CN104732100B - A kind of service of unidirectional chain form can sort the matching degree computational methods of exact properties - Google Patents
A kind of service of unidirectional chain form can sort the matching degree computational methods of exact properties Download PDFInfo
- Publication number
- CN104732100B CN104732100B CN201510156104.6A CN201510156104A CN104732100B CN 104732100 B CN104732100 B CN 104732100B CN 201510156104 A CN201510156104 A CN 201510156104A CN 104732100 B CN104732100 B CN 104732100B
- Authority
- CN
- China
- Prior art keywords
- attribute
- matching degree
- service
- demand
- node
- 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.)
- Active
Links
Abstract
Can sort the matching degree computational methods of exact properties the present invention relates to a kind of service of unidirectional chain form, and in traditional production, manufacture and electric business platform, the parameter of a part of attribute of service/product is accurate.The absolute unification of the parameter of both demand and attribute specification or compatibility.Some can also meet that the product of user's request or service are rejected from alternative sample set.The attribute of alternative services/product is ranked up according to the value of the parameter of these attributes, forms a unidirectional sequence, and provide the calculating formula of similarity of each node and destination node on the chain by the present invention.According to calculating formula of similarity, the matching degree between each attribute of demand and each attribute of service/product can be calculated, reuses weighting scheme, obtains demand and service/product overall matching degree.This method flexibility ratio is big, more meets the essential requirement of user, is a kind of efficient, quick, matching degree computational methods of close to the need.
Description
Technical field
A kind of service of unidirectional chain form of present invention design can sort the matching degree computational methods of exact properties, party's normal plane
It is accurate to some property parameters values, and the calculating of the matching degree for the attribute that can be ranked up.So as to solve with such
In the service of attribute and the matching process of product, its matching way is excessively strict and causes what sample set scale too reduced to ask
Topic, it belongs to manufacturing field.
Background technology
In today's society, reached its maturity in the form of the network service by cloud thought for core.All kinds of cloud computings, manufacture and
The electric business platform moment affects our life.These platform operations and the mainly all kinds of service/products of sale.These clothes
The description of business/product is extremely important, the operating scheme of rational description method meeting extreme influence to other modules of platform, while
Influence whether the selection of user.Most common at present and successful description method is structural method.This method is by the service/production
Each generic attribute, the parameter of attribute and the value of parameter of product describe one by one, are then stored by relational database.In order to realize
Its reasoning and combination etc. function, some also occur successively such as the describing mode of description language etc., such as in web services side
The WSDL and OWL-S in face etc..The information of one description service/product, its content is all properties and its ginseng of the service/product
What numerical value was formed.So each service/product can be indicated in table form.Service/production for same type
Product, it should have very close representation.Service/product description form can be Service=< Sd1, Sd2..., Sdn
>, wherein Sd1, Sd2..., SdnN attribute of service/product is represented respectively.
For current this service/product description form, the matching for demand and service/product, often use to every
Individual attribute matches one by one, finally carries out the mode of aggregative weighted.Matching for each attribute, typically there is based on semanteme, base
In reasoning from logic or method based on body etc..It is all fuzzy attribute that this kind of method, which is often applied to some parameter values, and
Some have the matching of the exact properties of clear and definite functional compatibility.For Fog property, the generic attribute occurrence includes certain language
Adopted information, but existence form is more random.Particularly, the value of these attributes has clear and definite implication, and the mankind read these
After value, the concrete condition of the attribute just can be more or less understood.Meanwhile these property values define without very specification or accurately
Rule, so its form may be varied.Such as the title of service.Same service, provided, may made by different suppliers
With different names.And the service that function is close, in fact it could happen that its name is also highly close or even identical.For some parameters
Value is accurate exact properties, and it is the exact properties for having clear and definite functional compatibility.Input/output such as in web services connects
Mouthful standard, it is assumed that certain input interface is required of double type data, often its can also compatible int types data.Such as
The matching that the method for body etc. corresponds to the generic attribute also has good effect.For some exact properties, such as carrying for service/product
For business, often without comparativity, typically matched using boolean's mode.Although this method is simple, meet user's
Demand, also there is preferable effect.But for a generic attribute, its parameter value be it is accurate, but entirely different two services/
Even if its generic attribute of product is different, the matching degree that can not utterly determine the generic attribute is 0.Meanwhile the generic attribute is past
It is past it to be ranked up according to different demand types again.Such as the CPU of computer, if user's request is one 4 cores
2.0GHz dominant frequency, and have two CPU, 1. 4 core 2.2GHz dominant frequency and 2. 2 core 1.6GHz dominant frequency.Substantially 1. number CPU can more approach use
Family demand, while if without user's request CPU when, user selection 1. a number CPU be also acceptable.Similarly for more category
Property requirement, as demand is CPU and video card is proposed the computer of design parameter value requirement, for this kind of service/product
Matching, it is excessively simple strict if using boolean's mode.
For this problem, the present invention devises a kind of service of unidirectional chain form and can sorted the matching degree meters of exact properties
Calculation method.Numerical value between the matching degree of this generic attribute can be changed into 0 to 1 by this method, is blurred.Propose as user
Demand provides the matching scheme of immediate service/product.
The purpose of the present invention is:Property parameters value is accurate and can be ranked up the matching of attribute, change into it is a kind of according to
The calculating of the fuzzy matching degree of user's request.By such calculating, the part attribute for needing accurately to match originally is changed into mould
Paste matching.The service/product in alternative sample set can be greatly protected using such method, while can also provide and be more conform with
The service/product of demand.
The content of the invention
The present invention is to solve the above problems, proposing a kind of service of unidirectional chain form can sort the matching degrees of exact properties
Computational methods.This method is established first can ordering attribute.Then be directed to per one kind can ordering attribute structure rule, form it into one
The directive unidirectional chain of bar.Then matching degree calculation formula is built, according to demand the position with the attribute of service/product in the chain
Put calculating matching degree.Other kinds of attribute (Fog property or other types exact properties) calculates matching using conventional method
Degree.The matching degree between demand and service/product is finally calculated using the mode of weighting
The technical solution adopted by the present invention is:A kind of service of unidirectional chain form can sort the matching degree calculating of exact properties
Method, this method include 3 steps:
Step 1), can ordering attribute establish
For a kind of service/product, its attribute that can be sorted is established first.I.e. which attribute is to go for this method
's.For the attribute for not being suitable for this method matching degree is calculated using other conventional methods.
Step 2), the ordering rule for building attribute
Attribute for going for this method, a rule is built, the rule can be according to the parameter of the generic attribute
And the value of parameter, the generic attribute is ranked up, obtains a directive attribute chain.A node on the chain is exactly such
A sample under attribute.
Step 3), structure attributes match degree calculation formula
By a matching degree calculation formula, node corresponding to the attribute of demand and the clothes in sample set on unidirectional chain are calculated
The similarity of node corresponding to the attribute of business/product, the similarity are defined as matching degree.
Step 4), calculate comprehensive matching degree
The matching degree of multiple attributes is calculated, by weighting scheme, calculates final matching degree.
Further, step 1 propose establish can ordering attribute thought, specific implementation is as follows:
In the description method of service/product, usually described one by one according to its attribute.Such as a calculating
The manufacturer of machine, CPU, display, introduce explanation etc..In these attributes, the attribute as introducing explanation etc., its content is often mould
Paste, be random and inaccurate.Method in terms of the general use information retrieval of this generic attribute calculates its matching degree.And as manufacturer
A kind of attribute, or demand and sample comply fully with, or not meeting completely, the matching degree of this generic attribute is Boolean type.
And often there are following several features for CPU or display a kind of attribute, this generic attribute:1. the demand of this generic attribute is past
It is past not to be strict.Even if i.e. some of samples can not be in strict conformity with the requirement of demand, but can also be connect by user
By;2. the parameter and parameter value of this generic attribute are often accurate, such as CPU, its parameter mainly includes core number and dominant frequency etc., this
Value corresponding to a little parameters and accurate numeral or word;3. simultaneously because the objective factor such as scientific and technological level, causing it can be with
It is ranked up.This property that sorts is service and product will meet scientific development requirement and it is objective caused by.Step 1
Seek to establish this kind of service.
Further, step 2 selects a kind of sortord on the basis of step 1, to this kind of category suitable for this method
Property, the rule of sequence is built, is ranked up these attributes.By taking the attribute CPU in computer as an example, it can collect at present
Common all CPU data a, there is provided rule, these CPU are connected into a unidirectional chain.The rule may have it is a variety of, such as
According to performance, according to some parameters etc..Specifically to see its applicable scene.
Further, step 3 builds the matching degree calculation formula of attribute on the basis of step 1 and 2.Specific implementation is to obtain
Take the position of the corresponding node in the unidirectional chain of the generic attribute of attribute corresponding to demand and sample set.Build a formula, the formula
According to the distance between node and direction, the matching degree of the attribute of demand and the attribute of each sample is calculated.
Further, in final step 4, other non-this method are applicable attribute and calculate matching degree using conventional method, finally
The matching degree of all properties is weighted, obtains final matching degree.
The present invention compared with prior art the advantages of be:
1. obtaining the attributes match degree of service/product relatively reasonablely, will not be caused because of conventional methods such as boolean's modes
Service/product quantity in alternative sample set is reduced, and the recall ratio of matching result can be guaranteed.
2. this method more meets the essence of user search demand, closer to the real demand of user.
3. this method is simple and quick, computation complexity is low and easy realization.It is complicated to the matching primitives of all kinds of service/products
Spend relatively low, matching speed is very fast.
Term is explained:
Sample set:Need the set of service/product being matched.The each specifically service/production of wherein each sample representation
Product.
Attribute:Description to service/product.In computer, main attribute includes CPU, video card, hard disk, display
Deng.
Parameter:Description to attribute.Such as CPU dominant frequency, core number, caching.
Parameter value:The occurrence of parameter.Dominant frequency such as CPU is 2.2GHz, and hard disk type is SSD etc..
Brief description of the drawings
Fig. 1 is this method overall schematic;
Fig. 2 is unidirectional chain schematic diagram;
Fig. 3 is comprehensive matching schematic diagram.
Embodiment
The present invention is described in further detail below in conjunction with the accompanying drawings.
The present invention is that a kind of service of unidirectional chain form can sort the matching degree computational methods of exact properties, as shown in figure 1,
Mainly include four parts:Can ordering attribute establish, build the ordering rule of attribute, build attributes match degree calculation formula and meter
Calculate comprehensive matching degree.
(1) can ordering attribute establishment
Establish which attribute meets using this method by the main rule of following three:
1. for demand, the generic attribute need not be definitely strict.I.e. user can receive the attribute can not be strict
The service/product matched somebody with somebody;
2. the parameter value of attribute is accurate, here accurate refers to that parameter value is fixed numeral or some standards.
Such as CPU core number, the type of hard disk etc.;
3. a rule can be found, this generic attribute can be ranked up.This rule meets demand trend.Such as
For CPU, the mode for comparing science is according to its actual performance ranking.And in practice, often user is concerned about it first
Core number, followed by dominant frequency, finally it is only other attributes.
(2) ordering rule of attribute is built
Ordering rule is key here, either matches purpose towards different demands, might have different sequences
Rule.Such as CPU, the mode for comparing science is according to its actual performance ranking.And in practice, often user is first
It is concerned about its core number, followed by dominant frequency, is finally only other attributes.Both ordering rules are all acceptables.For special
Industry user, first way can be used, it is more particularly suitable for domestic consumer, the second way.Explanation should in the second
Method:
By taking the CPU of mobile phone as an example, we often focus first on CPU core number, followed by dominant frequency during choosing,
Almost seldom consider as manufacture craft, framework mode etc..So we can be directed to the demand of user's inquiry in service library
The CPU models that all mobile phone manufacturing services provide are ranked up.For example user's search is n core number.That is just first with core
Number is standard, and core is few to be leaned on to the left side, being leaned on to the right more than core.If core number is identical, compare dominant frequency, dominant frequency is small
Keep left, dominant frequency is big to keep right.How the two all, then side by side.So all mobile phones are just made the CPU in service by we
A chain is lined up.All CPU with n core, it should there is identical matching degree, while this matching degree should be most
Big.Non- n CPU, distance have the CPU of n core its more remote matching degree of center should be lower.So this CPU line is just
Form CPU directed chain.Pass through this directed chain, it is possible to calculate two CPU gap.Its schematic diagram such as Fig. 2 institutes
Show
After user claims to CPU core number, system lines up a directive chain according to above-mentioned rule.If
Wherein No. 4 to No. 8 CPU core number meets the needs of user.Then No. 4 matching degrees to No. 8 CPU are 1.It is now in centre bit
The CPU put is No. 6 CPU, point centered on it, calculates other CPU and its distance.Get at other ungratified nodal distance centers
It is remote then its matching degree is lower.
(3) attributes match degree calculation formula is built
One formula for being used to calculate based on above-mentioned unidirectional chain of structure.The formula need to meet following several requirements:
1. matching degree section should be in [0,1];
2. the distance between two nodes are more remote, its matching degree is smaller.Two wherein farthest nodal communitys are close to 0,
And satisfactory attribute is 1;
3. the formula is specifically contemplated that directionality.Have on the destination node left side and the right with equidistant two nodes with
The distance of target should difference.The purpose for the arrangement is that consider some actual requirements.As requirement of the user to CPU be as
It is more preferable that fruit does not match then tendency suitably.So it should be slightly larger than according to above-mentioned unidirectional chain, the matching degree of right node
The left side.
Here it is as follows to provide a calculation formula 1:
Wherein m represents the node of Incomplete matching and represents Far Left node to rightmost to the centre distance of matched node, n
The distance of node, k >=0 are parameters so that the direction of demand tendency can obtain higher matching degree.
(4) comprehensive matching degree is calculated
Matching generally for service/product is the comprehensive matching to multiple attributes.Other kinds of attributes match uses
Currently used method, such as Fog property and non-this method, applicable attribute, can use such as reasoning, semanteme or body side
Method.Applicable attribute calculates matching degree using method proposed by the present invention.Matching degree corresponding to service/product should consider multiple
The integrated value of attribute.In general way is exactly to be weighted method.As shown in Equation 2:
Wherein, ω1, ω2, ω3..., ωmWeights corresponding to m attribute are represented, and meet ω1+ω2+ω3+…+ωm=
1。
Below by taking a two-dimentional Service Matching as an example, it is assumed that now with a service, it includes two exact properties, and
Exact properties are sorted, its schematic diagram is as shown in Figure 3:
In this drawing, Target Service are destination service, and No. 4 and No. 5 in wherein attribute i are objective attribute target attributes,
No. 7 and No. 8 are objective attribute target attribute in attribute j.It is more prone to simultaneously with configuring higher attribute, the higher first step is defaulted as 0.5.It is existing
There are 4 services, calculating this four services and the matching degree of destination service.
It can be drawn by upper figure, the central point position of attribute (i) No. 4 and No. 5, wherein, the n=8 of i attributes, make k=0.5 can
To draw the distance of attribute (i) as [3,2,1,0,0,0.5,1.5,1.5,2.5,3.5], matching degree for [0.625,0.75,
0.875,1,1,0.9375,0.8125,0.8125,0.6875,0.5625] the central point position of attributes (j) No. 7 and No. 8, j attributes
N=8, it is [5,4,3,2,1,0,0,0.5,1.5] to make k=0.5 obtain the distance of attribute (j), matching degree for [0.375,0.5,
0.625,0.75,0.75,0.875,1,1,0.9375,0.8124].Make ω1=0.7, ω2=0.3:
sim(serviceT, service1)=1 × 0.7+0.375 × 0.3=0.8125
sim(serviceT, service2)=0.75 × 0.7+0.625 × 0.3=0.7125
sim(serviceT, service3)=0.9375 × 0.7+0.75 × 0.3=0.8825
sim(serviceT, service4)=0.8125 × 0.7+0.875 × 0.3=0.83125
So can obtain matching degree order be:
sim(serviceT, service3) > sim (serviceT, service4) > sim (serviceT, service1)
> sim (serviceT, service2)
The techniques well known being related in the present invention does not elaborate.
Claims (1)
- A kind of matching degree computational methods of exact properties 1. service of unidirectional chain form can sort, it is characterised in that this method bag Include 3 steps:Step 1), can ordering attribute establishFor a kind of service/product, its attribute that can be sorted is established;Step 2), the ordering rule for building attributeAttribute for going for this method, a rule is built, the rule can be according to the parameter and ginseng of the generic attribute Several values, the generic attribute is ranked up, one directive attribute chain of acquisition, each node on the chain is exactly the generic A sample under property;Step 3), structure attributes match degree calculation formulaBy a matching degree calculation formula, calculate node corresponding to the attribute of demand and the service in sample set on unidirectional chain/ The similarity of node corresponding to the attribute of product, the similarity are defined as matching degree;Step 4), calculate comprehensive matching degreeThe matching degree of multiple attributes is calculated, by weighting scheme, calculates final matching degree;Wherein, step 1) propose establish can ordering attribute thought;Attribute has following several features:1. the demand of this generic attribute is not strict, though that is, some of samples can not in strict conformity with the requirement of demand, But it can also be easily accepted by a user;2. the parameter and parameter value of this generic attribute are often accurate, value corresponding to these parameters and accurate numeral or text Word;3. a rule can be found, this generic attribute can be ranked up;Step 1) seeks to establish this kind of service;Step 2) builds the ordering rule of attribute on the basis of step 1), and specific implementation is as follows:A kind of sortord is selected, to this kind of attribute suitable for this method, the rule of sequence is built, enters these attributes Row sequence;Step 3) in step 1) and 2) on the basis of build the matching degree calculation formula of attribute, specific implementation is as follows:The position of the corresponding node in the unidirectional chain of the generic attribute of attribute corresponding to acquisition demand and sample set, a formula is built, The formula calculates the matching degree of the attribute of demand and the attribute of each sample, finally according to the distance between node and direction In step 4), other non-this method are applicable attribute and calculate matching degree using conventional method, and finally the matching degree of all properties is entered Row weighting, obtains final matching degree;It is as follows to provide a calculation formula (1):Wherein m represents the node of Incomplete matching and represents Far Left node to rightmost node to the centre distance of matched node, n Distance, k >=0 is parameter so that the direction of demand tendency can obtain higher matching degree.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510156104.6A CN104732100B (en) | 2015-04-03 | 2015-04-03 | A kind of service of unidirectional chain form can sort the matching degree computational methods of exact properties |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510156104.6A CN104732100B (en) | 2015-04-03 | 2015-04-03 | A kind of service of unidirectional chain form can sort the matching degree computational methods of exact properties |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104732100A CN104732100A (en) | 2015-06-24 |
CN104732100B true CN104732100B (en) | 2018-01-23 |
Family
ID=53455982
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510156104.6A Active CN104732100B (en) | 2015-04-03 | 2015-04-03 | A kind of service of unidirectional chain form can sort the matching degree computational methods of exact properties |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104732100B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
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 |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101398852A (en) * | 2008-10-16 | 2009-04-01 | 北京邮电大学 | Semantic Web service matching method and system |
CN101820444A (en) * | 2010-03-24 | 2010-09-01 | 北京航空航天大学 | Descriptive information matching similarity-based resource service matching and searching method |
CN101833561A (en) * | 2010-02-12 | 2010-09-15 | 西安电子科技大学 | Natural language processing oriented Web service intelligent agent |
CN102760138A (en) * | 2011-04-26 | 2012-10-31 | 北京百度网讯科技有限公司 | Classification method and device for user network behaviors and search method and device for user network behaviors |
CN104035927A (en) * | 2013-03-05 | 2014-09-10 | 百度在线网络技术(北京)有限公司 | User behavior-based search method and system |
-
2015
- 2015-04-03 CN CN201510156104.6A patent/CN104732100B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101398852A (en) * | 2008-10-16 | 2009-04-01 | 北京邮电大学 | Semantic Web service matching method and system |
CN101833561A (en) * | 2010-02-12 | 2010-09-15 | 西安电子科技大学 | Natural language processing oriented Web service intelligent agent |
CN101820444A (en) * | 2010-03-24 | 2010-09-01 | 北京航空航天大学 | Descriptive information matching similarity-based resource service matching and searching method |
CN102760138A (en) * | 2011-04-26 | 2012-10-31 | 北京百度网讯科技有限公司 | Classification method and device for user network behaviors and search method and device for user network behaviors |
CN104035927A (en) * | 2013-03-05 | 2014-09-10 | 百度在线网络技术(北京)有限公司 | User behavior-based search method and system |
Also Published As
Publication number | Publication date |
---|---|
CN104732100A (en) | 2015-06-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107515938B (en) | Intelligent supply and demand matching method in cloud manufacturing environment | |
CN101496003B (en) | Compatibility scoring of users in a social network | |
WO2020207074A1 (en) | Information pushing method and device | |
CN110580649B (en) | Method and device for determining commodity potential value | |
CN107832338B (en) | Method and system for recognizing core product words | |
CN103902538A (en) | Information recommendation device and method based on decision-making tree | |
CN103946842A (en) | Search device, search method, search program, and storage medium | |
KR102601545B1 (en) | Geographic position point ranking method, ranking model training method and corresponding device | |
CN103294677A (en) | Searching method of electronic commerce search engine and electronic commerce search engine | |
CN108280234A (en) | Data query method and device | |
CN109636168A (en) | A kind of cloud platform scheduling resource intelligent matching process based on Ontology | |
CN111813951A (en) | Key point identification method based on technical map | |
CN104077327B (en) | The recognition methods of core word importance and equipment and search result ordering method and equipment | |
CN107016566A (en) | User model construction method based on body | |
CN109582868A (en) | The search recommended method of preference is clicked based on term vector weighting, support vector regression and user | |
CN113010752B (en) | Recall content determining method, apparatus, device and storage medium | |
CN104732100B (en) | A kind of service of unidirectional chain form can sort the matching degree computational methods of exact properties | |
CN114491236A (en) | Intelligent supply chain matching technology based on knowledge learning | |
CN103020083A (en) | Automatic mining method of requirement identification template, requirement identification method and corresponding device | |
CN112765490A (en) | Information recommendation method and system based on knowledge graph and graph convolution network | |
TWI684147B (en) | Cloud self-service analysis platform and analysis method thereof | |
CN106991122A (en) | A kind of film based on particle cluster algorithm recommends method | |
CN116228301A (en) | Method, device, equipment and medium for determining target user | |
CN114429384B (en) | Intelligent product recommendation method and system based on e-commerce platform | |
CN116229211A (en) | Sample generation method, model training method, object detection method and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |