CN112084322A - Airworthiness case recommendation method based on conformance vector - Google Patents

Airworthiness case recommendation method based on conformance vector Download PDF

Info

Publication number
CN112084322A
CN112084322A CN202010753933.3A CN202010753933A CN112084322A CN 112084322 A CN112084322 A CN 112084322A CN 202010753933 A CN202010753933 A CN 202010753933A CN 112084322 A CN112084322 A CN 112084322A
Authority
CN
China
Prior art keywords
node
similarity
compliance
conformance
airworthiness
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
CN202010753933.3A
Other languages
Chinese (zh)
Inventor
蔡喁
郝秀兰
申岳
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.)
China Civil Aviation Shanghai Aircraft Airworthiness Certification Center
Original Assignee
China Civil Aviation Shanghai Aircraft Airworthiness Certification Center
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 China Civil Aviation Shanghai Aircraft Airworthiness Certification Center filed Critical China Civil Aviation Shanghai Aircraft Airworthiness Certification Center
Priority to CN202010753933.3A priority Critical patent/CN112084322A/en
Publication of CN112084322A publication Critical patent/CN112084322A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/335—Filtering based on additional data, e.g. user or group profiles
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/31—Indexing; Data structures therefor; Storage structures
    • G06F16/316—Indexing structures
    • G06F16/322—Trees
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/3331—Query processing
    • G06F16/334—Query execution
    • G06F16/3344—Query execution using natural language analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
    • G06F16/367—Ontology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Databases & Information Systems (AREA)
  • Computational Linguistics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Computation (AREA)
  • Evolutionary Biology (AREA)
  • Animal Behavior & Ethology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Software Systems (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The invention relates to the technical field of airworthiness certification of aircrafts, in particular to an airworthiness case recommendation method based on a conformance vector, which comprises the steps of firstly obtaining a similarity measurement of a conformance object and conformance activities; then obtaining similarity measurement of the coincidence target; and finally, recommending the compliance case by using a Top-k algorithm according to the similarity measurement obtained in the last two steps, and establishing a similarity measurement relation among compliance objects, compliance activities and the compliance case in a body knowledge base on the basis of the body knowledge base in the airworthiness examination field, so that the values of three dimensional weight values of the compliance objects, the compliance objects and the compliance activities in the current test are reasonable, and the influence of the category of the compliance activity objects, the applicability of regulation terms and other influence factors on the recommendation of the compliance case is avoided.

Description

Airworthiness case recommendation method based on conformance vector
Technical Field
The invention relates to the technical field of airworthiness certification of aircrafts, in particular to an airworthiness case recommendation method based on a conformance vector.
Background
One key use of ontology knowledge bases in the field of airworthiness review is in the discovery of compliance experience. In an actual scene, due to the sparsity of case experience, the flexibility of design and the freedom of selection of a compliance strategy, the difficulty in accurately matching the case experience is high. For example, the electromagnetic compatibility test of the landing gear brake control system cannot necessarily find an accurate case match in a knowledge base, in the prior art, precious experience is obtained through similar electromagnetic compatibility tests developed by other systems or equipment for reference, but the used methods, judgment indexes, and used regulations and standards may be different for similar tests of different objects, so that the correlation between the category of the compliance activity object and the compliance case has a significant influence, and the applicability of the regulation clauses has a significant influence on the correlation of the relevant compliance activity experience case.
Therefore, it is necessary to design a method for recommending qualified cases after similarity measurement is performed on components in three dimensions of the compliance object, the compliance target and the compliance activity.
Disclosure of Invention
The invention breaks through the difficult problems in the prior art, and designs a method which can measure the similarity of the components in three dimensions of the conformance object, the conformance target and the conformance activity and then recommend the qualified case.
In order to achieve the purpose, the invention designs a airworthiness case recommendation method based on a conformance vector, which is characterized in that: the method comprises the following steps:
s1, obtaining similarity measurement of the coincidence object and the coincidence activity;
s2, obtaining similarity measurement of the coincidence target;
s3 recommendation of compliance case.
Further, the specific calculation method of the similarity measure of the compliance object and the compliance activity of S1 includes the following steps:
s11, acquiring activity similarity based on the shortest path length;
s12, obtaining node similarity based on depth;
s13, acquiring the specific degree of the node;
s14, calculating the node similarity based on the node concrete degree according to the node concrete degree;
s15 calculates the node similarity of the compliance object and the compliance activity.
Further, the calculation formula of the activity similarity based on the shortest path length obtained by S11 is:
Figure BDA0002610915140000021
wherein, SP (node)1，node2) Represents a node1And a node2The length of the shortest path between two nodes, height represents the maximum depth of the ontology tree, node represents node, SimSP(node1，node2) Represents a node1And a node2Active similarity of shortest path length between two nodes.
Further, the calculation formula of the depth-based node similarity obtained in S12 is as follows:
Figure BDA0002610915140000022
wherein, depth (node)1) Refers to a node1Depth in the ontology tree, Simdepth(node1，node2) Represents a node1And a node2And (4) activity similarity of two nodes in the depth of the ontology tree.
Further, in the above-mentioned case,
Figure BDA0002610915140000023
Figure BDA0002610915140000031
further, the calculation formula of the node specific degree obtained in S13 is as follows:
Figure BDA0002610915140000032
Figure BDA0002610915140000033
wherein Spec (node)1，node2) Representing a node1Relative node2The degree of specificity of (a).
Further, the calculation formula of S14 calculating the node similarity based on the node specificity according to the node specificity is as follows:
Figure BDA0002610915140000034
Figure BDA0002610915140000035
wherein NCA (node)1，node2) Representing concept node1And a node2Nearest common ancestor node, Sim, in the ontology treeSpec(node1，node2) Representing concept-based node1And a node2Node similarity of specific degrees.
Further, the calculation formula for calculating the node similarity in S15 is as follows:
Simonto(node1，node2)＝αSimSP(node1，node2)+βSimdepth(node1，node2)+γSimRF(node1，node2) Wherein, α, β, γ are parameter factors, α + β + γ is 1.
Further, the specific method for acquiring the similarity measure of the conformance target in S2 is as follows: firstly, selecting a regulation example, and measuring the similarity of the regulation example by using the Jaccard similarity; secondly, the inverse document frequency idf is used as weight, so that the influence of high-frequency regulations on the similarity of the examples is reduced; carrying out normalization processing; based on the weight of each regulatory act, a Jaccard similarity metric is performed.
Further, the algorithm for recommending the S3 compliance case adopts a top-k recommendation algorithm, which includes the following steps:
s31, acquiring a data set SD formed by the ordered list;
s32 performing sequential access to the data set SD;
s33 recording each dimension with a set to the position seen so far in sequential access;
s34, calculating a score function value of the optimal position;
s35 adds all the censored instances with score function values equal to the k instance score values to the set Ak;
and the output set Ak of S36 is a top-k set.
The invention also designs a airworthiness case recommendation system based on the conformance vector, which is characterized in that: the method comprises the following steps: the airworthiness examination ontology knowledge base comprises two ontology instances, wherein a conformance object and conformance activity are represented in a tree hierarchical relationship, and a conformance target is represented in a tag set;
an activity similarity calculation module for calculating a similarity measure of the compliance object and the compliance activity;
a target similarity calculation module for calculating a compliance target similarity measure;
a recommendation module for compliance case recommendation.
Further, the activity similarity calculation module includes: a first calculation unit for obtaining activity similarity based on the shortest path length;
a second calculation unit for acquiring node similarity based on depth;
a third calculating unit for obtaining the node concrete degree similarity;
and the fourth calculation unit is used for summarizing and calculating the node similarity.
The invention also designs a airworthiness case recommendation device, which is characterized in that: comprises a processor;
a memory for storing processor-executable instructions;
wherein the processor is configured to: the following method steps are carried out: firstly, acquiring similarity measurement of a conformance object and conformance activity; then obtaining similarity measurement of the coincidence target; and finally, recommending a coincidence case by using a Top-k algorithm according to the similarity measurement obtained in the last two steps.
The specific calculation method of the similarity measurement of the compliance object and the compliance activity is as follows: acquiring activity similarity based on the shortest path length; acquiring node similarity based on depth; acquiring specific degree of a node; calculating the node similarity based on the node concrete degree according to the node concrete degree; and calculating the node similarity of the compliance object and the compliance activity.
The present invention also contemplates a computer storage medium having computer program instructions stored thereon, characterized in that: the computer program instructions are executed by a processor in a method comprising: firstly, acquiring similarity measurement of a conformance object and conformance activity; then obtaining similarity measurement of the coincidence target; and finally, recommending a coincidence case by using a Top-k algorithm according to the similarity measurement obtained in the last two steps.
The specific calculation method of the similarity measurement of the compliance object and the compliance activity is as follows: acquiring activity similarity based on the shortest path length; acquiring node similarity based on depth; acquiring specific degree of a node; calculating the node similarity based on the node concrete degree according to the node concrete degree; and calculating the node similarity of the compliance object and the compliance activity.
Compared with the prior art, the method and the device have the advantages that the similarity measurement relation among the conformity objects, the conformity activities and the conformity cases in the ontology knowledge base is established on the basis of the ontology knowledge base in the airworthiness examination field, so that the values of the three-dimensional weight values of the conformity targets, the conformity objects and the conformity activities in the current test are reasonable, and the influence of the category of the conformity activity objects, the applicability of the regulation clauses and other influence factors on the recommendation of the conformity cases is avoided.
Drawings
Fig. 1 is a flow chart illustrating a airworthiness case recommendation method based on a conformance vector in an embodiment.
Fig. 2 is a schematic flow chart illustrating a process of calculating similarity measures of a compliance object and a compliance activity in the airworthiness case recommendation method based on a compliance vector in an embodiment.
Fig. 3 is a flowchart illustrating an algorithm of compliance case recommendation in the airworthiness vector-based airworthiness case recommendation method in an embodiment.
Fig. 4 is a schematic structural diagram of a airworthiness case recommendation system based on a conformance vector in an embodiment.
Detailed Description
The invention will be further described with reference to the accompanying drawings, but is not to be construed as being limited thereto.
The invention provides a airworthiness case recommendation method based on a conformity vector, which defines any case instance as components on three dimensions of a conformity object, a conformity target and a conformity activity:
one compliance case: (compliance object, compliance goal, compliance activity);
the consistency object and the consistency activity are two ontology instances and appear in the dimension of tree hierarchical relation, and the consistency target appears in the form of a tag set.
In a specific implementation, the invention comprises the following steps:
s1, obtaining similarity measurement of the coincidence object and the coincidence activity;
s2, obtaining similarity measurement of the coincidence target;
s3 recommendation of compliance case.
Preferably, for the similarity measure of the coincidence object and the coincidence activity in S1, the semantic similarity calculation of the node is mainly considered from the following points of view because of two conceptual dimensions obviously presenting a tree hierarchy relationship:
(1) number of branches from one node to another: the shortest path length measure can be used, namely, the relative value of the path length and the tree height and the tree width is used for representing; or by semantic distance measure, i.e. the relative value of path length and tree height; or measured in path distance, i.e. expressed in inverse path length;
(2) the hierarchy at which the two nodes are located: the similarity of two nodes with the same semantic distance is increased along with the increase of the sum of the levels where the two nodes are positioned, and the depth and the level depth can be used for measuring in the prior art;
(3) the number of the same upper concepts contained among the internal concepts of the ontology: the degree of refinement and the semantic overlap ratio can be used for measuring, and the depth information of the common ancestor of the two nodes are used;
(4) child brother node information of node and its ancestor: the node density can be used for measurement, and brother and child node information of two nodes is used; or may be measured as the ratio of the degree of the common ancestor node to the maximum node degree of the ontology tree;
(5) parameter factors: and setting a weight parameter for each similarity according to calculation requirements, thereby determining the similarity required by the system and adjusting the similarity value between the concepts according to the similarity.
In the embodiment, by analyzing the characteristics of the ontology knowledge base, the similarity measurement of the conformance object and the conformance activity is comprehensively analyzed by adopting three indexes of shortest path length, concept depth and concept specificity and combining parameter factors.
Setting alpha, beta and gamma as parameter factors, respectively, and using node1，node2Representing two nodes, in the calculation, if the object is a coincidence object, the node is represented by o1，o2Substitute for node1，node2(ii) a If it is a compliance activity, use a1，a2Substitute for node1，node2。
A. Shortest Path length (Shortest Path)
The path length refers to the number of branches that pass from one node to another. The shortest path between two nodes refers to the one having the shortest length among all paths connecting the two nodes in the ontology tree. The shorter the path between nodes, the higher the similarity between nodes.
In this embodiment, the height of the ontology tree is 5 levels, the width is about 900, and since the width is much larger than the height, the width is not suitable for being included in the calculation formula, and the specific calculation formula is as follows:
Figure BDA0002610915140000081
wherein, SP (node)1，node2) Represents a node1And a node2The length of the shortest path between two nodes, height represents the maximum depth of the ontology tree, node represents node, SimSP(node1，node2) Represents a node1And a node2Active similarity of shortest path length between two nodes.
B. Concept Depth (Depth)
The concept depth refers to the depth of the node in the ontology tree, and the similarity of the nodes increases with the increase of the sum of the levels of the nodes.
Therefore, in the invention, the similarity based on the node depth is calculated according to the following formula:
Figure BDA0002610915140000091
Figure BDA0002610915140000092
wherein, depth (node)1) Refers to a node1Depth in the ontology tree, Simdepth(node1，node2) Represents a node1And a node2And (4) activity similarity of two nodes in the depth of the ontology tree.
C. Concept details (Specification)
In an ontology tree, a concept's children are all materializations of its ancestor node concept at one time, any two concept nodes (nodes)1，node2) All have a nearest common ancestor node NCA (node) in the ontology tree1，node2) The farther the node is from the nearest common ancestor node, the higher the concrete degree of the two nodes is, and the smaller the similarity between the corresponding two nodes is.
In the calculation, the concrete degrees of two nodes are calculated firstly:
Figure BDA0002610915140000093
wherein Spec (node)1，node2) Representing a node1Relative node2The degree of specificity of (a).
Then, calculating the similarity between the two nodes based on the concrete degree:
Figure BDA0002610915140000101
wherein NCA (node)1，node2) Representing concept node1And a node2Nearest common ancestor node, Sim, in the ontology treeSpec(node1，node2) Representing concept-based node1And a node2Node similarity of specific degrees.
After the node similarity of the three different indexes is calculated, the similarity of the points is calculated by adopting the following formula:
Simonto(node1，node2)
＝αSimSP(node1，node2)+βSimdepth(node1，node2)
+γSimRF(node1，node2)
wherein, α, β, γ are parameter factors, α + β + γ is 1.
In particular implementations, the correspondence between the compliance targets and the "compliance objects" and "compliance activities" is not fixed in the compliance target dimension. Moreover, most of the rules and standards of the nearly thousand rules are in parallel relationship, and do not influence each other or influence weakly. Most activities under actual scrutiny will satisfy the requirements of more than one regulatory clause. Whereas high frequency "regulatory terms" have relatively less significance to the proximity calculation than low frequency "regulatory terms," therefore, similar weighting techniques as TF-IDF are used in making the proximity measure for the regulatory terms dimension.
Preferably, the specific method for obtaining the similarity measure of the compliance target includes: two regulatory examples are first chosen and labeled TA and TB, and the Jaccard similarity metric calculates the degree of similarity of TA and TB using the following formula:
Figure BDA0002610915140000111
secondly, using the inverse document frequency idf as a weight to reduce the influence of high-frequency rules on the similarity of the example, wherein the calculation formula is as follows:
Figure BDA0002610915140000112
where | A | is the total number of censorship activities in the dataset, | { k: t |, andi∈akis an inclusion rule tiThe number of censorship activities of (1) to prevent the occurrence of a regulation not occurring in the data set, results in wtiThe denominator of (1) is 0, so that in general, 1 is added to the denominator.
For an inspection activity akAnd normalizing the weight of the regulation i:
Figure BDA0002610915140000113
wherein m iskRepresenting an inspection Activity AkComprising mkAnd (4) regulating.
Based on the weight of each regulatory act, a Jaccard similarity metric is performed:
Figure BDA0002610915140000114
wherein m isk’Represents TA∩TBNumber of rules in, mk”Represents TA∪TBNumber of rules in (1).
Preferably, the top-k recommendation algorithm is adopted as the algorithm for compliance case recommendation, and comprises the following steps:
s31, acquiring a data set SD formed by 3 ordered lists (m elements);
s32 sequentially accesses the data set SD, and when a certain examination object S is seen in a certain list in the sequential access manner, finds its value and position on the corresponding attribute from other lists in the random access manner:
s321, saving the seen position and the corresponding attribute value;
s322, calculating the score value of an example S according to a formula F (S) ═ wo sos + wa sas + wt × sts, wherein wo, wa and wt are weights of an object dimension, an activity dimension and a target dimension respectively, wo + wa + wt equals to 1, wo is not less than 0, wa is not less than 0, wt is not less than 0, and the value is determined by a domain expert;
s323 saves the k objects with the highest score values seen so far to the set Ak.
S33 records, for each dimension Lo, La, Lt, the positions seen so far in sequential or random access in sets Po, Pa, Pt. Will opo、opa、ootRespectively called list Lo, La, Lt optimal position, opo、opa、ootPositions of Po, Pa and Pt are maximum respectively, so that 1 to op in LooAll 1 to op in Po, LaaAll in Pa and Lt from 1 to optAll in Pt. so(opo) In position op for dimension LooObject attribute value of sa(opa) In position op for dimension LaaObject attribute value of st(opt) At position op for dimension LttThe object property value of (1).
S34 calculates a score function value for the optimal position:
F＝wo×so(opo)+wa×sa(opa)+wt×st(opt)
if the score function values of all k objects in Ak are greater than or equal to the score function value of the optimal location, the sequential access to the list is stopped, otherwise, go to S31.
S35 adds all the censored instances with score function values equal to the k instance score values to the set Ak;
and the output set Ak of S36 is a top-k set.
The invention also designs a airworthiness case recommendation system based on the conformance vector, which comprises the following steps: the airworthiness examination ontology knowledge base comprises two ontology instances, wherein a conformance object and conformance activity are represented in a tree hierarchical relationship, and a conformance target is represented in a tag set;
an activity similarity calculation module for calculating a similarity measure of the compliance object and the compliance activity;
a target similarity calculation module for calculating a compliance target similarity measure;
a recommendation module for compliance case recommendation.
Preferably, the activity similarity calculation module includes: a first calculation unit for obtaining activity similarity based on the shortest path length;
a second calculation unit for acquiring node similarity based on depth;
a third calculating unit for obtaining the node concrete degree similarity;
and the fourth calculation unit is used for summarizing and calculating the node similarity.
The invention also designs a airworthiness case recommendation device, which comprises a processor;
a memory for storing processor-executable instructions;
wherein the processor is configured to: the following method steps are carried out: firstly, acquiring similarity measurement of a conformance object and conformance activity; then obtaining similarity measurement of the coincidence target; and finally, recommending a coincidence case by using a Top-k algorithm according to the similarity measurement obtained in the last two steps.
The specific calculation method of the similarity measurement of the compliance object and the compliance activity is as follows: acquiring activity similarity based on the shortest path length; acquiring node similarity based on depth; acquiring specific degree of a node; calculating the node similarity based on the node concrete degree according to the node concrete degree; and calculating the node similarity of the compliance object and the compliance activity.
The present invention also contemplates a computer storage medium having computer program instructions stored thereon for execution by a processor in a method comprising: firstly, acquiring similarity measurement of a conformance object and conformance activity; then obtaining similarity measurement of the coincidence target; and finally, recommending a coincidence case by using a Top-k algorithm according to the similarity measurement obtained in the last two steps.
The specific calculation method of the similarity measurement of the compliance object and the compliance activity is as follows: acquiring activity similarity based on the shortest path length; acquiring node similarity based on depth; acquiring specific degree of a node; calculating the node similarity based on the node concrete degree according to the node concrete degree; and calculating the node similarity of the compliance object and the compliance activity.

Claims (14)

1. The airworthiness case recommendation method based on the conformance vector is characterized by comprising the following steps: the method comprises the following steps:
s1, obtaining similarity measurement of the coincidence object and the coincidence activity;
s2, obtaining similarity measurement of the coincidence target;
s3 recommendation of compliance case.
2. The airworthiness-case recommendation method based on the conformance vector of claim 1, wherein: the specific calculation method of the similarity measure of the compliance object and the compliance activity of S1 includes the following steps:
s11, acquiring activity similarity based on the shortest path length;
s12, obtaining node similarity based on depth;
s13, acquiring the specific degree of the node;
s14, calculating the node similarity based on the node concrete degree according to the node concrete degree;
s15 calculates the node similarity of the compliance object and the compliance activity.
3. The airworthiness-case recommendation method based on the conformance vector of claim 2, wherein: the calculation formula of the activity similarity based on the shortest path length obtained by the step S11 is as follows:
Figure FDA0002610915130000011
wherein, SP (node)1，node2) Represents a node1And a node2The length of the shortest path between two nodes, height represents the maximum depth of the ontology tree, node represents node, SimSP(node1，node2) Represents a node1And a node2Active similarity of shortest path length between two nodes.
4. The airworthiness-case recommendation method based on the conformance vector of claim 2, wherein: the calculation formula of the node similarity based on the depth obtained by the step S12 is as follows:
Figure FDA0002610915130000012
wherein, depth (node)1) Refers to a node1Depth in the ontology tree, Simdepth(node1，node2) Represents a node1And a node2And (4) activity similarity of two nodes in the depth of the ontology tree.
5. The airworthiness-case-recommendation method based on the conformance vector of claim 4, wherein:
Figure FDA0002610915130000021
Figure FDA0002610915130000022
6. the airworthiness-case recommendation method based on the conformance vector of claim 2, wherein: the calculation formula of the specific degree of the acquired node of S13 is as follows:
Figure FDA0002610915130000023
Figure FDA0002610915130000024
wherein Spec (node)1，node2) Representing a node1Relative node2The degree of specificity of (a).
7. The airworthiness-case recommendation method based on the conformance vector of claim 2, wherein: s14, according to the node concrete degree, calculating the node similarity based on the node concrete degree by the calculation formula:
Figure FDA0002610915130000025
Figure FDA0002610915130000026
wherein NCA (node)1，node2) Representing concept node1And a node2Nearest common ancestor node, Sim, in the ontology treeSpec(node1，node2) Representing concept-based node1And a node2Node similarity of specific degrees.
8. The airworthiness-case recommendation method based on the conformance vector of claim 2, wherein: the calculation formula for calculating the node similarity in S15 is as follows: simonto(node1，node2)＝αSimSP(node1，node2)+βSimdepth(node1，node2)+γSimRF(node1，node2) Wherein, α, β, γ are parameter factors, α + β + γ is 1.
9. The airworthiness-case recommendation method based on the conformance vector of claim 1, wherein: s2, the specific method for obtaining the similarity measure of the coincidence target is: firstly, selecting a regulation example, and measuring the similarity of the regulation example by using the Jaccard similarity; secondly, the inverse document frequency idf is used as weight, so that the influence of high-frequency regulations on the similarity of the examples is reduced; carrying out normalization processing; based on the weight of each regulatory act, a Jaccard similarity metric is performed.
10. The airworthiness-case recommendation method based on the conformance vector of claim 1, wherein: the method for recommending the compliance case by the S3 adopts a top-k recommendation algorithm, and comprises the following steps:
s31, acquiring a data set SD formed by the ordered list;
s32 performing sequential access to the data set SD;
s33 recording each dimension with a set to the position seen so far in sequential access;
s34, calculating a score function value of the optimal position;
s35 adds all the censored instances with score function values equal to the k instance score values to the set Ak;
and the output set Ak of S36 is a top-k set.
11. A airworthiness case recommendation system based on a conformance vector is characterized in that: the method comprises the following steps: the airworthiness examination ontology knowledge base (1), wherein a conformance object and conformance activity are two ontology instances and appear in a tree hierarchical relationship, and a conformance target appears in a tag set form; an activity similarity calculation module (2) for calculating a similarity measure of the compliance object and the compliance activity;
a target similarity calculation module (3) for calculating a conformity target similarity measure;
a recommendation module (4) for compliance case recommendations.
12. The system of claim 11, wherein the airworthiness-vector-based airworthiness-case recommendation system is further configured to: the activity similarity calculation module (2) includes: a first calculation unit (2-1) for acquiring an activity similarity based on the shortest path length;
a second calculation unit (2-2) for acquiring a depth-based node similarity;
a third calculating unit (2-3) for acquiring the node concrete degree similarity;
and a fourth calculation unit (2-4) for calculating the node similarity in a summary manner.
13. A airworthiness case recommendation device is characterized in that: comprises a processor;
a memory for storing processor-executable instructions;
wherein the processor is configured to: performing the method of any one of claims 1 to 10.
14. A computer storage medium having computer program instructions stored thereon, characterized in that: the computer program instructions, when executed by a processor, implement the method of any one of claims 1 to 10.
CN202010753933.3A 2020-07-30 2020-07-30 Airworthiness case recommendation method based on conformance vector Pending CN112084322A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010753933.3A CN112084322A (en) 2020-07-30 2020-07-30 Airworthiness case recommendation method based on conformance vector

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010753933.3A CN112084322A (en) 2020-07-30 2020-07-30 Airworthiness case recommendation method based on conformance vector

Publications (1)

Publication Number Publication Date
CN112084322A true CN112084322A (en) 2020-12-15

Family

ID=73735095

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010753933.3A Pending CN112084322A (en) 2020-07-30 2020-07-30 Airworthiness case recommendation method based on conformance vector

Country Status (1)

Country Link
CN (1) CN112084322A (en)

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040034554A1 (en) * 2002-08-16 2004-02-19 Shirley David P. Request for conformity system
US20080243809A1 (en) * 2007-03-29 2008-10-02 Anand Ranganathan Information-theory based measure of similarity between instances in ontology
KR20140024533A (en) * 2012-08-20 2014-03-03 현대중공업 주식회사 Method for case based reasoning of integrated plant using the similarity and weight
CN105677740A (en) * 2015-12-29 2016-06-15 中国民用航空上海航空器适航审定中心 Method for matching entity-based text data and XML files
CN107562615A (en) * 2017-07-21 2018-01-09 北京航空航天大学 The seaworthiness compliance testing method of tree-model is proved based on the evaluation of target accordance
KR20180105921A (en) * 2017-03-16 2018-10-01 한국해양대학교 산학협력단 method for seakeeping quality assessment using ship monitoring system and system for seakeeping quality assessment
CN110647631A (en) * 2018-06-25 2020-01-03 阿里巴巴集团控股有限公司 Case recommendation method and device, storage medium and processor
US20200043022A1 (en) * 2018-08-06 2020-02-06 Walmart Apollo, Llc Artificial intelligence system and method for generating a hierarchical data structure
CN111143409A (en) * 2019-12-13 2020-05-12 中国航空工业集团公司西安飞机设计研究所 Aluminum alloy material design verification method for airworthiness certification

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040034554A1 (en) * 2002-08-16 2004-02-19 Shirley David P. Request for conformity system
US20080243809A1 (en) * 2007-03-29 2008-10-02 Anand Ranganathan Information-theory based measure of similarity between instances in ontology
KR20140024533A (en) * 2012-08-20 2014-03-03 현대중공업 주식회사 Method for case based reasoning of integrated plant using the similarity and weight
CN105677740A (en) * 2015-12-29 2016-06-15 中国民用航空上海航空器适航审定中心 Method for matching entity-based text data and XML files
KR20180105921A (en) * 2017-03-16 2018-10-01 한국해양대학교 산학협력단 method for seakeeping quality assessment using ship monitoring system and system for seakeeping quality assessment
CN107562615A (en) * 2017-07-21 2018-01-09 北京航空航天大学 The seaworthiness compliance testing method of tree-model is proved based on the evaluation of target accordance
CN110647631A (en) * 2018-06-25 2020-01-03 阿里巴巴集团控股有限公司 Case recommendation method and device, storage medium and processor
US20200043022A1 (en) * 2018-08-06 2020-02-06 Walmart Apollo, Llc Artificial intelligence system and method for generating a hierarchical data structure
CN111143409A (en) * 2019-12-13 2020-05-12 中国航空工业集团公司西安飞机设计研究所 Aluminum alloy material design verification method for airworthiness certification

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
吕欢欢;宋伟东;杨睿;: "基于领域本体的综合加权语义相似度算法研究", 计算机工程与设计, no. 12 *
索俊锋等: "基于地理本体的综合语义相似度算法", 《兰州大学学报(自然科学版)》, vol. 53, no. 1, pages 19 - 27 *

Similar Documents

Publication Publication Date Title
Zhu et al. Spatially varying coefficient model for neuroimaging data with jump discontinuities
Sim et al. Missing values and optimal selection of an imputation method and classification algorithm to improve the accuracy of ubiquitous computing applications
Houle et al. Generalized expansion dimension
CN105184103A (en) Virtual medical expert based on medical record database
Li et al. Time‐dependent ROC analysis under diverse censoring patterns
CN105512465B (en) Based on the cloud platform safety quantitative estimation method for improving VIKOR methods
Lee et al. Factor analysis in personality research
CN107239477B (en) Geographic data support vector regression method fusing spatial correlation
CN110717687A (en) Evaluation index acquisition method and system
Wicaksono et al. Metrics, user models, and satisfaction
Lahdelma et al. The shape of the utility or value function in stochastic multicriteria acceptability analysis
de Andrade Silva et al. An experimental study on the use of nearest neighbor-based imputation algorithms for classification tasks
Espinheira et al. Residual and influence analysis to a general class of simplex regression
Azimbagirad et al. Tsallis generalized entropy for Gaussian mixture model parameter estimation on brain segmentation application
Satoh Discrete Gompertz equation and model selection between Gompertz and logistic models
Falk et al. On Lagrange multiplier tests in multidimensional item response theory: Information matrices and model misspecification
Silva et al. Capturing provenance to improve the model training of pinns: first hand-on experiences with grid5000
CN112084322A (en) Airworthiness case recommendation method based on conformance vector
Ligtvoet et al. Latent class models for testing monotonicity and invariant item ordering for polytomous items
Guo et al. An improved density-based approach to risk assessment on railway investment
Coutinho et al. Calibrated hot-deck donor imputation subject to edit restrictions
Al Hosni et al. Assesing the stability and selection performance of feature selection methods under different data complexity.
Bai et al. Variable selection for ultra-high dimensional quantile regression with missing data and measurement error
Dou et al. The Measurement of Knowledge in Knowledge Graphs
Zhao et al. Extreme tail risk estimation with the generalized Pareto distribution under the peaks-over-threshold framework

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