CN115618948A - Knowledge information analysis method and system based on pervasive computing - Google Patents

Abstract

Translated fromChinese

本发明实施例提供了一种基于普适计算的知识信息分析方法及系统，对于线上行为抽取知识信息，依据一次遍历循环，通过基于惩罚项的特征提取模型的模型单元中的特征提取单元依据线上行为抽取知识信息中的第一行为实体、第二行为实体以及实体关系三个普适计算目标中的两个普适计算目标的行为抽取特征生成第三个普适计算目标对应的至少一个候选行为抽取特征，以确定一个候选行为抽取特征，根据获得的线上行为知识网络中各行为实体或实体关系的目标行为抽取特征对线上行为抽取知识信息进行普适计算。如此，能够有效对线上行为抽取知识信息进行普适计算，并且突出不同知识层级的实体关系。

The embodiment of the present invention provides a knowledge information analysis method and system based on pervasive computing. For online behavior extraction knowledge information, according to a traversal cycle, the feature extraction unit in the model unit of the feature extraction model based on penalty items is based on The first behavior entity, the second behavior entity, and the entity relationship in the online behavior extraction knowledge information generate at least one of the behavior extraction features corresponding to the third ubiquitous computing target in the three pervasive computing targets Candidate behavior extraction features to determine a candidate behavior extraction feature, and perform pervasive calculation of online behavior extraction knowledge information according to the obtained target behavior extraction features of each behavior entity or entity relationship in the online behavior knowledge network. In this way, it is possible to effectively perform ubiquitous computing on online behavior extraction knowledge information, and highlight entity relationships at different knowledge levels.

Description

Translated fromChinese

基于普适计算的知识信息分析方法及系统Knowledge information analysis method and system based on pervasive computing

技术领域technical field

本发明涉及计算机技术领域，具体而言，涉及一种基于普适计算的知识信息分析方法及系统。The present invention relates to the field of computer technology, in particular to a method and system for analyzing knowledge information based on pervasive computing.

背景技术Background technique

在普适计算过程中，如何有效对线上行为抽取知识信息进行普适计算，并且突出不同知识层级的实体关系，是本领域亟待解决的技术问题。In the process of ubiquitous computing, how to effectively perform ubiquitous computing on knowledge information extracted from online behaviors and highlight the entity relationship of different knowledge levels is a technical problem to be solved urgently in this field.

发明内容Contents of the invention

有鉴于此，本发明实施例的目的在于提供一种基于普适计算的知识信息分析方法及系统，能够有效对线上行为抽取知识信息进行普适计算，并且突出不同知识层级的实体关系。In view of this, the purpose of the embodiments of the present invention is to provide a pervasive computing-based knowledge information analysis method and system, which can effectively perform pervasive computing on online behavior extraction knowledge information, and highlight entity relationships at different knowledge levels.

根据本发明实施例的一个方面，提供获取一组线上行为抽取知识信息，所述线上行为抽取知识信息包括线上行为知识网络中相互关联的第一行为实体以及第二行为实体的各自的行为抽取特征，以及所述第一行为实体与所述第二行为实体之间的实体关系的行为抽取特征；According to an aspect of an embodiment of the present invention, it is provided to obtain a set of online behavior extraction knowledge information, the online behavior extraction knowledge information includes the first behavior entity and the second behavior entity related to each other in the online behavior knowledge network behavior extraction features, and behavior extraction features of the entity relationship between the first behavior entity and the second behavior entity;

对于所述线上行为抽取知识信息，依据一次遍历循环，通过基于惩罚项的特征提取模型的模型单元中的特征提取单元依据线上行为抽取知识信息中的第一行为实体、第二行为实体以及实体关系三个普适计算目标中的两个普适计算目标的行为抽取特征生成第三个普适计算目标对应的至少一个候选行为抽取特征；For the online behavior extraction knowledge information, according to a traversal cycle, the feature extraction unit in the model unit of the feature extraction model based on the penalty item extracts the first behavior entity, the second behavior entity and the The behavior extraction features of two ubiquitous computing objects among the three ubiquitous computing objects generate at least one candidate behavior extraction feature corresponding to the third ubiquitous computing object;

从所述至少一个候选行为抽取特征中确定一个候选行为抽取特征，作为第三个普适计算目标经过本轮遍历循环后其在线上行为知识网络中对应的行为抽取特征；Determining a candidate behavior extraction feature from the at least one candidate behavior extraction feature as the corresponding behavior extraction feature in the online behavior knowledge network after the third ubiquitous computing target passes through the current round of traversal cycle;

更新所述基于惩罚项的特征提取模型的模型单元中的所述特征提取单元以及特征解码单元；依此，进行多次遍历循环直到满足停止遍历循环条件，获得所述线上行为知识网络中各行为实体或实体关系的目标行为抽取特征，并根据所述目标行为抽取特征对所述线上行为抽取知识信息进行普适计算。Updating the feature extraction unit and the feature decoding unit in the model unit of the feature extraction model based on the penalty term; accordingly, performing multiple traversal loops until the condition for stopping the traversal loop is met, and obtaining each feature in the online behavior knowledge network. The target behavior extraction feature of the behavioral entity or entity relationship, and pervasive calculation is performed on the online behavior extraction knowledge information according to the target behavior extraction feature.

可选地，一次遍历循环过程中，所述更新所述基于惩罚项的特征提取模型的模型单元中的所述特征提取单元以及特征解码单元，包括：Optionally, during one traversal cycle, the updating of the feature extraction unit and the feature decoding unit in the model unit of the penalty-based feature extraction model includes:

依据所述两个普适计算目标的行为抽取特征以及第三个普适计算目标对应的至少一个候选行为抽取特征生成至少一个普适筛选特征，以及将所述线上行为抽取知识信息作为普适编码特征；Generate at least one pervasive screening feature based on the behavior extraction features of the two pervasive computing targets and at least one candidate behavior extraction feature corresponding to the third pervasive computing target, and use the online behavior extraction knowledge information as pervasive coded features;

依据所述特征解码单元，对所述至少一个普适筛选特征以及所述普适编码特征进行有效性验证以获取有效性验证结果；Perform validity verification on the at least one universal screening feature and the universal coding feature according to the feature decoding unit to obtain a validity verification result;

依据所述有效性验证结果，更新所述特征提取单元；updating the feature extraction unit according to the validity verification result;

依据所述有效性验证结果与各特征解码单元的真实属性更新所述特征解码单元。The feature decoding unit is updated according to the validity verification result and the real attribute of each feature decoding unit.

可选地，所述通过基于惩罚项的特征提取模型的模型单元中的特征提取单元依据线上行为抽取知识信息中的第一行为实体、第二行为实体以及实体关系三个普适计算目标中的两个普适计算目标的行为抽取特征生成第三个普适计算目标对应的至少一个候选行为抽取特征，包括：Optionally, the feature extraction unit in the model unit of the feature extraction model based on the penalty item extracts the first behavior entity, the second behavior entity, and the entity relationship in the knowledge information according to the online behavior in the three ubiquitous computing targets The behavior extraction features of the two ubiquitous computing targets generate at least one candidate behavior extraction feature corresponding to the third ubiquitous computing target, including:

获取至少一个随机行为抽取特征；Obtain at least one random behavior extraction feature;

通过所述特征提取单元依据所述两个普适计算目标的行为抽取特征以及所述至少一个随机行为抽取特征，获取对应于所述第三普适计算目标的至少一个候选行为抽取特征。At least one candidate behavior extraction feature corresponding to the third ubiquitous computing object is obtained by the feature extraction unit according to the behavior extraction features of the two ubiquitous computing objects and the at least one random behavior extraction feature.

可选地，所述通过所述特征提取单元依据所述两个普适计算目标的行为抽取特征以及所述至少一个随机行为抽取特征，获取对应于所述第三普适计算目标的至少一个候选行为抽取特征，包括：Optionally, the feature extraction unit acquires at least one candidate corresponding to the third ubiquitous computing target according to the behavior extraction features of the two pervasive computing targets and the at least one random behavior extraction feature. Behavior extraction features, including:

依据所述两个普适计算目标的行为抽取特征以及至少一个随机行为抽取特征，依据所述特征提取单元，获取对应于所述第三普适计算目标的至少一个候选特征；Acquiring at least one candidate feature corresponding to the third ubiquitous computing object according to the feature extraction unit according to the behavior extraction features of the two ubiquitous computing objects and at least one random behavior extraction feature;

对于每一个候选特征，依据所述线上行为知识网络，确定所述候选行为抽取特征。For each candidate feature, the candidate behavior extraction feature is determined according to the online behavior knowledge network.

可选地，所述对于每一个候选特征，依据所述线上行为知识网络，确定所述候选行为抽取特征，包括：Optionally, for each candidate feature, determining the candidate behavior extraction feature according to the online behavior knowledge network includes:

确定所述线上行为知识网络上与所述候选特征相似度满足预设条件的与第三普适计算目标同质的行为实体或实体关系的行为抽取特征；Determining the behavior extraction feature of the behavior entity or entity relationship that is homogeneous to the third ubiquitous computing target on the online behavior knowledge network and whose similarity with the candidate feature satisfies a preset condition;

将满足预设条件的与第三普适计算目标同质的行为实体或实体关系的行为抽取特征确定为候选行为抽取特征。The behavior extraction features of behavioral entities or entity relationships that meet the preset conditions and are homogeneous with the third ubiquitous computing target are determined as candidate behavior extraction features.

根据本发明实施例的另一方面，提供一种基于普适计算的知识信息分析系统，应用于服务器，所述系统包括：According to another aspect of the embodiments of the present invention, a knowledge information analysis system based on pervasive computing is provided, which is applied to a server, and the system includes:

获取模块，用于获取一组线上行为抽取知识信息，所述线上行为抽取知识信息包括线上行为知识网络中相互关联的第一行为实体以及第二行为实体的各自的行为抽取特征，以及所述第一行为实体与所述第二行为实体之间的实体关系的行为抽取特征；An acquisition module, configured to acquire a set of online behavior extraction knowledge information, the online behavior extraction knowledge information including the respective behavior extraction features of the first behavior entity and the second behavior entity that are interrelated in the online behavior knowledge network, and behavior extraction features of the entity relationship between the first behavior entity and the second behavior entity;

生成模块，用于对于所述线上行为抽取知识信息，依据一次遍历循环，通过基于惩罚项的特征提取模型的模型单元中的特征提取单元依据线上行为抽取知识信息中的第一行为实体、第二行为实体以及实体关系三个普适计算目标中的两个普适计算目标的行为抽取特征生成第三个普适计算目标对应的至少一个候选行为抽取特征；The generating module is used to extract knowledge information for the online behavior, and extract the first behavior entity, Generate at least one candidate behavior extraction feature corresponding to the third ubiquitous computing target from the behavior extraction features of two pervasive computing targets among the second behavior entity and entity relationship;

确定模块，用于从所述至少一个候选行为抽取特征中确定一个候选行为抽取特征，作为第三个普适计算目标经过本轮遍历循环后其在线上行为知识网络中对应的行为抽取特征；A determining module, configured to determine a candidate behavior extraction feature from the at least one candidate behavior extraction feature as the corresponding behavior extraction feature in the online behavior knowledge network of the third ubiquitous computing target after the current round of traversal cycle;

更新模块，用于更新所述基于惩罚项的特征提取模型的模型单元中的所述特征提取单元以及特征解码单元；依此，进行多次遍历循环直到满足停止遍历循环条件，获得所述线上行为知识网络中各行为实体或实体关系的目标行为抽取特征，并根据所述目标行为抽取特征对所述线上行为抽取知识信息进行普适计算。The update module is used to update the feature extraction unit and the feature decoding unit in the model unit of the feature extraction model based on the penalty item; accordingly, perform multiple traversal loops until the condition of stopping the traversal loop is satisfied, and obtain the online The target behavior extraction features of each behavior entity or entity relationship in the behavior knowledge network, and perform pervasive calculation on the online behavior extraction knowledge information according to the target behavior extraction features.

相较于现有技术而言，本发明实施例提供的基于普适计算的知识信息分析方法及系统，对于线上行为抽取知识信息，依据一次遍历循环，通过基于惩罚项的特征提取模型的模型单元中的特征提取单元依据线上行为抽取知识信息中的第一行为实体、第二行为实体以及实体关系三个普适计算目标中的两个普适计算目标的行为抽取特征生成第三个普适计算目标对应的至少一个候选行为抽取特征，以确定一个候选行为抽取特征，作为第三个普适计算目标经过本轮遍历循环后其在线上行为知识网络中对应的行为抽取特征，更新基于惩罚项的特征提取模型的模型单元中的特征提取单元以及特征解码单元，然后根据获得的线上行为知识网络中各行为实体或实体关系的目标行为抽取特征对线上行为抽取知识信息进行普适计算。如此，能够有效对线上行为抽取知识信息进行普适计算，并且突出不同知识层级的实体关系。Compared with the prior art, the ubiquitous computing-based knowledge information analysis method and system provided by the embodiments of the present invention, for online behavior extraction knowledge information, according to a traversal cycle, through the penalty item-based feature extraction model model The feature extraction unit in the unit generates the third ubiquitous computing target according to the behavior extraction features of the first behavior entity, the second behavior entity and the entity relationship in the online behavior extraction knowledge information of the two ubiquitous computing targets. At least one candidate behavior extraction feature corresponding to the adaptive computing target to determine a candidate behavior extraction feature, as the third ubiquitous computing target after the current round of traversal cycle, its corresponding behavior extraction feature in the online behavior knowledge network, update based on penalty The feature extraction unit and feature decoding unit in the model unit of the feature extraction model of the item, and then perform pervasive calculation of the online behavior extraction knowledge information according to the obtained target behavior extraction features of each behavior entity or entity relationship in the online behavior knowledge network . In this way, it is possible to effectively perform ubiquitous computing on online behavior extraction knowledge information, and highlight entity relationships at different knowledge levels.

为使本发明实施例的上述目的、特征和优点能更明显易懂，下面将结合实施例，并配合所附附图，作详细说明。In order to make the above objects, features and advantages of the embodiments of the present invention more comprehensible, the following will describe in detail with reference to the embodiments and accompanying drawings.

附图说明Description of drawings

为了更清楚地说明本发明实施例的技术方案，下面将对实施例中所需要使用的附图作简单地介绍，应当理解，以下附图仅示出了本发明的某些实施例，因此不应被看作是对范围的限定，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相关的附图。In order to illustrate the technical solutions of the embodiments of the present invention more clearly, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention, and thus It should be regarded as a limitation on the scope, and those skilled in the art can also obtain other related drawings based on these drawings without creative work.

图1示出了本发明实施例所提供的服务器的组件示意图；FIG. 1 shows a schematic diagram of components of a server provided by an embodiment of the present invention;

图2示出了本发明实施例所提供的基于普适计算的知识信息分析方法的流程示意图；FIG. 2 shows a schematic flowchart of a method for analyzing knowledge information based on pervasive computing provided by an embodiment of the present invention;

图3示出了本发明实施例所提供的基于普适计算的知识信息分析系统的功能模块框图。FIG. 3 shows a block diagram of functional modules of a knowledge information analysis system based on pervasive computing provided by an embodiment of the present invention.

具体实施方式detailed description

为了使本技术领域的学员更好地理解本发明方案，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，所描述的实施例仅是本发明一部分实施例，而不是全部的实施例。根据本发明中的实施例，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他实施例，都属于本发明保护的范围。In order to enable students in the technical field to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only It is a part of embodiments of the present invention, but not all embodiments. According to the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

本发明的说明书和权利要求书及上述附图中的术语“第一”、“第二”、“第三”等(如果存在)是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。应该理解这样使用的科技项目对象在适当情况下可以互换，以便这里描述的本发明的实施例例如能够以除了在这里图示或描述的那些以外的顺序实施。此外，术语“包括”和“具有”以及他们的任何变形，意图在于覆盖不排他的包含，例如，包含了一系列步骤或单元的过程、方法、系统、产品或设备不必限于清楚地列出的那些步骤或单元，而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备固有的其它步骤或单元。The terms "first", "second", "third", etc. (if any) in the description and claims of the present invention and the above drawings are used to distinguish similar objects and not necessarily to describe a specific order or sequentially. It is to be understood that the terms so used are interchangeable under appropriate circumstances such that the embodiments of the invention described herein are, for example, capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprising" and "having", as well as any variations thereof, are intended to cover a non-exclusive inclusion, for example, a process, method, system, product or device comprising a sequence of steps or elements is not necessarily limited to the expressly listed instead, may include other steps or elements not explicitly listed or inherent to the process, method, product or apparatus.

图1示出了服务器100的示例性组件示意图。服务器100可以包括一个或多个处理器104，诸如一个或多个中央处理单元(CPU)，每个处理单元可以实现一个或多个硬件线程。服务器100还可以包括任何存储介质106，其用于存储诸如代码、设置、数据等之类的任何种类的信息。非限制性的，比如，存储介质106可以包括以下任一项或多种组合：任何类型的RAM，任何类型的ROM，闪存设备，硬盘，光盘等。更一般地，任何存储介质都可以使用任何技术来存储信息。进一步地，任何存储介质可以提供信息的易失性或非易失性保留。进一步地，任何存储介质可以表示服务器100的固定或可移除部件。在一种情况下，当处理器104执行被存储在任何存储介质或存储介质的组合中的相关联的指令时，服务器100可以执行相关联指令的任一操作。服务器100还包括用于与任何存储介质交互的一个或多个驱动单元108，诸如硬盘驱动单元、光盘驱动单元等。FIG. 1 shows an exemplary component diagram of aserver 100 .Server 100 may include one ormore processors 104, such as one or more central processing units (CPUs), each of which may implement one or more hardware threads. Theserver 100 may also include anystorage medium 106 for storing any kind of information such as codes, settings, data, and the like. Without limitation, for example, thestorage medium 106 may include any one or a combination of the following: any type of RAM, any type of ROM, flash memory device, hard disk, optical disc, and so on. More generally, any storage medium can use any technology to store information. Further, any storage medium can provide for volatile or non-volatile retention of information. Further, any storage medium may represent a fixed or removable component of theserver 100 . In one instance, whenprocessor 104 executes the associated instructions stored in any storage medium or combination of storage media,server 100 may perform any operation of the associated instructions. Theserver 100 also includes one ormore drive units 108 for interfacing with any storage medium, such as a hard disk drive unit, an optical disk drive unit, and the like.

服务器100还包括输入/输出110（I/O），其用于接收各种输入(经由输入单元112)和用于提供各种输出(经由输出单元114))。一个具体输出机构可以包括呈现设备116和相关联的图形用户接口(GUI)118。服务器100还可以包括一个或多个网络接口120，其用于经由一个或多个通信单元122与其他设备交换数据。一个或多个通信总线124将上文所描述的部件耦合在一起。Server 100 also includes input/output 110 (I/O) for receiving various inputs (via input unit 112 ) and for providing various outputs (via output unit 114 )). One particular output mechanism may include a presentation device 116 and an associated graphical user interface (GUI) 118 . Theserver 100 may also include one ormore network interfaces 120 for exchanging data with other devices via one ormore communication units 122 . One ormore communication buses 124 couple together the components described above.

通信单元122可以以任何方式实现，例如，通过局域网、广域网(例如，因特网)、点对点连接等、或其任何组合。通信单元122可以包括由任何协议或协议组合支配的硬连线链路、无线链路、路由器、网关功能、名称服务器100等的任何组合。Communications unit 122 may be implemented in any manner, eg, via a local area network, a wide area network (eg, the Internet), a point-to-point connection, etc., or any combination thereof.Communications unit 122 may include any combination of hardwired links, wireless links, routers, gateway functions,name server 100, etc. governed by any protocol or combination of protocols.

图2示出了本发明实施例提供的基于普适计算的知识信息分析方法的流程示意图，该基于普适计算的知识信息分析方法可由图1中所示的服务器100执行，该基于普适计算的知识信息分析方法的详细步骤介绍如下。FIG. 2 shows a schematic flowchart of a method for analyzing knowledge information based on pervasive computing provided by an embodiment of the present invention. The method for analyzing knowledge information based on pervasive computing can be executed by theserver 100 shown in FIG. 1 . The detailed steps of the knowledge information analysis method are introduced as follows.

步骤S110，获取一组线上行为抽取知识信息，线上行为抽取知识信息包括线上行为知识网络中相互关联的第一行为实体以及第二行为实体的各自的行为抽取特征，以及第一行为实体与第二行为实体之间的实体关系的行为抽取特征。Step S110, obtaining a set of online behavior extraction knowledge information, the online behavior extraction knowledge information includes the respective behavior extraction features of the first behavior entity and the second behavior entity associated in the online behavior knowledge network, and the first behavior entity Behavior extraction features for entity relationships with second-behavior entities.

步骤S120，对于线上行为抽取知识信息，依据一次遍历循环，通过基于惩罚项的特征提取模型的模型单元中的特征提取单元依据线上行为抽取知识信息中的第一行为实体、第二行为实体以及实体关系三个普适计算目标中的两个普适计算目标的行为抽取特征生成第三个普适计算目标对应的至少一个候选行为抽取特征。Step S120, for online behavior extraction knowledge information, according to a traversal cycle, through the feature extraction unit in the model unit of the feature extraction model based on the penalty item, extract the first behavior entity and the second behavior entity in the knowledge information according to the online behavior And the behavior extraction features of two ubiquitous computing objects among the three ubiquitous computing objects of the entity relationship generate at least one candidate behavior extraction feature corresponding to the third ubiquitous computing object.

步骤S130，从至少一个候选行为抽取特征中确定一个候选行为抽取特征，作为第三个普适计算目标经过本轮遍历循环后其在线上行为知识网络中对应的行为抽取特征。Step S130, determine a candidate behavior extraction feature from at least one candidate behavior extraction feature, as the corresponding behavior extraction feature in the online behavior knowledge network after the third ubiquitous computing target goes through the current round of traversal.

步骤S130，更新基于惩罚项的特征提取模型的模型单元中的特征提取单元以及特征解码单元。依此，进行多次遍历循环直到满足停止遍历循环条件，获得线上行为知识网络中各行为实体或实体关系的目标行为抽取特征，并根据目标行为抽取特征对线上行为抽取知识信息进行普适计算。Step S130, updating the feature extraction unit and the feature decoding unit in the model unit of the feature extraction model based on the penalty item. In this way, multiple traversal loops are performed until the condition of stopping the traversal loop is met, and the target behavior extraction features of each behavior entity or entity relationship in the online behavior knowledge network are obtained, and the online behavior extraction knowledge information is universally adapted according to the target behavior extraction features. calculate.

依据上述设计，本实施例对于线上行为抽取知识信息，依据一次遍历循环，通过基于惩罚项的特征提取模型的模型单元中的特征提取单元依据线上行为抽取知识信息中的第一行为实体、第二行为实体以及实体关系三个普适计算目标中的两个普适计算目标的行为抽取特征生成第三个普适计算目标对应的至少一个候选行为抽取特征，以确定一个候选行为抽取特征，作为第三个普适计算目标经过本轮遍历循环后其在线上行为知识网络中对应的行为抽取特征，更新基于惩罚项的特征提取模型的模型单元中的特征提取单元以及特征解码单元，然后根据获得的线上行为知识网络中各行为实体或实体关系的目标行为抽取特征对线上行为抽取知识信息进行普适计算。如此，能够有效对线上行为抽取知识信息进行普适计算，并且突出不同知识层级的实体关系。According to the above-mentioned design, this embodiment extracts knowledge information from online behaviors, according to a traversal cycle, through the feature extraction unit in the model unit of the feature extraction model based on penalty items, extracts the first behavior entity in the knowledge information based on online behaviors, Generate at least one candidate behavior extraction feature corresponding to the third ubiquitous computing target from the behavior extraction features of two ubiquitous computing targets among the second behavior entity and entity relationship, so as to determine a candidate behavior extraction feature, As the third ubiquitous computing target, after the current round of traversal cycle, it extracts features corresponding to the behavior in the online behavior knowledge network, updates the feature extraction unit and feature decoding unit in the model unit of the feature extraction model based on penalty items, and then according to The obtained target behavior extraction features of each behavior entity or entity relationship in the online behavior knowledge network are used for pervasive calculation of online behavior extraction knowledge information. In this way, it is possible to effectively perform ubiquitous computing on online behavior extraction knowledge information, and highlight entity relationships at different knowledge levels.

作为一种示例，在一次遍历循环过程中，更新基于惩罚项的特征提取模型的模型单元中的特征提取单元以及特征解码单元的具体过程，可以是依据两个普适计算目标的行为抽取特征以及第三个普适计算目标对应的至少一个候选行为抽取特征生成至少一个普适筛选特征，以及将线上行为抽取知识信息作为普适编码特征，然后依据特征解码单元，对至少一个普适筛选特征以及普适编码特征进行有效性验证以获取有效性验证结果，依据有效性验证结果，更新特征提取单元，依据有效性验证结果与各特征解码单元的真实属性更新特征解码单元。As an example, in a traversal cycle process, the specific process of updating the feature extraction unit and the feature decoding unit in the model unit of the feature extraction model based on the penalty term can be based on the behavior of two universal computing targets. Feature extraction and At least one candidate behavior extraction feature corresponding to the third ubiquitous computing target generates at least one universal screening feature, and uses online behavior extraction knowledge information as a universal encoding feature, and then according to the feature decoding unit, at least one universal screening feature And the universal encoding feature is validated to obtain a validation result, the feature extraction unit is updated according to the validation result, and the feature decoding unit is updated according to the validity validation result and the real attributes of each feature decoding unit.

作为一种示例，在通过基于惩罚项的特征提取模型的模型单元中的特征提取单元依据线上行为抽取知识信息中的第一行为实体、第二行为实体以及实体关系三个普适计算目标中的两个普适计算目标的行为抽取特征生成第三个普适计算目标对应的至少一个候选行为抽取特征的过程中，可以获取至少一个随机行为抽取特征，然后通过特征提取单元依据两个普适计算目标的行为抽取特征以及至少一个随机行为抽取特征，获取对应于第三普适计算目标的至少一个候选行为抽取特征。As an example, the feature extraction unit in the model unit of the feature extraction model based on the penalty item extracts the first behavior entity, the second behavior entity and the entity relationship in the knowledge information according to the online behavior. Among the three ubiquitous computing targets In the process of generating at least one candidate behavior extraction feature corresponding to the third ubiquitous computing target from the behavior extraction features of two ubiquitous computing targets, at least one random behavior extraction feature can be obtained, and then the feature extraction unit can be used according to the two ubiquitous The behavior extraction feature of the computing target and at least one random behavior extraction feature are obtained to obtain at least one candidate behavior extraction feature corresponding to the third ubiquitous computing target.

作为一种示例，在通过特征提取单元依据两个普适计算目标的行为抽取特征以及至少一个随机行为抽取特征，获取对应于第三普适计算目标的至少一个候选行为抽取特征的过程中，可以依据两个普适计算目标的行为抽取特征以及至少一个随机行为抽取特征，依据特征提取单元，获取对应于第三普适计算目标的至少一个候选特征，对于每一个候选特征，依据线上行为知识网络，确定候选行为抽取特征。As an example, in the process of obtaining at least one candidate behavior extraction feature corresponding to the third ubiquitous computing target according to the behavior extraction features of two ubiquitous computing targets and at least one random behavior extraction feature by the feature extraction unit, it may be According to the behavior extraction features of two ubiquitous computing objects and at least one random behavior extraction feature, according to the feature extraction unit, obtain at least one candidate feature corresponding to the third ubiquitous computing object, and for each candidate feature, according to online behavior knowledge network to identify candidate behavior extraction features.

作为一种示例，对于每一个候选特征，依据线上行为知识网络，确定候选行为抽取特征的过程中，可以确定线上行为知识网络上与候选特征相似度满足预设条件的与第三普适计算目标同质的行为实体或实体关系的行为抽取特征，然后将满足预设条件的与第三普适计算目标同质的行为实体或实体关系的行为抽取特征确定为候选行为抽取特征。As an example, for each candidate feature, according to the online behavior knowledge network, in the process of determining the candidate behavior extraction feature, it can be determined that the similarity with the candidate feature on the online behavior knowledge network meets the preset conditions and the third universal Computing behavior extraction features of behavioral entities or entity relationships with homogeneous targets, and then determining behavior extraction features of behavioral entities or entity relationships that meet preset conditions and homogeneous with the third ubiquitous computing target as candidate behavior extraction features.

图3示出了本发明实施例提供的基于普适计算的知识信息分析系统200的功能模块图，该基于普适计算的知识信息分析系统200实现的功能可以对应上述方法执行的步骤。该基于普适计算的知识信息分析系统200可以理解为上述服务器100，或服务器100的处理器，也可以理解为独立于上述服务器100或处理器之外的在服务器100控制下实现本发明功能的组件，如图3所示，下面分别对该基于普适计算的知识信息分析系统200的各个功能模块的功能进行详细阐述。FIG. 3 shows a functional block diagram of a pervasive computing-based knowledgeinformation analysis system 200 provided by an embodiment of the present invention. The functions implemented by the pervasive computing-based knowledgeinformation analysis system 200 may correspond to the steps performed by the above method. The knowledgeinformation analysis system 200 based on ubiquitous computing can be understood as the above-mentionedserver 100, or the processor of theserver 100, and can also be understood as a system that realizes the functions of the present invention under the control of theserver 100 independently of the above-mentionedserver 100 or the processor Components, as shown in FIG. 3 , the functions of each functional module of the pervasive computing-based knowledgeinformation analysis system 200 will be described in detail below.

获取模块210，用于获取一组线上行为抽取知识信息，线上行为抽取知识信息包括线上行为知识网络中相互关联的第一行为实体以及第二行为实体的各自的行为抽取特征，以及第一行为实体与第二行为实体之间的实体关系的行为抽取特征；The obtainingmodule 210 is used to obtain a set of online behavior extraction knowledge information, the online behavior extraction knowledge information includes the respective behavior extraction features of the first behavior entity and the second behavior entity associated in the online behavior knowledge network, and the first Behavior extraction features of the entity relationship between a behavioral entity and a second behavioral entity;

生成模块220，用于对于线上行为抽取知识信息，依据一次遍历循环，通过基于惩罚项的特征提取模型的模型单元中的特征提取单元依据线上行为抽取知识信息中的第一行为实体、第二行为实体以及实体关系三个普适计算目标中的两个普适计算目标的行为抽取特征生成第三个普适计算目标对应的至少一个候选行为抽取特征；Thegeneration module 220 is used for extracting knowledge information for online behaviors. According to a traversal cycle, the feature extraction unit in the model unit of the feature extraction model based on penalty items is used to extract the first behavior entity and the second behavior entity in the knowledge information based on online behaviors. The second is to generate at least one candidate behavior extraction feature corresponding to the third ubiquitous computing target for the behavior extraction features of the two pervasive computing targets among the three ubiquitous computing targets of entity and entity relationship;

确定模块230，用于从至少一个候选行为抽取特征中确定一个候选行为抽取特征，作为第三个普适计算目标经过本轮遍历循环后其在线上行为知识网络中对应的行为抽取特征；Thedetermination module 230 is used to determine a candidate behavior extraction feature from at least one candidate behavior extraction feature, as the third ubiquitous computing target after the current round of traversal cycle and its corresponding behavior extraction feature in the online behavior knowledge network;

更新模块240，用于更新基于惩罚项的特征提取模型的模型单元中的特征提取单元以及特征解码单元；依此，进行多次遍历循环直到满足停止遍历循环条件，获得线上行为知识网络中各行为实体或实体关系的目标行为抽取特征，并根据目标行为抽取特征对线上行为抽取知识信息进行普适计算。Theupdate module 240 is used to update the feature extraction unit and the feature decoding unit in the model unit of the feature extraction model based on the penalty item; accordingly, perform multiple traversal loops until the condition of stopping the traversal loop is satisfied, and obtain the online behavior knowledge network. The target behavior extraction features of behavioral entities or entity relationships, and the pervasive calculation of online behavior extraction knowledge information according to the target behavior extraction features.

作为一种示例，一次遍历循环过程中，更新基于惩罚项的特征提取模型的模型单元中的特征提取单元以及特征解码单元，包括：As an example, during one traversal cycle, the feature extraction unit and the feature decoding unit in the model unit of the feature extraction model based on the penalty item are updated, including:

依据两个普适计算目标的行为抽取特征以及第三个普适计算目标对应的至少一个候选行为抽取特征生成至少一个普适筛选特征，以及将线上行为抽取知识信息作为普适编码特征；Generating at least one pervasive screening feature based on the behavior extraction features of two pervasive computing targets and at least one candidate behavior extraction feature corresponding to the third pervasive computing target, and using online behavior extraction knowledge information as pervasive coding features;

依据特征解码单元，对至少一个普适筛选特征以及普适编码特征进行有效性验证以获取有效性验证结果；Perform validity verification on at least one universal screening feature and universal coding feature according to the feature decoding unit to obtain a validity verification result;

依据有效性验证结果，更新特征提取单元；According to the validity verification result, the feature extraction unit is updated;

依据有效性验证结果与各特征解码单元的真实属性更新特征解码单元。The feature decoding unit is updated according to the validity verification result and the real attribute of each feature decoding unit.

作为一种示例，通过基于惩罚项的特征提取模型的模型单元中的特征提取单元依据线上行为抽取知识信息中的第一行为实体、第二行为实体以及实体关系三个普适计算目标中的两个普适计算目标的行为抽取特征生成第三个普适计算目标对应的至少一个候选行为抽取特征，包括：As an example, the feature extraction unit in the model unit of the feature extraction model based on the penalty item extracts the first behavior entity, the second behavior entity, and the entity relationship in the knowledge information according to the online behavior of the three ubiquitous computing targets The behavior extraction features of the two ubiquitous computing targets generate at least one candidate behavior extraction feature corresponding to the third ubiquitous computing target, including:

获取至少一个随机行为抽取特征；Obtain at least one random behavior extraction feature;

通过特征提取单元依据两个普适计算目标的行为抽取特征以及至少一个随机行为抽取特征，获取对应于第三普适计算目标的至少一个候选行为抽取特征。At least one candidate behavior extraction feature corresponding to the third ubiquitous computing target is acquired by the feature extraction unit according to the behavior extraction features of the two ubiquitous computing targets and at least one random behavior extraction feature.

作为一种示例，通过特征提取单元依据两个普适计算目标的行为抽取特征以及至少一个随机行为抽取特征，获取对应于第三普适计算目标的至少一个候选行为抽取特征，包括：As an example, the feature extraction unit obtains at least one candidate behavior extraction feature corresponding to the third ubiquitous computing target according to the behavior extraction features of two pervasive computing targets and at least one random behavior extraction feature, including:

依据两个普适计算目标的行为抽取特征以及至少一个随机行为抽取特征，依据特征提取单元，获取对应于第三普适计算目标的至少一个候选特征；Obtaining at least one candidate feature corresponding to the third ubiquitous computing target according to the behavior extraction features of the two ubiquitous computing targets and at least one random behavior extraction feature, according to the feature extraction unit;

对于每一个候选特征，依据线上行为知识网络，确定候选行为抽取特征。For each candidate feature, the candidate behavior extraction feature is determined according to the online behavior knowledge network.

作为一种示例，对于每一个候选特征，依据线上行为知识网络，确定候选行为抽取特征，包括：As an example, for each candidate feature, the candidate behavior extraction features are determined according to the online behavior knowledge network, including:

确定线上行为知识网络上与候选特征相似度满足预设条件的与第三普适计算目标同质的行为实体或实体关系的行为抽取特征；Determine the behavior extraction feature of the behavior entity or entity relationship that is homogeneous to the third ubiquitous computing target on the online behavior knowledge network and whose similarity with the candidate feature satisfies the preset condition;

将满足预设条件的与第三普适计算目标同质的行为实体或实体关系的行为抽取特征确定为候选行为抽取特征。The behavior extraction features of behavioral entities or entity relationships that meet the preset conditions and are homogeneous with the third ubiquitous computing target are determined as candidate behavior extraction features.

所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统，装置和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。Those skilled in the art can clearly understand that for the convenience and brevity of the description, the specific working process of the above-described system, device and unit can refer to the corresponding process in the foregoing method embodiment, which will not be repeated here.

对于本领域技术人员而言，显然本发明不限于上述示范性实施例的细节，而且在不背离本发明的精神或基本特征的情况下，能够以其它的具体形式实现本发明。因此，无论从哪一点来看，均应将实施例看作是示范性的，而且是非限制性的，本发明的范围由所附权利要求而不是上述说明限定，因此旨在将落在权利要求的等同要件的含义和范围内的所有变化囊括在本发明内。不应将权利要求中的任何附图进销存确认视为限制所涉及的权利要求。It will be apparent to those skilled in the art that the invention is not limited to the details of the above-described exemplary embodiments, but that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Accordingly, the embodiments should be regarded in all points of view as exemplary and not restrictive, the scope of the invention being defined by the appended claims rather than the foregoing description, and it is therefore intended that the scope of the invention be defined by the appended claims rather than by the foregoing description. All changes within the meaning and range of equivalents of the elements are embraced in the present invention. Any drawing confirmation in a claim should not be construed as limiting the claim involved.

Claims (10)

1. A knowledge information analysis method based on pervasive computing is characterized in that the method is applied to a server, and the method comprises the following steps:

acquiring a group of online behavior extraction knowledge information, wherein the online behavior extraction knowledge information comprises behavior extraction features of a first behavior entity and a second behavior entity which are associated with each other in an online behavior knowledge network, and behavior extraction features of an entity relationship between the first behavior entity and the second behavior entity;

extracting knowledge information of the online behaviors, and generating at least one candidate behavior extraction feature corresponding to a third pervasive computation object according to a traversal cycle and by means of a feature extraction unit in a model unit of a feature extraction model based on penalty terms and according to behavior extraction features of two pervasive computation objects in three pervasive computation objects of a first behavior entity, a second behavior entity and an entity relationship in the knowledge information of the online behaviors extraction;

determining a candidate behavior extraction feature from the at least one candidate behavior extraction feature, wherein the candidate behavior extraction feature is used as a corresponding behavior extraction feature in the online behavior knowledge network after the third pervasive computing target passes through the current round of traversal cycle;

updating the feature extraction unit and the feature decoding unit in the model unit of the feature extraction model based on the penalty term; and performing traversal circulation for multiple times until the condition of stopping traversal circulation is met, obtaining target behavior extraction features of each behavior entity or entity relationship in the online behavior knowledge network, and performing pervasive computation on the online behavior extraction knowledge information according to the target behavior extraction features.

2. The method of claim 1, wherein the updating the feature extraction unit and the feature decoding unit in the model unit of the penalty term based feature extraction model in a traversal loop process comprises:

generating at least one pervasive screening feature according to the behavior extraction features of the two pervasive computation targets and at least one candidate behavior extraction feature corresponding to the third pervasive computation target, and taking the online behavior extraction knowledge information as a pervasive coding feature;

according to the feature decoding unit, performing validity verification on the at least one pervasive screening feature and the pervasive coding feature to obtain a validity verification result;

updating the feature extraction unit according to the validity verification result;

and updating the feature decoding units according to the validity verification result and the real attributes of the feature decoding units.

3. The method as claimed in claim 1, wherein the generating of the at least one candidate behavior extraction feature corresponding to the third pervasive computation object by the feature extraction unit in the model unit of the feature extraction model based on the penalty term according to the behavior extraction features of two pervasive computation objects in the three pervasive computation objects of the first behavior entity, the second behavior entity and the entity relationship in the online behavior extraction knowledge information comprises:

acquiring at least one random behavior extraction feature;

and acquiring at least one candidate behavior extraction characteristic corresponding to the third pervasive computing target according to the behavior extraction characteristics of the two pervasive computing targets and the at least one random behavior extraction characteristic by the characteristic extraction unit.

4. The method according to claim 1, wherein the obtaining, by the feature extraction unit, at least one candidate behavior extraction feature corresponding to the third pervasive computing object according to the behavior extraction features of the two pervasive computing objects and the at least one random behavior extraction feature comprises:

acquiring at least one candidate feature corresponding to the third pervasive computation target according to the feature extraction unit and the behavior extraction features of the two pervasive computation targets and at least one random behavior extraction feature;

and for each candidate feature, determining the candidate behavior extraction feature according to the online behavior knowledge network.

5. The method of claim 1, wherein said determining, for each candidate feature, the candidate behavior extraction feature according to the knowledge network of on-line behaviors comprises:

determining behavior extraction characteristics of behavior entities or entity relations which are homogeneous with a third pervasive computing target and have similarity with the candidate characteristics on the online behavior knowledge network and meet preset conditions;

and determining the behavior extraction features of the behavior entities or entity relations which meet the preset conditions and are homogeneous with the third pervasive computing target as candidate behavior extraction features.

6. A knowledge information analysis system based on pervasive computing, which is applied to a server, and comprises:

the system comprises an acquisition module, a processing module and a display module, wherein the acquisition module is used for acquiring a group of online behavior extraction knowledge information, and the online behavior extraction knowledge information comprises behavior extraction characteristics of a first behavior entity and a second behavior entity which are associated with each other in an online behavior knowledge network and behavior extraction characteristics of an entity relationship between the first behavior entity and the second behavior entity;

the generating module is used for extracting knowledge information of the online behaviors, and generating at least one candidate behavior extraction feature corresponding to a third pervasive computation target according to behavior extraction features of two pervasive computation targets in the three pervasive computation targets of the first behavior entity, the second behavior entity and the entity relationship in the knowledge information extracted by a feature extraction unit in a model unit of a feature extraction model based on a penalty term according to a traversal cycle;

the determining module is used for determining a candidate behavior extraction feature from the at least one candidate behavior extraction feature, and the candidate behavior extraction feature is used as a corresponding behavior extraction feature in the online behavior knowledge network after the third pervasive computing target passes through the current round of traversal cycle;

the updating module is used for updating the feature extraction unit and the feature decoding unit in the model unit of the feature extraction model based on the penalty term; and performing traversal circulation for multiple times until the condition of stopping traversal circulation is met, obtaining target behavior extraction features of each behavior entity or entity relationship in the online behavior knowledge network, and performing pervasive computation on the online behavior extraction knowledge information according to the target behavior extraction features.

7. The system according to claim 6, wherein the updating the feature extraction unit and the feature decoding unit in the model unit of the penalty term based feature extraction model in a traversal loop process comprises:

generating at least one pervasive screening feature according to the behavior extraction features of the two pervasive computation targets and at least one candidate behavior extraction feature corresponding to the third pervasive computation target, and taking the online behavior extraction knowledge information as a pervasive coding feature;

according to the feature decoding unit, performing validity verification on the at least one pervasive screening feature and the pervasive coding feature to obtain a validity verification result;

updating the feature extraction unit according to the validity verification result;

and updating the feature decoding units according to the validity verification result and the real attribute of each feature decoding unit.

8. The system according to claim 6, wherein the generating of the at least one candidate behavior extraction feature corresponding to the third pervasive computation object by the feature extraction unit in the model unit of the feature extraction model based on the penalty term according to the behavior extraction features of two pervasive computation objects of the three pervasive computation objects of the first behavior entity, the second behavior entity and the entity relationship in the online behavior extraction knowledge information comprises:

acquiring at least one random behavior extraction feature;

and acquiring at least one candidate behavior extraction characteristic corresponding to the third pervasive computation target through the characteristic extraction unit according to the behavior extraction characteristics of the two pervasive computation targets and the at least one random behavior extraction characteristic.

9. The system according to claim 8, wherein the obtaining, by the feature extraction unit, at least one candidate behavior extraction feature corresponding to the third pervasive computing object according to the behavior extraction features of the two pervasive computing objects and the at least one random behavior extraction feature comprises:

according to the behavior extraction features of the two pervasive computing targets and at least one random behavior extraction feature, acquiring at least one candidate feature corresponding to the third pervasive computing target according to the feature extraction unit;

and for each candidate feature, determining the candidate behavior extraction feature according to the online behavior knowledge network.

10. The system of claim 9, wherein said determining, for each candidate feature, the candidate behavior extraction feature according to the knowledge network of on-line behaviors comprises:

determining behavior extraction characteristics of behavior entities or entity relations which are homogeneous with a third pervasive computing target and have similarity with the candidate characteristics on the online behavior knowledge network and meet preset conditions;

and determining the behavior extraction features of the behavior entities or entity relations which meet the preset conditions and are homogeneous with the third pervasive computing target as candidate behavior extraction features.

Images