CN114997621A - Scheme screening method and system based on trust and opinion similarity comprehensive relationship - Google Patents
Scheme screening method and system based on trust and opinion similarity comprehensive relationship Download PDFInfo
- Publication number
- CN114997621A CN114997621A CN202210573012.8A CN202210573012A CN114997621A CN 114997621 A CN114997621 A CN 114997621A CN 202210573012 A CN202210573012 A CN 202210573012A CN 114997621 A CN114997621 A CN 114997621A
- Authority
- CN
- China
- Prior art keywords
- evaluation
- individuals
- sub
- opinion
- network
- 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
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
Abstract
The embodiment of the invention provides a scheme screening method and system based on a trust and opinion similarity comprehensive relationship, and belongs to the technical field of data mining. The screening method comprises the following steps: acquiring a set of schemes to be screened, a set of attributes for evaluating the schemes, an evaluation opinion matrix of evaluation individuals and a trust matrix among the evaluation individuals; calculating the opinion similarity according to the evaluation opinion matrix; calculating a comprehensive horizontal basis matrix according to the opinion similarity and the trust matrix; constructing a comprehensive relationship network among the evaluation individuals; dividing the comprehensive relationship network into a plurality of sub-networks by adopting a network partitioning algorithm; calculating the feature weight of the feature individual of each sub-network aiming at each sub-network; calculating the comprehensive relation strength of each sub-network; directed edges are added between the sub-networks to get an overall characteristic individual common to all sub-networks. The screening method and the screening system can accurately screen out the optimal scheme.
Description
Technical Field
The invention relates to the technical field of data mining, in particular to a scheme screening method and system based on a trust and opinion similarity comprehensive relationship.
Background
Large-scale population decision methods are one of the common methods in the decision making field. In the prior art, for a plurality of solutions given to the same technical problem, due to different evaluation individuals or platforms and different reference systems and evaluation attributes of the evaluation individuals or platforms, when the solutions are screened, a situation that a solution with the optimal consensus is difficult to select often occurs. In the prior art, conventional algorithms such as a clustering method and the like are difficult to meet the accuracy requirement of solving the technical problem of large-scale group decision-making because the accuracy of sub-network division is not considered and the influence factors of network construction are not considered comprehensively.
Disclosure of Invention
The embodiment of the invention aims to provide a scheme screening method and system based on a comprehensive relationship between trust and opinion similarity.
In order to achieve the above object, an embodiment of the present invention provides a scheme screening method based on a comprehensive relationship between trust and opinion similarity, including:
acquiring a set of schemes to be screened, a set of attributes for evaluating the schemes, an evaluation opinion matrix of evaluation individuals and a trust matrix among the evaluation individuals;
calculating the opinion similarity according to the evaluation opinion matrix;
calculating a comprehensive horizontal basis matrix according to the opinion similarity and the trust matrix;
constructing a comprehensive relationship network among the evaluation individuals;
dividing the comprehensive relationship network into a plurality of sub-networks by adopting a network partitioning algorithm;
calculating the feature weight of the feature individual of each sub-network aiming at each sub-network;
calculating the comprehensive relation strength of each sub-network;
adding directed edges among the sub-networks to obtain a total characteristic individual common to all the sub-networks;
updating the opinion evaluation matrix of all the evaluated individuals according to the formula (1) and the formula (2),
wherein the content of the first and second substances,
for the kth assessment subject d of round t +1 k The element of the ith row and the jth column in the opinion evaluation matrix of (1), u is the number of evaluation individuals,
for the first evaluated individual d of the t round l Is evaluated in the ith row and jth column of the opinion matrix, beta k For the kth individual to be evaluated, b kl For determining evaluation individuals d k And evaluating the individuals d l Whether an indication vector of the directed edge exists between the two vectors;
calculating group consensus degree by adopting a formula (3) and a formula (4) according to the evaluation opinion matrix of the total characteristic individual,
wherein, GCD t Is the group consensus of the t-th round, θ k To evaluate individuals d k M is the number of currently available alternatives, n is the number of attributes used to evaluate the solution,
as evaluation individuals of the t-th round d k The element of the ith row and the jth column in the opinion evaluation matrix,
the element of the ith row and the jth column in the collective opinion matrix after weighted calculation by all the opinion evaluation matrixes for the tth round;
judging whether the group consensus degree is greater than or equal to a first preset threshold value or not;
under the condition that the group consensus degree is judged to be smaller than the first preset threshold value, returning to the step of updating the evaluation opinion matrixes of all the evaluation individuals according to the formula (1) and the formula (2);
and under the condition that the group consensus degree is judged to be greater than or equal to a first preset threshold value, the scheme is sorted by adopting the collective opinion matrix.
Optionally, calculating the opinion similarity according to the opinion evaluation matrix includes:
the opinion similarity is calculated according to formula (7),
wherein, SD l For the first evaluation of individuals d h And the l th assessment subject d l The degree of opinion similarity between, r, ij evaluating individuals for the h h The element of the ith row and the jth column in the opinion evaluation matrix.
Optionally, calculating a comprehensive level basis matrix according to the opinion similarity and the trust matrix includes:
calculating the comprehensive horizontal basis matrix according to formula (8),
wherein the content of the first and second substances,
evaluating individuals for the first h For the l th evaluated individual d l Alpha is a weight coefficient,
to evaluate individuals d h For evaluation of individuals d l The confidence score of (a) is calculated,
to evaluate an individual d h For the l th evaluated individual d l Directional opinion similarity.
Optionally, constructing the comprehensive relationship network between the evaluation individuals comprises:
in the case where there is an opinion similarity and a trust score between two of the evaluation individuals, determining a directed edge between each two of the evaluation individuals according to the formula (9), (10), (11) or (12),
wherein, the first and the second end of the pipe are connected with each other,
to evaluate individuals d h For evaluation of individuals d l Is provided with a directional edge of the frame,
to evaluate individuals d h For evaluation of individuals d l The basic value of the comprehensive relationship of (a),
is a threshold value for the underlying value of the synthetic relationship,
to evaluate individuals d h For evaluation of individuals d l The confidence score of (a) is calculated,
is a threshold value for the confidence score that,
to evaluate individuals d h For evaluation of individuals d l The degree of similarity of the opinions of (2),
a threshold value of opinion similarity;
in the case of equations (9) to (12), in the case where there is an opinion similarity between two of the evaluation individuals and there is no trust relationship, determining a directed edge for each two of the evaluation individuals according to equation (13),
optionally, dividing the integrated relationship network into a plurality of sub-networks by using a network partitioning algorithm includes:
randomly selecting an evaluation individual in the comprehensive relationship network;
taking the selected evaluation individual as a terminal point, and searching the evaluation individual which takes the evaluation individual as the terminal point and possibly as a starting point according to the direction of a directed edge in the comprehensive relationship network;
selecting all possible evaluation individuals as starting points and the selected evaluation individuals to join the same sub-network;
judging whether the unselected evaluation individuals exist in the comprehensive relationship network;
under the condition that the evaluation individuals which are not selected still exist in the comprehensive relationship network, returning to the step of randomly selecting one evaluation individual in the comprehensive relationship network;
and outputting a plurality of sub-networks when judging that the unselected evaluation individuals do not exist in the comprehensive relation network.
Optionally, calculating the feature weight of the feature individual of each sub-network includes:
calculating an out-of-network eigenvalue according to equation (14),
wherein, SD max To evaluate individuals as characteristic individuals h The value of the out-of-network characteristic of (c),
to evaluate individuals d h And evaluation of individuals d y The degree of similarity of the directional opinions between them,
is the set of characteristic individuals of the current sub-network,
set of characteristic individuals for any remaining sub-networks, G τ As the current subnetwork, G s Is any one of the rest sub-networks, M is the set of the sub-networks, and tau and s are sequence number indexes;
calculating the feature weight value according to formula (15) and formula (16),
RD＝TDC(d)·SD max , (15)
wherein RD is an evaluation individual d as a characteristic individual h The characteristic weight of (c), TDC (d) is an evaluation individual d h The intra-network eigenvalues in the located sub-network,
to evaluate individuals d h For evaluation of individuals d l A confidence score of.
Optionally, calculating the integrated relationship strength of each of the sub-networks comprises:
calculating the integrated relationship strength according to equation (17),
wherein the content of the first and second substances,
in order to be the strength of the comprehensive relationship,
for sub-network G τ The number of directed edges in, Q being the sub-network G τ The number of individuals was evaluated.
Optionally, adding a directed edge between the sub-networks to obtain a total feature individual common to all the sub-networks includes:
traversing each of the sub-networks, and randomly selecting one of the sub-networks;
determining an evaluation individual with the largest directional opinion similarity with the selected feature individual of the sub-network in each of the rest sub-networks;
the synthetic relationship base value is calculated according to the formula (18),
wherein, the first and the second end of the pipe are connected with each other,
evaluation of individuals as selected characteristic individuals d h And an evaluation individual d as a defined characteristic individual l The fundamental value of the comprehensive relationship between the two,
to determine the overall strength of the relationship of the sub-networks in which the characteristic individuals are located,
determining a directional opinion similarity between the selected evaluation individual and the determined feature individuals of the sub-network;
selecting the selected evaluation individual in the sub-network corresponding to the maximum comprehensive relation basic value and the determined characteristic individual of the sub-network, and adding directed edges;
calculating a comprehensive relation basic value of two reverse evaluation individuals of the two sub-networks corresponding to the directed edge;
judging whether the comprehensive relation basic value is larger than a threshold value or not;
and in the case that the comprehensive relation base value is judged to be larger than the threshold value, adding opposite directed edges between the two evaluation individuals of the two sub-networks again.
In another aspect, the present invention further provides an optimal solution screening system based on a comprehensive relationship between trust and opinion similarity, where the screening system includes a processor configured to execute any one of the screening methods described above.
According to the technical scheme, the scheme screening method and system based on the trust and opinion similarity comprehensive relationship divide a comprehensive relationship network between the evaluation individuals into a plurality of sub-networks, calculate the group identity degree aiming at the characteristic individuals of each sub-network, enable the group identity degree to reach the preset requirement through continuous iteration and updating, and finally carry out the sequencing of the scheme by combining the collective opinion matrixes of all the evaluation individuals. Compared with the prior art, the method and the system provided by the invention construct the comprehensive relationship network among all the evaluation individuals by combining the trust score and the opinion similarity, and enable the total characteristic individuals to exist in the comprehensive relationship network by adding the directed edges, and update the opinions of all the evaluation individuals in the network environment, so that the group consensus is obtained and the change of the initial opinions of all the evaluation individuals is reduced.
Additional features and advantages of embodiments of the invention will be set forth in the detailed description which follows.
Drawings
The accompanying drawings, which are included to provide a further understanding of the embodiments of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the embodiments of the invention without limiting the embodiments of the invention. In the drawings:
FIG. 1 is a flow diagram of a scenario screening method based on a comprehensive relationship of trust and opinion similarity according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a directed graph according to one embodiment of the present invention;
FIG. 3 is a flow diagram of a network partitioning algorithm according to one embodiment of the present invention;
FIG. 4 is a flow diagram of a method of adding directed edges according to one embodiment of the invention.
Detailed Description
The following detailed description of embodiments of the invention refers to the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating embodiments of the invention, are given by way of illustration and explanation only, not limitation.
Fig. 1 is a flowchart illustrating a scheme screening method based on a trust and opinion similarity integrated relationship according to an embodiment of the present invention. In fig. 1, the screening method may include:
in step S10, a set of the schemes to be screened, a set of attributes used for evaluating the schemes, a rating opinion matrix of the evaluation individuals, and a trust matrix between the evaluation individuals are acquired;
in step S11, calculating opinion similarity based on the opinion evaluation matrix;
in step S12, calculating a comprehensive level basis matrix according to the opinion similarity and the trust matrix;
in step S13, a comprehensive relationship network between the evaluation individuals is constructed;
in step S14, a network partitioning algorithm is used to divide the integrated relationship network into a plurality of sub-networks;
in step S15, for each sub-network, a feature weight of the feature individual for each sub-network is calculated;
in step S16, the integrated relationship strength of each sub-network is calculated;
in step S17, adding directed edges between the sub-networks to obtain total feature individuals common to all the sub-networks;
in step S18, the opinion evaluation matrix of all the evaluated individuals is updated according to the formula (1) and the formula (2),
wherein the content of the first and second substances,
for the kth assessment individual d of round t +1 k The element of the ith row and the jth column in the opinion evaluation matrix of (1), u is the number of evaluation individuals,
for the first evaluated individual d of the t round l Is evaluated in the ith row and jth column of the opinion matrix, beta k Evaluating the intrinsic weight of the individual for the kth, b kl For determining evaluation individuals d k And evaluation of individuals d l The vector indicating whether there is a directed edge between them, namely: b is a mixture of kl For the value of the (k, l) th position in the social matrix, there is a directed edge in the directed graph to evaluate individual d k To evaluation of individuals d l When b is greater than kl 1, otherwise 0;
in step S19, the group consensus degree is calculated using formula (3) and formula (4) based on the opinion evaluation matrix of the total feature individuals,
wherein, GCD t Is the group consensus of the t-th round, θ k To evaluate an individual d k M is the number of currently available alternatives, n is the number of attributes used to evaluate the solution,
evaluation individuals for round t d k The element of the ith row and the jth column in the opinion evaluation matrix,
weighting the elements of the ith row and the jth column in the calculated collective opinion matrix by all the opinion evaluation matrixes;
in step S21, determining whether the group identity is greater than or equal to a first preset threshold;
in step S22, when it is determined that the group identity is less than the first preset threshold, returning to the step of updating the evaluation opinion matrices of all the evaluation individuals according to the formula (1) and the formula (2);
in step S23, in the case where it is determined whether the group consensus degree is greater than or equal to a first preset threshold, the schemes are sorted using the collective opinion matrix.
In the method shown in fig. 1, step S10 may obtain a set of schemes to be screened, a set of attributes used for evaluating the schemes, a rating opinion matrix of an evaluation individual, and a trust matrix between the evaluation individuals. Wherein the set of schemes to be screened may be represented as x ═ x (x) 1 ，x 2 ，...，x m ) And m is more than or equal to 2. The set of evaluated individuals may be denoted as V ═ (d) 1 ，d 2 ，...，d u ). The set of attributes for the evaluation scheme may be denoted as a ═ (a) 1 ，a 2 ，...，a n )。
Step S11 may be used to calculate opinion similarity based on the opinion evaluation matrix. Wherein the opinion evaluation matrix can be represented as R l ＝(r l，ij ) m×n 。R l Evaluation of individuals for the first l Opinion evaluation matrix of r l，ij For the opinion matrix R l The ith row and the jth column of the element, and r l，ij ∈[0，1]Representing evaluation of individuals d l For scheme x i About attribute a j The evaluation of (1). r is l，ij The larger the value of (A), the evaluation individual d is represented l For scheme x i About attribute a j The more approved. Opinion similarity may be used to represent the degree of similarity in opinion ratings between two assessing individuals. The larger the value of the opinion similarity, the closer the opinion evaluation between the two evaluated individuals is. In one example of the present invention, the matrix of opinion similarity may be expressed as SM ═ SD hl ) u×u And l ≠ h, l, h ≠ 1, 2. In addition, for two individuals evaluated, two were usedThe opinion evaluation between the users may be the same or opposite, so in this embodiment, the opinion similarity may be directional, that is, directional opinion similarity, and the directional opinion similarity satisfies the following condition:
as for the calculation manner of the opinion similarity, although it may be in various forms known to those skilled in the art, in a preferred example of the present invention, the calculation manner of the opinion similarity may be to calculate the opinion similarity according to formula (5),
wherein, SD l Evaluating individuals for the first h And the l th assessment individual d l The degree of opinion similarity between, r, ij evaluation of individuals for the h-th h The element of the ith row and the jth column in the opinion evaluation matrix.
Steps S12 and S13 may be used to construct a directed graph between the evaluated individuals. In the directed graph, the directed graph may be, for example, as shown in fig. 2. In this fig. 2, directed edges (arrows) represent the tendency relationship between the evaluated individuals. The knowledge system or the evaluation angle of the evaluation individual corresponding to the start point of the directed edge depends on the knowledge system or the evaluation angle of the evaluation individual at the end of the directed edge. The synthetic horizontal basis matrix calculated at step S12 may be used to determine whether a directed edge exists between any two evaluated individuals. The specific manner of calculation for the comprehensive horizontal basis matrix, although many may be known to those skilled in the art. In a preferred example of the present invention, however, the manner of calculating the fundamental value of the synthetic relationship may be calculated according to formula (6),
wherein the content of the first and second substances,
for the first evaluation of individuals d h For the l th evaluated individual d l Alpha is a weight coefficient used for representing the accepting or rejecting relationship of the evaluated individuals to the trust score and the directional opinion similarity,
to evaluate individuals d h For the l th evaluated individual d l The confidence score of (a) is calculated,
to evaluate an individual d h For the l th evaluated individual d l Directional opinion similarity.
The comprehensive horizontal basis matrix can be expressed as equation (7),
based on the representation of the synthetic relationship basis values in the synthetic horizontal basis matrix, a directed graph is obtained as shown in fig. 2. For the specific method of determining the directed edge, although many ways known to those skilled in the art are possible. In a preferred example of the present invention, however, any of the four types of synthetic relationship constraints shown in table 1 may be employed for determination,
TABLE 1
The constraint types shown in this table 1 can also be expressed as: in the case where there is an opinion similarity and a trust score between two evaluation individuals, a directional edge is determined for each of the two evaluation individuals according to the formula (8), (9), (10) or (11),
wherein the content of the first and second substances,
to evaluate individuals d h For evaluation of individuals d l Is provided with a directional edge of the frame,
to evaluate individuals d h For evaluation of individuals d l The basic value of the comprehensive relationship of (a),
is a threshold value for the underlying value of the synthetic relationship,
to evaluate individuals d h For evaluation of individuals d l The confidence score of (a) is calculated,
is a threshold value for the confidence score,
to evaluate individuals d h For evaluation of individuals d l The degree of similarity of the opinions of (2),
is a threshold value of opinion similarity.
For the case that only opinion similarity exists between two evaluation individuals and no trust relationship exists (trust score is 0), the directed edge can be established directly through threshold judgment of opinion similarity. Namely: under the condition that the opinion similarity exists between the two evaluation individuals and the trust relationship does not exist, determining the directed edge of each two evaluation individuals according to the formula (12),
after the directed graph is generated in step S13, it is necessary to find the evaluation individuals pointed by all arrows in the directed graph, which can be the total feature individuals. However, since the arrows between the individual evaluation individuals point relatively randomly, if the overall characteristic individual cannot be found directly in the directed graph, the comprehensive relationship network needs to be first divided into a plurality of sub-networks by using a network partitioning algorithm through step S14. In the step S14, the network partitioning algorithm may be any number of algorithms known to those skilled in the art. In a preferred example of the present invention, the network partitioning algorithm may include the steps as shown in FIG. 3. In fig. 3, the step S14 may include:
in step S30, an evaluation individual is randomly selected in the integrated relationship network;
in step S31, with the selected evaluation individual as the end point, the evaluation individual with the evaluation individual as the end point and possibly as the starting point is searched in the direction of the directional edge in the integrated relationship network. Wherein, the selected evaluation individual is the characteristic individual of the current sub-network;
in step S32, all possible evaluation individuals as starting points and the selected evaluation individuals are selected to join the same sub-network;
in step S33, it is determined whether or not there is any unselected evaluation individual in the integrated relationship network;
under the condition that the unselected evaluation individuals exist in the comprehensive relationship network, returning to the step of randomly selecting one evaluation individual in the comprehensive relationship network;
in step S34, when it is determined that there is no unselected evaluation individual in the integrated relationship network, a plurality of subnetworks are output.
After dividing a plurality of sub-networks in step S14, it is necessary to add directed edges between the evaluation individuals of the respective sub-networks, so as to obtain the total feature individual. Therefore, steps S15 to S17 may be adopted, thereby obtaining the total characteristic individual. On the other hand, in the directed graph obtained in step S13, even if the directed graph is divided into a plurality of sub-networks, it is obviously impossible to find the total feature individuals by directly searching the total feature individuals according to the method in step S15. Therefore, step S15 is required to first calculate the feature weight of the feature individual of each sub-network, then calculate the comprehensive relationship strength by combining the directional edge number of the sub-network and the characteristics of the number of the evaluation individuals, and finally add a directional edge between different evaluation individuals in different sub-networks by combining the comprehensive relationship strength, the comprehensive relationship base value, and the directional opinion similarity, thereby obtaining the total feature individual.
Specifically, step S15 may be used to calculate a feature weight of the feature individuals of each sub-network. In this embodiment, the feature weight value can be used to indicate the importance degree (intra-network feature value) of the feature expression of the current feature individual in the sub-network; on the other hand, the feature weight value can also be used to indicate the degree of correlation (extranet feature value) between the individual features in other subnetworks. Then, the method for calculating the feature weight may be to calculate the out-of-network feature value according to formula (13), calculate the feature weight according to formula (14) and formula (15),
wherein, SD max To evaluate individuals as characteristic individuals h The value of the out-of-network characteristic of (c),
to evaluate an individual d h And evaluating the individuals d y The degree of similarity of the directional opinions between them,
is the set of characteristic individuals of the current sub-network,
set of characteristic individuals for any remaining sub-networks, G τ As the current subnetwork, G s Is any one of the rest sub-networks, M is the set of the sub-networks, and tau and s are sequence number indexes;
RD＝TDC(d)·SD max ， (14)
wherein RD is an evaluation individual d as a characteristic individual h The characteristic weight of (d), TDC (d) is an evaluation unit d h The intra-network eigenvalues in the located sub-network,
to evaluate individuals d l And evaluating the subject d k A confidence score between.
Step S16 may be to calculate the integrated relationship strength according to equation (17),
wherein the content of the first and second substances,
in order to integrate the strength of the relationship,
as a sub-network G τ The number of directed edges in, Q being the sub-network G τ The number of individuals was evaluated.
For the method of adding the directed edge in step S17, although many ways are known to those skilled in the art. However, in a preferred example of the present invention, the step S17 may be a method including the method as shown in fig. 4. In the fig. 4, the step S17 may include:
in step S40, each sub-network is traversed, and one sub-network is randomly selected;
in step S41, an evaluation individual having the greatest opinion similarity with the characteristic individual of the selected sub-network is determined in each of the remaining sub-networks;
in step S42, a comprehensive relationship base value is calculated according to the formula (18),
wherein the content of the first and second substances,
evaluation of individuals as selected characteristic individuals d h And an evaluation individual d as a defined characteristic individual l The basic value of the comprehensive relationship between the two,
to determine the overall strength of the relationship of the sub-networks in which the characteristic individuals are located,
a directional opinion similarity between the selected evaluation individual and the determined feature individuals of the sub-network;
in step S43, selecting the selected evaluation individual in the sub-network corresponding to the maximum synthetic relationship base value and the determined feature individual of the sub-network to add directed edges;
in step S44, a comprehensive relationship base value of two evaluation individual inversions of two subnetworks corresponding to the directed edge is calculated;
in step S45, it is determined whether the integrated relationship base value is greater than a threshold value;
in step S46, in the case where it is determined that the integrated relationship base value is greater than the threshold value, an opposite directed edge is added again between the two evaluation individuals of the two subnetworks.
After the directed edges are added, the method for further obtaining the total feature individuals can be implemented by using the network partitioning algorithm in step S14. Since the foregoing has been described in detail, it is not repeated here.
After the total characteristic individuals are obtained, the evaluation opinion matrix of all the evaluation individuals is further updated through step S18 and step S19, thereby completing the evolution of collective opinions. However, the evolved result does not mean that the current collective opinion matrix can satisfy the requirement of the consensus degree of all the evaluation individuals, and the group consensus degree needs to be further calculated through steps S19 and S20, and a corresponding threshold judgment is performed, that is, whether the group consensus degree is greater than a first preset threshold is judged. If the group identity is determined to be less than the first predetermined threshold, the step S18 is executed again. On the contrary, under the condition that the group consensus degree is judged to be larger than the first preset threshold, the result of the current evolution can already represent the opinions of all the evaluation individuals, so that the scheme can be sorted by directly adopting the collective opinion matrix, and finally, the sorted result is screened.
In another aspect, the present invention further provides an optimal solution screening system based on a comprehensive relationship between trust and opinion similarity, where the screening system includes a processor configured to execute any one of the screening methods described above.
According to the technical scheme, the scheme screening method and system based on the trust and opinion similarity comprehensive relationship divide the comprehensive relationship network among the evaluation individuals into a plurality of sub-networks, calculate the group consensus degree aiming at the characteristic individuals of each sub-network, enable the group consensus degree to reach the preset requirement through continuous iteration and updating, and finally carry out the scheme sequencing by combining the collective opinion matrixes of all the evaluation individuals. Compared with the prior art, the method and the system provided by the invention construct the comprehensive relationship network among all the evaluation individuals by combining the trust score and the opinion similarity, and enable the total characteristic individuals to exist in the comprehensive relationship network by adding the directed edges, and update the opinions of all the evaluation individuals in the network environment, so that the group consensus is obtained and the change of the initial opinions of all the evaluation individuals is reduced.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
The memory may include forms of volatile memory in a computer readable medium, Random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). The memory is an example of a computer-readable medium.
Computer-readable media, including both permanent and non-permanent, removable and non-removable media, may implement the information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Disks (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in the process, method, article, or apparatus that comprises the element.
The above are merely examples of the present application and are not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.
Claims (9)
1. A scheme screening method based on a trust and opinion similarity comprehensive relationship is characterized in that the screening method comprises the following steps:
acquiring a set of schemes to be screened, a set of attributes for evaluating the schemes, an evaluation opinion matrix of evaluation individuals and a trust matrix among the evaluation individuals;
calculating the opinion similarity according to the evaluation opinion matrix;
calculating a comprehensive horizontal basis matrix according to the opinion similarity and the trust matrix;
constructing a comprehensive relationship network among the evaluation individuals;
dividing the comprehensive relationship network into a plurality of sub-networks by adopting a network partitioning algorithm;
calculating the feature weight of the feature individual of each sub-network aiming at each sub-network;
calculating the comprehensive relation strength of each sub-network;
adding directed edges among the sub-networks to obtain a total characteristic individual common to all the sub-networks;
updating the opinion evaluation matrix of all the evaluation individuals according to the formula (1) and the formula (2),
wherein the content of the first and second substances,
for the kth assessment individual d of round t +1 k The element of the ith row and the jth column in the opinion evaluation matrix of (1), u is the number of evaluation individuals,
for the first evaluation individual d of the t-th round l Is evaluated in the ith row and jth column of the opinion matrix, beta k For the kth individual to be evaluated, b kl For determining evaluation individuals d k And evaluating the individuals d l Whether an indication vector of a directed edge exists between the two vectors;
calculating group consensus degree by adopting a formula (3) and a formula (4) according to the evaluation opinion matrix of the total characteristic individuals,
wherein, GCD t Is the group consensus of the t-th round, θ k To evaluate individuals d k M is the number of currently available alternatives, n is the number of attributes used to evaluate the solution,
as evaluation individuals of the t-th round d k The element of the ith row and the jth column in the opinion evaluation matrix,
the element of the ith row and the jth column in the collective opinion matrix after weighted calculation by all the opinion evaluation matrixes for the tth round;
judging whether the group consensus degree is greater than or equal to a first preset threshold value or not;
under the condition that the group consensus degree is judged to be smaller than the first preset threshold value, returning to the step of updating the evaluation opinion matrixes of all the evaluation individuals according to the formula (1) and the formula (2);
and under the condition that the group consensus degree is judged to be greater than or equal to a first preset threshold value, the scheme is sorted by adopting the collective opinion matrix.
2. The screening method according to claim 1, wherein calculating the opinion similarity according to the opinion evaluation matrix comprises:
the opinion similarity is calculated according to formula (7),
wherein, SD l Evaluating individuals for the first h And the l th assessment individual d l The opinion similarity between r ,ij Evaluating individuals for the h h The element of the ith row and the jth column in the opinion evaluation matrix.
3. The screening method of claim 1, wherein computing a synthetic level basis matrix from the opinion similarity and trust matrix comprises:
calculating the synthetic horizontal basis matrix according to equation (8),
wherein the content of the first and second substances,
evaluating individuals for the first h For the l th evaluated individual d l Alpha is a weight coefficient,
to evaluate individuals d h For evaluation of individuals d l The confidence score of (a) is calculated,
to evaluate individuals d h For the l th evaluated individual d l Directional opinion similarity of (c).
4. The screening method of claim 1, wherein constructing the comprehensive relationship network between the individuals under evaluation comprises:
in the case where there is an opinion similarity and a trust score between two of the evaluation individuals, determining a directed edge between each two of the evaluation individuals according to the formula (9), (10), (11) or (12),
wherein the content of the first and second substances,
to evaluate an individual d h For evaluation of individuals d l Is provided with a directional edge of the frame,
to evaluate individuals d h For evaluation of individuals d l The basic value of the comprehensive relationship of (a),
is a threshold value for the underlying value of the synthetic relationship,
to evaluate an individual d h For evaluation of individuals d l The confidence score of (a) is calculated,
is a threshold value for the confidence score,
to evaluate an individual d h For evaluation of individuals d l The degree of similarity of the opinions of (2),
a threshold value for opinion similarity;
in the case of equations (9) to (12), in the case where there is an opinion similarity between two of the evaluation individuals and there is no trust relationship, determining a directed edge between each two of the evaluation individuals according to equation (13),
5. the screening method of claim 1, wherein partitioning the integrated relationship network into a plurality of sub-networks using a network partitioning algorithm comprises:
randomly selecting an evaluation individual in the comprehensive relationship network;
taking the selected evaluation individual as a terminal point, and searching the evaluation individual which takes the evaluation individual as the terminal point and possibly as a starting point according to the direction of a directed edge in the comprehensive relationship network;
selecting all possible evaluation individuals as starting points and the selected evaluation individuals to join the same sub-network;
judging whether the unselected evaluation individuals exist in the comprehensive relationship network;
under the condition that the evaluation individuals which are not selected still exist in the comprehensive relationship network, returning to the step of randomly selecting one evaluation individual in the comprehensive relationship network;
and under the condition that the evaluation individuals which are not selected do not exist in the comprehensive relation network, outputting a plurality of sub-networks.
6. The screening method according to claim 1, wherein calculating the feature weight of the feature individuals of each of the sub-networks comprises:
calculating an out-of-network eigenvalue according to equation (14),
wherein, SD max To evaluate individuals as characteristic individuals h The value of the out-of-network characteristic of (c),
to evaluate individuals d h And evaluating the individuals d y The degree of similarity of the directional opinions between them,
is the set of characteristic individuals of the current sub-network,
set of characteristic individuals for any remaining sub-networks, G τ As the current subnetwork, G s Is any one of the rest sub-networks, M is the set of the sub-networks, and tau and s are sequence number indexes;
calculating the feature weight value according to formula (15) and formula (16),
RD＝TDC(d)·SD max ， (15)
wherein RD is an evaluation individual d as a characteristic individual h The characteristic weight of (d), TDC (d) is an evaluation unit d h The intra-network eigenvalues in the sub-network in which they are located,
to evaluate individuals d h For evaluation of individuals d l A confidence score of.
7. The screening method of claim 1, wherein calculating the composite relationship strength for each of the sub-networks comprises:
calculating the integrated relationship strength according to equation (17),
wherein the content of the first and second substances,
in order to be the strength of the comprehensive relationship,
for sub-network G τ The number of directed edges in, Q being the sub-network G τ The number of individuals was evaluated.
8. The screening method of claim 7, wherein adding directed edges between the sub-networks to obtain total feature individuals common to all sub-networks comprises:
traversing each of the sub-networks, and randomly selecting one of the sub-networks;
determining the evaluation individual with the largest directional opinion similarity with the selected characteristic individual of the sub-network in each of the rest sub-networks;
the synthetic relationship base value is calculated according to the formula (18),
wherein the content of the first and second substances,
evaluation of individuals as selected characteristic individuals d h And an evaluation individual d as a defined characteristic individual l The basic value of the comprehensive relationship between the two,
for determining the comprehensive relationship strength of the sub-networks in which the characteristic individuals are located,
determining a directional opinion similarity between the selected evaluation individual and the determined feature individuals of the sub-network;
selecting the selected evaluation individual in the sub-network corresponding to the maximum comprehensive relation base value and the determined characteristic individual of the sub-network, and adding a directed edge;
calculating a comprehensive relation basic value of two reverse evaluation individuals of the two sub-networks corresponding to the directed edge;
judging whether the comprehensive relation basic value is larger than a threshold value or not;
and in the case that the comprehensive relation base value is judged to be larger than the threshold value, adding opposite directed edges between the two evaluation individuals of the two sub-networks again.
9. An optimal solution screening system based on a trust and opinion similarity comprehensive relationship, characterized in that the screening system comprises a processor for executing the screening method according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210573012.8A CN114997621A (en) | 2022-05-25 | 2022-05-25 | Scheme screening method and system based on trust and opinion similarity comprehensive relationship |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210573012.8A CN114997621A (en) | 2022-05-25 | 2022-05-25 | Scheme screening method and system based on trust and opinion similarity comprehensive relationship |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114997621A true CN114997621A (en) | 2022-09-02 |
Family
ID=83028797
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210573012.8A Pending CN114997621A (en) | 2022-05-25 | 2022-05-25 | Scheme screening method and system based on trust and opinion similarity comprehensive relationship |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114997621A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116484113A (en) * | 2023-04-12 | 2023-07-25 | 烟台大学 | Group view prediction method and system based on dynamic trust perception |
-
2022
- 2022-05-25 CN CN202210573012.8A patent/CN114997621A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116484113A (en) * | 2023-04-12 | 2023-07-25 | 烟台大学 | Group view prediction method and system based on dynamic trust perception |
| CN116484113B (en) * | 2023-04-12 | 2023-09-19 | 烟台大学 | Group view prediction method and system based on dynamic trust perception |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zahra et al. | Novel centroid selection approaches for KMeans-clustering based recommender systems | |
| US10031945B2 (en) | Automated outlier detection | |
| CN110570111A (en) | Enterprise risk prediction method, model training method, device and equipment | |
| CN110991474A (en) | Machine learning modeling platform | |
| CN110033097B (en) | Method and device for determining association relation between user and article based on multiple data fields | |
| US10628412B2 (en) | Iterative visualization of a cohort for weighted high-dimensional categorical data | |
| US20180307743A1 (en) | Mapping method and device | |
| CN109189876A (en) | A kind of data processing method and device | |
| CN112470131A (en) | Apparatus and method for detecting anomalies in a data set and computer program products corresponding thereto | |
| CN114997621A (en) | Scheme screening method and system based on trust and opinion similarity comprehensive relationship | |
| CN111178435B (en) | Classification model training method and system, electronic equipment and storage medium | |
| WO2019184480A1 (en) | Item recommendation | |
| CN112801231A (en) | Decision model training method and device for business object classification | |
| CN113158022A (en) | Service recommendation method, device, server and storage medium | |
| CN112733035A (en) | Knowledge point recommendation method and device based on knowledge graph, storage medium and electronic device | |
| CN109993338B (en) | Link prediction method and device | |
| US7805667B2 (en) | System and method for identifying target node graphs from predetermined seed node subsets | |
| CN107423319B (en) | Junk web page detection method | |
| KR102323424B1 (en) | Rating Prediction Method for Recommendation Algorithm Based on Observed Ratings and Similarity Graphs | |
| US20030037016A1 (en) | Method and apparatus for representing and generating evaluation functions in a data classification system | |
| JP2010073195A (en) | Collaborative filtering processing method and collaborative filtering processing program | |
| US11295229B1 (en) | Scalable generation of multidimensional features for machine learning | |
| US10698910B2 (en) | Generating cohorts using automated weighting and multi-level ranking | |
| CN111125541B (en) | Method for acquiring sustainable multi-cloud service combination for multiple users | |
| KR102136222B1 (en) | System and method for clustering graph data and computer program for the same |
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 |