WO2020001616A1 - 本体构建方法及装置 - Google Patents

本体构建方法及装置 Download PDF

Info

Publication number
WO2020001616A1
WO2020001616A1 PCT/CN2019/093683 CN2019093683W WO2020001616A1 WO 2020001616 A1 WO2020001616 A1 WO 2020001616A1 CN 2019093683 W CN2019093683 W CN 2019093683W WO 2020001616 A1 WO2020001616 A1 WO 2020001616A1
Authority
WO
WIPO (PCT)
Prior art keywords
attribute information
relationship
ontology
instance attribute
target
Prior art date
Legal status (The legal status 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 status listed.)
Ceased
Application number
PCT/CN2019/093683
Other languages
English (en)
French (fr)
Inventor
姜伟浩
葛挺
邵勇
闫春
浦世亮
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hangzhou Hikvision Digital Technology Co Ltd
Original Assignee
Hangzhou Hikvision Digital Technology Co Ltd
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 Hangzhou Hikvision Digital Technology Co Ltd filed Critical Hangzhou Hikvision Digital Technology Co Ltd
Publication of WO2020001616A1 publication Critical patent/WO2020001616A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/21Design, administration or maintenance of databases
    • G06F16/215Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/23Updating
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/24Querying
    • G06F16/245Query processing
    • G06F16/2455Query execution

Definitions

  • the present disclosure relates to the field of computer technology, and in particular, to a method and device for constructing an ontology.
  • Ontology is a description of entities (such as people or things) in the real world, events involved in the entities, and relationships between entities. People can query the objects described by the ontology in the ontology.
  • each object described by the ontology has a variety of attributes, such as the attributes of a person: name, gender, and ID number.
  • the ontology can describe the object by recording the attribute information of the object.
  • a computer device when constructing an ontology, a computer device needs to instantiate a specified data set, thereby obtaining instance attribute information of each object (such as an entity, event, or relationship), and performing instance attribute information of each object. storage.
  • the present disclosure provides a method and device for constructing an ontology.
  • the technical solution is as follows:
  • a method for constructing an ontology includes:
  • the instance attribute information of the plurality of objects is stored to construct the ontology.
  • the multiple objects include entities, and the time-dimensional attribute information of the entities includes one or more of the following:
  • a surviving identification used to record whether the entity is dead
  • the persistence event identifier of the entity and the persistence event identifier is used to record a trigger event when the entity's time-dimensional attribute information changes.
  • storing the instance attribute information of the multiple objects includes:
  • the instance attribute information of the i-th entity in the plurality of objects it is detected whether the historical instance attribute information of the i-th entity is stored locally, 1 ⁇ i ⁇ k, where k is the number of entities in the plurality of objects. number;
  • the instance attribute information of the i-th entity is used to cover its historical instance attribute information.
  • the multiple objects include relationships between entities, and the time-dimensional attribute information of the relationships includes one or more of the following:
  • a version identifier of the relationship wherein each of the relationships has at least one version identifier, and each of the version identifiers uniquely corresponds to a time period maintained by the relationship;
  • the source event identifier of the relationship and the source event identifier is used to record an event related to each version identifier of the relationship.
  • storing the instance attribute information of the multiple objects includes:
  • the cleanable historical instance attribute information of the hth relationship stored locally is moved to a peripheral for recycling So that the number of remaining historical instance attribute information of the h-th relationship is less than or equal to the preset number threshold; wherein the historical instance attribute information of the h-th relationship includes: the h-th Cleanable historical instance attribute information of the relationship and remaining historical instance attribute information of the h-th relationship;
  • the relationship version in the remaining historical instance attribute information is a remaining relationship version
  • the relationship version in the cleanable historical instance attribute information is a cleanable relationship version
  • the time period corresponding to the remaining relationship version is after the time period corresponding to the cleanable relationship version.
  • the method further includes: obtaining an ontology template and an instantiation rule, wherein the ontology template is used to record attribute information of multiple object groups, each Each object group includes objects belonging to the same type, and the multiple object groups are related in a hierarchical form;
  • the instantiating the specified data set includes: performing the instantiating process on the data set according to the ontology template and the instantiation rule;
  • the instance attribute information of each of the plurality of objects is information extracted in the data set based on the attribute information recorded in the ontology template.
  • the method further includes:
  • the data to be processed is instantiated according to the current ontology template and the current instantiation rule to obtain multiple update objects Instance attribute information, wherein the data to be processed is data to be instantiated indicated by the current instantiation rule in the current data set;
  • the stored instance attribute information is overwritten by the instance attribute information of the multiple update objects to update the ontology.
  • the method further includes:
  • the target instance attribute information exists in the ontology
  • the target instance attribute information is acquired.
  • the acquiring target query conditions include:
  • the initial query condition includes the time condition, determining that the initial query condition is the target query condition;
  • a preset time condition is added to the initial query condition to obtain the target query condition.
  • the method further includes:
  • the time condition is used to record a specified time period, and the target data is data generated in the data set within the specified time period.
  • an ontology construction device in a second aspect, includes:
  • a first processing module configured to instantiate a specified data set, and obtain instance attribute information of each of a plurality of objects
  • a storage module configured to store instance attribute information of the plurality of objects according to a specified storage structure to construct the ontology
  • the instance attribute information of each object includes: time-dimensional attribute information of each object in a time dimension.
  • the multiple objects include entities, and the time-dimensional attribute information of the entities includes one or more of the following:
  • a surviving identification used to record whether the entity is dead
  • the persistence event identifier of the entity and the persistence event identifier is used to record a trigger event when the entity's time-dimensional attribute information changes.
  • the storage module is used for:
  • attribute information of the i-th entity in the plurality of objects it is detected whether historical instance attribute information of the i-th entity is stored locally, 1 ⁇ i ⁇ k, where k is the number of entities in the plurality of objects. number;
  • the instance attribute information of the i-th entity is used to cover its historical instance attribute information.
  • the multiple objects include relationships between entities, and the time-dimensional attribute information of the relationships includes one or more of the following:
  • a version identifier of the relationship wherein each of the relationships has at least one version identifier, and each of the version identifiers uniquely corresponds to a time period maintained by the relationship;
  • the source event identifier of the relationship and the source event identifier is used to record an event related to each version identifier of the relationship.
  • the storage module is used for:
  • the cleanable historical instance attribute information of the hth relationship stored locally is moved to a peripheral for recycling So that the number of remaining historical instance attribute information of the h-th relationship is less than or equal to the preset number threshold; wherein the historical instance attribute information of the h-th relationship includes: the h-th Cleanable historical instance attribute information of the relationship and remaining historical instance attribute information of the h-th relationship;
  • the relationship version in the remaining historical instance attribute information is a remaining relationship version
  • the relationship version in the cleanable historical instance attribute information is a cleanable relationship version
  • the time period corresponding to the remaining relationship version is after the time period corresponding to the cleanable relationship version.
  • the ontology construction device further includes:
  • a first obtaining module is used to obtain an ontology template and an instantiation rule, wherein the ontology template is used to record attribute information of multiple object groups, and each object group includes objects of the same type, and the multiple object groups are based on Hierarchical formal association
  • the first processing module is configured to perform the instantiation processing on the data set according to the ontology template and the instantiation rule;
  • the instance attribute information of each of the plurality of objects is information extracted in the data set based on the attribute information recorded in the ontology template.
  • the ontology construction device further includes:
  • a detection module configured to detect whether the ontology template, the instantiation rule, and the data set are updated respectively
  • a second processing module configured to instantiate the data to be processed according to the current ontology template and the current instantiation rule when one or more of the ontology template, the instantiation rule, and the data set are updated To obtain instance attribute information of multiple update objects, where the data to be processed is data to be instantiated indicated by the current instantiation rule in the current data set;
  • An overlay module is configured to overwrite the stored instance attribute information of the multiple update objects to update the ontology.
  • the ontology construction device further includes:
  • a second obtaining module configured to obtain a target query condition, where the target query condition includes: a time condition;
  • a first query module configured to query, according to the target query condition, whether there is target instance attribute information in the ontology that meets the target query condition;
  • a third obtaining module is configured to obtain the target instance attribute information when the target instance attribute information exists in the ontology.
  • the second obtaining module is configured to:
  • the initial query condition includes the time condition, determining that the initial query condition is the target query condition;
  • a preset time condition is added to the initial query condition to obtain the target query condition.
  • the ontology construction device further includes:
  • a second query module configured to query whether the target instance attribute information exists in the peripheral recycle bin when the ontology does not include the target instance attribute information
  • the third obtaining module is further configured to perform the step of obtaining the target instance attribute information when the target instance attribute information exists in the peripheral recycle bin;
  • a third processing module configured to instantiate the target data in the data set according to the current ontology template and the current instantiation rule when the target instance attribute information does not exist in the peripheral recycle bin, To obtain the target instance attribute information;
  • the time condition is used to record a specified time period, and the target data is data generated in the data set within the specified time period.
  • a computer device including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory complete communication with each other through the bus; the memory is used to store a computer program; the processor For executing a program stored in a memory to implement the method described in the first aspect.
  • a computer-readable storage medium stores a computer program, and the computer program implements the method according to the first aspect when executed by a processor.
  • a computer program product containing instructions is provided, and when the computer program product is run on a computer, the computer is caused to execute the method described in the first aspect.
  • FIG. 1 is a method flowchart of an ontology construction method according to an embodiment of the present disclosure
  • FIG. 3 is a schematic diagram of an ontology template provided by an embodiment of the present disclosure.
  • FIG. 4 is a schematic diagram of another ontology template provided by an embodiment of the present disclosure.
  • FIG. 5 is a flowchart of a method for obtaining an ontology template according to an embodiment of the present disclosure
  • FIG. 6 is a schematic structural diagram of an ontology construction device according to an embodiment of the present disclosure.
  • FIG. 7 is a schematic structural diagram of another ontology construction device according to an embodiment of the present disclosure.
  • Ontology is a description of entities in the real world, the events involved in the entities, and the relationships between entities.
  • the entities, events, and relationships are the objects that the ontology describes.
  • an entity can be a student, teacher, police, car, etc.
  • the entity-related events can be: entity-related events such as student attending a class, police arresting a thief, and the relationship between entities can be: teachers and students between students and teachers Relationship, etc.
  • Each object (such as an entity, event, or relationship) described by an ontology has multiple attributes.
  • the attributes of a student include: name, gender, grade, ID number, etc.
  • the attributes of a car include: nameplate and number of the car.
  • the ontology can describe the object by recording the instance attribute information of the object.
  • the embodiment of the present disclosure provides a method for constructing an ontology.
  • the method can be applied to a computer device.
  • the computer device may include a device or a device cluster composed of multiple devices, which is not limited in the embodiments of the present disclosure.
  • the computer device may include: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the bus; the computer device may communicate with other external devices through the communication interface; the memory It can be used to store computer programs; the processor is used to execute the programs stored on the memory to implement the ontology construction method.
  • FIG. 1 is a method flowchart of an ontology construction method according to an embodiment of the present disclosure. As shown in FIG. 1, the ontology construction method may include:
  • Step 101 Instantiate a specified data set to obtain instance attribute information of each object in a plurality of objects.
  • Step 102 Store instance attribute information of the plurality of objects according to a specified storage structure to construct the ontology.
  • the instance attribute information of each object includes: time-dimensional attribute information of each object in a time dimension.
  • the instance attribute information obtained in the embodiment of the present disclosure includes time-dimensional attribute information of the time dimension
  • the ontology constructed based on the instance attribute information can describe the relationship between the object and time. In this way, the object described by the ontology Being able to unify with the objects in the real world improves the accuracy of the ontology.
  • FIG. 2 is a method flowchart of another ontology construction method according to an embodiment of the present disclosure. As shown in FIG. 2, the ontology construction method may include:
  • Step 201 Obtain an ontology template and instantiation rules.
  • the ontology template may be used to record attributes of multiple object groups, and the attributes of each object group may include time-dimensional attributes in a time dimension.
  • Each object group includes objects belonging to the same type, such as an object group composed of students belonging to the type of students, an object group composed of objects belonging to the doctor type, etc., an object group composed of objects belonging to the kinship type, and objects belonging to public welfare events Group of objects.
  • the attributes of the next-level object group include the attributes of the previous-level object group, that is, the next-level object.
  • a group has the attributes of a group of objects one level up.
  • the object group on the upper layer is a person, and the object group on the lower layer may be a student, a police officer, or the like.
  • the lower-level object group can be referred to as a child object group of the upper-level object group
  • the upper-level object group can be referred to as the parent object group of the lower-level object group
  • the lower-level object group inherits the previous-level object group.
  • the top-level object group can be called the ancestor object group of other layer object groups.
  • the ontology template can record the attribute information of each object group in multiple implementable ways, and the embodiments of the present disclosure will explain two of the implementable ways.
  • the ontology template records: attribute information corresponding to each object group, and the attribute information corresponding to each object group is attribute information of each object group.
  • FIG. 3 is a schematic diagram of an ontology template provided by an embodiment of the present disclosure.
  • the plurality of object groups may include: a top-level object group, a person object group, a student object group, a police object group, and a car object group.
  • the person object group and the vehicle object group are both associated with the top object group and are both located one level below the top object group.
  • the student object group and the police object group are both associated with the person object group and both are located one level below the person object group.
  • the top-level object group is located at the top level of these object groups.
  • the student object group, police object group, and car object group are all at the bottom level of these object groups, that is, the object group not associated with the next-level object group is the bottom-level object group.
  • the attribute information corresponding to the top-level object group includes: an identification code, a birth time, and a name.
  • Each object in the real world is a type of the top-level object group, such as a person object group and a car object group, respectively Two types in an object group.
  • the attribute information corresponding to the human object group includes: an ID number (the identification code in the attribute information corresponding to the top-level object group is used to describe the same attribute), the date of birth (the birth time in the attribute information corresponding to the top-level object group is used for Describe the same attribute), name (same name in attribute information corresponding to the top-level object group), gender (each attribute information corresponding to the top-level object group is used to describe different attributes), level (corresponding to the top-level object group) Each attribute information is used to describe different attributes).
  • the attribute information corresponding to the vehicle object group includes: the license plate number (the identification code in the attribute information corresponding to the top-level object group is used to describe the same attribute), the production date (the birth time in the attribute information corresponding to the top-level object group is used to describe The same attribute), name (the name in the attribute information corresponding to the top-level object group is the same), mobile (each attribute information corresponding to the top-level object group is used to describe different attributes), brand (each corresponding to the top-level object group) Each attribute information is used to describe different attributes).
  • the human object group may include two types: a student object group and a police object group.
  • the attribute information corresponding to the student object group may include: an ID number (the ID number in the attribute information corresponding to the human object group is the same), a date of birth (the date of birth in the attribute information corresponding to the human object group is the same), First name (same name in attribute information corresponding to person object group), gender (same gender in attribute information corresponding to person object group), grade (level in attribute information corresponding to person object group) is used to describe the same Attributes), school name (each attribute information corresponding to a person object group is used to describe a different attribute).
  • the attribute information corresponding to the police object group may include: ID number (identical number in the attribute information corresponding to the person object group), date of birth (same birth date in the attribute information corresponding to the person object group), and name ( The name in the attribute information corresponding to the person object group is the same), the gender (the same gender in the attribute information corresponding to the person object group), the job title (the level in the attribute information corresponding to the person object group is used to describe the same attribute), The name of the police station (each attribute information corresponding to the person object group is used to describe different attributes).
  • the ontology template records: attribute information corresponding to each object group.
  • the attribute information of the top-level object group in the multiple object groups includes: the attribute information corresponding to the top-level object group; in each two-level object group, the types of the attribute information of the upper-level object group include: direct inheritance classes and indirect inheritance classes; lower-level inheritance classes; The types of attribute information corresponding to an object group include: implementation classes.
  • the attribute information of the lower-level object group includes attribute information of the directly inherited class corresponding to the upper-level object group, and attribute information corresponding to the lower-level object group.
  • the attribute information of the indirect inheritance class corresponding to the upper-level object group is different from the attribute information of the implementation class corresponding to the lower-level object group and is used to describe the same attribute; the attribute information of the direct-inherited class corresponding to the upper-level object group corresponds to the lower-level object group.
  • the property information of the implementation class is used to describe different properties.
  • FIG. 4 is a schematic diagram of another ontology template provided by an embodiment of the present disclosure.
  • the attribute information corresponding to an object group is the same as attribute information corresponding to an upper-level object group associated with it, the attribute information corresponding to the lower-level object group needs to delete the certain attribute information, so that an ontology template as shown in FIG. 4 can be obtained.
  • the attribute information of the person object group includes: ID number, date of birth, gender, and level, and because the attribute information corresponding to the top-level object group includes a name, the attribute information corresponding to the person object group does not include: a name.
  • the attribute information of the car object group includes: the license plate number, the production date, the motor, and the brand, and because the attribute information corresponding to the top-level object group includes the name, the attribute information corresponding to the car object group does not include: the name.
  • the attribute information of the student object group may include: grade, school name, and because the attribute information corresponding to the human object group includes ID card number, date of birth, name, and gender, the attribute information corresponding to the student object group does not include: ID number , Date of birth, name, gender.
  • the attribute information of the police object group may include: the job title, the name of the police station where it belongs, and because the attribute information corresponding to the person object group includes the ID number, date of birth, name, and gender, the attribute information corresponding to the police object group does not include: identity ID number, date of birth, name, gender.
  • the ontology template can record the attribute information of each object group by recording the attribute information corresponding to each object group, and the attribute information of each non-top-level object group can be partially omitted, which can reduce the ontology template
  • the attribute information recorded in the simplifies the ontology template, which can save more space for storing the ontology template, and can effectively reduce the storage cost.
  • the types of attribute information corresponding to the lower-level object group include: implementation classes and custom classes
  • the attribute information corresponding to the upper-level object group includes: the same attributes described in the attribute information of each implementation class corresponding to the lower-level object group Attribute information, and does not include the same attribute information described in the attribute information of each custom class corresponding to the lower-level object group.
  • the ID number and the date of birth in the attribute information corresponding to the person object group are attribute information of the implementation class
  • the gender and level are attribute information of the custom class.
  • the attribute information corresponding to the top-level object group includes: the same attribute information (identification code) as the attribute described by the ID number, and the same attribute (birth time) as the attribute described on the date of birth, excluding the gender and level Describes attributes with the same attribute information.
  • a top-level object group a person object group, a car object group, a student object group, and a police object group.
  • multiple object groups may also be It includes other object groups, such as a teacher object group, a building object group, a kinship object group, and a teacher-student relationship object group, which are not limited in the embodiments of the present disclosure.
  • the attribute information of each object may also include information not mentioned in the above embodiments, such as the instance attribute information of a person may also include: ethnicity; the attribute information of the relationship may also include: the type of relationship (such as kinship, teacher-student relationship And other types of relationships), which are not described in the embodiment of the present disclosure.
  • the computer device may obtain the ontology template in multiple ways in step 201, and the embodiment of the present disclosure will take the way shown in FIG. 5 as an example.
  • FIG. 5 is a flowchart of a method for obtaining an ontology template according to an embodiment of the present disclosure. As shown in FIG. 5, the process of obtaining the ontology template may include:
  • Step 2011 Determine attribute information of each object group in the multiple object groups.
  • the ontology construction method may be applied to computer equipment.
  • the computer equipment may use one or more of historical data and expert experience values (such as historical data, expert experience values, or historical data). And expert experience) to determine the attribute information of each object group.
  • expert experience values such as historical data, expert experience values, or historical data.
  • step 2012 it is detected whether the i-th object group in the multiple object groups meets the pull-up condition and the sinking condition.
  • the multiple object groups need to be related in a hierarchical form, and then, it is checked whether the i-th object group in the multiple object groups meets the pull-up condition and the sinking condition.
  • the pull-up condition includes: the attribute described in the attribute information of each lower-level object group associated with the i-th object group includes the first attribute, and the attribute described in the attribute information of the i-th object group does not include the first An attribute, 1 ⁇ i ⁇ j, where j is the number of object groups in multiple object groups.
  • the attributes described by the attribute information of the student object group and the police object group both include level attributes.
  • the grade in the attribute information of the student object group is used to describe the level attribute
  • the title in the attribute information of the police object group is used to describe Level attributes. If the attribute described by the attribute information of the person object group does not include the level attribute, it can be determined that the person object group satisfies the pull-up condition.
  • the sinking condition includes: the attribute described in the attribute information of the i-th object group includes a second attribute, and the attribute described in the attribute information of each lower-layer object group associated with the i-th object group does not include the first Two attributes. For example, suppose the level in the attribute information of the person object group is used to describe the level attribute. If the attributes described by the attribute information of the student object group and the police object group do not include the level attribute, for example, the attribute information of the student object group does not include the grade, The attribute information of the police object group does not include the job title, it can be determined that the human object group meets the sinking condition.
  • Step 2013 When the i-th object group satisfies the pull-up condition, attribute information describing the first attribute is added to the attribute information of the i-th object group to update the attribute information of the plurality of object groups.
  • Attribute information describing the first attribute may be added to the attribute information of the i-th object group to update the attribute information of the i-th object group. That is, the attributes required by the upper-level object group are obtained from the attribute information of the multiple lower-level object groups, so that the attribute information of the upper-level object group can be adjusted according to the attributes of the multiple lower-level object groups to avoid lack of experience or experts Carelessness, missing some attribute information of the upper object group.
  • Step 2014 When the i-th object group meets the sinking condition, remove the attribute information describing the second attribute from the attribute information of the i-th object group to update the attribute information of the plurality of object groups.
  • the attribute information describing the second attribute in the attribute information of the i-th object group may be removed to update the attribute information of the i-th object group. That is, attribute information that is not required by the upper-level object group is derived from the attribute information of the multiple lower-level object groups, so that the attribute information of the upper-level object group can be adjusted according to the attribute information of the multiple lower-level object groups to avoid lack of experience. Or the negligence of the expert, it is determined that the attribute information of the upper-level object group is redundant.
  • Step 2015 Determine an ontology template according to the updated attribute information of multiple object groups.
  • Step 202 Instantiate the specified data set according to the ontology template and the instantiation rules to obtain instance attribute information of multiple objects.
  • the specified data set is data that needs to be instantiated. Processing the data set can obtain instance attribute information of multiple objects.
  • the computer device needs to first determine attribute information of each underlying object group according to the ontology template.
  • the computer device can directly obtain the attribute information corresponding to each underlying object group from the ontology template, and correspondingly obtain each object group obtained.
  • the attribute information is determined as the attribute information of each object group.
  • the computer device needs to first obtain the attribute information corresponding to each underlying object group from the ontology template, and obtain the previous layer object associated with each underlying object group. Attribute information of the group; after that, the obtained attribute information (the attribute information corresponding to each underlying object group and the attribute information of the upper-level object group associated with each object group) can be determined as each underlying layer The attribute information of the object group.
  • the attribute information of the upper-level object group is attribute information corresponding to the upper-level object group; when the upper-level object group is not a top-level object group, For the process of obtaining the attribute information of the upper-level object group, refer to the process of obtaining the attribute information of each lower-level object group.
  • a preset instantiation rule can also be obtained. Then, based on the attribute information of each underlying object group, combined with the instantiation rules, instantiate the data collection to extract instance attribute information of multiple objects in the data collection.
  • the instance attribute information of Xiaohong can include: 6XXXX ... XXX (instantiation of ID number), 198X years December X (instantiation of birth date), Xiaohong (instantiation of name), female (instantiation of gender), third grade (instantiation of grade), third elementary school (instantiation of school name). It should be noted that the "X" means "some”.
  • the instance attribute information of each object obtained by the final instantiation process may include: each object Time-dimensional attribute information in the time dimension.
  • the objects described by the ontology can be: entities, relationships between entities, or entity-related events.
  • the time-dimensional attribute information in the instance attribute information of these three objects will be explained below. Instructions.
  • the time-dimensional attribute information of the entity may include: a surviving identifier for recording whether the entity is deceased.
  • the time-dimensional attribute information of the entity may further include indication information of one or more of an entity's birth time and an entity's extinction time.
  • the time-dimensional attribute information of the entity may further include: a surviving event identifier of the entity, and the surviving event identifier is used to record a triggering event when the time-dimensional attribute information of the entity changes.
  • the time-dimensional attribute information of a person may include: a surviving identifier 1, a date of birth (instruction information for recording a person's birth date), and a date of death (instruction information for recording a person's demise time) , Survival event identification.
  • the surviving mark 1 is used to record that the person is alive
  • the surviving mark 0 is used to record that the person has died.
  • the time-dimensional attribute information of a car (an entity) may include: survival identifier 0, production date (instruction information used to record the birth date of the car), scrap date (instruction information used to record the death time of the car), and survival events Logo.
  • the surviving mark 1 is used to record that the car has not been scrapped
  • the surviving mark 0 is used to record that the car has been scrapped.
  • the time-dimensional attribute information of the event may include: whether it is directed, active, passive, event start time, and event end time. It should be noted that the event directed means that the event is initiated by one entity and the other entity is passively involved; the event non-directed means that the event is initiated and participated by all participating entities.
  • the time-dimensional attribute information of the relationship includes: a version identifier of the relationship, wherein each relationship has one or more version identifiers, and each version identifier uniquely corresponds to a time period maintained by the relationship.
  • the time-dimensional attribute information of the relationship further includes one or more of a start time and an end time of the relationship.
  • the time-dimensional attribute information of the relationship further includes: a source event identifier of the relationship, and the source event identifier is used to record an event related to each version identifier of the relationship.
  • Zhang San and Li Si had a living-dwelling relationship.
  • the relationship version of Zhang San's and Li Si's living-dwelling relationship could be relationship version 1.
  • the time period corresponding to this relationship version 1 was September 2, 2017 ⁇ September 5, 2017;
  • Zhang San and Li Si had a cohabitation relationship again.
  • the relationship version of Zhang San and Li Si's cohabitation relationship can be relationship version 2, the time corresponding to this relationship version 2.
  • the segment is from October 2, 2017 to October 5, 2017.
  • the attribute information of the object recorded in the ontology template includes attribute information for describing the attribute
  • the instantiation of the attribute information of the object may be left blank, or the attribute information may not be instantiated.
  • the indication information of the extinction event in the entity's time-dimensional attribute information can be left blank, or the attribute information is not instantiated, so that the data After the collection is instantiated, the indication information of the extinction event will not be obtained.
  • Step 203 Store instance attribute information of multiple objects according to the specified storage structure to construct an ontology.
  • instance attribute information may be stored in different ways according to the type of the object.
  • the computer device can detect whether the historical instance attribute information of the i-th entity is stored locally, 1 ⁇ i ⁇ k, where k is multiple The number of entities in each object; when historical instance attribute information is stored locally with the i-th entity, the computer device may use the instance attribute information of the i-th entity to cover its historical instance attribute information. That is, once the computer device has the latest instance attribute information of an entity, it can overwrite the latest instance attribute information of the entity to ensure that the computer device always stores the latest instance attribute information locally, so that the computer The instance attribute information stored locally on the device is consistent with the attribute information of the entities in the real world. When the historical instance attribute information of the i-th entity is not stored locally, the computer device may directly store the instance attribute information of the i-th entity.
  • the computer equipment can directly store the instance attribute information of the i-th event among the instance attribute information of multiple objects.
  • the computer device may first determine the number of historical instance attribute information of the h-th relationship stored locally, 1 ⁇ h ⁇ p, p is the number of relations in the plurality of objects. For example, the computer device may determine the number of relationship versions of the h-th relationship stored locally as the number of historical instance attribute information of the h-th relationship stored locally.
  • the computer device needs to determine whether the number of historical instance attribute information of the h-th relationship stored locally is greater than a preset number threshold; and the number of historical instance attribute information of the h-th relationship stored locally is greater than a preset number
  • the peripheral recycle bin periodically deletes the data stored in it.
  • the peripheral recycle bin can periodically delete data according to the instructions of the computer equipment.
  • the relationship version in the remaining historical instance attribute information is the remaining relationship version
  • the relationship version in the cleanable historical instance attribute information is the cleanable relationship version
  • the time period indicated by the remaining relationship version is in the cleanable relationship version. After the time period.
  • the time period indicated by the remaining relationship version is: October 2, 2017-October 5, 2017, and the relationship can be cleared
  • the version is relationship version 1.
  • the time period indicated by the cleanable relationship version is from September 2, 2017 to September 5, 2017.
  • Step 204 Update the ontology when the target parameters are updated.
  • the target parameter may include one or more of an ontology template, an instantiation rule, and a data set.
  • the target parameters include an ontology template, an instantiation rule, and a data set.
  • the ontology template, instantiation rules, and data collection can be stored on a peripheral storage medium of a computer device, and the ontology template, instantiation rules, and data collection can be stored on different peripheral storage media.
  • the disclosed embodiments are not limited thereto.
  • the staff can update the ontology template on the peripheral storage medium.
  • the computer device may detect whether the ontology template, the instantiation rule, and the data set are updated.
  • the computer device may The current instantiation rule instantiates the data to be processed to obtain instance attribute information of multiple update objects, where the data to be processed is the data to be instantiated indicated by the current instantiation rule in the current data set .
  • the computer device may overwrite the stored instance attribute information with the instance attribute information of multiple update objects to update the ontology.
  • the ontology created in the embodiment of the present disclosure can be changed with the target parameters and time, so that the ontology created in the embodiment of the present disclosure is not static but is a dynamically changing ontology. Dynamic Ontology.
  • Step 205 Obtain a target query condition.
  • the target query condition includes: a time condition.
  • the initial conditions may include: time conditions (September 1, 2017-September 5, 2017), relationship type (cohabitation relationship), active party (Zhang San), passive party (Li Si).
  • the computer equipment needs to determine whether the initial query condition includes a time condition, and when the initial query condition includes a time condition, determine that the initial query condition is a target query condition. It should be noted that if the initial query condition does not include a time condition, the computer device may add a preset time condition to the initial query condition to obtain a target query condition.
  • the preset time condition may be used to record a recent time period, such as a time period from a certain historical time to a current time. For example, if the current time is October 1, 2018, the preset time condition can be used to record: the period from September 1, 2018 to October 1, 2018.
  • Step 206 Query whether there is target instance attribute information that satisfies the target query condition in the ontology according to the target query condition.
  • step 207 is performed; when the target instance attribute information does not exist in the ontology, step 208 is performed.
  • Step 207 Obtain attribute information of the target instance.
  • the target query conditions include: time condition (September 1, 2017-September 5, 2017), relationship type (cohabitation relationship), active party (Zhang San), passive party (Li Si), then According to the target query condition, the target instance attribute information can be found in the ontology: the instance attribute information of Zhang San and Li Si's cohabitation relationship from September 1, 2017 to September 5, 2017.
  • Step 208 Query whether there is target instance attribute information in the peripheral recycle bin.
  • step 207 is performed; when the target instance attribute information does not exist in the peripheral recycle bin, step 209 is performed.
  • the target instance attribute information may be moved to the peripheral recycle bin. Therefore, the peripheral recycle bin can continue to query whether the target instance attribute information exists.
  • the target instance attribute information in the peripheral recycle bin can be obtained.
  • Step 209 Process the target data in the data set by using a preset processing method to obtain target instance attribute information.
  • the processing of the target data in the data set by using a preset processing method may include: instantiating the target data in the data set according to the current ontology template and the current instantiation rule to obtain target instance attribute information.
  • the computer device can obtain the target data in the data set, the target data is data generated in the data set within a specified time period, and the specified time period is time The time period indicated by the condition. Thereafter, the computer device may process the target data according to the current ontology template (ie, the latest ontology template) and the current instantiation rule (ie, the latest instantiation rule) to obtain target instance attribute information.
  • the current ontology template ie, the latest ontology template
  • the current instantiation rule ie, the latest instantiation rule
  • the instance attribute information obtained in the embodiment of the present disclosure includes time-dimensional attribute information of the time dimension
  • the ontology constructed based on the instance attribute information can describe the relationship between the object and time. In this way, the object described by the ontology Being able to unify with the objects in the real world improves the accuracy of the ontology.
  • an ontology template is first obtained, and then instance attribute information of multiple objects is obtained according to the ontology template. Because the ontology template can associate multiple object groups in a hierarchical form according to attributes, it can sort and group multiple and disordered objects in the real world in an orderly manner, so that the objects that the ontology needs to describe can be classified. The ordering of the sequence therefore reduces the complexity of the ontology construction process.
  • multiple object groups are associated in a hierarchical manner according to attributes, so that multiple and disordered objects in the real world can be ordered and classified into multiple layers, so that the attribute information of the objects in the ontology is The accuracy is higher.
  • Grouping in the related art is based on the type of the object, so the granularity level of this method of grouping is type, and the granularity level of the method of grouping in the embodiments of the present disclosure is an attribute. Therefore, the embodiment of the present disclosure realizes the transition from grouping by type to grouping by attribute, realizes the transition from grouping at coarse grain level to grouping at fine grain level, and refines grouping of objects.
  • FIG. 6 is a schematic structural diagram of an ontology construction device according to an embodiment of the present disclosure. As shown in FIG. 6, the ontology construction device may include:
  • a first processing module 601 configured to instantiate a specified data set to obtain instance attribute information of each of a plurality of objects
  • a storage module 602 configured to store instance attribute information of multiple objects according to a specified storage structure to construct an ontology
  • the instance attribute information of each object includes: time-dimensional attribute information of each object in the time dimension.
  • the instance attribute information obtained by the first processing module in the embodiment of the present disclosure includes time-dimensional attribute information of the time dimension
  • the ontology constructed by the storage module based on the instance attribute information can describe the relationship between the object and time.
  • the objects described by the ontology can be unified with those in the real world, which improves the accuracy of the ontology.
  • multiple objects include entities, and the time-dimensional attribute information of the entities includes one or more of the following:
  • Survival event ID of the entity is used to indicate the trigger event when the entity's time-dimensional attribute information changes.
  • the storage module 602 is configured to:
  • the instance attribute information of the i-th entity is used to cover its historical instance attribute information.
  • the multiple objects include relationships between entities, and the time-dimensional attribute information of the relationship includes one or more of the following:
  • a version identifier of the relationship wherein each relationship has at least one version identifier, and each version identifier uniquely corresponds to a time period maintained by the relationship;
  • the source event ID of the relationship is used to indicate events related to each version ID of the relationship.
  • the storage module is used for:
  • attribute information of the h-th relationship in multiple objects determine the number of historical instance attribute information of the h-th relationship stored locally, 1 ⁇ h ⁇ p, where p is the number of relationships in multiple objects;
  • the cleanable historical instance attribute information of the hth relationship stored locally is moved to the peripheral recycle bin, so that the hth relationship
  • the number of remaining historical instance attribute information of each relationship is less than or equal to a preset number threshold; wherein the historical instance attribute information of the h-th relationship includes: cleanable historical instance attribute information of the h-th relationship and the information of the h-th relationship Remaining historical instance attribute information;
  • the relationship version in the remaining historical instance attribute information is the remaining relationship version
  • the relationship version in the cleanable historical instance attribute information is the cleanable relationship version
  • the time period corresponding to the remaining relationship version is after the time period corresponding to the cleanable relationship version.
  • FIG. 7 is a schematic structural diagram of another ontology construction device according to an embodiment of the present disclosure. As shown in FIG. 7, based on FIG. 6, the ontology construction device further includes:
  • a first obtaining module 603 is configured to obtain an ontology template and an instantiation rule.
  • the ontology template is used to record attribute information of multiple object groups.
  • Each object group includes objects of the same type, and multiple object groups are related in a hierarchical form.
  • the first processing module 601 is configured to: instantiate the data set according to the ontology template and the instantiation rules; wherein the instance attribute information of each object in the multiple objects is based on the attribute information recorded in the ontology template. Extracted information.
  • a detection module 604 configured to detect whether an ontology template, an instantiation rule, and a data set are updated
  • the second processing module 605 is configured to instantiate the data to be processed according to the current ontology template and the current instantiation rule when one or more of the ontology template, the instantiation rule, and the data set are updated, so as to obtain multiple Update the instance attribute information of the object, where the data to be processed is the data to be instantiated indicated by the current instantiation rule in the current data set;
  • the overwriting module 606 is configured to overwrite the stored instance attribute information of the multiple update objects to update the ontology.
  • the second obtaining module 607 is configured to obtain a target query condition, and the target query condition includes: a time condition;
  • a first query module 608, configured to query whether there is target instance attribute information that satisfies the target query condition in the ontology according to the target query condition;
  • the third obtaining module 609 is configured to obtain target instance attribute information when the target instance attribute information exists in the ontology.
  • a second query module 610 configured to query whether the target instance attribute information exists in the peripheral recycle bin when the ontology does not include the target instance attribute information
  • the third obtaining module 609 is further configured to execute the step of obtaining the target instance attribute information when the target instance attribute information exists in the peripheral recycle bin;
  • the third processing module 611 is configured to instantiate the target data in the data set according to the current ontology template and the current instantiation rule when the target instance attribute information does not exist in the peripheral recycle bin to obtain the target instance. Attribute information
  • the time condition is used to indicate a specified time period
  • the target data is data generated in the data set within the specified time period
  • the second obtaining module 608 is configured to: receive the initial query condition; determine whether the initial query condition includes a time condition; when the initial query condition includes the time condition, determine that the initial query condition is the target query condition; when the initial query condition does not include the time condition, The preset time condition is added to the initial query condition to obtain the target query condition.
  • the instance attribute information obtained by the first processing module in the embodiment of the present disclosure includes time-dimensional attribute information of the time dimension
  • the ontology constructed by the storage module based on the instance attribute information can describe the relationship between the object and time.
  • the objects described by the ontology can be unified with those in the real world, which improves the accuracy of the ontology.
  • An embodiment of the present disclosure also provides a computer-readable storage medium.
  • a computer program is stored in the storage medium, and the computer program is executed by a processor to implement the ontology construction method.
  • An embodiment of the present disclosure also provides a computer program product containing instructions, which causes the computer to execute the ontology construction method described above when the computer program product runs on the computer.
  • the embodiment of the ontology construction method and the embodiment of the ontology construction device provided by the embodiments of the present disclosure may refer to each other, and the embodiments of the present disclosure do not limit this.
  • the order of the steps of the method embodiments provided by the embodiments of the present invention can be appropriately adjusted, and the steps can be increased or decreased according to the situation. Any person skilled in the art can easily think of changes within the technical scope disclosed by the present invention. The methods should all be covered by the protection scope of the present invention, so they will not be described again.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Quality & Reliability (AREA)
  • Computational Linguistics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

一种本体构建方法及装置,属于本体构建技术领域。所述方法包括:对指定的数据集合进行实例化处理,得到多个对象中每个对象的实例属性信息(101);其中,所述每个对象的实例属性信息包括:所述每个对象在时间维度上的时维属性信息;按照指定的存储结构,存储所述多个对象的实例属性信息,以构建所述本体(102)。本方案解决了本体的准确度较低的问题,提高了本体的准确度。

Description

本体构建方法及装置
本公开要求于2018年06月29日提交的申请号为201810701221.X、发明名称为“本体构建方法及装置”的中国专利申请的优先权,其全部内容通过引用结合在本公开中。
技术领域
本公开涉及计算机技术领域,特别涉及一种本体构建方法及装置。
背景技术
本体是对真实世界中的实体(如人或物)、实体涉及的事件以及实体之间的关系的描述,人们可以在本体中对被本体描述的对象进行查询。
示例的,本体描述的每个对象均具有多种属性,如人的属性包括:姓名、性别和身份证号等。本体可以通过记载对象的属性信息对该对象进行描述。相关技术中,在构建本体时,计算机设备需要对指定的数据集合进行实例化处理,从而得到每个对象(如实体、事件或关系)的实例属性信息,并将每个对象的实例属性信息进行存储。
发明内容
本公开提供了一种本体构建方法及装置,所述技术方案如下:
第一方面,提供了一种本体构建方法,所述方法包括:
对指定的数据集合进行实例化处理,得到多个对象中每个对象的实例属性信息;其中,所述每个对象的实例属性信息包括:所述每个对象在时间维度上的时维属性信息;
按照指定的存储结构,存储所述多个对象的实例属性信息,以构建所述本体。
可选的,所述多个对象包括实体,所述实体的时维属性信息包括以下一种或多种:
用于记载所述实体是否消亡的存续标识;
所述实体的诞生时间的指示信息;
所述实体的消亡时间的指示信息;
所述实体的存续事件标识,所述存续事件标识用于记载所述实体的时维属性信息发生变化的触发事件。
可选的,所述存储所述多个对象的实例属性信息,包括:
对于所述多个对象中第i个实体的实例属性信息,检测本地是否存储有所述第i个实体的历史实例属性信息,1≤i≤k,k为所述多个对象中实体的个数;
在本地存储有与所述第i个实体的历史实例属性信息时,采用所述第i个实体的实例属性信息覆盖其历史实例属性信息。
可选的,所述多个对象包括实体间的关系,所述关系的时维属性信息包括以下一种或多种:
所述关系的版本标识,其中,每个所述关系具有至少一个版本标识,每个所述版本标识唯一对应所述关系所维持的一个时间段;
所述关系的起始时间的指示信息;
所述关系的结束时间的指示信息;
所述关系的来源事件标识,所述来源事件标识用于记载与所述关系的每个版本标识相关的事件。
可选的,所述存储所述多个对象的实例属性信息,包括:
对于所述多个对象中第h个关系的实例属性信息,确定本地存储的所述第h个关系的历史实例属性信息的个数,1≤h≤p,p为所述多个对象中关系的个数;
判断本地存储的所述第h个关系的历史实例属性信息的个数是否大于预设个数阈值;
在本地存储的所述第h个关系的历史实例属性信息的个数大于所述预设个数阈值时,将本地存储的所述第h个关系的可清理历史实例属性信息移动至外设回收站,以使得所述第h个关系的剩余历史实例属性信息的个数小于或等于所述预设个数阈值;其中,所述第h个关系的历史实例属性信息包括:所述第h个关系的可清理历史实例属性信息和所述第h个关系的剩余历史实例属性信息;
存储所述第h个关系的实例属性信息。
可选的,所述剩余历史实例属性信息中的关系版本为剩余关系版本,所述可清理历史实例属性信息中的关系版本为可清理关系版本;
所述剩余关系版本所对应的时间段在所述可清理关系版本所对应的时间段之后。
可选的,在所述对指定的数据集合进行实例化处理之前,所述方法还包括:获取本体模板和实例化规则,其中,所述本体模板用于记载多个对象组的属性信息,每个对象组包括属于同一类型的对象,所述多个对象组以层级的形式关联;
所述对指定的数据集合进行实例化处理,包括:根据所述本体模板和所述实例化规则,对所述数据集合进行所述实例化处理;
其中,所述多个对象中每个对象的实例属性信息为基于所述本体模板所记载的属性信息,在所述数据集合中提取的信息。
可选的,在存储所述多个对象的实例属性信息之后,所述方法还包括:
分别检测所述本体模板、所述实例化规则以及所述数据集合是否更新;
当所述本体模板、所述实例化规则以及所述数据集合中的一个或多个更新时,根据当前的本体模板和当前的实例化规则对待处理数据进行实例化处理,以得到多个更新对象的实例属性信息,其中,所述待处理数据为当前的数据集合中所述当前的实例化规则所指示的待进行实例化处理的数据;
将所述多个更新对象的实例属性信息覆盖已存储的实例属性信息,以更新所述本体。
可选的,在基于所述多个对象的实例属性信息构建所述本体之后,所述方法还包括:
获取目标查询条件,所述目标查询条件包括:时间条件;
根据所述目标查询条件,查询所述本体中是否存在满足所述目标查询条件的目标实例属性信息;
在所述本体中存在所述目标实例属性信息时,获取所述目标实例属性信息。
可选的,所述获取目标查询条件,包括:
接收初始查询条件;
判断所述初始查询条件是否包括所述时间条件;
在所述初始查询条件包括所述时间条件时,确定所述初始查询条件为所述目标查询条件;
在所述初始查询条件不包括所述时间条件时,将预设的时间条件加入所述初始查询条件,以得到所述目标查询条件。
可选的,所述方法还包括:
在所述本体不包括所述目标实例属性信息时,查询外设回收站中是否存在 所述目标实例属性信息;
在所述外设回收站中存在所述目标实例属性信息时,执行获取所述目标实例属性信息的步骤;
在所述外设回收站中不存在所述目标实例属性信息时,根据当前的本体模板和当前的实例化规则对所述数据集合中的目标数据进行实例化处理,以得到所述目标实例属性信息;
其中,所述时间条件用于记载指定时间段,所述目标数据为所述数据集合中在所述指定时间段内生成的数据。
第二方面,提供了一种本体构建装置,所述本体构建装置包括:
第一处理模块,用于对指定的数据集合进行实例化处理,得到多个对象中每个对象的实例属性信息;
存储模块,用于按照指定的存储结构,存储所述多个对象的实例属性信息,以构建所述本体;
其中,所述每个对象的实例属性信息包括:所述每个对象在时间维度上的时维属性信息。
可选的,所述多个对象包括实体,所述实体的时维属性信息包括以下一种或多种:
用于记载所述实体是否消亡的存续标识;
所述实体的诞生时间的指示信息;
所述实体的消亡时间的指示信息;
所述实体的存续事件标识,所述存续事件标识用于记载所述实体的时维属性信息发生变化的触发事件。
可选的,所述存储模块用于:
对于所述多个对象中第i个实体的实例属性信息,检测本地是否存储有所述第i个实体的历史实例属性信息,1≤i≤k,k为所述多个对象中实体的个数;
在本地存储有与所述第i个实体的历史实例属性信息时,采用所述第i个实体的实例属性信息覆盖其历史实例属性信息。
可选的,所述多个对象包括实体间的关系,所述关系的时维属性信息包括以下一种或多种:
所述关系的版本标识,其中,每个所述关系具有至少一个版本标识,每个所述版本标识唯一对应所述关系所维持的一个时间段;
所述关系的起始时间的指示信息;
所述关系的结束时间的指示信息;
所述关系的来源事件标识,所述来源事件标识用于记载与所述关系的每个版本标识相关的事件。
可选的,所述存储模块用于:
对于所述多个对象中第h个关系的实例属性信息,确定本地存储的所述第h个关系的历史实例属性信息的个数,1≤h≤p,p为所述多个对象中关系的个数;
判断本地存储的所述第h个关系的历史实例属性信息的个数是否大于预设个数阈值;
在本地存储的所述第h个关系的历史实例属性信息的个数大于所述预设个数阈值时,将本地存储的所述第h个关系的可清理历史实例属性信息移动至外设回收站,以使得所述第h个关系的剩余历史实例属性信息的个数小于或等于所述预设个数阈值;其中,所述第h个关系的历史实例属性信息包括:所述第h个关系的可清理历史实例属性信息和所述第h个关系的剩余历史实例属性信息;
存储所述第h个关系的实例属性信息。
可选的,所述剩余历史实例属性信息中的关系版本为剩余关系版本,所述可清理历史实例属性信息中的关系版本为可清理关系版本;
所述剩余关系版本所对应的时间段在所述可清理关系版本所对应的时间段之后。
可选的,所述本体构建装置还包括:
第一获取模块,用于获取本体模板和实例化规则,其中,所述本体模板用于记载多个对象组的属性信息,每个对象组包括属于同一类型的对象,所述多个对象组以层级的形式关联;
所述第一处理模块用于:根据所述本体模板和所述实例化规则,对所述数据集合进行所述实例化处理;
其中,所述多个对象中每个对象的实例属性信息为基于所述本体模板所记载的属性信息,在所述数据集合中提取的信息。
可选的,所述本体构建装置还包括:
检测模块,用于分别检测所述本体模板、所述实例化规则以及所述数据集合是否更新;
第二处理模块,用于在所述本体模板、所述实例化规则以及所述数据集合 中的一个或多个更新时,根据当前的本体模板和当前的实例化规则对待处理数据进行实例化处理,以得到多个更新对象的实例属性信息,其中,所述待处理数据为当前的数据集合中所述当前的实例化规则所指示的待进行实例化处理的数据;
覆盖模块,用于将所述多个更新对象的实例属性信息覆盖已存储的实例属性信息,以更新所述本体。
可选的,所述本体构建装置还包括:
第二获取模块,用于获取目标查询条件,所述目标查询条件包括:时间条件;
第一查询模块,用于根据所述目标查询条件,查询所述本体中是否存在满足所述目标查询条件的目标实例属性信息;
第三获取模块,用于在所述本体中存在所述目标实例属性信息时,获取所述目标实例属性信息。
可选的,所述第二获取模块用于:
接收初始查询条件;
判断所述初始查询条件是否包括所述时间条件;
在所述初始查询条件包括所述时间条件时,确定所述初始查询条件为所述目标查询条件;
在所述初始查询条件不包括所述时间条件时,将预设的时间条件加入所述初始查询条件,以得到所述目标查询条件。
可选的,所述本体构建装置还包括:
第二查询模块,用于在所述本体不包括所述目标实例属性信息时,查询外设回收站中是否存在所述目标实例属性信息;
所述第三获取模块还用于在所述外设回收站中存在所述目标实例属性信息时,执行获取所述目标实例属性信息的步骤;
第三处理模块,用于在所述外设回收站中不存在所述目标实例属性信息时,根据当前的本体模板和当前的实例化规则对所述数据集合中的目标数据进行实例化处理,以得到所述目标实例属性信息;
其中,所述时间条件用于记载指定时间段,所述目标数据为所述数据集合中在所述指定时间段内生成的数据。
第三方面,提供了一种计算机设备,包括处理器、通信接口、存储器和通 信总线,其中,处理器,通信接口,存储器通过总线完成相互间的通信;存储器,用于存放计算机程序;处理器,用于执行存储器上所存放的程序,实现第一方面所述的方法。
第四方面,提供了一种计算机可读存储介质,所述存储介质内存储有计算机程序,所述计算机程序被处理器执行时实现第一方面所述的方法。
第五方面,提供了一种包含指令的计算机程序产品,当所述计算机程序产品在计算机上运行时,使得计算机执行第一方面所述的方法。
附图说明
为了更清楚地说明本公开实施例中的技术方案,下面将对实施例描述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公开的一些实施例,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他的附图。
图1为本公开实施例提供的一种本体构建方法的方法流程图;
图2为本公开实施例提供的另一种本体构建方法的方法流程;
图3为本公开实施例提供的一种本体模板的示意图;
图4为本公开实施例提供的另一种本体模板的示意图;
图5为本公开实施例提供的一种获取本体模板的方法流程图;
图6为本公开实施例提供的一种本体构建装置的结构示意图;
图7为本公开实施例提供的另一种本体构建装置的结构示意图。
具体实施方式
为使本公开的目的、技术方案和优点更加清楚,下面将结合附图对本公开实施方式作进一步地详细描述。
本体是对真实世界中的实体、实体涉及的事件以及实体之间的关系的描述,该实体、事件和关系均为本体描述的对象。例如,实体可以为学生、老师、警察、车等,实体涉及的事件可以为:学生上课、警察抓小偷等与实体相关的事件,实体之间的关系可以为:学生与老师之间的师生关系等。
本体描述的每个对象(如实体、事件或关系)均具有多种属性。例如,学生的属性包括:姓名、性别、年级和身份证号等,车的属性包括:车的名牌和车牌号等。本体可以通过记载对象的实例属性信息对该对象进行描述。
本公开实施例提供了一种本体构建方法,该方法可以用于计算机设备,该计算机设备可以包括一个设备或多个设备组成的设备集群,本公开实施例对此不作限定。该计算机设备可以包括:处理器、通信接口、存储器和通信总线,其中,处理器、通信接口、存储器通过总线完成相互间的通信;该计算机设备可以通过该通信接口与外部的其他设备通信;存储器可以用于存放计算机程序;处理器,用于执行存储器上所存放的程序,实现该本体构建方法。
图1为本公开实施例提供的一种本体构建方法的方法流程图,如图1所示,该本体构建方法可以包括:
步骤101、对指定的数据集合进行实例化处理,得到多个对象中每个对象的实例属性信息。
步骤102、按照指定的存储结构,存储所述多个对象的实例属性信息,以构建所述本体。
其中,每个对象的实例属性信息包括:该每个对象在时间维度上的时维属性信息。
综上所述,由于本公开实施例得到的实例属性信息包括时间维度的时维属性信息,因此,基于该实例属性信息构建的本体能够描述对象与时间的关系,这样一来,本体描述的对象能够与真实世界中的对象统一,提高了本体的准确度。
图2为本公开实施例提供的另一种本体构建方法的方法流程图,如图2所示,该本体构建方法可以包括:
步骤201、获取本体模板和实例化规则。
示例的,该本体模板可以用于记载多个对象组的属性,每个对象组的属性可以包括在时间维度上的时维属性。每个对象组包括属于同一类型的对象,如:属于学生类型的对象组成的对象组,属于医生类型的对象组成的对象组等,属于亲属关系类型的对象组成的对象组,属于公益事件的对象组成的对象组。
多个对象组以层级的形式关联,对于多个对象组相关联且相邻的每两层对象组,下一层对象组的属性包括上一层对象组的属性,也即,下一层对象组具有上一层对象组的属性。如,上一层对象组为人,该下一层对象组可以为学生、警察等。该下一层对象组可以称为上一层对象组的子对象组,上一层对象组可以称为下一层对象组的父对象组,下一层对象组继承了上一层对象组的部分属 性,顶层对象组可以称为其他层对象组的祖先对象组。
需要说明的是,本体模板可以通过多种可实现方式对每个对象组的属性信息进行记载,本公开实施例将对其中的两种可实现方式进行解释说明。
在本体模板的一种可实现方式中,本体模板中记载有:每个对象组对应的属性信息,并且每个对象组对应的属性信息为每个对象组的属性信息。
图3为本公开实施例提供的一种本体模板的示意图,如图3所示,该多个对象组可以包括:顶层对象组、人对象组、学生对象组、警察对象组和车对象组。人对象组和车对象组均与顶层对象组关联,且均位于顶层对象组的下一层,学生对象组和警察对象组均与人对象组关联,且均位于人对象组的下一层。顶层对象组位于这些对象组中的顶层,学生对象组、警察对象组以及车对象组均位于这些对象组中的底层,也即,未关联有下一层对象组的对象组为底层对象组。
其中,顶层对象组对应的属性信息包括:身份标识码、诞生时间以及名字,真实世界中的每种对象均为该顶层对象组中的一种类型,如人对象组和车对象组分别为顶层对象组中的两种类型。
人对象组对应的属性信息包括:身份证号码(与顶层对象组对应的属性信息中的身份标识码用于描述同一属性)、出生日期(与顶层对象组对应的属性信息中的诞生时间用于描述同一属性)、名字(与顶层对象组对应的属性信息中的名字相同)、性别(与顶层对象组对应的每个属性信息均用于描述不同的属性)、级别(与顶层对象组对应的每个属性信息均用于描述不同的属性)。
车对象组对应的属性信息包括:车牌号码(与顶层对象组对应的属性信息中的身份标识码用于描述同一属性)、生产日期(与顶层对象组对应的属性信息中的诞生时间用于描述同一属性)、名字(与顶层对象组对应的属性信息中的名字相同)、机动(与顶层对象组对应的每个属性信息均用于描述不同的属性)、品牌(与顶层对象组对应的每个属性信息均用于描述不同的属性)。
进一步地,人对象组又可以包括:学生对象组和警察对象组这两种类型。
其中,学生对象组对应的属性信息可以包括:身份证号码(与人对象组对应的属性信息中的身份证号码相同)、出生日期(与人对象组对应的属性信息中的出生日期相同)、名字(与人对象组对应的属性信息中的名字相同)、性别(与人对象组对应的属性信息中的性别相同)、年级(与人对象组对应的属性信息中的级别用于描述为同一属性)、学校名称(与人对象组对应的每个属 性信息均用于描述不同的属性)。
警察对象组对应的属性信息可以包括:身份证号码(与人对象组对应的属性信息中的身份证号码相同)、出生日期(与人对象组对应的属性信息中的出生日期相同)、名字(与人对象组对应的属性信息中的名字相同)、性别(与人对象组对应的属性信息中的性别相同)、职称(与人对象组对应的属性信息中的级别用于描述同一属性)、所在警局名称(与人对象组对应的每个属性信息均用于描述不同的属性)。
在本体模板的另一种可实现方式中,本体模板中记载有:每个对象组对应的属性信息。并且,多个对象组中顶层对象组的属性信息包括:顶层对象组对应的属性信息;在每两层对象组中,上层对象组的属性信息的类型包括:直接继承类和间接继承类;下层对象组对应的属性信息的类型包括:实现类。下层对象组的属性信息包括:上层对象组对应的直接继承类的属性信息,以及该下层对象组对应的属性信息。其中,上层对象组对应的间接继承类的属性信息与下层对象组对应的实现类的属性信息不同,且用于描述相同的属性;上层对象组对应的直接继承类的属性信息与下层对象组对应的实现类的属性信息用于描述不同的属性。
图4为本公开实施例提供的另一种本体模板的示意图,如图4所示,在图3的基础上,对于多个对象组中相关联且相邻的每两层对象组,当下层对象组对应的某一属性信息与其关联的上层对象组对应的属性信息相同时,该下层对象组对应的属性信息需要删除该某一属性信息,从而能够得到如图4所述的本体模板。
示例的,人对象组的属性信息包括:身份证号码、出生日期、性别、级别,而由于顶层对象组对应的属性信息包括名字,因此人对象组对应的属性信息并不包括:名字。车对象组的属性信息包括:车牌号码、生产日期、机动、品牌,而由于顶层对象组对应的属性信息包括名字,因此车对象组对应的属性信息并不包括:名字。学生对象组的属性信息可以包括:年级、学校名称,而由于人对象组对应的属性信息包括身份证号码、出生日期、名字、性别,因此学生对象组对应的属性信息并不包括:身份证号码、出生日期、名字、性别。警察对象组的属性信息可以包括:职称、所在警局名称,而由于人对象组对应的属性信息包括身份证号码、出生日期、名字、性别,因此警察对象组对应的属性信息并不包括:身份证号码、出生日期、名字、性别。
这样一来,本体模板可以通过记载每个对象组对应的属性信息,对每个对象组的属性信息进行记载,并且每个非顶层对象组的属性信息均可以进行部分省略,从而能够减少本体模板中记载的属性信息,简化了本体模板,从而可以节省更多的用于存储本体模板的空间,可以有效地降低存储成本。
可选的,下层对象组对应的属性信息的类型包括:实现类和自定义类,该上层对象组对应的属性信息包括:与下层对象组对应的每个实现类的属性信息所描述的属性相同的属性信息,且不包括与下层对象组对应的每个自定义类的属性信息所描述的属性相同的属性信息。示例的,请继续参考图4,人对象组对应的属性信息中的身份证号码和出生日期均为实现类的属性信息,而性别和级别均为自定义类的属性信息。顶层对象组对应的属性信息包括:与身份证号码所描述的属性相同的属性信息(身份标识码),以及与出生日期所描述的属性相同的(诞生时间),而不包括与性别和级别所描述的属性相同的属性信息。
需要说明的是,本公开实施例中仅以多个对象组包括:顶层对象组、人对象组、车对象组、学生对象组和警察对象组为例,实际应用中,多个对象组还可以包括其他对象组,如:老师对象组、建筑对象组、亲属关系对象组、师生关系对象组等,本公开实施例对此不做限定。每个对象的属性信息还可以包括上述实施例中并未提到信息,如人的实例属性信息还可以包括:民族;关系的属性信息还可以包括:关系的种类(如亲属关系、师生关系等多种关系种类),本公开实施例在此不做赘述。
步骤201中计算机设备可以通过多种方式获取该本体模板,本公开实施例将以图5所示的方式为例。示例地,图5为本公开实施例提供的一种获取本体模板的方法流程图。如图5所示,获取上述本体模板的过程可以包括:
步骤2011、确定多个对象组中每个对象组的属性信息。
示例的,该本体构建方法可以用于计算机设备,在步骤2011中,计算机设备可以通过历史数据和专家的经验值中的一种或多种(比如历史数据,专家的经验值,或,历史数据和专家的经验值)确定每个对象组的属性信息。
步骤2012、检测多个对象组中的第i个对象组是否满足上拢条件以及下沉条件。
在步骤2012前,还需要将该多个对象组通过层级的形式关联,之后,再检测该多个对象组中的第i个对象组是否满足上拢条件和下沉条件。
其中,上拢条件包括:第i个对象组关联的每个下一层对象组的属性信息所 描述的属性均包括第一属性,且第i个对象组的属性信息所描述的属性不包括第一属性,1≤i≤j,j为多个对象组中对象组的个数。例如,假设学生对象组和警察对象组的属性信息所描述的属性均包括级别属性,如学生对象组的属性信息中的年级用于描述级别属性,警察对象组的属性信息中的职称用于描述级别属性。若人对象组的属性信息所描述的属性不包括级别属性,则可以确定人对象组满足上拢条件。
下沉条件包括:第i个对象组的属性信息所描述的属性包括第二属性,且所述第i个对象组关联的每个下一层对象组的属性信息所描述的属性均不包括第二属性。例如,假设人对象组的属性信息中的级别用于描述级别属性,若学生对象组和警察对象组的属性信息所描述的属性均不包括级别属性,如学生对象组的属性信息不包括年级,警察对象组的属性信息不包括职称,则可以确定人对象组满足下沉条件。
步骤2013、在第i个对象组满足上拢条件时,在第i个对象组的属性信息中增加用于描述第一属性的属性信息,以更新多个对象组的属性信息。
在确定第i个对象组关联的每个下一层对象组的属性信息所描述的属性均包括第一属性,而该第i个对象组的属性信息所描述的属性却不包括第一属性时,可以在该第i个对象组的属性信息中增加用于描述第一属性的属性信息,以对该第i个对象组的属性信息进行更新。也即,从多个下层对象组的属性信息中提取得到上层对象组需要具有的属性,从而能够根据多个下层对象组的属性对上层对象组的属性信息进行调整,以避免由于经验缺失或者专家的疏忽大意,漏掉上层对象组的部分属性信息。
步骤2014、在第i个对象组满足下沉条件时,将第i个对象组的属性信息中用于描述第二属性的属性信息去除,以更新多个对象组的属性信息。
在确定第i个对象组关联的每个下一层对象组的属性信息所描述的属性均不包括第二属性,而该第i个对象组的属性信息所描述的属性包括第二属性时,可以确定该第i个对象组实质上也并不具有第二属性。此时,可以将第i个对象组的属性信息中用于描述第二属性的属性信息进行去除,以对该第i个对象组的属性信息进行更新。也即,从多个下层对象组的属性信息中推出上层对象组不需要具有的属性信息,从而能够根据多个下层对象组的属性信息对上层对象组的属性信息进行调整,以避免由于经验缺失或者专家的疏忽大意,确定出上层对象组的属性信息多余的情况。
步骤2015、根据更新后的多个对象组的属性信息确定本体模板。
步骤202、根据本体模板和实例化规则,对指定的数据集合进行实例化处理,得到多个对象的实例属性信息。
该指定的数据集合为需要进行实例化处理的数据,对该数据集合进行处理能够得到多个对象的实例属性信息。在步骤202中,计算机设备需要首先根据本体模板确定每个底层对象组的属性信息。
示例的,当本体模板的实现方式为如图3所示的实现方式时,计算机设备可以直接从本体模板中获取每个底层对象组对应的属性信息,并将获取到的每个对象组对应的属性信息,确定为该每个对象组的属性信息。当本体模板的实现方式为如图4所示的实现方式时,计算机设备需要首先从本体模板中获取每个底层对象组对应的属性信息,以及获取该每个底层对象组关联的上一层对象组的属性信息;之后,可以将获取到的属性信息(该每个底层对象组对应的属性信息,以及该每个对象组关联的上一层对象组的属性信息)均确定为该每个底层对象组的属性信息。其中,当该上一层对象组为顶层对象组时,该上一层对象组的属性信息为该上一层对象组对应的属性信息;当该上一层对象组并不是顶层对象组时,获取该上一层对象组的属性信息的过程,可以参考获取该每个底层对象组的属性信息的过程。
在确定每个底层对象组的属性信息后,还可以获取预设的实例化规则。之后,就可以基于该每个底层对象组的属性信息,结合实例化规则对数据集合进行实例化处理,以在数据集合中提取多个对象的实例属性信息。示例的,假设底层对象组为学生,则实例化处理之后,可以得到小红(小红是一个学生)的实例属性信息可以包括:6XXXX...XXX(身份证号码的实例化)、198X年12月X日(出生日期的实例化)、小红(姓名的实例化)、女(性别的实例化)、三年级(年级的实例化)、第三小学(学校名称的实例化)。需要说明的是,上述“X”表示“某”。
另外,由于该本体模板记载的每个对象组的属性信息包括该对象组在时间维度上的时维属性信息,因此,最终实例化处理得到的每个对象的实例属性信息可以包括:每个对象在时间维度上的时维属性信息。需要说明的是,本体描述的对象可以为:实体、实体间的关系或实体相关的事件,以下将对这三种对象(实体、关系和事件)的实例属性信息中的时维属性信息进行解释说明。
第一方面,实体的时维属性信息可以包括:用于记载实体是否消亡的存续 标识。可选的,实体的时维属性信息还可以包括:实体的诞生时间和实体的消亡时间中的一种或多种的指示信息。可选的,实体的时维属性信息还可以包括:实体的存续事件标识,存续事件标识用于记载实体的时维属性信息发生变化的触发事件。
示例的,人(一种实体)的时维属性信息可以包括:存续标识1、出生日期(用于记载人的诞生日期的指示信息)、死亡日期(用于记载人的消亡时间的指示信息)、存续事件标识。其中,存续标识1用于记载人活着,存续标识0用于记载人已死亡。车(一种实体)的时维属性信息可以包括:存续标识0、生产日期(用于记载车的诞生日期的指示信息)、报废日期(用于记载车的消亡时间的指示信息)、存续事件标识。其中,存续标识1用于记载车并未报废,存续标识0用于记载车已报废。
第二方面,事件(与实体相关的时间)的时维属性信息可以包括:是否有向、主动方、被动方、事件起始时间和事件结束时间。需要说明的是,事件有向是指:该事件由一方实体发起,且另一方实体被动参与;而事件无向是指:该事件由参与的所有实体发起并参与。
第三方面,关系(实体间的关系)的时维属性信息包括:关系的版本标识,其中,每个关系具有一个或多个版本标识,每个版本标识唯一对应关系所维持的一个时间段。可选的,关系的时维属性信息还包括:关系的起始时间和结束时间中的一种或多种。可选的,关系的时维属性信息还包括:关系的来源事件标识,来源事件标识用于记载与关系的每个版本标识相关的事件。
以同住关系为例,假设多个实体在交叉时间范围内入住同一个房间,即定义为该多个实体具有同住关系。例如,张三在2017年9月1日~2017年9月5日入住A酒店602房间,李四在2017年9月2日~2017年9月5日入住A酒店602房间,张三在2017年10月1日~2017年10月5日入住B酒店402房间,李四在2017年10月2日~2017年10月5日入住B酒店402房间。
显然在2017年9月,张三和李四存在同住关系,张三和李四的同住关系的关系版本可以为关系版本1,该关系版本1对应的时间段为2017年9月2日~2017年9月5日;在2017年10月张三和李四再次存在同住关系,张三和李四的这个同住关系的关系版本可以为关系版本2,该关系版本2对应的时间段为2017年10月2日~2017年10月5日。
另外需要说明的是,当某一对象实质上不具有某一属性,且本体模板中记 载的该对象的属性信息包括用于描述该属性的属性信息时,在对数据集合进行实例化处理时,可以将该对象的该属性信息的实例化(也即实例属性信息)置空,或不对该属性信息进行实例化。如,在对数据集合进行实例化处理时,若实体并未消亡,则可以将实体的时维属性信息中消亡事件的指示信息置空,或者,不对该属性信息进行实例化,从而在对数据集合进行实例化处理后不会得到该消亡事件的指示信息。
步骤203、按照指定的存储结构,存储多个对象的实例属性信息,以构建本体。
需要说明的是,在存储多个对象的实例属性信息时,可以根据对象的类型采用不同的方式对其实例属性信息进行存储。
第一方面,对于多个对象的实例属性信息中,第i个实体的实例属性信息,计算机设备可以检测本地是否存储有第i个实体的历史实例属性信息,1≤i≤k,k为多个对象中实体的个数;在本地存储有与第i个实体的历史实例属性信息时,计算机设备可以采用第i个实体的实例属性信息覆盖其历史实例属性信息。也即,计算机设备一旦最新得到某一实体的实例属性信息,则可以将最新的该实例属性信息覆盖该实体的历史实例属性信息,以保证计算机设备本地始终存储有最新的实例属性信息,使得计算机设备本地存储的实例属性信息与真实世界中实体的属性信息一致。在本地未存储有与第i个实体的历史实例属性信息时,计算机设备可以直接存储该第i个实体的实例属性信息。
第二方面,对于多个对象的实例属性信息中,第i个事件的实例属性信息,计算机设备可以直接进行存储。
第三方面,对于多个对象的实例属性信息中,第h个关系的实例属性信息,计算机设备可以首先确定本地存储的第h个关系的历史实例属性信息的个数,1≤h≤p,p为该多个对象中关系的个数。示例的,计算机设备可以将本地存储的第h个关系的关系版本的个数,确定为本地存储的第h个关系的历史实例属性信息的个数。之后,计算机设备需要判断本地存储的第h个关系的历史实例属性信息的个数是否大于预设个数阈值;并在本地存储的第h个关系的历史实例属性信息的个数大于预设个数阈值时,将本地存储的第h个关系的可清理历史实例属性信息移动至外设回收站,以使得第h个关系的剩余历史实例属性信息的个数小于或等于预设个数阈值;之后就可以存储该第h个关系的实例属性信息。该外设回收站会对其中存储的数据进行定时删除,可选的,该外设回收站 可以根据计算机设备的指示定时删除数据。
可选的,剩余历史实例属性信息中的关系版本为剩余关系版本,可清理历史实例属性信息中的关系版本为可清理关系版本;剩余关系版本所指示的时间段在可清理关系版本所指示的时间段之后。例如,假设该第h个关系为上述同住关系,且剩余关系版本为关系版本2,剩余关系版本所指示的时间段为:2017年10月2日~2017年10月5日,可清理关系版本为关系版本1,可清理关系版本所指示的时间段为2017年9月2日~2017年9月5日。这样一来,能够保证计算机设备本地存储的同一关系的实例属性信息的个数总是较少,从而减轻了计算机设备的存储压力;并且,能够保证最新发生的关系能够被保存在计算机设备本地,便于后续对最新发生的关系的查询。
步骤204、在目标参数更新时对本体进行更新。
可选的,该目标参数可以包括本体模板、实例化规则以及数据集合中的一种或多种。本公开实施例中以目标参数包括本体模板、实例化规则以及数据集合为例。需要说明的是,该本体模板、实例化规则以及数据集合均可以存储在计算机设备的外设存储介质上,且本体模板、实例化规则以及数据集合均可以存储在不同的外设存储介质,本公开实施例对此不做限定。工作人员可以对外设存储介质上的本体模板进行更新。
在步骤204中,计算机设备可以检测本体模板、实例化规则以及数据集合是否更新;当本体模板、实例化规则以及数据集合中的一种或多种更新时,计算机设备可以根据当前的本体模板和当前的实例化规则对待处理数据进行实例化处理,以得到多个更新对象的实例属性信息,其中,待处理数据为当前的数据集合中当前的实例化规则所指示的待进行实例化处理的数据。之后,计算机设备可以将多个更新对象的实例属性信息覆盖已存储的实例属性信息,以更新本体。
可以看出,本公开实施例中创建的本体能够随目标参数和时间进行改变,从而使得本公开实施例中创建的本体并不是一成不变的,而是动态变化的本体,该本体可以称为时维动态本体。
步骤205、获取目标查询条件,目标查询条件包括:时间条件。
在创建得到本体后,人们就可以在计算机设备上对本体进行查询,以获取相应的实例属性信息。
示例的,当人们在计算机设备上对本体进行查询时,人们可以在计算机设 备中输入初始查询条件,使得计算机设备能够接收到初始查询条件。示例的,该初始条件可以包括:时间条件(2017年9月1日~2017年9月5日)、关系类型(同住关系)、主动方(张三)、被动方(李四)。
之后,计算机设备就需要判断初始查询条件是否包括时间条件,并在初始查询条件包括时间条件时,确定初始查询条件为目标查询条件。需要说明的是,若该初始查询条件不包括时间条件,则计算机设备可以将预设的时间条件加入初始查询条件,以得到目标查询条件。其中,该预设的时间条件可以用于记载最近的一个时间段,如某一历史时间到当前时间的时间段。比如,当前时间为2018年10月1日,则该预设的时间条件可以用于记载:2018年9月1日至2018年10月1日这个时间段。
步骤206、根据目标查询条件,查询本体中是否存在满足目标查询条件的目标实例属性信息。在本体中存在目标实例属性信息时,执行步骤207;在本体中不存在目标实例属性信息时,执行步骤208。
步骤207、获取目标实例属性信息。
示例的,假设目标查询条件包括:时间条件(2017年9月1日~2017年9月5日)、关系类型(同住关系)、主动方(张三)、被动方(李四),则可以根据该目标查询条件,在本体中查找到目标实例属性信息为:张三和李四在2017年9月1日~2017年9月5日的同住关系的实例属性信息。
步骤208、查询外设回收站中是否存在目标实例属性信息。在外设回收站中存在目标实例属性信息时,执行步骤207;在外设回收站中不存在目标实例属性信息时,执行步骤209。
当本体中不存在目标实例属性信息时,由于该目标实例属性信息有可能是被移动至外设回收站中,因此,可以在该外设回收站中继续查询是否存在目标实例属性信息。在外设回收站中存在目标实例属性信息时,可以获取该外设回收站中的目标实例属性信息。
步骤209、采用预设处理方式对数据集合中的目标数据进行处理,以得到目标实例属性信息。
其中,采用预设处理方式对数据集合中的目标数据进行处理可以包括:根据当前的本体模板和当前的实例化规则对数据集合中的目标数据进行实例化处理,以得到目标实例属性信息。
当本体和外设回收站中均不存在目标实例属性信息时,计算机设备可以获 取数据集合中的目标数据,该目标数据为数据集合中在指定时间段内生成的数据,该指定时间段为时间条件所指示的时间段。之后,计算机设备可以根据当前的本体模板(也即最新的本体模板)和当前的实例化规则(也即最新的实例化规则)对该目标数据进行处理,以得到目标实例属性信息。
综上所述,由于本公开实施例得到的实例属性信息包括时间维度的时维属性信息,因此,基于该实例属性信息构建的本体能够描述对象与时间的关系,这样一来,本体描述的对象能够与真实世界中的对象统一,提高了本体的准确度。
并且,在构建本体时,首先获取本体模板,之后再根据该本体模板得到多个对象的实例属性信息。由于本体模板能够根据属性以层级的形式对多个对象组进行关联,从而能够对真实世界中多且杂乱的对象进行了有序的多层归类分组,使得对本体需要描述的对象能够进行有序的罗列,因此,降低了本体构建过程的复杂度。
另外,相关技术中,虽然存在按照对象的类型对多个对象进行分组的概念,但相关技术中复杂多样的对象有时候并不能很好的进行分组,如某一对象有可能同时被分配到多个对象组中,这就会对本体中对象的属性信息的准确度产生影响。而本公开实施例中,根据属性以层级的形式对多个对象组进行关联,从而能够对真实世界中多且杂乱的对象进行了有序的多层归类分组,使得本体中对象的属性信息的准确度较高。
相关技术中的分组是基于对象的类型,所以这种分组的方法的粒度级别是类型,而本公开实施例中分组的方法的粒度级别为属性。因此,本公开实施例实现了从按类型分组到按属性分组的转变,实现了从粗粒度级别的分组到细粒度级别的分组的转变,细化了对象的分组。
图6为本公开实施例提供的一种本体构建装置的结构示意图,如图6所示,该本体构建装置可以包括:
第一处理模块601,用于对指定的数据集合进行实例化处理,得到多个对象中每个对象的实例属性信息;
存储模块602,用于按照指定的存储结构,存储多个对象的实例属性信息,以构建本体;
其中,每个对象的实例属性信息包括:每个对象在时间维度上的时维属性 信息。
综上所述,由于本公开实施例中第一处理模块得到的实例属性信息包括时间维度的时维属性信息,因此,存储模块基于该实例属性信息构建的本体能够描述对象与时间的关系,这样一来,本体描述的对象能够与真实世界中的对象统一,提高了本体的准确度。
可选的,多个对象包括实体,实体的时维属性信息包括以下一种或多种:
用于指示实体是否消亡的存续标识;
实体的诞生时间和/或实体的消亡时间的指示信息;
实体的存续事件标识,存续事件标识用于指示实体的时维属性信息发生变化的触发事件。
可选的,存储模块602用于:
对于多个对象中第i个实体的实例属性信息,检测本地是否存储有第i个实体的历史实例属性信息,1≤i≤k,k为多个对象中实体的个数;
在本地存储有与第i个实体的历史实例属性信息时,采用第i个实体的实例属性信息覆盖其历史实例属性信息。
可选的,多个对象包括实体间的关系,该关系的时维属性信息包括以下一种或多种:
关系的版本标识,其中,每个关系具有至少一个版本标识,每个版本标识唯一对应关系所维持的一个时间段;
关系的起始时间和/或关系的结束时间的指示信息;
关系的来源事件标识,来源事件标识用于指示与关系的每个版本标识相关的事件。
可选的,存储模块用于:
对于多个对象中第h个关系的实例属性信息,确定本地存储的第h个关系的历史实例属性信息的个数,1≤h≤p,p为多个对象中关系的个数;
判断本地存储的第h个关系的历史实例属性信息的个数是否大于预设个数阈值;
在本地存储的第h个关系的历史实例属性信息的个数大于预设个数阈值时,将本地存储的第h个关系的可清理历史实例属性信息移动至外设回收站,以使得第h个关系的剩余历史实例属性信息的个数小于或等于预设个数阈值;其中,第h个关系的历史实例属性信息包括:第h个关系的可清理历史实例属性信息 和第h个关系的剩余历史实例属性信息;
存储第h个关系的实例属性信息。
可选的,剩余历史实例属性信息中的关系版本为剩余关系版本,可清理历史实例属性信息中的关系版本为可清理关系版本;
剩余关系版本所对应的时间段在可清理关系版本所对应的时间段之后。
可选的,图7为本公开实施例提供的另一种本体构建装置的结构示意图,如图7所示,在图6的基础上,该本体构建装置还包括:
第一获取模块603,用于获取本体模板和实例化规则,本体模板用于记载多个对象组的属性信息,每个对象组包括属于同一类型的对象,多个对象组以层级的形式关联;
第一处理模块601用于:根据本体模板和实例化规则,对数据集合进行实例化处理;其中,多个对象中每个对象的实例属性信息为基于本体模板所记载的属性信息,在数据集合中提取的信息。
检测模块604,用于分别检测本体模板、实例化规则以及数据集合是否更新;
第二处理模块605,用于在本体模板、实例化规则以及数据集合中的一个或多个更新时,根据当前的本体模板和当前的实例化规则对待处理数据进行实例化处理,以得到多个更新对象的实例属性信息,其中,待处理数据为当前的数据集合中当前的实例化规则所指示的待进行实例化处理的数据;
覆盖模块606,用于将多个更新对象的实例属性信息覆盖已存储的实例属性信息,以更新本体。
第二获取模块607,用于获取目标查询条件,目标查询条件包括:时间条件;
第一查询模块608,用于根据目标查询条件,查询本体中是否存在满足目标查询条件的目标实例属性信息;
第三获取模块609,用于在本体中存在目标实例属性信息时,获取目标实例属性信息。
第二查询模块610,用于在本体不包括目标实例属性信息时,查询外设回收站中是否存在目标实例属性信息;
第三获取模块609还用于在外设回收站中存在目标实例属性信息时,执行获取目标实例属性信息的步骤;
第三处理模块611,用于在外设回收站中不存在目标实例属性信息时,根据当前的本体模板和当前的实例化规则对所述数据集合中的目标数据进行实例化 处理,以得到目标实例属性信息;
其中,时间条件用于指示指定时间段,目标数据为数据集合中在指定时间段内生成的数据。
第二获取模块608用于:接收初始查询条件;判断初始查询条件是否包括时间条件;在初始查询条件包括时间条件时,确定初始查询条件为目标查询条件;在初始查询条件不包括时间条件时,将预设的时间条件加入初始查询条件,以得到目标查询条件。
综上所述,由于本公开实施例中第一处理模块得到的实例属性信息包括时间维度的时维属性信息,因此,存储模块基于该实例属性信息构建的本体能够描述对象与时间的关系,这样一来,本体描述的对象能够与真实世界中的对象统一,提高了本体的准确度。
本公开实施例还提供了一种计算机可读存储介质,存储介质内存储有计算机程序,计算机程序被处理器执行时实现上述本体构建方法。
本公开实施例还提供了一种包含指令的计算机程序产品,当该计算机程序产品在计算机上运行时,使得计算机执行上述本体构建方法。
需要说明的是,本公开实施例提供的本体构建方法实施例与本体构建装置实施例可以互相参考,本公开实施例对此不做限定。本发明实施例提供的方法实施例步骤的先后顺序能够进行适当调整,步骤也能够根据情况进行相应增减,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,可轻易想到变化的方法,都应涵盖在本发明的保护范围之内,因此不再赘述。
在本公开中,术语“第一”、“第二”、“第三”和“第四”仅用于描述目的,而不能理解为指示或暗示相对重要性。术语“多个”指两个或两个以上,除非另有明确的限定。
以上所述仅为本公开的可选实施例,并不用以限制本公开,凡在本公开的精神和原则之内,所作的任何修改、等同替换、改进等,均应包含在本公开的保护范围之内。

Claims (24)

  1. 一种本体构建方法,其特征在于,所述方法包括:
    对指定的数据集合进行实例化处理,得到多个对象中每个对象的实例属性信息;其中,所述每个对象的实例属性信息包括:所述每个对象在时间维度上的时维属性信息;
    按照指定的存储结构,存储所述多个对象的实例属性信息,以构建所述本体。
  2. 根据权利要求1所述的方法,其特征在于,所述多个对象包括实体,所述实体的时维属性信息包括以下一种或多种:
    用于记载所述实体是否消亡的存续标识;
    所述实体的诞生时间的指示信息;
    所述实体的消亡时间的指示信息;
    所述实体的存续事件标识,所述存续事件标识用于记载所述实体的时维属性信息发生变化的触发事件。
  3. 根据权利要求2所述的方法,其特征在于,所述存储所述多个对象的实例属性信息,包括:
    对于所述多个对象中第i个实体的实例属性信息,检测本地是否存储有所述第i个实体的历史实例属性信息,1≤i≤k,k为所述多个对象中实体的个数;
    在本地存储有与所述第i个实体的历史实例属性信息时,采用所述第i个实体的实例属性信息覆盖其历史实例属性信息。
  4. 根据权利要求1至3任一所述的方法,其特征在于,所述多个对象包括实体间的关系,所述关系的时维属性信息包括以下一种或多种:
    所述关系的版本标识,其中,每个所述关系具有至少一个版本标识,每个所述版本标识唯一对应所述关系所维持的一个时间段;
    所述关系的起始时间的指示信息;
    所述关系的结束时间的指示信息;
    所述关系的来源事件标识,所述来源事件标识用于记载与所述关系的每个 版本标识相关的事件。
  5. 根据权利要求4所述的方法,其特征在于,所述存储所述多个对象的实例属性信息,包括:
    对于所述多个对象中第h个关系的实例属性信息,确定本地存储的所述第h个关系的历史实例属性信息的个数,1≤h≤p,p为所述多个对象中关系的个数;
    判断本地存储的所述第h个关系的历史实例属性信息的个数是否大于预设个数阈值;
    在本地存储的所述第h个关系的历史实例属性信息的个数大于所述预设个数阈值时,将本地存储的所述第h个关系的可清理历史实例属性信息移动至外设回收站,以使得所述第h个关系的剩余历史实例属性信息的个数小于或等于所述预设个数阈值;其中,所述第h个关系的历史实例属性信息包括:所述第h个关系的可清理历史实例属性信息和所述第h个关系的剩余历史实例属性信息;
    存储所述第h个关系的实例属性信息。
  6. 根据权利要求5所述的方法,其特征在于,
    所述剩余历史实例属性信息中的关系版本为剩余关系版本,所述可清理历史实例属性信息中的关系版本为可清理关系版本;
    所述剩余关系版本所对应的时间段在所述可清理关系版本所对应的时间段之后。
  7. 根据权利要求1至6任一所述的方法,其特征在于,
    在所述对指定的数据集合进行实例化处理之前,所述方法还包括:获取本体模板和实例化规则,其中,所述本体模板用于记载多个对象组的属性信息,每个对象组包括属于同一类型的对象,所述多个对象组以层级的形式关联;
    所述对指定的数据集合进行实例化处理,包括:根据所述本体模板和所述实例化规则,对所述数据集合进行所述实例化处理;
    其中,所述多个对象中每个对象的实例属性信息为基于所述本体模板所记载的属性信息,在所述数据集合中提取的信息。
  8. 根据权利要求7所述的方法,其特征在于,在存储所述多个对象的实例属性信息之后,所述方法还包括:
    分别检测所述本体模板、所述实例化规则以及所述数据集合是否更新;
    当所述本体模板、所述实例化规则以及所述数据集合中的一个或多个更新时,根据当前的本体模板和当前的实例化规则对待处理数据进行实例化处理,以得到多个更新对象的实例属性信息,其中,所述待处理数据为当前的数据集合中所述当前的实例化规则所指示的待进行实例化处理的数据;
    将所述多个更新对象的实例属性信息覆盖已存储的实例属性信息,以更新所述本体。
  9. 根据权利要求1至8任一所述的方法,其特征在于,在基于所述多个对象的实例属性信息构建所述本体之后,所述方法还包括:
    获取目标查询条件,所述目标查询条件包括:时间条件;
    根据所述目标查询条件,查询所述本体中是否存在满足所述目标查询条件的目标实例属性信息;
    在所述本体中存在所述目标实例属性信息时,获取所述目标实例属性信息。
  10. 根据权利要求9所述的方法,其特征在于,所述获取目标查询条件,包括:
    接收初始查询条件;
    判断所述初始查询条件是否包括所述时间条件;
    在所述初始查询条件包括所述时间条件时,确定所述初始查询条件为所述目标查询条件;
    在所述初始查询条件不包括所述时间条件时,将预设的时间条件加入所述初始查询条件,以得到所述目标查询条件。
  11. 根据权利要求9所述的方法,其特征在于,所述方法还包括:
    在所述本体不包括所述目标实例属性信息时,查询外设回收站中是否存在所述目标实例属性信息;
    在所述外设回收站中存在所述目标实例属性信息时,执行获取所述目标实 例属性信息的步骤;
    在所述外设回收站中不存在所述目标实例属性信息时,根据当前的本体模板和当前的实例化规则对所述数据集合中的目标数据进行实例化处理,以得到所述目标实例属性信息;
    其中,所述时间条件用于记载指定时间段,所述目标数据为所述数据集合中在所述指定时间段内生成的数据。
  12. 一种本体构建装置,其特征在于,所述本体构建装置包括:
    第一处理模块,用于对指定的数据集合进行实例化处理,得到多个对象中每个对象的实例属性信息;其中,所述每个对象的实例属性信息包括:所述每个对象在时间维度上的时维属性信息;
    存储模块,用于按照指定的存储结构,存储所述多个对象的实例属性信息,以构建所述本体。
  13. 根据权利要求12所述的本体构建装置,其特征在于,所述多个对象包括实体,所述实体的时维属性信息包括以下一种或多种:
    用于记载所述实体是否消亡的存续标识;
    所述实体的诞生时间的指示信息;
    所述实体的消亡时间的指示信息;
    所述实体的存续事件标识,所述存续事件标识用于记载所述实体的时维属性信息发生变化的触发事件。
  14. 根据权利要求13所述的本体构建装置,其特征在于,所述存储模块用于:
    对于所述多个对象中第i个实体的实例属性信息,检测本地是否存储有所述第i个实体的历史实例属性信息,1≤i≤k,k为所述多个对象中实体的个数;
    在本地存储有与所述第i个实体的历史实例属性信息时,采用所述第i个实体的实例属性信息覆盖其历史实例属性信息。
  15. 根据权利要求12至14任一所述的本体构建装置,其特征在于,所述 多个对象包括实体间的关系,所述关系的时维属性信息包括以下一种或多种:
    所述关系的版本标识,其中,每个所述关系具有至少一个版本标识,每个所述版本标识唯一对应所述关系所维持的一个时间段;
    所述关系的起始时间的指示信息;
    所述关系的结束时间的指示信息;
    所述关系的来源事件标识,所述来源事件标识用于记载与所述关系的每个版本标识相关的事件。
  16. 根据权利要求15所述的本体构建装置,其特征在于,所述存储模块用于:
    对于所述多个对象中第h个关系的实例属性信息,确定本地存储的所述第h个关系的历史实例属性信息的个数,1≤h≤p,p为所述多个对象中关系的个数;
    判断本地存储的所述第h个关系的历史实例属性信息的个数是否大于预设个数阈值;
    在本地存储的所述第h个关系的历史实例属性信息的个数大于所述预设个数阈值时,将本地存储的所述第h个关系的可清理历史实例属性信息移动至外设回收站,以使得所述第h个关系的剩余历史实例属性信息的个数小于或等于所述预设个数阈值;其中,所述第h个关系的历史实例属性信息包括:所述第h个关系的可清理历史实例属性信息和所述第h个关系的剩余历史实例属性信息;
    存储所述第h个关系的实例属性信息。
  17. 根据权利要求16所述的本体构建装置,其特征在于,
    所述剩余历史实例属性信息中的关系版本为剩余关系版本,所述可清理历史实例属性信息中的关系版本为可清理关系版本;
    所述剩余关系版本所对应的时间段在所述可清理关系版本所对应的时间段之后。
  18. 根据权利要求12至17任一所述的本体构建装置,其特征在于,
    所述本体构建装置还包括:
    第一获取模块,用于获取本体模板和实例化规则,其中,所述本体模板用 于记载多个对象组的属性信息,每个对象组包括属于同一类型的对象,所述多个对象组以层级的形式关联;
    所述第一处理模块用于:根据所述本体模板和所述实例化规则,对所述数据集合进行所述实例化处理;
    其中,所述多个对象中每个对象的实例属性信息为基于所述本体模板所记载的属性信息,在所述数据集合中提取的信息。
  19. 根据权利要求18所述的本体构建装置,其特征在于,所述本体构建装置还包括:
    检测模块,用于分别检测所述本体模板、所述实例化规则以及所述数据集合是否更新;
    第二处理模块,用于在所述本体模板、所述实例化规则以及所述数据集合中的一个或多个更新时,根据当前的本体模板和当前的实例化规则对待处理数据进行实例化处理,以得到多个更新对象的实例属性信息,其中,所述待处理数据为当前的数据集合中所述当前的实例化规则所指示的待进行实例化处理的数据;
    覆盖模块,用于将所述多个更新对象的实例属性信息覆盖已存储的实例属性信息,以更新所述本体。
  20. 根据权利要求12至19任一所述的本体构建装置,其特征在于,所述本体构建装置还包括:
    第二获取模块,用于获取目标查询条件,所述目标查询条件包括:时间条件;
    第一查询模块,用于根据所述目标查询条件,查询所述本体中是否存在满足所述目标查询条件的目标实例属性信息;
    第三获取模块,用于在所述本体中存在所述目标实例属性信息时,获取所述目标实例属性信息。
  21. 根据权利要求20所述的本体构建装置,其特征在于,所述第二获取模块用于:
    接收初始查询条件;
    判断所述初始查询条件是否包括所述时间条件;
    在所述初始查询条件包括所述时间条件时,确定所述初始查询条件为所述目标查询条件;
    在所述初始查询条件不包括所述时间条件时,将预设的时间条件加入所述初始查询条件,以得到所述目标查询条件。
  22. 根据权利要求20所述的本体构建装置,其特征在于,所述本体构建装置还包括:
    第二查询模块,用于在所述本体不包括所述目标实例属性信息时,查询外设回收站中是否存在所述目标实例属性信息;
    所述第三获取模块还用于在所述外设回收站中存在所述目标实例属性信息时,执行获取所述目标实例属性信息的步骤;
    第三处理模块,用于在所述外设回收站中不存在所述目标实例属性信息时,根据当前的本体模板和当前的实例化规则对所述数据集合中的目标数据进行实例化处理,以得到所述目标实例属性信息;
    其中,所述时间条件用于记载指定时间段,所述目标数据为所述数据集合中在所述指定时间段内生成的数据。
  23. 一种计算机设备,其特征在于,包括处理器、通信接口、存储器和通信总线,其中,处理器,通信接口,存储器通过总线完成相互间的通信;存储器,用于存放计算机程序;处理器,用于执行存储器上所存放的程序,实现权利要求1至11任一所述的方法。
  24. 一种计算机可读存储介质,其特征在于,所述存储介质内存储有计算机程序,所述计算机程序被处理器执行时实现权利要求1至11任一所述的方法。
PCT/CN2019/093683 2018-06-29 2019-06-28 本体构建方法及装置 Ceased WO2020001616A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810701221.X 2018-06-29
CN201810701221.XA CN109033208B (zh) 2018-06-29 2018-06-29 本体构建方法及装置

Publications (1)

Publication Number Publication Date
WO2020001616A1 true WO2020001616A1 (zh) 2020-01-02

Family

ID=65521010

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/093683 Ceased WO2020001616A1 (zh) 2018-06-29 2019-06-28 本体构建方法及装置

Country Status (2)

Country Link
CN (1) CN109033208B (zh)
WO (1) WO2020001616A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109033208B (zh) * 2018-06-29 2021-07-02 杭州海康威视数字技术股份有限公司 本体构建方法及装置
CN111598116B (zh) * 2019-02-21 2024-01-23 杭州海康威视数字技术股份有限公司 数据分类方法、装置、电子设备及可读存储介质
CN116204554B (zh) * 2023-05-04 2023-08-08 浙江康勒工业软件有限公司 数据处理方法、系统、电子设备和存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1866250A (zh) * 2005-10-12 2006-11-22 华为技术有限公司 对系统数据进行管理的方法和系统
CN107844600A (zh) * 2017-11-23 2018-03-27 浪潮软件集团有限公司 一种数据存储方法和装置
CN108062379A (zh) * 2017-12-13 2018-05-22 北京天广汇通科技有限公司 数据处理方法、平台、装置和计算机可读存储介质
CN108205593A (zh) * 2017-12-29 2018-06-26 华为技术有限公司 一种查询的方法及装置
CN109033208A (zh) * 2018-06-29 2018-12-18 杭州海康威视数字技术股份有限公司 本体构建方法及装置

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8059635B1 (en) * 2006-05-05 2011-11-15 Workday, Inc. Non-destructive data storage
CN105528204A (zh) * 2015-01-27 2016-04-27 北京理工大学 领域本体的构建方法和装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1866250A (zh) * 2005-10-12 2006-11-22 华为技术有限公司 对系统数据进行管理的方法和系统
CN107844600A (zh) * 2017-11-23 2018-03-27 浪潮软件集团有限公司 一种数据存储方法和装置
CN108062379A (zh) * 2017-12-13 2018-05-22 北京天广汇通科技有限公司 数据处理方法、平台、装置和计算机可读存储介质
CN108205593A (zh) * 2017-12-29 2018-06-26 华为技术有限公司 一种查询的方法及装置
CN109033208A (zh) * 2018-06-29 2018-12-18 杭州海康威视数字技术股份有限公司 本体构建方法及装置

Also Published As

Publication number Publication date
CN109033208A (zh) 2018-12-18
CN109033208B (zh) 2021-07-02

Similar Documents

Publication Publication Date Title
US11514054B1 (en) Supervised graph partitioning for record matching
US20190370587A1 (en) Attention-based explanations for artificial intelligence behavior
JP2018200685A (ja) 完全教師あり学習用のデータセットの形成
CN108874783A (zh) 电力信息运维知识模型构建方法
WO2020001616A1 (zh) 本体构建方法及装置
Gang et al. Recognition of honeycomb lung in CT images based on improved MobileNet model
CN116935188B (zh) 模型训练方法、图像识别方法、装置、设备及介质
WO2021184776A1 (zh) 图像识别方法、装置、计算机设备和存储介质
CN113779358A (zh) 一种事件检测方法和系统
WO2020001614A1 (zh) 本体构建方法及装置
CN110109908B (zh) 基于社会基础信息挖掘人物潜在关系的分析系统及方法
CN108241867A (zh) 一种分类方法及装置
CN114298679A (zh) 基于大数据的工业ai平台及其应用方法、介质及电子设备
Gao Design and implementation of 3D animation data processing development platform based on artificial intelligence
Myers-Dean et al. Spin: Hierarchical segmentation with subpart granularity in natural images
CN120747670B (zh) 基于视觉语言模型的图像解释方法和装置
US20250371379A1 (en) Method of managing personal knowledge graph and user device using the method
EP3306540A1 (en) System and method for content affinity analytics
CN116957036A (zh) 伪造多媒体检测模型的训练方法、装置和计算设备
CN120353802A (zh) 一种基于分层知识图谱的湿地语义虚拟地理环境构建方法
Yuan et al. A novel user scenario and behavior sequence recognition approach based on vision-context fusion architecture
CN111797856B (zh) 建模方法、装置、存储介质及电子设备
Smarandache et al. An Operational View in Computational Construction of Information
CN112434722B (zh) 基于类别相似度的标签平滑计算的方法、装置、电子设备及介质
CN116977770A (zh) 图像分类模型的构建方法、图像分类方法、装置和设备

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: 19827525

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: 19827525

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 19827525

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 30/06/2021)