WO2015070683A1 - 推理社会关系的方法及装置 - Google Patents

推理社会关系的方法及装置 Download PDF

Info

Publication number
WO2015070683A1
WO2015070683A1 PCT/CN2014/088437 CN2014088437W WO2015070683A1 WO 2015070683 A1 WO2015070683 A1 WO 2015070683A1 CN 2014088437 W CN2014088437 W CN 2014088437W WO 2015070683 A1 WO2015070683 A1 WO 2015070683A1
Authority
WO
WIPO (PCT)
Prior art keywords
activity
users
events
event
determining
Prior art date
Application number
PCT/CN2014/088437
Other languages
English (en)
French (fr)
Inventor
丁强
王娜敏
余辰
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Publication of WO2015070683A1 publication Critical patent/WO2015070683A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/20—Services signaling; Auxiliary data signalling, i.e. transmitting data via a non-traffic channel
    • H04W4/21—Services signaling; Auxiliary data signalling, i.e. transmitting data via a non-traffic channel for social networking applications
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/335—Filtering based on additional data, e.g. user or group profiles
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/24—Querying
    • G06F16/245—Query processing
    • G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
    • G06F16/2462—Approximate or statistical queries
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9535—Search customisation based on user profiles and personalisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9537—Spatial or temporal dependent retrieval, e.g. spatiotemporal queries
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/50—Network services
    • H04L67/535—Tracking the activity of the user
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02—Services making use of location information
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02—Services making use of location information
    • H04W4/029—Location-based management or tracking services

Definitions

  • the present invention relates to the field of communication technologies, and in particular, to a method and apparatus for inferring social relations.
  • the social network service can provide users with personalized services according to the social relationship between users, bringing a lot of convenience to users.
  • the social relationship between users can be obtained through the reasoning of the daily activities of the user.
  • Various methods are used in the prior art, for example, according to the frequency of meetings between users in a specific time period, or the identification of personal photos of users, or The communication information of the virtual network, the distribution of user interests, and the like are used to estimate the social relationship between users.
  • the social relationship of the user is inferred through characteristics such as user interest, communication messages in the virtual network, and personal photos, all of which are based on a certain feature of the user's daily activities. . Because the relationship features used for reasoning are single, the categories are few, and the relationship in the virtual network is not true, the photo recognition process is prone to errors, etc., so the inferred social relationship has one-sidedness and unreality, and can not correctly derive the true between users. The social relationship, so as not to provide users with the correct social network services, brings a lot of inconvenience to users.
  • Embodiments of the present invention provide a method and apparatus for inferring a social relationship, which can solve the problem that the social relationship inferred is not true when inferring the social relationship between users.
  • an embodiment of the present invention provides a method of inferring a social relationship, comprising:
  • the determining the common activity event according to the activity event includes:
  • the determining, according to the common activity event, the social relationship of the at least two users includes:
  • the total number of common activity events of the two users in the important user pair is greater than a quantity threshold, and the cumulative duration of the common activity events of the two users in the important user pair is greater than a time threshold;
  • Extracting semantic features of the common activity event of the important user pair the semantic feature including a time period, a location type, and an activity type;
  • the social relationship is determined according to at least one of the semantic features and classification criteria, the classification criteria being determined based on the activity parameters.
  • the method before the calculating an activity type relevance score of the activity events of the two users according to the activity data, the method further includes:
  • the activity type relevance scores of the activity events of the two users are stopped according to the activity data.
  • the determining two Whether the user's activity events meet the preset conditions include:
  • the geographic distance is less than the distance threshold
  • the intersection time is greater than the time threshold
  • the sensor data similarity is greater than the similar threshold
  • an embodiment of the present invention provides an apparatus for inferring a social relationship, including:
  • An acquiring unit configured to acquire activity data of at least two users, and determine activity events of the at least two users according to the activity data, where the activity data includes location location, time, and sensor data;
  • a first determining unit configured to determine a common activity event according to the activity event determined by the acquiring unit
  • a second determining unit configured to determine a social relationship of the at least two users according to the common activity event determined by the first determining unit.
  • the first determining unit includes:
  • a first acquiring subunit configured to acquire an activity type relevance score of an activity event of two users according to the activity event determined by the acquiring unit;
  • a first determining subunit configured to determine, when the activity type relevance score of the activity events of the two users acquired by the first acquiring subunit is greater than an association threshold, determining that the activity events of the two users are a common activity event
  • the first determining subunit is further configured to determine activity of the two users when an activity type relevance score of an activity event of the two users acquired by the first acquiring subunit is not greater than an association threshold Events are non-common events.
  • the second determining unit includes:
  • a second acquiring subunit configured to acquire an important user pair, where a total number of common activity events of the two users of the important user pair is greater than a quantity threshold, and a total of the common activity events of the two users of the important user pair The duration is greater than the time threshold;
  • Extracting a sub-unit configured to extract a semantic feature of a common activity event of the important user pair acquired by the second acquiring sub-unit, where the semantic feature includes a time period, a location type, and an activity type;
  • a second determining subunit configured to determine the social relationship according to the semantic feature and the classification criterion extracted by the at least one of the extracting subunits, where the classification criterion is determined according to the activity parameter.
  • the first determining unit further includes:
  • a determining subunit configured to determine whether the activity events of the two users meet the preset condition
  • the first obtaining sub-unit is further configured to acquire an activity type relevance score of the activity events of the two users according to the activity data if the activity events of the two users meet the preset condition;
  • the first obtaining sub-unit further stops acquiring an activity type relevance score of the activity events of the two users according to the activity data if the activity events of the two users do not satisfy the preset condition.
  • the determining subunit is specifically configured to:
  • the geographic distance is less than the distance threshold
  • the intersection time is greater than the time threshold
  • the sensor data similarity is greater than the similar threshold
  • the relationship characteristics for reasoning are single, the categories are few, and the relationship in the virtual network is not true, and the inferred social relationship has one-sidedness and unreality, and the true social relationship between users cannot be correctly obtained.
  • the method and device for inferring social relations provided by the embodiment of the present invention, compared with the prior art, the terminal device in the present invention acquires activity data of at least two users, according to the activity The dynamic data determines the user's activity events, and then the common activity events are determined by the multi-faceted features of each user activity event, thereby inferring the user's social relationship.
  • the terminal device infers the user's social relationship according to the characteristics of the location, time, environment and type of the event, and combines the time and space information, which can effectively filter out the interference, improve the accuracy of the user's common event mining, and improve the social relationship classification. Granularity and precision give users a lot of convenience.
  • FIG. 1 is a flowchart of a method according to an embodiment of the present invention
  • FIG. 3 is a social relationship diagram provided by another embodiment of the present invention.
  • FIG. 4 and FIG. 5 are schematic structural diagrams of a device according to another embodiment of the present invention.
  • FIG. 6 is a schematic structural diagram of a device according to another embodiment of the present invention.
  • An embodiment of the present invention provides a method for inferring a social relationship, which is used in a terminal device. As shown in FIG. 1, the method includes:
  • the terminal device acquires activity data of at least two users, and determines at least two according to the activity data. Activity events for users.
  • the activity data includes location location, time and sensor data.
  • the terminal device determines a common activity event according to the activity event.
  • the terminal device determines whether the activity events of any two users meet the preset condition that the geographical distance is less than the distance threshold, the intersection time length is greater than the time threshold, and the sensor data similarity is greater than the similar threshold. If the activity events of the two users meet the preset conditions at the same time, the activity type relevance scores of the activity events of the two users are obtained according to the activity data; if the preset events are met when the activity events of the two users are different, the data acquisition according to the activity data is stopped. Activity type affinity scores for activity events for two users.
  • the terminal device obtains the activity type relevance scores of the activity events of the two users according to the activity event, it is determined whether the activity type relevance score is greater than the association threshold.
  • the activity type relevance score of the activity events of the two users is greater than the association threshold, it is determined that the activity events of the two users are common activity events; when the activity type relevance scores of the activity events of the two users are not greater than the association At the threshold, it is determined that the activity events of the two users are non-common activity events.
  • the terminal device determines a social relationship of at least two users according to a common activity event.
  • the total number of common activity events for the two users is greater than the number threshold, and the cumulative duration of the common activity events of the two users in the important user pair is greater than the time threshold.
  • the terminal device After acquiring the important user pair, extracts the semantic features of the common activity event of the important user pair, and then determines the social relationship according to at least one semantic feature and the classification criterion. Semantic features include time period, location type, activity type, and classification criteria are determined based on activity parameters.
  • the terminal device acquires activity data of at least two users, determines an activity event of the user according to the activity data, and then determines a common activity event by using various aspects of each user activity event, and further Infer the user's social relationship.
  • the terminal device infers the user's social relationship according to the characteristics of the location, time, environment and type of the event event.
  • a further embodiment of the present invention provides a method for inferring a social relationship, which acquires user data through a user terminal device and performs inference analysis on the acquired data to obtain a social relationship of the user, wherein the terminal device takes a mobile phone as an example, as shown in the figure.
  • the method includes:
  • the mobile phone collects activity data of the specified user.
  • the mobile phone needs to periodically measure the trajectory information, time, and surrounding environment of the specified user for a period of time when estimating the social relationship of the designated user.
  • the mobile phone extracts an activity event of each user in the specified user from the activity data.
  • the location type of the event event includes a semantic location "home or office” that is meaningful to the user, and also includes a place with public attributes such as "shop, restaurant, park, hotel", which can be accessed by a third-party map service or user.
  • the points of interest are obtained; the location level of the event can be classified according to the importance of the society, the grade of the store, the per capita consumption of the restaurant, the star rating of the hotel, etc. It can also be a general subjective evaluation through the third-party life service website query; the activity type of the activity event can be comprehensively reasoned based on the location type, the event recorded by the calendar in the mobile phone, the sensor data, the environmental background sound, and the image.
  • the mobile phone infers its social relationship through the connection between the activity events of the two users, and the two users in the following steps are any two users of the designated users.
  • MA Matual Activity
  • the mobile phone determines whether the geographical distance of the two user activity events is less than a distance threshold. If the geographic distance of the two user activity events is less than the distance threshold, step 204 is performed; otherwise, the determination is ended.
  • the mobile phone can calculate the distance between two active events according to the latitude and longitude of the geographic location of the two user activity events, for example It is then compared to the set distance threshold dist_TH. If the distance between two active events is less than the threshold, ie Then, the distance between the two users is very close, and step 204 is performed to continue to compare other information of the two user activity events; if the distance between the two activity events is not less than the threshold, This means that the two users are not active in the same area, and the process of ending this judgment is ended.
  • the mobile phone determines whether the intersection time length of the two user activity events is greater than a time threshold. If the intersection time length of the two user activity events is greater than the time threshold, step 205 is performed; otherwise, the determination is ended.
  • the mobile phone can know the time range of the activity event according to the time information of the two user activity events, that is, with And then the length of the intersection of the event events It is then compared to the set time threshold time_TH. If the intersection time of two active events is greater than the time threshold, ie The two users are in the same area for a period of time, and there may be a social relationship. Step 205 is performed to continue comparing other information of the two user activity events; if the intersection time of the two activity events is not greater than the time threshold, This means that the two users are not active in the same area at the same time, ending the process of this judgment.
  • the mobile phone determines whether the sensor data similarity of the two user activity events is greater than a similar threshold. If the sensor data similarity of the two user activity events is greater than the similar threshold, step 206 is performed; otherwise, the determination is ended.
  • the mobile phone calculates the similarity of the two active event environments, that is, the similarity, according to sensor data of two user activity events, such as pressure atm, temperature tem, humidity hun, volume decibel voi, etc. by using Euclidean distance and the like. It is then compared to the set similar threshold Sim_TH. If the sensor data similarity of two user activity events is greater than the similarity threshold, ie It means that the two users are in the same area and environment for a period of time, and there is a high probability of social relations, and step 206 is performed; if the sensor data similarity of the two user activity events is not greater than the similar threshold, It means that two users perform different activities in the same area at the same time, and the process of ending the judgment is ended.
  • sensor data of two user activity events such as pressure atm, temperature tem, humidity hun, volume decibel voi, etc. by using Euclidean distance and the like. It is then compared to the set similar threshold Sim_TH. If the sensor data similarity of two user activity events
  • the mobile phone obtains an activity type relevance score of two user activity events.
  • the activity type relevance score indicates the possibility that the two activity types are the same, and can be set according to experience and actual scenarios.
  • the same type of activity is set to 1 point
  • similar activities are set to 0.8 points, such as running and yoga
  • low similarity activities are set to 0.3 points, such as eating and shopping, or watching movies and cleaning.
  • the mobile phone determines whether the activity type relevance score of the two user activity events is greater than the association threshold. If the activity type relevance score of the two user activity events is greater than the association threshold, step 208 is performed; otherwise, two are determined. The activity events of the users are non-common event events, and the judgment is ended.
  • the mobile phone compares with the set similar threshold after acquiring the activity type relevance score of the two user activity events. If the sensor data similarity of the two user activity events is greater than the similarity threshold, it means that the two users are in the same area and environment for a period of time, performing the same or similar types of activities, performing step 208; if two user activity events Sensor data is similar If the sex is not greater than the similarity threshold, it means that two users perform different types of activities in the same area at the same time, and it is determined that the activity events of the two users are non-common event events.
  • the mobile phone determines that the activity events of the two users are common event events.
  • the two user activity events satisfy the geographical distance less than the distance threshold and the intersection time length is greater than the time threshold. Since both users are running, the surrounding environment is the same, and the sensor data similarity is greater than the similar threshold, the activity type correlation scores of the two user activity events are considered to be 1 (ie, the activities are the same). .
  • the association threshold is set to 0.5, and the activity type relevance score is greater than the association threshold, and the activity events of the two users are determined to be the common activity event.
  • the mobile phone counts the number of events and cumulative time of the two users' common activities.
  • the mobile phone collects the activity data of the designated user within one week, and it is concluded that the two users A and B run for one hour each afternoon in the same gym, and the activity event is a common activity event of the two users A and B.
  • the number of activities is 7, and the accumulated time is 7 hours.
  • the mobile phone finds important user pairs according to the number of common event events and the accumulated time.
  • the mobile phone sets a quantity threshold and a time threshold of two user common activity events, and compares the number of the common activity events and the accumulated time counted in step 209 with the quantity threshold and the time threshold, and the total activity event is total. Two users whose number is greater than the number threshold and whose accumulated time is greater than the time threshold are important user pairs.
  • the mobile phone extracts characteristics of a common activity event of an important user pair.
  • common event events such as the time period of the common event, the type of the event, the type of the event, etc., thereby obtaining the common activity within a certain time period or a certain type of event.
  • the two activities of A and B users during the week are for dinner at home every night, two After shopping in the mall one afternoon and walking in the park one night, the activities of the two users A and B at home in the evening accounted for 70% of the total activities, or the joint activities of the two users A and B in one week.
  • the proportion of A and B users working in the office is 100% of the total activities, or the proportion of A and B users in the week.
  • the location type of the highest activity event is hospital and the activity type is medical treatment; or the location of the activity event with the highest proportion of the two activities of A and B users in the week is advanced.
  • the mobile phone infers the social relationship of the designated user according to the semantic feature and the classification criterion.
  • the mobile phone infers the social relationship of the important user pairs according to the semantic features and classification criteria obtained in step 211, and the classification criteria are determined according to specific activity parameters and experience.
  • the classification criteria are: the time period is night and the location type is home.
  • the proportion of common activity events is greater than or equal to 70%, and the social relationship is inferred to be family members; the time period is working time and the location type is common to the office.
  • the proportion of active events is greater than or equal to 80%, and the social relationship is inferred to be a colleague; the highest proportion of activities in the same event is the type of hospital and the type of activity is medical treatment, and the social relationship is inferred to be doctors and patients; the highest proportion of activities in the same event
  • the location type of the event is the office location and the activity type is meeting, visiting or explaining, and the social relationship is inferred as the partner; the highest proportion of the event event in the same event event is the movie theater, park or shopping mall and the activity type is watching movies, playing. Or shopping, inferring social relations as a couple or close friends; the highest proportion of activities in the same event event is the type of restaurant or sports field and the type of activity is eating or playing, inferring social relations as friends.
  • the social relationship between users is inferred according to the classification criteria.
  • the social relationship can be inferred according to the classification criteria.
  • the social relationship of the user pair can be inferred according to the priority as shown in FIG. 3, so that a social relationship with high priority among users is obtained, for example, The two users A and B are an important user pair, both family and partner, and the inference results only have high priority family members, and the lower priority partners no longer infer; alternatively, the social relationship is inferred.
  • the mobile phone acquires activity data of at least two users, determines an activity event of the user according to the activity data, and then uses a geographic location, an intersection time length, and sensor data similar to each user activity event.
  • Sexuality and activity type association characteristics identify common activity events, thereby identifying important user pairs and inferring the user's social relationships.
  • the terminal device infers the user's social relationship according to the characteristics of the location, time, environment and type of the event, combines the time and space information, improves the accuracy of the user's common event event mining, and can solve the social relationship between the inferred users.
  • the reasoning of the social relationship is not true.
  • a further embodiment of the present invention provides a device 30 for inferring a social relationship. As shown in FIG. 4, the device 30 includes:
  • the acquiring unit 31 is configured to acquire activity data of at least two users, and determine activity events of the at least two users according to the activity data, where the activity data includes location location, time, and sensor data;
  • a first determining unit 32 configured to determine a common activity event according to the activity event determined by the acquiring unit 31;
  • the second determining unit 33 is configured to determine the social relationship of the at least two users according to the common activity event determined by the first determining unit 32.
  • the first determining unit 32 includes:
  • the first obtaining sub-unit 321 is configured to acquire an activity type relevance score of the activity events of the two users according to the activity event determined by the acquiring unit 31;
  • a first determining sub-unit 322, configured to determine, when the activity type relevance score of the activity events of the two users acquired by the first acquiring sub-unit 321 is greater than an association threshold, determine an activity event of the two users For the joint activity event;
  • the first determining sub-unit 322 is further configured to: when the activity type relevance score of the activity events of the two users acquired by the first acquiring sub-unit 321 is not greater than an association threshold, determine the two users The event is a non-common event.
  • the second determining unit 33 includes:
  • a second obtaining sub-unit 331, configured to acquire an important user pair, where a total number of common activity events of the two users is greater than a quantity threshold, and a common activity event of the two users of the important user pair The accumulated duration is greater than the time threshold;
  • the extraction sub-unit 332 is configured to extract a semantic feature of a common activity event of the important user pair acquired by the second acquisition sub-unit 331, where the semantic feature includes a time period, a location type, and an activity type;
  • the second determining sub-unit 333 is configured to determine the social relationship according to the semantic feature and the classification criterion extracted by the at least one extraction sub-unit 332, where the classification criterion is determined according to the activity parameter.
  • the first determining unit 32 further includes:
  • the determining sub-unit 323 is configured to determine whether the activity events of the two users meet the preset condition
  • the first obtaining sub-unit 321 is further configured to: if the activity events of the two users meet the preset condition, acquire an activity type relevance score of the activity events of the two users according to the activity data;
  • the first obtaining sub-unit 321 further stops acquiring an activity type relevance score of the activity events of the two users according to the activity data if the activity events of the two users do not satisfy the preset condition.
  • the determining subunit 323 is specifically configured to:
  • the geographic distance is less than the distance threshold
  • the intersection time is greater than the time threshold
  • the sensor data similarity is greater than the similar threshold
  • the relationship characteristics for reasoning are single, the categories are few, and the relationship in the virtual network is not true, and the inferred social relationship has one-sidedness and unreality, and the true social relationship between users cannot be correctly obtained.
  • the device 30 in the embodiment of the present invention acquires activity data of at least two users, determines an activity event of the user according to the activity data, and then determines a common activity event by multi-faceted features of each user activity event, and further Infer the user's social relationship.
  • the terminal device infers the user's social relationship according to the characteristics of the location, time, environment and type of the event, combines the time and space information, improves the accuracy of the user's common event event mining, and can solve the social relationship between the inferred users.
  • the reasoning of the social relationship is not true.
  • a further embodiment of the present invention provides a device 40 for inferring a social relationship. As shown in FIG. 6, the device 40 includes:
  • the processor 41 is configured to acquire activity data of at least two users, and determine activity events of the at least two users according to the activity data, where the activity data includes location location, time, and sensor data; and, The activity event determines a common activity event; and is configured to determine a social relationship of the at least two users based on the common activity event.
  • the processor 41 is further configured to acquire an activity type relevance score of the activity events of the two users according to the activity event; and an activity type relevance score for the activity events of the two users When the value is greater than the association threshold, determining that the activity events of the two users are the common activity event; and, when the activity type relevance score of the activity events of the two users is not greater than the association threshold, Determining that the activity events of the two users are non-common activity events.
  • the processor 41 is further configured to acquire an important user pair, the total number of common activity events of the two users of the important user pair is greater than a quantity threshold, and the common activities of the two users of the important user pair The cumulative duration of the event is greater than a time threshold; and a semantic feature for extracting a common activity event of the important user pair, the semantic feature including a time period, a location type, an activity type; and, for, according to at least one The semantic features and classification criteria determine the social relationship, the classification criteria being determined based on the activity parameters.
  • the processor 41 is further configured to determine whether the activity events of the two users meet the preset condition; and, if the activity events of the two users meet the preset condition, acquire two according to the activity data.
  • the activity type relevance score of the activity event of the user if the activity events of the two users do not satisfy the preset condition, stop acquiring the activity type relevance score of the activity events of the two users according to the activity data.
  • the processor 41 is further configured to: when the activity events of the two users satisfy the geographic distance less than the distance threshold, the intersection time length is greater than the time threshold, and the sensor data similarity is greater than the similar threshold, determining the two The activity event of the user satisfies a preset condition; and, when the activity events of the two users are different, when the geographical distance is less than the distance threshold, the intersection time length is greater than the time threshold, and the sensor data similarity is greater than the similar threshold, the determination is performed.
  • the activity events of the two users do not satisfy the preset conditions.
  • the relationship characteristics for reasoning are single, the categories are few, and the relationship in the virtual network is not true, and the inferred social relationship has one-sidedness and unreality, and the true social relationship between users cannot be correctly obtained.
  • the device 40 in the embodiment of the present invention acquires activity data of at least two users, determines an activity event of the user according to the activity data, and then determines a common activity event by multi-faceted features of each user activity event, and further Infer the user's social relationship.
  • the terminal device infers the user's social relationship according to the characteristics of the location, time, environment and type of the event, combines the time and space information, improves the accuracy of the user's common event event mining, and can solve the social relationship between the inferred users.
  • the reasoning of the social relationship is not true.
  • An apparatus for inferring a social relationship provided by the embodiment of the present invention may implement the foregoing method embodiment.
  • a method and apparatus for inferring a social relationship provided by an embodiment of the present invention may be applied to a terminal device, but is not limited thereto.
  • the storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Computational Linguistics (AREA)
  • Computer Hardware Design (AREA)
  • Fuzzy Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

本发明实施例公开了一种推理社会关系的方法及装置，涉及通信技术领域，能够在推理用户之间的社会关系时，推理出真实的社会关系。本发明的方法包括：获取至少两个用户的活动数据，根据所述活动数据确定所述至少两个用户的活动事件，所述活动数据包括地点位置、时间范围和传感器数据；根据所述活动事件确定共同活动事件；根据所述共同活动事件确定所述至少两个用户的社会关系。本发明适用于终端设备。

Description

推理社会关系的方法及装置 技术领域
本发明涉及通信技术领域，尤其涉及一种推理社会关系的方法及装置。
背景技术
随着近年来互联网技术与应用的快速发展，社会化网络服务已经逐渐成熟。社会化网络服务可以根据各用户之间的社会关系，为用户提供个性化的服务，给用户带来很多便利。各用户之间的社会关系可以通过对用户的日常活动的推理得到，现有技术中采用了多种方法，例如根据用户之间在特定时间段内见面的频次，或者用户个人照片的识别，或者虚拟网络的通信消息、用户兴趣的分布等来推测用户间的社会关系。
现有技术中至少存在如下问题：上述方案中，通过用户兴趣、虚拟网络中的通信消息、个人照片等特征对用户的社会关系进行推理，都是根据用户日常活动的某一项特征进行的推理。由于用于推理的关系特征单一、类别少，且虚拟网络中的关系不真实，照片识别过程容易产生错误等，所以推理出的社会关系具有片面性和不真实性，不能正确得出用户之间真实的社会关系，从而不能为用户提供正确的社会化网络服务，给用户带来很多不便。
发明内容
本发明的实施例提供一种推理社会关系的方法及装置，能够解决在推理用户之间的社会关系时，推理出的社会关系不真实的问题。
为达到上述目的，本发明的实施例采用如下技术方案：
第一方面，本发明的实施例提供一种推理社会关系的方法，包括：
获取至少两个用户的活动数据，根据所述活动数据确定所述至少两个用户的活动事件，所述活动数据包括地点位置、时间和传感器数据；
根据所述活动事件确定共同活动事件；
根据所述共同活动事件确定所述至少两个用户的社会关系。
结合第一方面，在第一种可能的实现方式中，所述根据所述活动事件确定共同活动事件包括：
根据所述活动事件获取两个用户的活动事件的活动类型关联性分值；
当所述两个用户的活动事件的活动类型关联性分值大于关联性阈值时，确定所述两个用户的活动事件为所述共同活动事件；
当所述两个用户的活动事件的活动类型关联性分值不大于关联性阈值时，确定所述两个用户的活动事件为非共同活动事件。
结合第一方面或第一种可能的实现方式，在第二种可能的实现方式中，所述根据所述共同活动事件确定所述至少两个用户的社会关系包括：
获取重要用户对，所述重要用户对中两个用户的共同活动事件总数量大于数量门限值，且所述重要用户对中两个用户的共同活动事件的累计时长大于时间门限值；
抽取所述重要用户对的共同活动事件的语义特征，所述语义特征包括时间段、地点类型、活动类型；
根据至少一个所述语义特征和分类标准确定所述社会关系，所述分类标准根据所述活动参数确定。
结合第一种可能的实现方式，在第三种可能的实现方式中，所述根据所述活动数据计算两个用户的活动事件的活动类型关联性分值之前，所述方法还包括：
判断两个用户的活动事件是否满足预设条件；
如果所述两个用户的活动事件满足预设条件，根据所述活动数据获取两个用户的活动事件的活动类型关联性分值；
如果所述两个用户的活动事件不满足预设条件，停止根据所述活动数据获取两个用户的活动事件的活动类型关联性分值。
结合第三种可能的实现方式，在第四种可能的实现方式中，所述判断两 个用户的活动事件是否满足预设条件包括：
当所述两个用户的活动事件同时满足地理位置距离小于距离阈值、交集时间长度大于时间阈值和传感器数据相似性大于相似阈值时，判定所述两个用户的活动事件满足预设条件；
当所述两个用户的活动事件不同时满足地理位置距离小于距离阈值、交集时间长度大于时间阈值和传感器数据相似性大于相似阈值时，判定所述两个用户的活动事件不满足预设条件。
第二方面，本发明的实施例提供一种推理社会关系的装置，包括：
获取单元，用于获取至少两个用户的活动数据，根据所述活动数据确定所述至少两个用户的活动事件，所述活动数据包括地点位置、时间和传感器数据；
第一确定单元，用于根据所述获取单元确定的所述活动事件确定共同活动事件；
第二确定单元，用于根据所述第一确定单元确定的所述共同活动事件确定所述至少两个用户的社会关系。
结合第二方面，在第一种可能的实现方式中，所述第一确定单元包括：
第一获取子单元，用于根据所述获取单元确定的所述活动事件获取两个用户的活动事件的活动类型关联性分值；
第一确定子单元，用于当所述第一获取子单元获取的所述两个用户的活动事件的活动类型关联性分值大于关联性阈值时，确定所述两个用户的活动事件为所述共同活动事件；
所述第一确定子单元还用于当所述第一获取子单元获取的所述两个用户的活动事件的活动类型关联性分值不大于关联性阈值时，确定所述两个用户的活动事件为非共同活动事件。
结合第二方面或第一种可能的实现方式，在第二种可能的实现方式中，所述第二确定单元包括：
第二获取子单元，用于获取重要用户对，所述重要用户对中两个用户的共同活动事件总数量大于数量门限值，且所述重要用户对中两个用户的共同活动事件的累计时长大于时间门限值；
抽取子单元，用于抽取所述第二获取子单元获取的所述重要用户对的共同活动事件的语义特征，所述语义特征包括时间段、地点类型、活动类型；
第二确定子单元，用于根据至少一个所述抽取子单元抽取的所述语义特征和分类标准确定所述社会关系，所述分类标准根据所述活动参数确定。
结合第一种可能的实现方式，在第三种可能的实现方式中，所述第一确定单元还包括：
判断子单元，用于判断两个用户的活动事件是否满足预设条件；
所述第一获取子单元还用于如果所述两个用户的活动事件满足预设条件，根据所述活动数据获取两个用户的活动事件的活动类型关联性分值；
所述第一获取子单元还用如果所述两个用户的活动事件不满足预设条件，停止根据所述活动数据获取两个用户的活动事件的活动类型关联性分值。
结合第三种可能的实现方式，在第四种可能的实现方式中，所述判断子单元具体用于：
当所述两个用户的活动事件同时满足地理位置距离小于距离阈值、交集时间长度大于时间阈值和传感器数据相似性大于相似阈值时，判定所述两个用户的活动事件满足预设条件；
当所述两个用户的活动事件不同时满足地理位置距离小于距离阈值、交集时间长度大于时间阈值和传感器数据相似性大于相似阈值时，判定所述两个用户的活动事件不满足预设条件。
现有技术中，用于推理的关系特征单一、类别少，且虚拟网络中的关系不真实，推理出的社会关系具有片面性和不真实性，不能正确得出用户之间真实的社会关系。本发明实施例提供的一种推理社会关系的方法及装置，与现有技术相比，本发明中终端设备在获取至少两个用户的活动数据，根据活 动数据确定用户的活动事件，然后通过对各用户活动事件的多方面特征确定共同活动事件，进而推断用户的社会关系。终端设备根据活动事件的地点、时间、环境和类型等特征来推断用户的社会关系，将时间和空间信息结合，可有效滤除干扰，提高用户共同活动事件挖掘的准确性，提升社会关系分类的粒度和精度，给用户带来很多便利。
附图说明
为了更清楚地说明本发明实施例中的技术方案，下面将对实施例中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其它的附图。
图1为本发明一实施例提供的方法流程图；
图2为本发明又一实施例提供的方法流程图；
图3为本发明又一实施例提供的社会关系图；
图4、图5为本发明又一实施例提供的装置结构示意图；
图6为本发明又一实施例提供的装置结构示意图。
具体实施方式
下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其它实施例，都属于本发明保护的范围。
本发明一实施例提供一种推理社会关系的方法，用于终端设备，如图1所示，所述方法包括：
101、终端设备获取至少两个用户的活动数据，根据活动数据确定至少两 个用户的活动事件。
其中，活动数据包括地点位置、时间和传感器数据。
102、终端设备根据活动事件确定共同活动事件。
其中，终端设备判断任意两个用户的活动事件是否同时满足地理位置距离小于距离阈值、交集时间长度大于时间阈值和传感器数据相似性大于相似阈值的预设条件。如果两个用户的活动事件同时满足预设条件，根据活动数据获取两个用户的活动事件的活动类型关联性分值；如果两个用户的活动事件不同时满足预设条件，停止根据活动数据获取两个用户的活动事件的活动类型关联性分值。
进一步的，在终端设备根据活动事件获取两个用户的活动事件的活动类型关联性分值后，判断活动类型关联性分值是否大于关联性阈值。当两个用户的活动事件的活动类型关联性分值大于关联性阈值时，确定两个用户的活动事件为共同活动事件；当两个用户的活动事件的活动类型关联性分值不大于关联性阈值时，确定两个用户的活动事件为非共同活动事件。
103、终端设备根据共同活动事件确定至少两个用户的社会关系。
其中，重要用户对为两个用户的共同活动事件总数量大于数量门限值，且重要用户对中两个用户的共同活动事件的累计时长大于时间门限值。终端设备获取重要用户对后，抽取重要用户对的共同活动事件的语义特征，进而根据至少一个语义特征和分类标准确定社会关系。语义特征包括时间段、地点类型、活动类型，分类标准根据活动参数确定。
现有技术中，用于推理的关系特征单一、类别少，且虚拟网络中的关系不真实，推理出的社会关系具有片面性和不真实性，不能正确得出用户之间真实的社会关系。与现有技术相比，本发明实施例中终端设备在获取至少两个用户的活动数据，根据活动数据确定用户的活动事件，然后通过对各用户活动事件的多方面特征确定共同活动事件，进而推断用户的社会关系。终端设备根据活动事件的地点、时间、环境和类型等特征来推断用户的社会关系， 将时间和空间信息结合，提高用户共同活动事件挖掘的准确性，能够解决在推理用户之间的社会关系时，推理出的社会关系不真实的问题。通过解决上述技术问题，能够有效滤除干扰，提升推断社会关系的真实性，给用户带来很多便利。
本发明又一实施例提供一种推理社会关系的方法，通过用户的终端设备获取用户数据并对获取的数据进行推理分析，得出用户的社会关系，其中，终端设备以手机为例，如图2所示，所述方法包括：
201、手机采集指定用户的活动数据。
其中，手机在推测指定用户的社会关系时，需要周期性的测量指定用户一段时间内的活动的轨迹信息、时间和周围环境的数据。手机可以通过GPS(Global Positioning System，全球定位系统)、Wi-Fi(Wireless Fidelity，无线保真)、Cell ID(Cell Identification，小区标识)等定位方式得到的用户所在位置的经纬度P＝(lat,lng)，然后根据经纬度的变化得出用户活动的轨迹信息；周围环境的数据Ctx可以有手机中的传感器测得，包括气压、温度、湿度、音量等，即Ctx＝(atm,tem,hum,voi)。
202、手机从活动数据中抽取指定用户中每个用户的活动事件。
其中，手机从用户的轨迹信息中确定用户的活动区域，活动区域的中心点为活动事件的地理位置，进而对活动事件的相关信息进行分析，例如，用户活动区域定位点集合{Pj,Pj+1,…,Pj+L}，定位点对应的传感器数据集合{Ctxj,Ctxj+1,…,Ctxj+L}，活动事件对应的中心点SPk＝(latavg,lngavg)，活动事件对应的传感器数据为
Figure PCTCN2014088437-appb-000001
例如，活动事件的地点类型包括对用户个人有意义的语义地点“家或办公地”，也包括“商场、餐馆、公园、酒店”等具有公共属性的地点，可借助第三方地图服务或用户的兴趣点获得；活动事件的地点等级可以根据社会重要程度、商场档次、餐馆人均消费、酒店星级等进行分类，分类标准可以 通过第三方生活服务类网站查询获得，也可以是普遍的主观评价；活动事件的活动类型可以结合地点类型、手机中日历记录的事件、传感器数据、环境背景音、图像等方式综合推理获得。
需要说明的是，手机通过两个用户的活动事件之间的联系来推断其的社会关系，则下述步骤中的两个用户为指定用户中的任意两个用户。为了确定这两个用户之间的共同活动事件，定义为MA(Mutual Activity，共同活动)，其中，通过
Figure PCTCN2014088437-appb-000002
表示这两个用户间的共同活动事件。
203、手机判断两个用户活动事件的地理位置距离是否小于距离阈值，若两个用户活动事件的地理位置距离小于距离阈值，则执行步骤204；否则，则结束判断。
其中，手机根据两个用户活动事件的地理位置的经纬度，可以计算出两个活动事件的距离，例如
Figure PCTCN2014088437-appb-000003
然后与设定的距离阈值dist_TH进行比较。如果两个活动事件的距离小于阈值，即
Figure PCTCN2014088437-appb-000004
则表示两个用户的距离很近，执行步骤204，继续比较两个用户活动事件的其他信息；如果两个活动事件的距离不小于阈值，即
Figure PCTCN2014088437-appb-000005
则表示两个用户没有在同一区域内活动，结束此判断的过程。
204、手机判断两个用户活动事件的交集时间长度是否大于时间阈值，若两个用户活动事件的交集时间长度大于时间阈值，则执行步骤205；否则，则结束判断。
其中，手机根据两个用户活动事件的时间信息中可知活动事件的时间范围，即
Figure PCTCN2014088437-appb-000006
和
Figure PCTCN2014088437-appb-000007
进而得出活动事件的交集时间长度
Figure PCTCN2014088437-appb-000008
然后与设定的时间阈值time_TH进行比较。如果两个活动事件的交集时间长度大于时间阈值，即
Figure PCTCN2014088437-appb-000009
则表示两个用户在一段时间内，处于同一区域，可能存在社会关系，执行步骤205，继续比较两个用户活动事件的其他信息；如果两个活动事件的交集时间长度不大于时间阈值，即
Figure PCTCN2014088437-appb-000010
则表示两个用户没有同时在同 一区域内活动，结束此判断的过程。
205、手机判断两个用户活动事件的传感器数据相似性是否大于相似阈值，若两个用户活动事件的传感器数据相似性大于相似阈值，则执行步骤206；否则，则结束判断。
其中，手机根据两个用户活动事件的传感器数据，例如压力atm、温度tem、湿度hun、音量分贝voi等，利用欧氏距离等方法，计算两个活动事件环境的相似度，即相似度
Figure PCTCN2014088437-appb-000011
然后与设定的相似阈值Sim_TH进行比较。如果两个用户活动事件的传感器数据相似性大于相似阈值，即
Figure PCTCN2014088437-appb-000012
则表示两个用户在一段时间内，处于同一区域和环境，存在社会关系的概率较大，执行步骤206；如果两个用户活动事件的传感器数据相似性不大于相似阈值，即
Figure PCTCN2014088437-appb-000013
则表示两个用户同时在同一区域内进行不同的活动，结束此判断的过程。
206、手机获取两个用户活动事件的活动类型关联性分值。
其中，活动类型关联性分值表示两种活动类型相同的可能性，可以根据经验和实际场景设置。例如，相同类型的活动设为1分、相似的活动设为0.8分，如跑步和瑜伽，而相似度低的活动设为0.3分，如吃饭和逛街、或者看电影和打扫卫生等。
207、手机判断两个用户活动事件的活动类型关联性分值是否大于关联性阈值，若两个用户活动事件的活动类型关联性分值大于关联性阈值，则执行步骤208；否则，则确定两个用户的活动事件为非共同活动事件，结束判断。
其中，手机在获取两个用户活动事件的活动类型关联性分值后，与设定的相似阈值进行比较。如果两个用户活动事件的传感器数据相似性大于相似阈值，则表示两个用户在一段时间内，处于同一区域和环境下，进行相同或者相似类型的活动，执行步骤208；如果两个用户活动事件的传感器数据相似 性不大于相似阈值，则表示两个用户同时在同一区域内进行不同类型的活动，确定两个用户的活动事件为非共同活动事件。
208、手机确定两个用户的活动事件为共同活动事件。
例如，两个用户同时在同一家健身房跑步健身，且时间长度大于时间阈值，则两个用户活动事件满足地理位置距离小于距离阈值和交集时间长度大于时间阈值。由于两个用户都在进行跑步运动，周围环境相同，则传感器数据相似性大于相似阈值，则由上述三个条件认为这两个用户活动事件的活动类型关联性分值为1(即活动相同)。一般而言可以设置关联性阈值为0.5，则活动类型关联性分值大于关联性阈值，确定两个用户的活动事件为共同活动事件。
209、手机统计两个用户共同活动事件的数量和累计时间。
例如，手机采集了指定用户一周内的活动数据，从中得出，A和B两个用户每天下午，在同一健身房同时跑步健身一个小时，则此活动事件为A和B两个用户的共同活动事件，活动数量为7，累计时间为7小时。
210、手机根据共同活动事件的数量和累计时间找出重要用户对。
其中，手机设置两个用户共同活动事件的数量门限值和时间门限值，将步骤209统计的共同活动事件的数量和累计时间与数量门限值和时间门限值比较，共同活动事件总数量大于数量门限值，且累计时长大于时间门限值的两个用户为重要用户对。
211、手机抽取重要用户对的共同活动事件的特征。
其中，对于每个重要用户对，通过共同活动事件获取其多种特征，如共同活动事件的时间段、地点类型、活动类型等，进而得出某个时间段内或者某个地点类型的共同活动事件占总共同活动事件的比例，以及在共同活动事件中比例最高的活动事件的地点类型和活动类型等，以此来推断用户的社会关系。
例如，A和B两个用户在一周内的共同活动事件为每天晚上在家吃饭、两 次下午在商场购物和一次晚上在公园散步，则A和B两个用户晚上在家吃饭的活动事件占总共同活动事件的比例为70％；或者A和B两个用户在一周内的共同活动事件为周一到周五工作时间再办公室工作，A和B两个用户工作时间在办公室的活动事件占总共同活动事件的比例为100％；或者A和B两个用户在一周内共同活动事件中比例最高的活动事件的地点类型为医院和活动类型为看病；或者A和B两个用户在一周内共同活动事件中比例最高的活动事件的地点等级为高级等。
212、手机根据语义特征和分类标准推断出指定用户的社会关系。
其中，手机根据步骤211中得出的语义特征和分类标准来推断重要用户对的社会关系，分类标准根据具体的活动参数和经验来确定。
例如，如图3所示，分类标准：时间段为晚上且地点类型为家的共同活动事件比例大于或等于70％，推断社会关系为家人；时间段为工作时间且地点类型为办公地的共同活动事件比例大于或等于80％，推断社会关系为同事；同活动事件中比例最高的活动事件的地点类型为医院且活动类型为看病，推断社会关系为医患；同活动事件中比例最高的活动事件的地点类型为办公地且活动类型为开会、参观或讲解，推断社会关系为合作伙伴；同活动事件中比例最高的活动事件的地点类型为电影院、公园或商场且活动类型为看电影、游玩或购物，推断社会关系为情侣或密友；同活动事件中比例最高的活动事件的地点类型为餐馆或运动场且活动类型为吃饭或打球，推断社会关系为朋友等。根据分类标准推断出用户之间的社会关系。
需要说明的是，根据分类标准推断社会关系可以如图3中所示方式，对用户对的社会关系按照优先级进行推断，如此得出用户之间的一种优先级高的社会关系，例如，A和B两个用户为一个重要用户对，既是家人又是合作伙伴，而推断结果只有优先级高的家人，优先级低的合作伙伴不再进行推断；可选的，社会关系的推断方式还可以为根据分类标准对用户的每一种社会关系都进行推断，例如，A和B两个用户为一个重要用户对，既是家人又是合作伙伴， 将A和B用户对根据分类标准中的每一项都进行判断，推断出A和B两个用户的社会关系为家人和合作伙伴，进而可以得出用户对的多种社会关系。
现有技术中，用于推理的关系特征单一、类别少，且虚拟网络中的关系不真实，推理出的社会关系具有片面性和不真实性，不能正确得出用户之间真实的社会关系。与现有技术相比，本发明实施例中手机在获取至少两个用户的活动数据，根据活动数据确定用户的活动事件，然后通过对各用户活动事件的地理位置、交集时间长度、传感器数据相似性和活动类型关联性特征确定共同活动事件，进而找出重要用户对，推断用户的社会关系。终端设备根据活动事件的地点、时间、环境和类型等特征来推断用户的社会关系，将时间和空间信息结合，提高用户共同活动事件挖掘的准确性，能够解决在推理用户之间的社会关系时，推理出的社会关系不真实的问题。通过解决上述技术问题，能够有效滤除干扰，提升推断社会关系的真实性，给用户带来很多便利。
本发明又一实施例提供一种推理社会关系的装置30，如图4所示，所述装置30包括：
获取单元31，用于获取至少两个用户的活动数据，根据所述活动数据确定所述至少两个用户的活动事件，所述活动数据包括地点位置、时间和传感器数据；
第一确定单元32，用于根据所述获取单元31确定的所述活动事件确定共同活动事件；
第二确定单元33，用于根据所述第一确定单元32确定的所述共同活动事件确定所述至少两个用户的社会关系。
进一步的，如图5所示，所述第一确定单元32包括：
第一获取子单元321，用于根据所述获取单元31确定的所述活动事件获取两个用户的活动事件的活动类型关联性分值；
第一确定子单元322，用于当所述第一获取子单元321获取的所述两个用户的活动事件的活动类型关联性分值大于关联性阈值时，确定所述两个用户的活动事件为所述共同活动事件；
所述第一确定子单元322还用于当所述第一获取子单元321获取的所述两个用户的活动事件的活动类型关联性分值不大于关联性阈值时，确定所述两个用户的活动事件为非共同活动事件。
进一步的，如图5所示，所述第二确定单元33包括：
第二获取子单元331，用于获取重要用户对，所述重要用户对中两个用户的共同活动事件总数量大于数量门限值，且所述重要用户对中两个用户的共同活动事件的累计时长大于时间门限值；
抽取子单元332，用于抽取所述第二获取子单元331获取的所述重要用户对的共同活动事件的语义特征，所述语义特征包括时间段、地点类型、活动类型；
第二确定子单元333，用于根据至少一个所述抽取子单元332抽取的所述语义特征和分类标准确定所述社会关系，所述分类标准根据所述活动参数确定。
进一步的，如图5所示，所述第一确定单元32还包括：
判断子单元323，用于判断两个用户的活动事件是否满足预设条件；
所述第一获取子单元321还用于如果所述两个用户的活动事件满足预设条件，根据所述活动数据获取两个用户的活动事件的活动类型关联性分值；
所述第一获取子单元321还用如果所述两个用户的活动事件不满足预设条件，停止根据所述活动数据获取两个用户的活动事件的活动类型关联性分值。
进一步的，如图5所示，所述判断子单元323具体用于：
当所述两个用户的活动事件同时满足地理位置距离小于距离阈值、交集时间长度大于时间阈值和传感器数据相似性大于相似阈值时，判定所述两个 用户的活动事件满足预设条件；
当所述两个用户的活动事件不同时满足地理位置距离小于距离阈值、交集时间长度大于时间阈值和传感器数据相似性大于相似阈值时，判定所述两个用户的活动事件不满足预设条件。
现有技术中，用于推理的关系特征单一、类别少，且虚拟网络中的关系不真实，推理出的社会关系具有片面性和不真实性，不能正确得出用户之间真实的社会关系。与现有技术相比，本发明实施例中装置30在获取至少两个用户的活动数据，根据活动数据确定用户的活动事件，然后通过对各用户活动事件的多方面特征确定共同活动事件，进而推断用户的社会关系。终端设备根据活动事件的地点、时间、环境和类型等特征来推断用户的社会关系，将时间和空间信息结合，提高用户共同活动事件挖掘的准确性，能够解决在推理用户之间的社会关系时，推理出的社会关系不真实的问题。通过解决上述技术问题，能够有效滤除干扰，提升推断社会关系的真实性，给用户带来很多便利。
本发明又一实施例提供一种推理社会关系的装置40，如图6所示，所述装置40包括：
处理器41，用于获取至少两个用户的活动数据，根据所述活动数据确定所述至少两个用户的活动事件，所述活动数据包括地点位置、时间和传感器数据；以及，用于根据所述活动事件确定共同活动事件；以及，用于根据所述共同活动事件确定所述至少两个用户的社会关系。
进一步的，所述处理器41还用于根据所述活动事件获取两个用户的活动事件的活动类型关联性分值；以及，用于当所述两个用户的活动事件的活动类型关联性分值大于关联性阈值时，确定所述两个用户的活动事件为所述共同活动事件；以及，用于当所述两个用户的活动事件的活动类型关联性分值不大于关联性阈值时，确定所述两个用户的活动事件为非共同活动事件。
进一步的，所述处理器41还用于获取重要用户对，所述重要用户对中两个用户的共同活动事件总数量大于数量门限值，且所述重要用户对中两个用户的共同活动事件的累计时长大于时间门限值；以及，用于抽取所述重要用户对的共同活动事件的语义特征，所述语义特征包括时间段、地点类型、活动类型；以及，用于根据至少一个所述语义特征和分类标准确定所述社会关系，所述分类标准根据所述活动参数确定。
进一步的，所述处理器41还用于判断两个用户的活动事件是否满足预设条件；以及，用于如果所述两个用户的活动事件满足预设条件，根据所述活动数据获取两个用户的活动事件的活动类型关联性分值；如果所述两个用户的活动事件不满足预设条件，停止根据所述活动数据获取两个用户的活动事件的活动类型关联性分值。
进一步的，所述处理器41还用于当所述两个用户的活动事件同时满足地理位置距离小于距离阈值、交集时间长度大于时间阈值和传感器数据相似性大于相似阈值时，判定所述两个用户的活动事件满足预设条件；以及，用于当所述两个用户的活动事件不同时满足地理位置距离小于距离阈值、交集时间长度大于时间阈值和传感器数据相似性大于相似阈值时，判定所述两个用户的活动事件不满足预设条件。
现有技术中，用于推理的关系特征单一、类别少，且虚拟网络中的关系不真实，推理出的社会关系具有片面性和不真实性，不能正确得出用户之间真实的社会关系。与现有技术相比，本发明实施例中装置40在获取至少两个用户的活动数据，根据活动数据确定用户的活动事件，然后通过对各用户活动事件的多方面特征确定共同活动事件，进而推断用户的社会关系。终端设备根据活动事件的地点、时间、环境和类型等特征来推断用户的社会关系，将时间和空间信息结合，提高用户共同活动事件挖掘的准确性，能够解决在推理用户之间的社会关系时，推理出的社会关系不真实的问题。通过解决上述技术问题，能够有效滤除干扰，提升推断社会关系的真实性，给用户带来 很多便利。
本发明实施例提供的一种推理社会关系的装置可以实现上述提供的方法实施例，具体功能实现请参见方法实施例中的说明，在此不再赘述。本发明实施例提供的一种推理社会关系的方法及装置可以适用于终端设备，但不仅限于此。
本说明书中的各个实施例均采用递进的方式描述，各个实施例之间相同相似的部分互相参见即可，每个实施例重点说明的都是与其他实施例的不同之处。尤其，对于设备实施例而言，由于其基本相似于方法实施例，所以描述得比较简单，相关之处参见方法实施例的部分说明即可。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通过计算机程序来指令相关的硬件来完成，所述的程序可存储于一计算机可读取存储介质中，该程序在执行时，可包括如上述各方法的实施例的流程。其中，所述的存储介质可为磁碟、光盘、只读存储记忆体(Read-Only Memory，ROM)或随机存储记忆体(Random Access Memory，RAM)等。
以上所述，仅为本发明的具体实施方式，但本发明的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本发明揭露的技术范围内，可轻易想到的变化或替换，都应涵盖在本发明的保护范围之内。因此，本发明的保护范围应该以权利要求的保护范围为准。

Claims (10)

  1. 一种推理社会关系的方法，其特征在于，包括：
    获取至少两个用户的活动数据，根据所述活动数据确定所述至少两个用户的活动事件，所述活动数据包括地点位置、时间和传感器数据；
    根据所述活动事件确定共同活动事件；
    根据所述共同活动事件确定所述至少两个用户的社会关系。
  2. 根据权利要求1所述的方法，其特征在于，所述根据所述活动事件确定共同活动事件包括：
    根据所述活动事件获取两个用户的活动事件的活动类型关联性分值；
    当所述两个用户的活动事件的活动类型关联性分值大于关联性阈值时，确定所述两个用户的活动事件为所述共同活动事件；
    当所述两个用户的活动事件的活动类型关联性分值不大于关联性阈值时，确定所述两个用户的活动事件为非共同活动事件。
  3. 根据权利要求1或2所述的方法，其特征在于，所述根据所述共同活动事件确定所述至少两个用户的社会关系包括：
    获取重要用户对，所述重要用户对中两个用户的共同活动事件总数量大于数量门限值，且所述重要用户对中两个用户的共同活动事件的累计时长大于时间门限值；
    抽取所述重要用户对的共同活动事件的语义特征，所述语义特征包括时间段、地点类型、活动类型；
    根据至少一个所述语义特征和分类标准确定所述社会关系，所述分类标准根据所述活动参数确定。
  4. 根据权利要求2所述的方法，其特征在于，在所述根据所述活动数据计算两个用户的活动事件的活动类型关联性分值之前，所述方法还包括：
    判断两个用户的活动事件是否满足预设条件；
    如果所述两个用户的活动事件满足预设条件，根据所述活动数据获取两个用户的活动事件的活动类型关联性分值；
    如果所述两个用户的活动事件不满足预设条件，停止根据所述活动数据获取两个用户的活动事件的活动类型关联性分值。
  5. 根据权利要求4所述的方法，其特征在于，所述判断两个用户的活动事件是否满足预设条件包括：
    当所述两个用户的活动事件同时满足地理位置距离小于距离阈值、交集时间长度大于时间阈值和传感器数据相似性大于相似阈值时，判定所述两个用户的活动事件满足预设条件；
    当所述两个用户的活动事件不同时满足地理位置距离小于距离阈值、交集时间长度大于时间阈值和传感器数据相似性大于相似阈值时，判定所述两个用户的活动事件不满足预设条件。
  6. 一种推理社会关系的装置，其特征在于，包括：
    获取单元，用于获取至少两个用户的活动数据，根据所述活动数据确定所述至少两个用户的活动事件，所述活动数据包括地点位置、时间和传感器数据；
    第一确定单元，用于根据所述获取单元确定的所述活动事件确定共同活动事件；
    第二确定单元，用于根据所述第一确定单元确定的所述共同活动事件确定所述至少两个用户的社会关系。
  7. 根据权利要求6所述的装置，其特征在于，所述第一确定单元包括：
    第一获取子单元，用于根据所述获取单元确定的所述活动事件获取两个用户的活动事件的活动类型关联性分值；
    第一确定子单元，用于当所述第一获取子单元获取的所述两个用户的活动事件的活动类型关联性分值大于关联性阈值时，确定所述两个用户的活动事件为所述共同活动事件；
    所述第一确定子单元还用于当所述第一获取子单元获取的所述两个用户的活动事件的活动类型关联性分值不大于关联性阈值时，确定所述两个用户的活动事件为非共同活动事件。
  8. 根据权利要求6或7任一项所述的装置，其特征在于，所述第二确定单元包括：
    第二获取子单元，用于获取重要用户对，所述重要用户对中两个用户的共同活动事件总数量大于数量门限值，且所述重要用户对中两个用户的共同活动事件的累计时长大于时间门限值；
    抽取子单元，用于抽取所述第二获取子单元获取的所述重要用户对的共同活动事件的语义特征，所述语义特征包括时间段、地点类型、活动类型；
    第二确定子单元，用于根据至少一个所述抽取子单元抽取的所述语义特征和分类标准确定所述社会关系，所述分类标准根据所述活动参数确定。
  9. 根据权利要求7所述的装置，其特征在于，所述第一确定单元还包括：
    判断子单元，用于判断两个用户的活动事件是否满足预设条件；
    所述第一获取子单元还用于如果所述两个用户的活动事件满足预设条件，根据所述活动数据获取两个用户的活动事件的活动类型关联性分值；
    所述第一获取子单元还用如果所述两个用户的活动事件不满足预设条件，停止根据所述活动数据获取两个用户的活动事件的活动类型关联性分值。
  10. 根据权利要求9所述的装置，其特征在于，所述判断子单元具体用于：
    当所述两个用户的活动事件同时满足地理位置距离小于距离阈值、交集时间长度大于时间阈值和传感器数据相似性大于相似阈值时，判定所述两个用户的活动事件满足预设条件；
    当所述两个用户的活动事件不同时满足地理位置距离小于距离阈值、交集时间长度大于时间阈值和传感器数据相似性大于相似阈值时，判定所述两 个用户的活动事件不满足预设条件。
PCT/CN2014/088437 2013-11-13 2014-10-13 推理社会关系的方法及装置 WO2015070683A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201310566158.0 2013-11-13
CN201310566158.0A CN104636382B (zh) 2013-11-13 2013-11-13 推理社会关系的方法及装置

Publications (1)

Publication Number Publication Date
WO2015070683A1 true WO2015070683A1 (zh) 2015-05-21

Family

ID=53056744

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2014/088437 WO2015070683A1 (zh) 2013-11-13 2014-10-13 推理社会关系的方法及装置

Country Status (2)

Country Link
CN (1) CN104636382B (zh)
WO (1) WO2015070683A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10123169B2 (en) 2015-12-14 2018-11-06 International Business Machines Corporation Group inference based upon venue zone events
CN111104468A (zh) * 2019-09-25 2020-05-05 西安交通大学 一种基于语义轨迹推断用户活动的方法
CN111324677A (zh) * 2018-12-13 2020-06-23 中国移动通信集团山西有限公司 用户位置数据的获取方法、装置、设备及介质

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104765873B (zh) * 2015-04-24 2019-03-26 百度在线网络技术（北京）有限公司 用户相似度确定方法和装置
CN106445942B (zh) * 2015-08-05 2020-07-03 腾讯科技（北京）有限公司 一种用户跨屏识别方法和装置
CN105069145A (zh) * 2015-08-20 2015-11-18 中国科学院计算技术研究所 用于确定社交网络用户关系强度的方法及系统
CN106557942B (zh) * 2015-09-30 2020-07-10 百度在线网络技术（北京）有限公司 一种用户关系的识别方法和装置
CN105354749A (zh) * 2015-10-16 2016-02-24 重庆邮电大学 一种基于社会网络的移动终端用户分组方法
CN105824921A (zh) 2016-03-16 2016-08-03 广州彩瞳网络技术有限公司 用户社会关系识别装置和方法
CN106980644B (zh) * 2017-02-20 2019-08-13 浙江大学 一种异构城市数据的个体人际关系可视推理方法
CN108319647A (zh) * 2017-12-27 2018-07-24 福建工程学院 一种基于浮动车技术的社交关系发现方法及终端
CN110895760A (zh) * 2018-09-05 2020-03-20 北京京东金融科技控股有限公司 数据处理方法和装置
CN109543078A (zh) * 2018-10-18 2019-03-29 深圳云天励飞技术有限公司 社会关系确定方法、装置、设备及计算机可读存储介质
CN111027781A (zh) * 2019-12-24 2020-04-17 北京明略软件系统有限公司 一种社会关系的预测方法、装置、存储介质和电子设备
CN111680077B (zh) * 2020-06-17 2023-10-27 郑州市中之易科技有限公司 一种通过关联度评分和模型比对确定相互关系的方法
CN112085542B (zh) * 2020-10-23 2024-01-26 北京金堤科技有限公司 用户筛选方法和装置、计算机可读存储介质、电子设备
CN114582120B (zh) * 2022-02-11 2023-01-06 北京中交兴路信息科技有限公司 基于车辆轨迹的隐藏社区发现方法、装置、设备及介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101819572A (zh) * 2009-09-15 2010-09-01 电子科技大学 一种用户兴趣模型的建立方法
CN102663047A (zh) * 2012-03-29 2012-09-12 中国科学院计算技术研究所 移动阅读中的社交关系挖掘方法及装置
CN102831206A (zh) * 2012-08-06 2012-12-19 吴迪 基于浏览器的微博社交方法及装置
US20130054433A1 (en) * 2011-08-25 2013-02-28 T-Mobile Usa, Inc. Multi-Factor Identity Fingerprinting with User Behavior
CN103279533A (zh) * 2013-05-31 2013-09-04 北京华悦博智科技有限责任公司 一种社交关系推荐方法及系统

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102541886B (zh) * 2010-12-20 2015-04-01 郝敬涛 一种识别用户群和用户之间关系的系统和方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101819572A (zh) * 2009-09-15 2010-09-01 电子科技大学 一种用户兴趣模型的建立方法
US20130054433A1 (en) * 2011-08-25 2013-02-28 T-Mobile Usa, Inc. Multi-Factor Identity Fingerprinting with User Behavior
CN102663047A (zh) * 2012-03-29 2012-09-12 中国科学院计算技术研究所 移动阅读中的社交关系挖掘方法及装置
CN102831206A (zh) * 2012-08-06 2012-12-19 吴迪 基于浏览器的微博社交方法及装置
CN103279533A (zh) * 2013-05-31 2013-09-04 北京华悦博智科技有限责任公司 一种社交关系推荐方法及系统

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10123169B2 (en) 2015-12-14 2018-11-06 International Business Machines Corporation Group inference based upon venue zone events
US10306409B2 (en) 2015-12-14 2019-05-28 International Business Machines Corporation Group inference based upon venue zone events
CN111324677A (zh) * 2018-12-13 2020-06-23 中国移动通信集团山西有限公司 用户位置数据的获取方法、装置、设备及介质
CN111104468A (zh) * 2019-09-25 2020-05-05 西安交通大学 一种基于语义轨迹推断用户活动的方法
CN111104468B (zh) * 2019-09-25 2023-03-28 西安交通大学 一种基于语义轨迹推断用户活动的方法

Also Published As

Publication number Publication date
CN104636382B (zh) 2018-03-06
CN104636382A (zh) 2015-05-20

Similar Documents

Publication Publication Date Title
WO2015070683A1 (zh) 推理社会关系的方法及装置
US11128979B2 (en) Inferring user availability for a communication
US8473512B2 (en) Dynamic profile slice
CN106605418B (zh) 使用基于位置的服务的移动客户端的电力管理
JP6761417B2 (ja) スケジュール検出に基づく動的ウェアラブルデバイス挙動
US10032233B2 (en) Social context in augmented reality
JP6423461B2 (ja) 動的デバイス構成のための周期的周囲波形分析
US8983976B2 (en) Dynamically expiring crowd-sourced content
US20120066067A1 (en) Fragmented advertisements for co-located social groups
US10255285B2 (en) Predicting acoustic features for geographic locations
EP3418967A1 (en) Method relating to presence granularity with augmented reality
US20170017928A1 (en) Inferring physical meeting location
US11089453B2 (en) Creation and consumption of transient user profiles
US20170116285A1 (en) Semantic Location Layer For User-Related Activity
KR101693429B1 (ko) 대인 관계 유형 파악을 통한 코칭 정보 제공 시스템
WO2014004550A1 (en) Personal mode contextual presence
CN117591751B (zh) 基于图嵌入的上下文忠诚度融合的兴趣点推荐方法及系统
EP3091497A1 (en) Recommendation system and method for a mobile device based on raw data which is collected from sensors of the mobile device
EP3136306A1 (en) Methods of providing information related to activities
CN107786899A (zh) 节目推荐方法及装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 14862203

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 14862203

Country of ref document: EP

Kind code of ref document: A1