CN116320021A - Hierarchical service matching method based on intention in Internet of things scene - Google Patents

Hierarchical service matching method based on intention in Internet of things scene Download PDF

Info

Publication number
CN116320021A
CN116320021A CN202310329502.8A CN202310329502A CN116320021A CN 116320021 A CN116320021 A CN 116320021A CN 202310329502 A CN202310329502 A CN 202310329502A CN 116320021 A CN116320021 A CN 116320021A
Authority
CN
China
Prior art keywords
service
intention
matching
qos
similarity
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.)
Pending
Application number
CN202310329502.8A
Other languages
Chinese (zh)
Inventor
杨鲲
杨涛
严康
梅海波
王洋
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.)
University of Electronic Science and Technology of China
Original Assignee
University of Electronic Science and Technology of China
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 University of Electronic Science and Technology of China filed Critical University of Electronic Science and Technology of China
Priority to CN202310329502.8A priority Critical patent/CN116320021A/en
Publication of CN116320021A publication Critical patent/CN116320021A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/50—Network services
    • H04L67/60—Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources
    • H04L67/61—Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources taking into account QoS or priority requirements
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/50—Network services
    • H04L67/60—Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources
    • H04L67/63—Routing a service request depending on the request content or context
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The application discloses an intention-based layered service matching method in an Internet of things scene, which aims at the problems that a traditional service description model is used for the deficiency of matching information in the Internet of things, and is based on fuzzy matching grade division, low matching success rate and lack of nonfunctional attribute judgment in an OWL-S service matching algorithm, so that the method is not applicable to the Internet of things.

Description

Hierarchical service matching method based on intention in Internet of things scene
Technical Field
The invention relates to the technical field of the Internet of things, in particular to an intention-based hierarchical service matching method in an Internet of things scene.
Background
The internet of things is regarded as a product of the third revolution of information technology, and the purpose of the internet of things is to connect the 'things' on the edge to the network, so as to realize the internet of everything, thereby facilitating the life of people. With the maturity of communication technology, the internet of things is one of the most commonly mentioned technologies, and can provide many automated intelligent scenes, and a very good experience is provided for users. The Internet of things now covers almost all industries, and has a countless application scene such as smart home, industrial automation, smart agriculture and the like.
In the internet of things, a user usually only focuses on describing a target state to be achieved, expresses a user requirement or a user intention, and does not focus on how to achieve the target, so that an internet of things application system is required to automatically match a service capable of achieving the user intention according to the user intention.
However, with the continuous development of the internet of things technology, the types of terminal devices in the internet of things are continuously increased, and the types and the number of services which can be provided are also continuously increased, so that the accuracy of service matching is continuously reduced. In addition, service input and output information in the internet of things cannot fully embody a service function, and a large number of services with the same input and output information but different calling effects exist, so that the services are generally difficult to match by virtue of the input and output information. The user demands usually contain service context information such as service places and service QoS information such as service response speed, while the traditional service descriptions such as OWL-S-based service ontology does not contain service context and service QoS information, and the traditional service matching method only considers input and output information of the service, for example, the matching grade division is fuzzy, so that the service descriptions and the user demands are difficult to match, the accuracy of service matching is reduced, and the traditional service description method and the service matching method cannot be suitable for the scene of the Internet of things.
At present, for the service matching technology in the scene of the internet of things, mainly the traditional service matching is mainly adopted, although the prior papers make reference to the purpose of matching, the traditional method is still adopted in the aspect of calculating the service matching degree, and the service context and the service QoS matching are not completely considered. Therefore, the hierarchical service matching method based on the intention in the scene of the Internet of things is designed, and the existing research blank is made up.
Disclosure of Invention
In order to solve the problems in the background art, the invention provides an intention-based hierarchical service matching method in an Internet of things scene, which aims to solve the problems of insufficient content and low service matching accuracy in the traditional service description in the existing Internet of things environment.
In order to achieve the above purpose, the present invention provides the following technical solutions:
an intent-based hierarchical service matching method in an internet of things scene, the method comprising:
step 1: expanding a service ontology in the Internet of things to obtain at least one service S comprising service input and output information, intention actions, intention targets, service context information and multi-index service QoS information i All services S i Forming a candidate service ontology set S, and evaluating all indexes of service QoS information by using fuzzy language;
step 2: acquiring a complete user intention by using an input device;
step 3: analyzing the user intention by using a Stanford NLP tool package to acquire an intention action, an intention target, a service context requirement and a service QoS requirement;
step 4: calculating the similarity between the intention action and the intention target in the intention of the user and the intention action and the intention target in the candidate service ontology set S by using a concept similarity calculation method to obtain a weighted similarity value and a user-defined threshold omega 1 Filtering out the weighted similarity value in the service ontology set S to be smaller than the threshold omega 1 Obtain the service set S 1 ；
Step 5: the utilized concept similarity calculation method calculates each attribute and service set S in the service context requirement in the user intention 1 The attribute of each service context information in the network is subjected to similarity calculation to obtain a weighted similarity value and a self-defined threshold omega 2 Filtering out the service set S 1 The medium weighted similarity value is less than a threshold omega 2 Obtain the service set S 2 ；
Step 6: converting the user' S service QoS requirement into user preference vector P, and converting service set S by triangle fuzzy number 2 The evaluation of each index of the medium service QoS information is converted into a matrix Q in a numerical form, and the matrix Q' is obtained after linear normalization;
step 7: selecting an optimal QoS evaluation index b from each column in the matrix Q i And constitutes an optimal QoS index vector b= (B) 1 ,b 2 ,b 3 ,b 4 ,b 5 ,b 6 )；
Step 8: euclidean proximity is selected as a QoS similarity measurement index, and a service set S is calculated 2 Each service S in (a) i The similarity between the QoS index and B is calculated as
Figure BDA0004154435330000031
Wherein x 'is the corresponding element in matrix Q';
step 9: according to sim Si For service set S 2 The elements in the service set S are sorted in a descending order to obtain a final matching result 3 。
Preferably, the specific operation in step 1 is: expanding a Service body in the Internet of things, and representing the Service body by four components of service= { IO, intion, context, qoS }, wherein:
IO represents input and output information of a service;
the interaction= { Action, target } represents the Intention information of the service, and contains Action Intention Action and Target Intention Target;
context= { Type, pos, time } represents service Context information, including Type service Type, pos service callable location, and Time service callable Time;
qos= { Sta, sec, res, ene, ava, int } represents QoS information of a service, and includes five levels of Sta service stability, sec service security, response speed of Res service, ene service energy consumption, ava service availability and Int service interoperability, and each index of QoS is available or unavailable, low, medium, high and extremely high.
Preferably, the specific method for calculating the similarity in the step 4 and the step 5 is as follows: introducing depth concepts of the tree to optimize similarity algorithm based on semantic distance, adding weights for paths among connection concepts in the tree, wherein a weight calculation formula is as follows
Figure BDA0004154435330000041
Wherein n is the layer number of concept nodes in the ontology tree, and at the moment, two concepts C with inheritance relation exist 1 And C 2 Inter-semantic distance Dis (C) 1 ,C 2 ) Can be represented by Sigma W (n), concept C if the same concept semantic distance is 0 1 And C 2 The similarity calculation formula between the two is +.>
Figure BDA0004154435330000042
Preferably, the step 4 specifically comprises the following sub-steps:
b1: each service S from the candidate service ontology set S i Extracting service intention actions and service intention targets;
b2: by means of
Figure BDA0004154435330000043
Respectively calculating semantic similarity of a user and a service intention action and an intention target, and obtaining weighted similarity sim;
b3: setting a phase matching threshold value as omega 1 Discard sim<ω 1 Obtain service set S 1 。
Preferably, step 5 specifically comprises the following sub-steps:
c1: from service set S 1 Each of the services S i Extracting all the context information UserContext;
c2: determining whether each attribute in the user context requirements can be in S i Finding corresponding attributes in the context information, and if the corresponding attributes cannot be found completely, thenConsider the match to fail, discard S i ；
And C3: by means of
Figure BDA0004154435330000044
Respectively calculating the matching degree of each attribute in the user context demand and the service context information to obtain weighted similarity sim;
and C4: setting the two-stage matching threshold value as omega 1 Discard sim<ω 2 Obtain service set S 2 。
Compared with the prior art, the invention has the beneficial effects that:
1. the service intention, the service context and the service QoS information are expanded in the service body, the object service is more comprehensively described, and the method is more suitable for complex Internet of things environments.
2. The method for calculating the similarity of the ontology concepts by fusing semantic distance and hierarchy information of the ontology tree is designed, so that the similarity information is more accurate, matching grade division is not adopted, matching judgment is carried out through specific similarity values, and a matching result is more accurate.
3. The intention matching and the service context matching are used for matching the specific user intention to meet the service of the functional requirement, and more accurate matching results are achieved compared with the matching based on the service input and output information.
4. The service which can best meet the current requirements of the users is selected for the users through QoS matching, and the customized and nonfunctional requirements of different users can be met.
5. The method is suitable for service matching under different service numbers, and higher service accuracy is obtained in the whole stage.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a schematic diagram of a device type ontology tree of the present invention;
FIG. 3 is a diagram of an intent matching method of the present invention;
FIG. 4 is a diagram of a context matching method of the present invention;
FIG. 5 is a schematic diagram of membership of triangular fuzzy numbers;
FIG. 6 is a graph of triangle blur number conversion results;
FIG. 7 is a simulation diagram of the precision of the hierarchical matching method according to the number of services;
fig. 8 is a simulation diagram of the precision ratio by only intent matching in the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1
Fig. 1 is a flow chart of an intent-based service hierarchical matching method of the internet of things. In the matching method, the matching method can be divided into three different stages, each stage reduces the matching range according to the threshold value of the matching degree, and the next stage continues to match in the reduced range of the previous stage, so that the service is prevented from being matched one by one, and the service matching efficiency is improved. The specific operation steps are as follows:
step 1: expanding a service ontology in the Internet of things to obtain at least one service S comprising service input and output information, intention actions, intention targets, service context information and multi-index service QoS information i All services S i Forming a candidate service ontology set S, and evaluating all indexes of service QoS information by using fuzzy language;
step 2: acquiring a complete user intention by using an input device;
step 3: analyzing the user intention by using a Stanford NLP tool package to acquire an intention action, an intention target, a service context requirement and a service QoS requirement;
step 4: using concept similarity calculation method to compute intent actions in user intent, intent targets and intent in candidate service ontology set SSimilarity calculation is carried out between the graph action and the intention target to obtain a weighted similarity value and a self-defined threshold omega 1 Filtering out the weighted similarity value in the service ontology set S to be smaller than the threshold omega 1 Obtain the service set S 1 ；
Step 5: the utilized concept similarity calculation method calculates each attribute and service set S in the service context requirement in the user intention 1 The attribute of each service context information in the network is subjected to similarity calculation to obtain a weighted similarity value and a self-defined threshold omega 2 Filtering out the service set S 1 The medium weighted similarity value is less than a threshold omega 2 Obtain the service set S 2 ；
Step 6: converting the user' S service QoS requirement into user preference vector P, and converting service set S by triangle fuzzy number 2 The evaluation of each index of the medium service QoS information is converted into a matrix Q in a numerical form, and the matrix Q' is obtained after linear normalization;
step 7: selecting an optimal QoS evaluation index b from each column in the matrix Q i And constitutes an optimal QoS index vector b= (B) 1 ,b 2 ,b 3 ,b 4 ,b 5 ,b 6 )；
Step 8: euclidean proximity is selected as a QoS similarity measurement index, and a service set S is calculated 2 Each service S in (a) i The similarity between the QoS index and B is calculated as
Figure BDA0004154435330000071
Wherein x 'is the corresponding element in matrix Q';
step 9: according to
Figure BDA0004154435330000072
For service set S 2 The elements in the service set S are sorted in a descending order to obtain a final matching result 3 。
In this embodiment, the service intention, the service context and the service QoS information are expanded in the service ontology, so that the object service is more comprehensively described, the method is more suitable for a complex internet of things environment, and the intention matching and the service context matching are used for matching the specific user intention to meet the service of the functional requirement, so that a more accurate matching result is achieved compared with the matching based on the service input and output information. The service which can best meet the current requirements of the users is selected for the users through QoS matching, and the customized and nonfunctional requirements of different users can be met. The method is suitable for service matching under different service numbers, and higher service accuracy is obtained in the whole stage.
Example 2
The difference between this embodiment and embodiment 1 is that, as shown in fig. 2, the root node is regarded as layer 1, and the weight calculation formula of the connection conceptual path in the n-layer node is:
Figure BDA0004154435330000073
after introducing this concept, the semantic distance calculation formula is:
Figure BDA0004154435330000074
wherein Sigma W (n) represents the concept C 1 And C 2 And the sum of all path weights.
After obtaining the semantic distance between the optimized concepts, a proper similarity conversion function is also required to be constructed, and the similarity between the concepts is used as [0,1 ]]Numerical quantization expressions within the interval. In the service matching of the Internet of things, the matching of semantic type attributes is included, the matching of numerical type and Boolean type attributes is also included, and if the service maximum working time attribute is the numerical type attribute in the service context, the corresponding numerical comparison is needed during the matching. Therefore, the characteristics of these attributes need to be fully considered in calculating the similarity. Suppose C 1 To request attributes, C 2 Corresponding attributes are in the candidate service set.
For semantic attributes, the similarity between concepts is:
Figure BDA0004154435330000081
and alpha is a mediation factor, and the influence degree of the semantic distance on the semantic similarity result is determined.
When C 1 Inheritance C 2 When the attribute concept in the requirement is indicated to be more specific than the attribute concept in the service; when C 2 Inheritance C 1 When this indicates that the in-service attribute is more specific than the in-demand attribute. The latter is generally considered to be more matched, as the lighting device is required in the request, the lighting device can be provided with a switch in the service, which is satisfactory, but the exchange of locations between the two is not necessarily satisfactory, since the lighting device also comprises other types of lighting devices. Therefore, in order to distinguish this situation and to facilitate calculation, the mediation factor is only valued (1, 2 when the demand attribute is a sub-concept of the service attribute]In other cases, the value is 1.
For numerical attributes, the similarity between concepts is:
Figure BDA0004154435330000082
C 2 satisfy C 1 Two situations can be distinguished: numerical comparison type and range type. And the numerical comparison type service response time delay is satisfied only by the fact that the response time delay in service is smaller than the time delay in request. The callable time period in the service is larger than the time range required to be called in the request, and the callable time period in the service is satisfied. Of course, both cases are more than satisfied and less than satisfied, which is only a simple example.
For boolean attributes:
Figure BDA0004154435330000083
because the Boolean type attribute has only two values, the calculation is simpler, the similarity is 1 only when the two Boolean type attributes are identical, and the similarity is 0 when the two Boolean type attributes are different.
Example 3
The difference between this embodiment and embodiment 2 is that, as shown in fig. 3, fig. 3 is a diagram of the method for matching intention of the present invention. Aiming at the matching between the user requirement and the service intention, the method is the first step of the service matching method flow, and the service set is initially selected through the matching of the intention action and the target. Before the service matching method is performed, the equipment body is required to be acquired from a body library, and intention, context and QoS information are acquired through analysis; at the same time, the user intention needs to be resolved through a Stanford-like NLP toolkit, mapped into the corresponding concept of the domain ontology, and then matched. In the present invention, it is assumed that these tasks have been completed, i.e., the user's intent appears in the same form as the domain ontology concept, and the device ontology can directly match the user's intent needs. Obtaining a threshold omega passing through a stage after the method is operated 1 And the filtered service set. And takes it as input for the second stage service matching. The existing ontology in the Internet does not have service intention, service context and QoS information, in order to enable the method to be compatible with original data, the existing ontology in the Internet is matched in a mode of carrying out semantic similarity calculation on service input and output information, and filtering is carried out according to a set threshold value.
Example 4
The difference between this embodiment and embodiment 3 is that, as shown in fig. 4, the service context matching is the second stage of the hierarchical matching method designed by the present invention, and this stage is the service set S obtained by the intentional matching method 1 On the basis of the above, performing secondary screening on the service set according to the context demands in the demands of the user and the similarity of the service context information in the equipment body, wherein the screened similarity is smaller than a set second stage threshold omega 2 Is a service of (a). Unlike service intent matching, in matching of service contexts, the related attributes have richer types, so that when calculating similarity, the corresponding calculation mode introduced above needs to be selected according to the attribute types. It should be noted that when the request attribute is a sub-concept of the service attribute, a mediation factor α needs to be introduced, since the service context contains concepts such as the service typeThis will more lead to this situation. Since the user's requirements for the context generally include a plurality of different attributes, the service to be selected must be able to satisfy each attribute in the user's context requirements to be considered as successfully matching the user's requirements, otherwise, the matching is failed, which may also indicate that the number of attributes in the device service context ontology is not less than the number of attributes in the user's context requirements. By utilizing the characteristics, the user context requirements and the service context attributes can be judged before similarity calculation, so that the service which cannot be matched with the requirements can be found and discarded before matching calculation, and the calculated amount is reduced to a certain extent. At the same time, the feature also indicates that if the service context attribute to be selected cannot completely contain the attribute in the user context requirement, that is, the service context attribute cannot find the same attribute as one attribute in the user requirement, the matching should be immediately determined to fail, so that the current service is discarded, and no additional computing resource is consumed.
Since the third-stage service QoS matching does not involve attributes affecting the service function hierarchy, the resulting service set after service context matching can already be considered as a service set that can meet the user's needs.
Example 5
The present embodiment differs from embodiment 4 in that, as shown in fig. 5, it is assumed that the user's requirement for QoS is input in the form of a user preference vector, and the difference between the two is eliminated as much as possible in this matching manner. The user preferences may be expressed as:
P＝{η 1 ,η 2 ,η 3 ,η 4 ,η 5 ,η 6 }
wherein the sum of the individual items is 1.
The triangle ambiguity is a model in the field of fuzzy mathematics, with which it is possible to represent information that is difficult to describe with numerical accuracy. The concept of definition domain similar to each mathematical domain in the fuzzy number is a domain, and a fuzzy set A is arranged:
A＝{(x,μ A (x)|x∈X}
where X is called the argument of fuzzy set A, μ A Is the membership function of fuzzy set A, mu A The membership of the fuzzy set a, called point x, is used to indicate some degree of proximity between the point and the set concept. For the triangle blur number, its membership function μ A The value range of (2) is [0,1 ]]The following formula is satisfied:
Figure BDA0004154435330000111
wherein r is l Is a fuzzy lower bound; r is (r) m The fuzzy median value is the value with the largest occurrence probability of the fuzzy number in the fuzzy interval; r is (r) u Is a fuzzy upper bound. The closer the membership is to 1, the closer the element represented by the point is to the fuzzy median, and the corresponding relation between the membership of the triangular fuzzy number and the interval probability is shown in fig. 5.
Example 6
The difference between this embodiment and embodiment 5 is that, as shown in fig. 6, five fuzzy language levels { none, low, medium, high, extremely high } are adopted in the QoS model designed in the present invention to evaluate QoS indexes, and the result is shown in fig. 6 by triangle fuzzy number conversion and vector representation of fuzzy numbers. Because only common parameters are discussed in the QoS model of the invention, and the input service set is established on the basis of filtering in two matching stages, the calculated amount is not great, the result can be obtained through matrix operation, and the QoS matching flow is as follows:
a1, matching the context to obtain a result S 2 QoS indexes of all services are extracted to form a QoS index matrix; the indices are then converted to a triangular blur description based on the results in fig. 6, resulting in a matrix Q.
A2, performing linear normalization processing on the matrix Q to obtain a matrix Q'. According to the QoS description model of the invention, the linear normalization formula is as follows:
Figure BDA0004154435330000112
a3, selecting the optimal QoS evaluation index from each column of the matrix QLabel b i And marks it as an optimal QoS index vector:
B＝(b 1 ,b 2 ,b 3 ,b 4 ,b 5 ,b 6 )
the vector is used to calculate the closeness of each service QoS indicator to the optimal indicator, which can be regarded as the similarity between QoS in the present method.
And A4, extracting preference vectors from QoS demands of users, and introducing similarity between each column of elements in the preference vector calculation matrix Q' and the optimal QoS index vector B. Because the elements in the matrix Q' are normalized descriptions of QoS indexes after triangle fuzzy number conversion, euclidean proximity is selected as a measure of similarity. Service S, in combination with the parameters and procedures described above i The similarity calculation formula with the optimal QoS index vector is as follows:
Figure BDA0004154435330000121
wherein P is the corresponding element in the user preference, B is the corresponding element in the optimal QoS index vector B, and x 'is the corresponding element in the normalized QoS description matrix Q'.
A5, according to S 2 The similarity calculation result of each service and B in the service set is used for sorting the service sets, and the sorted service sets S are output 3 . After the matching is finished, a service set S which meets the requirements of users and is arranged from high to low according to the matching degree is obtained 3 。
Example 7
The difference between the embodiment and the embodiment 6 is that, as shown in fig. 7, the comparison object of the present invention is an OWLS-M0 and an OWLS-M3 matching method in the traditional service matching method OWLS-MX, where OWLS-M0 is a method for matching based on only logic reasoning, and similarity between the two is obtained by judging whether there is a containment relationship between each ontology concept; the OWLS-M3 combines text similarity matching between the service to be selected and the service requirement on the basis of the former, and is a hybrid method. Selecting 400 services and 20 service requests in OWLS-TC4, and setting a text similarity thresholdIs omega text =0.75, the integrated semantic similarity threshold is set to ω 1 =0.6. 400 services in the OWL-TC 4 are expanded by using Prot g as the expansion based on the equipment body of the invention, wherein the expansion comprises an intention body, a context body and a QoS body, and corresponding intention information is added to 20 service requests, and the matching accuracy is influenced by the combination of the matching flow designed by the invention and the inter-concept semantic similarity method by executing the intention-based hierarchical matching method and comparing with the OWL-M0 and OWL-M3 methods, as shown in figure 7. The results in fig. 7 can obtain the average precision of 65.8%, 77.4% and 86.1% respectively, and it can be seen that the matching precision can be improved by 8.7% by adopting the method of the invention.
Example 8
The difference between this embodiment and embodiment 7 is that, as shown in fig. 8, considering that there are many services with simple functions in the environment of the internet of things, there is no environment context or QoS class requirement for such services, so that the accuracy result of the first stage of the method is compared with the OWLS-M0 and OWLS-M3 methods, and the result is shown in fig. 8. From the results of fig. 8, it can be seen that the precision is still higher than the OWLS-MX method considering only the intended match.
Those of ordinary skill in the art will recognize that the embodiments described herein are for the purpose of aiding the reader in understanding the principles of the present invention and should be understood that the scope of the invention is not limited to such specific statements and embodiments. Various modifications and variations of the present invention will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.

Claims (5)

1. An intention-based hierarchical service matching method in an internet of things scene is characterized by comprising the following steps:
step 1: expanding a service body in the Internet of things to obtain information including service input and output, intention actions, intention targets and serviceAt least one service S of context information, multi-index service QoS information i All services S i Forming a candidate service ontology set S, and evaluating all indexes of service QoS information by using fuzzy language;
step 2: acquiring a complete user intention by using an input device;
step 3: analyzing the user intention by using a Stanford NLP tool package to acquire an intention action, an intention target, a service context requirement and a service QoS requirement;
step 4: calculating the similarity between the intention action and the intention target in the intention of the user and the intention action and the intention target in the candidate service ontology set S by using a concept similarity calculation method to obtain a weighted similarity value and a user-defined threshold omega 1 Filtering out the weighted similarity value in the service ontology set S to be smaller than the threshold omega 1 Obtain the service set S 1 ；
Step 5: the utilized concept similarity calculation method calculates each attribute and service set S in the service context requirement in the user intention 1 The attribute of each service context information in the network is subjected to similarity calculation to obtain a weighted similarity value and a self-defined threshold omega 2 Filtering out the service set S 1 The medium weighted similarity value is less than a threshold omega 2 Obtain the service set S 2 ；
Step 6: converting the user' S service QoS requirement into user preference vector P, and converting service set S by triangle fuzzy number 2 The evaluation of each index of the medium service QoS information is converted into a matrix Q in a numerical form, and the matrix Q' is obtained after linear normalization;
step 7: selecting an optimal QoS evaluation index b from each column in the matrix Q i And constitutes an optimal QoS index vector b= (B) 1 ,b 2 ,b 3 ,b 4 ,b 5 ,b 6 )；
Step 8: euclidean proximity is selected as a QoS similarity measurement index, and a service set S is calculated 2 Each service S in (a) i The similarity between the QoS index and B is calculated as
Figure FDA0004154435310000021
Wherein x 'is the corresponding element in matrix Q';
step 9: according to sim Si For service set S 2 The elements in the service set S are sorted in a descending order to obtain a final matching result 3 。
2. The hierarchical service matching method based on intention in the scene of internet of things according to claim 1, wherein the specific operation in step 1 is as follows: expanding a Service body in the Internet of things, and representing the Service body by four components of service= { IO, intion, context, qoS }, wherein:
IO represents input and output information of a service;
the interaction= { Action, target } represents the Intention information of the service, and contains Action Intention Action and Target Intention Target;
context= { Type, pos, time } represents service Context information, including Type service Type, pos service callable location, and Time service callable Time;
qos= { Sta, sec, res, ene, ava, int } represents QoS information of a service, and includes five levels of Sta service stability, sec service security, response speed of Res service, ene service energy consumption, ava service availability and Int service interoperability, and each index of QoS is available or unavailable, low, medium, high and extremely high.
3. The hierarchical service matching method based on intention in the scene of the internet of things according to claim 1, wherein the specific method for calculating the similarity between the step 4 and the step 5 is as follows: introducing depth concepts of the tree to optimize similarity algorithm based on semantic distance, adding weights for paths among connection concepts in the tree, wherein a weight calculation formula is as follows
Figure FDA0004154435310000022
Wherein n is the layer number of concept nodes in the ontology tree, and at the moment, two concepts C with inheritance relation exist 1 And C 2 Inter-semantic distance Dis (C) 1 ,C 2 ) Can be usedSigma W (n), concept C if the same concept semantic distance is 0 1 And C 2 The similarity calculation formula between the two is
Figure FDA0004154435310000023
Figure FDA0004154435310000024
4. The hierarchical service matching method based on intention in the scene of internet of things according to claim 3, wherein the step 4 specifically comprises the following sub-steps:
b1: each service S from the candidate service ontology set S i Extracting service intention actions and service intention targets;
b2: by means of
Figure FDA0004154435310000031
Respectively calculating semantic similarity of a user and a service intention action and an intention target, and obtaining weighted similarity sim;
b3: setting a phase matching threshold value as omega 1 Discard sim<ω 1 Obtain service set S 1 。
5. The hierarchical service matching method based on intention in the scene of internet of things according to claim 4, wherein the step 5 specifically comprises the following sub-steps:
c1: from service set S 1 Each of the services S i Extracting all the context information UserContext;
c2: determining whether each attribute in the user context requirements can be in S i If the corresponding attributes are found in the context information and cannot be found completely, the matching is considered to be failed, and S is discarded i ；
And C3: by means of
Figure FDA0004154435310000032
Respectively calculating the matching degree of each attribute in the user context demand and the service context information to obtain weighted similarity sim;
and C4: setting the two-stage matching threshold value as omega 1 Discard sim<ω 2 Obtain service set S 2 。
CN202310329502.8A 2023-03-30 2023-03-30 Hierarchical service matching method based on intention in Internet of things scene Pending CN116320021A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310329502.8A CN116320021A (en) 2023-03-30 2023-03-30 Hierarchical service matching method based on intention in Internet of things scene

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310329502.8A CN116320021A (en) 2023-03-30 2023-03-30 Hierarchical service matching method based on intention in Internet of things scene

Publications (1)

Publication Number Publication Date
CN116320021A true CN116320021A (en) 2023-06-23

Family

ID=86824013

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310329502.8A Pending CN116320021A (en) 2023-03-30 2023-03-30 Hierarchical service matching method based on intention in Internet of things scene

Country Status (1)

Country Link
CN (1) CN116320021A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117151433A (en) * 2023-10-30 2023-12-01 浙江大学高端装备研究院 Cloud-based intelligent manufacturing service supply and demand matching evaluation method and device

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117151433A (en) * 2023-10-30 2023-12-01 浙江大学高端装备研究院 Cloud-based intelligent manufacturing service supply and demand matching evaluation method and device
CN117151433B (en) * 2023-10-30 2024-01-30 浙江大学高端装备研究院 Cloud-based intelligent manufacturing service supply and demand matching evaluation method and device

Similar Documents

Publication Publication Date Title
Karim et al. Decision tree and naive bayes algorithm for classification and generation of actionable knowledge for direct marketing
Wilson et al. Improved heterogeneous distance functions
CN107391772A (en) A kind of file classification method based on naive Bayesian
Dudas et al. Integration of data mining and multi-objective optimisation for decision support in production systems development
Febriantono et al. Classification of multiclass imbalanced data using cost-sensitive decision tree C5. 0
CN116320021A (en) Hierarchical service matching method based on intention in Internet of things scene
CN104216993A (en) Tag-co-occurred tag clustering method
CN102426585A (en) Webpage automatic classification method based on Bayesian network
Momin et al. Prediction of customer churn using machine learning
Qin et al. Enhanced neural gas network for prototype-based clustering
Ali et al. Fake accounts detection on social media using stack ensemble system
CN112486467B (en) Interactive service recommendation method based on dual interaction relation and attention mechanism
Mishra et al. A Survey on Data mining classification approaches
Prince et al. An Imbalanced Dataset and Class Overlapping Classification Model for Big Data.
CN103425716B (en) Main characteristic analysis method based on propositional logic and system in a kind of data mining
CN102193928A (en) Method for matching lightweight ontologies based on multilayer text categorizer
CN114896514B (en) Web API label recommendation method based on graph neural network
Elisa et al. A study of the necessity of signal categorisation in dendritic cell algorithm
CN114268625B (en) Feature selection method, device, equipment and storage medium
CN114708015A (en) Service demand prediction method adopting big data deep mining and big data system
Sellami et al. An implicit approach for building communities of web service registries
Alyahyan et al. Feature level ensemble method for classifying multi-media data
Chen et al. Modelling of experienced-based data in linked data environment
Liu et al. Pattern classification based on the combination of the selected sources of evidence
Le et al. Current practices in measuring ontological concept similarity

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination