> / set LOCATION Entrance examination Through Stuck, passing <PERSON> <this> <OBJECT> <> passed on <DAY> <TIME> OBJECT Diet Exercise Worked from <LOCATION> to <PERSON> <and> <OBJECT> <> on <DAY> <TIME> $ holiday I have a date with my girlfriend on a holiday. $ anger I was angry today ... ... ... ... ...

In Table 11, the first column indicates a topic. In this case, ＄ represents a sentence template used for generating a basic sentence. The second column represents verbs, and the third column represents similar verbs with the same meaning as the verbs in the second column, but with different expressions. The fourth column is a sentence template, organized by verb and topic. The fifth column shows entity name information that is essential when a sentence template is used. For example, if the topic of the episode is "study" and the verb is "tell", the story generating device may change <OBJECT> <from <LOCATION> to "<DAY> <TIME> according to <Table 11>. Recognize a sentence template, such as> / ", as the middle sentence of the episode. In this case, the story generating device essentially requires any one of the <DAY>, <TIME>, <LOCATION>, and <OBJECT>. For example, if it is necessary to have an <OBJECT>, then an intermediate statement is created, including the <OBJECT> object name. If there is no required <OBJECT> entity name, the story generating device does not form an episode because the story template generation condition is not satisfied.

Additionally, the story generation device may generate intermediate sentences with similar verbs rather than verbs of episodes to provide diversity of sentence generation. For example, if the topic of the episode is "study" and the verb is "tell", the story generating device uses "Kun-na-ju", which is a significant verb of "tell", to "<DAY> <TIME>". Sentence templates such as <OBJECT> <and / handed> in <LOCATION> may be recognized as intermediate sentences of the episode.

After generating the intermediate sentence for the episode, the story generating apparatus proceeds to step 703 to remove the duplicate sentence. For example, if a middle sentence between “going to / from school” and “to go to / from school today” is generated, the story generating device may be in the middle of “to / from / going to school with today”. Delete "go to school" because the sentence covers the middle sentence of "go / go".

After removing the duplicate sentences, the story generating apparatus proceeds to step 705 to convert the intermediate sentences into complete sentences. That is, the story generating device converts the intermediate sentence into a complete sentence using the mother information and the sentence connection information. For example, if there is an intermediate sentence of “today at school / tested / tested” and “withdrawn / basketballed / basketballed” and sentence linking information is conjoined, the story generating device may interpret the two sentences. Concatenate one sentence. In this case, the story generating apparatus selects a sentence connecting ending to connect two sentences using the sentence connecting ending dictionary of <Table 7>. The story generating device then selects the ending mother to generate a complete sentence, "tested at school today, withdrawn and basketball." In this case, the story generating device may select the ending ending in consideration of at least one of region, gender and age. If the ending ending is set differently for each age, the story generating device may select the ending ending using the following Table 12.

age sex eomi 30 ha south I'm eating it. 30 ha female I'm using it. 30 phases south Yes, yes, yes, yes. 30 phases female I'm buying it.

On the other hand, when the ending ending is set differently for each region, the story generating device may select the ending ending using the following <Table 13>.

loc eomi Gangwon All game All Jeolla All done Chungcheong Dashu fish oil Ordinary It's all daimyeo Jeju All

After converting the intermediate sentence into a complete sentence in step 705, the story generating apparatus proceeds to step 417 to display the sentence generated by the user on the display unit for checking.

8 illustrates a procedure for generating a story for each topic according to an embodiment of the present invention.

Referring to FIG. 8, if it is determined in step 419 that a story is generated for each topic, the story generating device collects the quality of episodes related to the main topic in step 801. In this case, the story generating apparatus merges episodes of the same topic but having different occurrence times into one episode.

In operation 803, the story generating apparatus generates an intermediate sentence using the collected features. For example, the story generating device identifies a topic and a verb from the collected qualities. Subsequently, the story generating device selects a sentence template corresponding to the topic and a verb from the <Table 11>. If the topic of the episode is "Study" and the verb is "tell", the story generating device uses <OBJECT> <<> in <LOCATION> at << DAY> <TIME> according to <Table 11>. Sentence template, such as "received" as the middle sentence of the episode. In this case, the story generating device essentially requires any one of the <DAY>, <TIME>, <LOCATION>, and <OBJECT>. For example, if it is necessary to have an <OBJECT>, then an intermediate statement is created, including the <OBJECT> object name. If there is no required <OBJECT> entity name, the story generating device does not form an episode because the story template generation condition is not satisfied.

Additionally, the story generation device may generate intermediate sentences with similar verbs rather than verbs of episodes to provide diversity of sentence generation. For example, if the topic of the episode is "study" and the verb is "tell", the story generating device uses "Kun-na-ju", which is a significant verb of "tell", to "<DAY> <TIME>". Sentence templates such as <OBJECT> <and / handed> in <LOCATION> may be recognized as intermediate sentences of the episode.

After generating the intermediate sentence for the episode, the story generating device proceeds to step 805 to remove the duplicate sentence. For example, if a middle sentence between “going to / from school” and “to go to / from school today” is generated, the story generating device may be in the middle of “to / from / going to school with today”. Delete "go to school" because the sentence covers the middle sentence of "go / go".

After removing the duplicate sentences, the story generating apparatus proceeds to step 807 and converts the intermediate sentences into complete sentences. That is, the story generating device converts the intermediate sentence into a complete sentence using the mother information and the sentence connection information. For example, if there is an intermediate sentence of “today at school / tested / tested” and “withdrawn / basketballed / basketballed” and sentence linking information is conjoined, the story generating device may interpret the two sentences. Concatenate one sentence. In this case, the story generating apparatus selects a sentence connecting ending to connect two sentences using the sentence connecting ending dictionary of <Table 7>. The story generating device then selects the ending mother to generate a complete sentence, "tested at school today, withdrawn and basketball." In this case, the story generating device may select the ending ending in consideration of at least one of region, gender and age. If the ending ending is set differently for each age, the story generating device may select the ending ending using the following Table 12.

After converting the intermediate sentence into a complete sentence in step 807, the story generating apparatus proceeds to step 417 and displays the sentence generated by the subject on the display unit for the user to check.

As described above, when generating a story by theme, the story generating device may generate a story related to weather and user emotion as well as a main topic.

In the above-described embodiment, the mobile device performs user information collection, user information analysis, and story generation. In this case, the mobile device may estimate user's behavior pattern by accumulating user information. Accordingly, the mobile device can generate a story based on topic information that can summarize a user's behavior pattern.

In another embodiment, user information may be collected by the mobile device, and information analysis and stories may be generated by an information analysis server. For example, the mobile device provides the collected user information to the information analysis server. The information analysis server analyzes user information provided from the mobile device to generate a story. Thereafter, the information analysis server provides the generated story to the mobile device. The mobile device displays a story generated by the information analysis server on a display unit for the user to check. That is, the information analysis server performs the same operations as the analysis unit 120 and the story generation unit 130 of FIG. 1. In this case, the information analysis server may estimate user's behavior pattern by accumulating user information. Accordingly, the information analysis server may generate a story based on topic information that can summarize the behavior pattern of the user.

In another embodiment, the mobile device may collect user information, analyze information in an information analysis server, and generate a story in a story generation server. For example, the mobile device provides the collected user information to the information analysis server. The information analysis server analyzes the user information provided from the mobile device and provides it to the story generation server. The story generation server generates a story using the analysis information on the user information provided from the information analysis server, and provides the generated story to the mobile device. The mobile device displays a story generated by the information analysis server on a display unit for the user to check. That is, the information analysis server performs the same operation as that of the analysis unit 120 of FIG. 1, and the story generation server performs the same operation as the story generation unit 130 of FIG. 1. In this case, the information analysis server may estimate user's behavior pattern by accumulating user information. Accordingly, the story generation server may generate a story based on topic information that can summarize the behavior pattern of the user provided from the information analysis server.

In another embodiment, the mobile device may collect user information and generate a story, and analyze the information in the information analysis server. For example, the mobile device provides the collected user information to the information analysis server. The information analysis server analyzes the user information provided from the mobile device and provides it to the mobile device. The mobile device generates a story using analysis information on user information provided from the information analysis server, and displays the story on a display unit for the user to check. That is, the information analysis server performs the same operation as the analysis unit 120 of FIG. In this case, the information analysis server may estimate user's behavior pattern by accumulating user information. Accordingly, the mobile device can generate a story based on topic information that can summarize the behavior pattern of the user provided from the information analysis server.

In the above-described embodiment, the story generating device generates a story by selecting a topic related to user daily information. In this case, the story generating device may select a topic using a topic classification list as shown in Table 14 below.

Main Category Small classification Situation Definition Include in all topics Credit card Credit card usage history currency Call on your phone letter Everyday letters, text on your phone, email life religion Religious Activities finance Activities related to equity investment and finance Plan / Promise When making appointments or planning is the main thing housework Helping or carrying out household work Diet When work related to diet activity is main pain If you get sick and go to the hospital or care Anniversary Personal holidays from everyday life such as wedding anniversary and hundred days Special holiday Holidays on the calendar with special attributes, such as Christmas, New Year, New Year's Day, and Chuseok human relationship human relationship An activity that occurs between family, friends, or employees, such as a meal or a meal, and does not belong to any other topic. Hobby Exercise When performing for the purpose of enjoying sports, not for diet, such as soccer or baseball Reading Borrow books or read books in libraries fishing music Activities related to listening to music game Computer games, board games Drawing A series of activities to draw or watch Picture Taking photos or activities related to photography party Birthday party, celebration party Cooking Making or cooking Pets Washing or playing with dogs TV Watching TV Outdoor activities Travel Work in zoo, botanical garden, amusement park
Overseas travel, domestic travel Preview Theater
musical
concert shopping Buying something and shopping school life class Acts related to classes other than study study Self-development, study related activities Athletic Competition Athletic meet, athletic meet festival School festival exam Midterm exam, Final exam, Note test Presentation Presentation Seminars, lectures, academic conferences Exam, Notice Exam / Certification Activities related to midterm, final exams, various examinations, and certification Entrance examination Interview with a university, activities related to joining a company job Entertainment Hospitality Related to the Contract conference Weekly meetings, monthly meetings, business meetings Business trip task What happens in the workplace on a daily basis, not related to topics related to other jobs

In addition, the story generating apparatus may store and analyze log information based on a preset environment setting file as shown in Table 15 below.

Line number Configuration Information Configuration example 1 line Topic Information Life, class, study, athletic meet, ..., picnic, ... 2 lines Distance Information by Topic 0, 0, 0, 0, ..., 100, ... 3 lines Entity Name Information PERSON, ORG, DAY, TIME, LOCATION, OBJECT 4 lines
...
n lines Log-specific setting information SMS \ tS \ t \ t0.2 \ t000010 \ t010000
Event \ t
GPS \ tG \ t1: Walking | Walking | Travel: 0 | 100 | 50000 \ t0.1 \ t100100 \ t100200
Call Log \ tC \ t2: Call To: 20 \ t0.1 \ t020000 \ t100002

The environment setting information of the GPS configured using the <Table 15> may be configured as shown in the following <Table 16>.

Item Configuration Information Configuration example Item 1 Log name GPS Item 2 Log code G Item 3 Basic verb information 1: Walking | Walking | Travel: 0 | 100 | 50000 Item 4 Log weight 0.1 Item 5
...
Item n Log data
Processing method information 100100 \ t100200

In Table 16, basic verb information defines basic verbs used according to distance. That is, when the distance is 0m or more and less than 100m, the story generating device selects "walk" as the basic verb, and when the distance is 100m or more and less than 50km, the "verb" is selected as the basic verb.

Meanwhile, in the detailed description of the present invention, specific embodiments have been described, but various modifications may be made without departing from the scope of the present invention. Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined not only by the scope of the following claims, but also by the equivalents of the claims.

Claims (24)

An apparatus for summarizing daily information of a user,
An information collecting unit collecting log information including daily information of a user from at least one electronic device;
An analysis unit for analyzing at least one log information collected by the information collection unit to determine at least one topic representing daily information of a user;
And a story generator for generating at least one sentence representing daily information of a user using at least one topic determined by the analyzer.
The method of claim 1,
The analysis unit,
A topic segment configuration unit constituting at least one topic segment including at least one candidate topic using log information collected by the information collection unit;
An extraction unit for extracting sentence components for constructing sentences from each topic segment;
A topic determiner which determines a topic for each topic segment in consideration of the frequency of occurrence of the topic according to the sentence component extracted by the extractor;
And an episode determination unit for selecting one topic segment for a reference time in consideration of the frequency of occurrence of the topic included in each topic segment.
The method of claim 2,
The topic segment configuration unit configures at least one topic segment using any one of the plurality of log information using reference log information, and maps auxiliary log information to each topic segment in consideration of an occurrence time of the auxiliary log information. Dummy segments are constructed using secondary log information that is not mapped to topic segments.
The auxiliary log information, characterized in that the log information of the plurality of log information except the reference log information, characterized in that the log information.
The method of claim 2,
The extractor is characterized in that it comprises an entity name recognizer.
The method of claim 2,
The extractor, characterized in that for updating the frequency of occurrence of the topic for the sentence component extracted from each topic segment.
The method of claim 2,
The episode extraction unit, characterized in that for selecting any one of the reference topic in consideration of the frequency of occurrence of the topic of the selected topic segment at each reference time.
The method according to claim 6,
The episode determiner, when the reference topics of the consecutive topic segments are the same, characterized in that the consolidation of consecutive topic segments including the same reference topic into one topic segment.
The method of claim 1,
The analysis unit, characterized in that for analyzing the log information collected by the information collecting unit according to the preset environment setting information.
The method of claim 1,
The story generator comprises at least one intermediate sentence using sentence components included in an episode set at a reference time, removes an intermediate sentence having the same meaning, and adds a sentence linking ending and ending ending information to the intermediate sentence. Device for generating a sentence for the episode.
The method of claim 9,
The story generation unit, characterized in that for setting the ending ending in consideration of at least one of gender, region and age.
The method of claim 1,
The story generation unit, characterized in that for generating at least one sentence representing the user's daily information in consideration of at least one of the reference topic, weather, and emotion that is representative of the user's daily information.
The method of claim 1,
And a display unit for displaying at least one sentence generated by the story generator.
In a method for summarizing daily information of a user,
Collecting log information including daily information of a user from at least one electronic device;
Determining at least one topic representing daily information of a user by analyzing log information collected by the information collector;
And generating at least one sentence representing daily information of a user using at least one topic determined by the analysis unit.
The method of claim 13,
The process of determining the topic,
Constructing at least one topic segment including at least one candidate topic using the collected log information;
Extracting sentence components for constructing sentences from each topic segment;
Determining a topic for each topic segment in consideration of the frequency of occurrence of the topic according to the extracted sentence component;
Selecting one topic segment for a reference time in consideration of the frequency of occurrence of the topic included in each topic segment.
The method of claim 14,
The process of configuring the topic segment,
Configuring at least one topic segment by using any one of the plurality of log information based on the reference log information;
Mapping the auxiliary log information to each topic segment in consideration of the occurrence time of the auxiliary log information;
If there is auxiliary log information not mapped to the topic segment, including a step of configuring a dummy segment by using the auxiliary log information not mapped to the topic segment;
And the auxiliary log information indicates log information excluding the reference log information among the plurality of log information.
The method of claim 14,
Extracting the sentence component; extracting the sentence component from textual data using an entity name recognizer.
The method of claim 14,
And after extracting the sentence component, updating the frequency of occurrence of the topic for the sentence component extracted from each topic segment.
The method of claim 14,
The process of selecting the topic segment,
And selecting one reference topic in consideration of the frequency of occurrence of the topic of the selected topic segment at each reference time.
19. The method of claim 18,
Selecting the reference topic and checking whether the reference topics of consecutive topic segments are the same;
If the reference topics of consecutive segments are the same, further comprising merging consecutive topic segments comprising the same reference topic into one topic segment.
The method of claim 13,
Determining the at least one topic,
And determining at least one topic by analyzing the collected log information according to preset environment setting information.
The method of claim 13,
The process of generating the sentence,
Constructing at least one intermediate sentence using sentence components included in the episode set at the reference time;
Removing the middle sentence of the same meaning,
And adding a sentence linking ending and ending ending information to the middle sentence to generate a sentence for the episode.
22. The method of claim 21,
The ending ending is set in consideration of at least one of gender, region and age.
The method of claim 13,
The process of generating the sentence,
And generating at least one sentence representing the user's daily information in consideration of at least one of reference topics, weather, and emotions that are representative of the user's daily information.
The method of claim 13,
And displaying the generated at least one sentence.

Images