CN113159235B - Knowledge collaborative clustering method for product green design - Google Patents
Knowledge collaborative clustering method for product green design Download PDFInfo
- Publication number
- CN113159235B CN113159235B CN202110565342.8A CN202110565342A CN113159235B CN 113159235 B CN113159235 B CN 113159235B CN 202110565342 A CN202110565342 A CN 202110565342A CN 113159235 B CN113159235 B CN 113159235B
- Authority
- CN
- China
- Prior art keywords
- knowledge
- clustering
- similarity
- semantic
- distance
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- 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
- G06F16/35—Clustering; Classification
- G06F16/355—Class or cluster creation or modification
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/10—Text processing
- G06F40/12—Use of codes for handling textual entities
- G06F40/14—Tree-structured documents
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/237—Lexical tools
- G06F40/247—Thesauruses; Synonyms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- 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
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Databases & Information Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to the technical field of knowledge clustering, and discloses a knowledge collaborative clustering method for product green design, which comprises the steps of restraining knowledge required in the product green design process in two aspects of functions and attributes, establishing a knowledge model of characteristic restraint, limiting the research range of the knowledge, distinguishing data types of different knowledge, respectively selecting corresponding distance calculation formulas based on similarity calculation consideration, and carrying out one-layer retrieval clustering on different functional characteristics by combining a semantic tree according to new design requirements. The knowledge collaborative clustering method for the green design of the product can effectively manage the green design knowledge, achieve the purposes of reducing research and development cost and improving the quality of product design, and can also process fragmented and isomerized knowledge, so that complicated and complicated data can be rapidly and accurately clustered in a collaborative mode.
Description
Technical Field
The invention relates to the technical field of knowledge clustering, in particular to a knowledge collaborative clustering method for product green design.
Background
Under the background of the current big data age, knowledge clustering plays an important and active role in the development process of knowledge engineering, belongs to a product with rapidly rising science and technology, acquires new knowledge with higher value or usability by organizing and managing a dispersive data source and a knowledge source and processing, converting and integrating knowledge elements according to the requirements of customers, provides subsequent services based on knowledge by optimally integrating knowledge objects in aspects of functions, attributes, structures and the like and needs a large amount of design knowledge at each stage of design activity like product green design of the conceptual design, but in the design process, engineers usually spend a large amount of time processing the dispersive data source to acquire required data.
Although the concept of green design knowledge is widely known at present, a systematic method is still lacked for researching the green design knowledge of the product, the green design knowledge cannot be effectively managed, the purposes of reducing research and development cost and improving the quality of product design are achieved, and in addition, the fragmentized and isomerized knowledge lacks connection, is not easy to process, and cannot perform rapid and accurate clustering processing on complex and complicated data.
Disclosure of Invention
Technical problem to be solved
Aiming at the defects of the prior art, the invention provides a knowledge collaborative clustering method for product green design, which can effectively manage green design knowledge, achieve the purposes of reducing research and development cost and improving the quality of product design, can also process fragmented and isomerized knowledge, can realize the advantages of fast and accurate collaborative clustering of complex and frequent data, and the like, and solves the problems that the green design knowledge cannot be effectively managed, the research and development cost is reduced, and the quality of product design is improved.
(II) technical scheme
In order to achieve the purpose, the invention provides the following technical scheme: a knowledge collaborative clustering method for product green design comprises the following steps;
step one, constraining knowledge required in the green design process of a product in two aspects of functions and attributes, establishing a knowledge model of characteristic constraint, and limiting the research range of the knowledge;
step two, distinguishing data types of different knowledge, calculating and considering based on similarity, and respectively selecting corresponding distance calculation formulas;
step three, aiming at new design requirements, firstly, combining a semantic tree to perform one-layer retrieval clustering on different functional characteristics to obtain one-layer functional similarity (FR), and retrieving a certain amount of samples from a sample library according to the FR;
step four, using the deeper second-layer attribute to screen and cluster to make a decision of a feasible clustering scheme;
and step five, calculating the distances among the attributes with different data types, introducing a cooperative operator, and calculating cooperative similarity (CR) by combining different distance formulas and an improved classical Fuzzy C-means (FCM) algorithm to realize the cooperative clustering effect of hierarchical iteration and the centralized processing of discrete knowledge.
Preferably, in the step one, a knowledge model of characteristic constraints is established, a research range of knowledge is limited, knowledge required in a product green design process is constrained in two aspects of functions and attributes, and a knowledge model based on the characteristic constraints is established.
Preferably, the calculation of the similarity in the second step is the core of the clustering calculation, and the calculation of the similarity is highly dependent on the distance between each sample, the similarity and the sample distance are in an inverse relationship in terms of quantity, the greater the distance, the lower the sample similarity is, and the smaller the distance, the higher the sample similarity is, therefore, the calculation of the similarity can be obtained by the reciprocal of the distance.
Preferably, the Semantic Tree (ST) in the three steps analyzes and identifies a text, selects feature words and feature words with synonyms, peers or membership of the feature words to be represented in a tree form, measures attributes of the Semantic tree mainly including Semantic tree height, Semantic element path, Semantic element speed and Semantic element depth, when functional knowledge is retrieved from a sample library in input design requirements, FR is a key factor, according to the analysis of the Semantic tree, when the difference of the depth d(s) between two Semantic elements is small, the degree of association is high, when some functional features are in the same layer or adjacent layers, whether the functional features are equally needed can be determined through the similarity between the Semantic elements, where the inverse of the depth difference is used to measure the correlation between the two Semantic elements, as shown in the following formula:
preferably, when the deeper second-layer attribute screening clustering is used in the fourth step to make a decision on a feasible clustering scheme, the feature constraint is characterized in that the data types of constraint variables are not uniform, the data types are classified into four types, namely a semantic type, a numerical type, a fuzzy interval type and a tuple type, and the types appearing in the feature constraint include several types to all types.
Preferably, the classic fuzzy C-means clustering algorithm in the fifth step is calculated by adopting the following formula;
as described above, equation (1) defines the equation for the FCM objective function, where | x | (R |) j -c i I represents a sample x j To class center c i A distance of c i As a cluster center, x j As data samples, u ij Is degree of membership, m is degree of membership factor, J is objective function, formula (2) is constraint formula, and formula (3) is degree of membership u ij Calculating a formula, wherein the formula (4) is a clustering center iteration formula;
when the algorithm is started to run, giving u randomly ij Or c i One of the values is assigned, as long as the value is assigned under a full condition, then through repeated iteration, the objective function J gradually tends to a stable value, and finally a solution meeting the requirement is obtained, wherein different functional characteristics comprise N attribute characteristics, the N attribute characteristics have different data types, and each type of attribute characteristics are connected in series to form the functional characteristics in a knowledge model of characteristic constraint, wherein the work to be done by collaborative clustering is to generate clustering subsamples of each group of functional characteristics on the premise of dividing the functional characteristics into N fuzzy classes, and then through collaborative cooperation among the subsamples, a collaborative operator is introduced to perform appropriate strengthening or weakening on the different characteristics to generate an integral clustering sample of the characteristic constraint;
let Ui be a membership matrix, U i ＝{u ij ∈[0,1]}，p i Is a layer of samples after functional clustering, CO [ i, j]Is a co-operator, d ij (i 1,2, …, n, j 1,2, p) is a distance function, an objective function form in a classic FCM clustering function is applied and improved, and after a synergistic operator is added, the following synergistic function is obtainedImproved objective function J with same effect i ；
Referring to an iteration mode in a classical FCM algorithm, introducing a Lagrange multiplier lambda, converting a cooperation problem into a non-constrained optimization problem, and obtaining:
minimizing the requirements according to the objective function:
obtaining a membership matrix element iterative formula:
and obtaining the collaborative clustering algorithm after determining the main parameter iteration mode.
Preferably, the collaborative clustering algorithm is specifically as follows;
inputting: input sample set X 1 ,X 2 ,..,X n ；
Selecting: selecting corresponding distance function, cluster number c and cooperative operator CO [ i, j ] according to different data types, wherein the algorithm termination criterion satisfies the formula:
D i+1 -D i ≤δ
initialization: initializing a membership matrix and an algorithm termination threshold delta, and carrying out independent clustering by adopting different distance formulas according to specific data types of attribute characteristics contained in the membership matrix and the algorithm termination threshold delta to obtain an initial clustering sample related to design requirements;
and (3) calculating: and continuously iterating the membership degree matrix until a termination criterion is met.
(III) advantageous effects
Compared with the prior art, the invention provides a knowledge collaborative clustering method for product green design, which has the following beneficial effects:
when the invention is used, the knowledge required in the green design process of the product is restricted in two aspects of function and attribute, a knowledge model of characteristic restriction is established, the knowledge research is limited to be in a certain specific condition set according to characteristic parameters, the data types of different knowledge are distinguished, then the similarity of different knowledge is obtained through the reciprocal distance obtained by a distance calculation formula, different functional characteristics are subjected to one-layer searching and clustering by combining a semantic tree to obtain one-layer functional similarity, a certain amount of samples are searched from a sample library, a text is analyzed and identified through the semantic tree, the selected characteristic words and the characteristic words with synonyms, peers or subordination relations are expressed in a tree form, when certain functional characteristics are in the same layer or adjacent layers, whether the functional characteristics are equally required can be judged through the similarity between semantic elements, at the moment, the correlation of two semantic elements is measured by utilizing the reciprocal of the depth difference, the knowledge model characteristic attributes identified by semantic tree analysis are calculated by utilizing a formula, and then the cooperative similarity can be obtained, finally, the distance between the attributes with different data types is calculated by introducing a cooperative operator, and the cooperative similarity is calculated by combining different distance formulas and an improved FCM algorithm, so that the cooperative clustering effect of hierarchical iteration is realized, the centralized processing of discrete knowledge is realized, and the rapid and accurate clustering of complicated data is realized.
Drawings
FIG. 1 is a knowledge model diagram of feature constraints in a knowledge collaborative clustering method for product green design according to the present invention;
FIG. 2 is a logic expression diagram of a semantic tree in the knowledge collaborative clustering method for product green design according to the present invention;
FIG. 3 is a flow chart of collaborative clustering calculation in the knowledge collaborative clustering method for green product design according to the present invention;
FIG. 4 is a collaborative clustering meaning diagram in the knowledge collaborative clustering method for product green design according to the present invention;
FIG. 5 is a flow chart of a collaborative clustering algorithm in the knowledge collaborative clustering method for product green design according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example (b):
referring to fig. 1-5, a knowledge collaborative clustering method for product green design includes the following steps;
step one, constraining knowledge required in the green design process of a product in two aspects of functions and attributes, establishing a knowledge model of characteristic constraint, limiting the research range of the knowledge, establishing the knowledge model of the characteristic constraint in step one, limiting the research range of the knowledge, constraining the knowledge required in the green design process of the product in two aspects of functions and attributes, and establishing a knowledge model based on the characteristic constraint, wherein in the knowledge model of the characteristic constraint, the characteristic constraint is composed of characteristic parameters of functional characteristics, meanwhile, different functional characteristics comprise different attributes and corresponding values thereof, and the characteristic parameters limit the research of the knowledge in a certain specific condition set;
and secondly, distinguishing data types of different knowledge, and selecting corresponding distance calculation formulas based on similarity calculation consideration, wherein the data types of various attributes contained in the characteristic parameters are different, the data types of some attributes are semantic types, and the data types of some attributes are numerical types. Therefore, when sample clustering is performed, two or three or more functional features and attribute features doubled after decomposition may be faced, and the features and the attributes also have different data types, the calculation of the similarity is the core of the clustering calculation, and the calculation of the similarity is highly dependent on the distance between each sample, the similarity and the sample distance are in an inverse relation in number, the greater the distance is, the lower the sample similarity is, the smaller the distance is, the higher the sample similarity is, and therefore, the calculation of the similarity can be obtained by the reciprocal of the distance;
let the feature constraint of knowledge be FC i Where i is 1, 2.. and n, different feature constraints correspond to different data types, and there are corresponding distance calculation formulas, as shown in table 1,
TABLE 1
Step three, aiming at new design requirements, firstly, combining a Semantic tree to perform one-layer retrieval clustering on different functional characteristics to obtain one-layer functional similarity (FR), retrieving a certain amount of samples from a sample library according to FR, wherein the Semantic Tree (ST) is used for analyzing and identifying texts, selecting characteristic words and characteristic words with synonymy, same level or membership of the characteristic words and expressing the characteristic words in a tree form, and measuring the attributes of the Semantic tree mainly comprising Semantic tree height, Semantic element path, Semantic element high speed and Semantic element depth, when the design requirements are input to retrieve functional knowledge from the sample library, FR is a key factor, according to the analysis of the Semantic tree, when the difference of the depth d(s) between two Semantic elements is small, the association degree is high, and when some functional characteristics are in the same layer or an adjacent layer, whether the functional characteristics are equally needed or not can be judged through the similarity between the Semantic elements, the inverse of the depth difference is used here to measure the correlation between two semantic elements, as shown in the following equation:
in the above, the semantic tree height, i.e. the total number of layers in the semantic tree, is represented by h (ST),
semantic Meta Path, i.e. the shortest path from semantic Meta i to semantic Meta j, using p(s) i ,s j ) It is shown that the process of the present invention,
the semantic element height, i.e. the length of the semantic element path, is called the height of the semantic element, denoted h(s),
the semantic element depth is the length of the longest path from the root of the semantic tree to a specific semantic element, namely the semantic element depth and is represented by d(s);
step four, using the deeper second-layer attribute screening cluster to make a decision of a feasible clustering scheme, when using the deeper second-layer attribute screening cluster to make a decision of the feasible clustering scheme, the characteristic constraint is characterized in that the data types of the constraint variables are not uniform, the data types are divided into four types of semantic type, numerical type, fuzzy interval type and group type, and the types appearing in the characteristic constraint include several types to all types, therefore, when calculating the sample similarity, the distance relation between the characteristics needs to be considered firstly, the influence of different data types needs to be faced, and the attribute characteristic constraint in the knowledge model of the characteristic constraint is set as AC i (i-1, 2, …, n), after determining the data type of the corresponding feature, selecting the corresponding distance formula in table 1 to calculate the distance, and setting the attribute feature constraint AC i (i is 1,2, …, n) is d i (i ═ 1,2, …, n) and a cooperative distance D, acting on D i The CO-operator (Cooperator) above is CO i (i＝1,2,…,n),COi∈[0,1]The following formula is used to express the cooperative distance:
the reciprocal of the cooperative distance is the cooperative similarity:
calculating the distances among the attributes with different data types, introducing a cooperative operator, and calculating cooperative similarity (CR) by combining different distance formulas and an improved classical Fuzzy C-means (FCM) algorithm to realize the cooperative clustering effect of hierarchical iteration and the centralized processing of discrete knowledge, wherein the classical Fuzzy C-means clustering algorithm adopts the following formula for calculation;
as described above, equation (1) defines the equation for the FCM objective function, where | x | (R |) j -c i I represents a sample x j To class center c i A distance of c i As a cluster center, x j As data samples, u ij Is the degree of membership, m is the degree of membership factor, J is the objective function, equation (2) is the constraint equation, and equation (3) is the degree of membership u ij Calculating a formula, wherein the formula (4) is a clustering center iterative formula;
when the algorithm is started to run, giving u randomly ij Or c i One assignment is only required to be assigned under a condition, then the objective function J gradually tends to a stable value through repeated iteration, and finally a solution meeting the requirement is obtained, wherein different functional characteristics comprise N attribute characteristics, and the N attribute characteristics have different numbersAccording to types, each type of attribute features are connected in series to form functional features in a feature constrained knowledge model, wherein the collaborative clustering is to do work, namely, on the premise of dividing the functional features into n fuzzy classes, clustering subsamples of each group of functional features are generated, then, through collaborative cooperation among the subsamples, a collaborative operator is introduced to carry out appropriate amount of strengthening or weakening on different features, and an integral clustering sample of feature constraint is generated;
let Ui be a membership matrix, U i ＝{u ij ∈[0,1]}，p i Is a layer of samples after functional clustering, CO [ i, j ]]Is a co-operator, d ij The method is characterized in that (i is 1,2, …, n, J is 1,2, p) is a distance function, an objective function form in a classic FCM clustering function is applied and improved, and after a synergistic operator is added, an improved objective function J with a synergistic effect is obtained i ；
Referring to an iteration mode in a classical FCM algorithm, introducing a Lagrange multiplier lambda, converting a cooperation problem into a non-constrained optimization problem, and obtaining:
minimizing the requirements according to the objective function:
obtaining a membership matrix element iterative formula:
after the main parameter iteration mode is determined, the obtained collaborative clustering algorithm is concretely as follows;
inputting: input sample set X 1 ,X 2 ,..,X n ；
Selecting: selecting corresponding distance function, cluster number c and cooperative operator CO [ i, j ] according to different data types, wherein the algorithm termination criterion satisfies the formula:
D i+1 -D i ≤δ
the difference value before and after the cooperative distance is smaller than a threshold value delta, the termination criterion depends on the change of a membership matrix in the continuous iteration process of the clustering method, when different cooperative operators are taken, the distance formula can be used according to different distance formulas, for example, the Mahalanobis distance can be used as the measurement standard of the final cooperative distance, and along with continuous iteration, when the error between the current measured value and the measured value of the last two times is smaller than a given threshold value, the iteration is terminated;
initialization: initializing a membership matrix and an algorithm termination threshold delta, and carrying out independent clustering by adopting different distance formulas according to specific data types of attribute characteristics contained in the membership matrix and the algorithm termination threshold delta to obtain an initial clustering sample related to design requirements;
and (3) calculating: and continuously iterating the membership degree matrix until a termination criterion is met.
When the invention is used, the knowledge required in the green design process of the product is restricted in two aspects of function and attribute, a knowledge model of characteristic restriction is established, the knowledge research is limited in a certain specific condition set according to characteristic parameters, the data types of different knowledge are distinguished, then the similarity of different knowledge is obtained through the reciprocal distance obtained by a distance calculation formula, different functional characteristics are subjected to one-layer searching clustering by combining a semantic tree to obtain one-layer functional similarity, a certain amount of samples are searched from a sample library, a text is analyzed and identified through the semantic tree, the selected characteristic words and the characteristic words with synonymy, sibling or subordination relation are expressed in a tree form, and when certain functional characteristics are in the same layer or adjacent layers, whether the functional characteristics are equally required can be judged through the similarity between semantic elements, at the moment, the correlation of two semantic elements is measured by utilizing the reciprocal of the depth difference, the knowledge model characteristic attributes identified by semantic tree analysis are calculated by utilizing a formula, and then the cooperative similarity can be obtained, finally, the distance between the attributes with different data types is calculated by introducing a cooperative operator, and the cooperative similarity is calculated by combining different distance formulas and an improved FCM algorithm, so that the cooperative clustering effect of hierarchical iteration is realized, the centralized processing of discrete knowledge is realized, and the rapid and accurate clustering of complicated data is realized.
It is to be noted that the term "comprises" or any other variation thereof is intended to cover a non-exclusive inclusion, such that a process or method including a list of elements does not include only those elements but also other elements not expressly listed or inherent to such process or method. Without further limitation, an element defined by the phrase "comprising … …" does not exclude the presence of additional like elements in a process or method that includes the element.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (4)
1. A knowledge collaborative clustering method for product green design is characterized in that: comprises the following steps;
step one, constraining knowledge required in the green design process of a product in two aspects of functions and attributes, establishing a knowledge model of characteristic constraint, and limiting the research range of the knowledge;
step two, distinguishing data types of different knowledge, calculating and considering based on similarity, and respectively selecting corresponding distance calculation formulas;
step three, aiming at new design requirements, firstly, combining a semantic tree to perform one-layer retrieval clustering on different functional features to obtain one-layer functional similarity (FR), and retrieving a certain amount of samples from a sample library according to the FR;
the Semantic Tree (ST) analyzes and identifies a text, selects feature words and feature words with synonyms, peers or membership of the feature words and represents the feature words in a tree form, measures attributes of the Semantic tree mainly including Semantic tree height, Semantic element path, Semantic element high speed and Semantic element depth, when functional knowledge is retrieved from a sample library in input design requirements, FR is a key factor, according to the analysis of the Semantic tree, when the difference of the depth d(s) between two Semantic elements is small, the degree of association is high, when some functional features are in the same layer or adjacent layer, whether the functional features are required to be equal can be determined through the similarity between the Semantic elements, the reciprocal of the depth difference is used to measure the correlation between the two Semantic elements, as shown in the following formula:
in the formula, R f Being the degree of association between two semantic elements, d i (s) and d j (s) are two semantic elements;
step four, using the deeper second-layer attribute to screen the clustering to make a decision on the feasible clustering scheme, when using the deeper second-layer attribute to screen the clustering to make a decision on the feasible clustering scheme, the characteristic constraint is characterized in that the data types of the constraint variables are not uniform, the data types are divided into four types, namely a semantic type, a numerical type, a fuzzy interval type and a tuple type, and the types appearing in the characteristic constraint include several types to all types, so when calculating the sample similarity, the distance relation among the characteristics needs to be considered at first, and then the influence of different data types needs to be faced;
calculating distances among attributes with different data types, introducing a cooperative operator, and calculating cooperative similarity (CR) by combining different distance formulas and an improved classical Fuzzy C-means (FCM) algorithm to realize a cooperative clustering effect of hierarchical iteration and centralized processing of discrete knowledge;
calculating by using the following formula in a classical fuzzy C-means clustering algorithm in the step five;
above, equation (1) defines the equation for the FCM target function, where | x j -c i II denotes sample x j To class center c i A distance of c i As a cluster center, x j As data samples, u ij Is degree of membership, m is degree of membership factor, J is objective function, formula (2) is constraint formula, and formula (3) is degree of membership u ij Calculating a formula, wherein the formula (4) is a clustering center iterative formula;
when the algorithm is started to run, giving u randomly ij Or c i One of the values is assigned, as long as the value is assigned under a full condition, then through repeated iteration, the objective function J gradually tends to a stable value, and finally a solution meeting the requirement is obtained, wherein different functional characteristics comprise N attribute characteristics, the N attribute characteristics have different data types, and each type of attribute characteristics are connected in series to form the functional characteristics in a knowledge model of characteristic constraint, wherein the work to be done by collaborative clustering is to generate clustering subsamples of each group of functional characteristics on the premise of dividing the functional characteristics into N fuzzy classes, and then through collaborative cooperation among the subsamples, a collaborative operator is introduced to perform appropriate strengthening or weakening on the different characteristics to generate an integral clustering sample of the characteristic constraint;
let Ui be a membership matrix, U i ＝{u ij ∈[0,1]}，p i Is a layer of samples after functional clustering, CO [ i, j]Is a co-operator, d ij (i 1,2, …, n; J1, 2,., p) is a distance function, an objective function form in a classic FCM clustering function is applied and improved, and after a synergistic operator is added, an improved objective function J with a synergistic effect is obtained i ；
Referring to an iteration mode in a classical FCM algorithm, introducing a Lagrange multiplier lambda, converting a cooperation problem into a non-constrained optimization problem, and obtaining:
minimizing the requirements according to the objective function:
obtaining a membership matrix element iterative formula:
and obtaining the collaborative clustering algorithm after determining the main parameter iteration mode.
2. The knowledge collaborative clustering method for green design of products according to claim 1, characterized in that: in the first step, a knowledge model of characteristic constraint is established, the research range of knowledge is limited, the knowledge required in the green design process of products is constrained in two aspects of functions and attributes, and the knowledge model based on the characteristic constraint is established.
3. The knowledge collaborative clustering method for green design of products according to claim 1, characterized in that: the calculation of the similarity in the second step is the core of the clustering calculation, the calculation of the similarity is highly dependent on the distance between each sample, the similarity and the sample distance are in an inverse relationship in quantity, the greater the distance is, the lower the sample similarity is, the smaller the distance is, the higher the sample similarity is, and therefore, the calculation of the similarity can be obtained through the reciprocal of the distance.
4. The product green design-oriented knowledge collaborative clustering method according to claim 1, characterized in that: the collaborative clustering algorithm is concretely as follows;
inputting: input sample set X 1 ,X 2 ,..,X n ；
Selecting: selecting corresponding distance function, cluster number c and cooperative operator CO [ i, j ] according to different data types, wherein the algorithm termination criterion satisfies the formula:
D i+1 -D i ≤δ
initialization: initializing a membership matrix and an algorithm termination threshold delta, and carrying out independent clustering by adopting different distance formulas according to specific data types of attribute characteristics contained in the membership matrix and the algorithm termination threshold delta to obtain an initial clustering sample related to design requirements;
and (3) calculating: and continuously iterating the membership degree matrix until a termination criterion is met.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110565342.8A CN113159235B (en) | 2021-05-24 | 2021-05-24 | Knowledge collaborative clustering method for product green design |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110565342.8A CN113159235B (en) | 2021-05-24 | 2021-05-24 | Knowledge collaborative clustering method for product green design |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113159235A CN113159235A (en) | 2021-07-23 |
| CN113159235B true CN113159235B (en) | 2022-09-30 |
Family
ID=76877096
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110565342.8A Active CN113159235B (en) | 2021-05-24 | 2021-05-24 | Knowledge collaborative clustering method for product green design |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113159235B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101404073A (en) * | 2008-10-21 | 2009-04-08 | 浙江大学 | Complex product fuzz hierarchy collocation method |
| CN101477521A (en) * | 2008-12-18 | 2009-07-08 | 四川大学 | Non-standard knowledge acquisition method used for constructing mechanical product design knowledge base |
| US7644052B1 (en) * | 2006-03-03 | 2010-01-05 | Adobe Systems Incorporated | System and method of building and using hierarchical knowledge structures |
| CN107491814A (en) * | 2017-07-12 | 2017-12-19 | 浙江大学 | A kind of process case layering knowledge model construction method for knowledge push |
| CN109977572A (en) * | 2019-04-02 | 2019-07-05 | 合肥工业大学 | A kind of product parts blocks of knowledge relational network construction method |
| CN110232058A (en) * | 2019-05-30 | 2019-09-13 | 河海大学常州校区 | A kind of product appearance construction of knowledge base method |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7725299B2 (en) * | 2004-03-01 | 2010-05-25 | Purdue Research Foundation | Multi-tier and multi-domain distributed rapid product configuration and design system |
| CN112424828B (en) * | 2019-07-15 | 2024-02-02 | 广东工业大学 | Nuclear fuzzy C-means quick clustering algorithm integrating space constraint |
-
2021
- 2021-05-24 CN CN202110565342.8A patent/CN113159235B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7644052B1 (en) * | 2006-03-03 | 2010-01-05 | Adobe Systems Incorporated | System and method of building and using hierarchical knowledge structures |
| CN101404073A (en) * | 2008-10-21 | 2009-04-08 | 浙江大学 | Complex product fuzz hierarchy collocation method |
| CN101477521A (en) * | 2008-12-18 | 2009-07-08 | 四川大学 | Non-standard knowledge acquisition method used for constructing mechanical product design knowledge base |
| CN107491814A (en) * | 2017-07-12 | 2017-12-19 | 浙江大学 | A kind of process case layering knowledge model construction method for knowledge push |
| CN109977572A (en) * | 2019-04-02 | 2019-07-05 | 合肥工业大学 | A kind of product parts blocks of knowledge relational network construction method |
| CN110232058A (en) * | 2019-05-30 | 2019-09-13 | 河海大学常州校区 | A kind of product appearance construction of knowledge base method |
Non-Patent Citations (4)
| Title |
|---|
| Constrained ant brood clustering algorithm with adaptive radius: A case study on aspect based sentiment analysis;Mohammed Qasem et.al;《2017 IEEE Symposium Series on Computational Intelligence (SSCI)》;IEEE;20180205;第1-8页 * |
| 基于属性映射的产品绿色设计方案优化方法;张雷 等;《机械工程学报》;20190715;第55卷(第17期);第153-161页 * |
| 基于粗糙集的产品协同设计知识推送方法研究;杨洁等;《中国机械工程》;20091025(第20期);第62-66页 * |
| 基于语义网络的知识协作关键技术研究;代印唐;《中国博士学位论文全文数据库 信息科技辑》;20091215;第I140-7页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113159235A (en) | 2021-07-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106649260B (en) | Product characteristic structure tree construction method based on comment text mining | |
| CN105975488B (en) | A kind of keyword query method based on theme class cluster unit in relational database | |
| US8224805B2 (en) | Method for generating context hierarchy and system for generating context hierarchy | |
| EP2045739A2 (en) | Modeling topics using statistical distributions | |
| CN110543564B (en) | Domain label acquisition method based on topic model | |
| CN111914162B (en) | Method for guiding personalized learning scheme based on knowledge graph | |
| CN109241278B (en) | Scientific research knowledge management method and system | |
| CN113051370B (en) | Similarity measurement method for language evaluation based on mathematical expression | |
| CN108647729B (en) | User portrait acquisition method | |
| CN108304382A (en) | Mass analysis method based on manufacturing process text data digging and system | |
| CN109165351B (en) | Service component search recommendation method based on semantics | |
| JP5780036B2 (en) | Extraction program, extraction method and extraction apparatus | |
| CN113159235B (en) | Knowledge collaborative clustering method for product green design | |
| CN108197295B (en) | Application method of attribute reduction in text classification based on multi-granularity attribute tree | |
| Yu et al. | Research on text categorization of KNN based on K-means for class imbalanced problem | |
| Arevalillo-Herráez et al. | A relevance feedback CBIR algorithm based on fuzzy sets | |
| Wedashwara et al. | Combination of genetic network programming and knapsack problem to support record clustering on distributed databases | |
| CN114511027B (en) | Method for extracting English remote data through big data network | |
| CN114564594A (en) | Knowledge graph user preference entity recall method based on double-tower model | |
| CN115062600A (en) | Code plagiarism detection method based on weighted abstract syntax tree | |
| JP4125951B2 (en) | Text automatic classification method and apparatus, program, and recording medium | |
| Phamtoan et al. | Improving fuzzy clustering algorithm for probability density functions and applying in image recognition | |
| CN116578611B (en) | Knowledge management method and system for inoculated knowledge | |
| CN117556118B (en) | Visual recommendation system and method based on scientific research big data prediction | |
| CN115470332B (en) | Intelligent question-answering system for content matching based on matching degree |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |