CN112734154A - Multi-factor public opinion risk assessment method based on fuzzy number similarity - Google Patents
Multi-factor public opinion risk assessment method based on fuzzy number similarity Download PDFInfo
- Publication number
- CN112734154A CN112734154A CN202011279436.0A CN202011279436A CN112734154A CN 112734154 A CN112734154 A CN 112734154A CN 202011279436 A CN202011279436 A CN 202011279436A CN 112734154 A CN112734154 A CN 112734154A
- Authority
- CN
- China
- Prior art keywords
- risk
- public opinion
- evaluation
- factor
- influence degree
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 56
- 238000012502 risk assessment Methods 0.000 title claims abstract description 19
- 238000011156 evaluation Methods 0.000 claims abstract description 80
- 230000008569 process Effects 0.000 claims abstract description 11
- 238000012545 processing Methods 0.000 claims abstract description 8
- 238000006243 chemical reaction Methods 0.000 claims abstract description 7
- 238000013507 mapping Methods 0.000 claims abstract description 6
- 239000011159 matrix material Substances 0.000 claims description 32
- 238000010191 image analysis Methods 0.000 claims description 3
- 230000009286 beneficial effect Effects 0.000 abstract description 6
- 238000007670 refining Methods 0.000 abstract description 3
- 238000004458 analytical method Methods 0.000 description 16
- 230000008451 emotion Effects 0.000 description 8
- 238000007726 management method Methods 0.000 description 8
- 238000010276 construction Methods 0.000 description 4
- 238000013210 evaluation model Methods 0.000 description 4
- 238000012163 sequencing technique Methods 0.000 description 4
- 230000007547 defect Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000002265 prevention Effects 0.000 description 3
- 238000004445 quantitative analysis Methods 0.000 description 3
- BTCSSZJGUNDROE-UHFFFAOYSA-N gamma-aminobutyric acid Chemical compound NCCCC(O)=O BTCSSZJGUNDROE-UHFFFAOYSA-N 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012854 evaluation process Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000004451 qualitative analysis Methods 0.000 description 1
- 230000007480 spreading Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
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
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- 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
- 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
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
- G06F18/24155—Bayesian classification
-
- 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
- G06F40/216—Parsing using statistical methods
-
- 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—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/18—Legal services
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Evolutionary Computation (AREA)
- Tourism & Hospitality (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Probability & Statistics with Applications (AREA)
- Life Sciences & Earth Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- General Health & Medical Sciences (AREA)
- General Business, Economics & Management (AREA)
- Development Economics (AREA)
- Primary Health Care (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Technology Law (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Educational Administration (AREA)
- Computational Linguistics (AREA)
- Quality & Reliability (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a multi-factor public opinion risk assessment method based on fuzzy number similarity. Firstly, obtaining the influence degree of risk factors by using an extended good-bad solution distance method, and simultaneously obtaining an index model structure and the arrangement sequence of each layer of elements by using an analytic hierarchy process; secondly, processing the input of different data types to obtain risk factor fact coefficient evaluation values and fusing the risk factor fact coefficient evaluation values; and finally, mapping the evaluation comprehensive value to a risk grade and fuzzy number conversion system based on fuzzy number similarity to obtain a public opinion risk grade. According to the invention, risk factors in a workflow can be fused from multiple sources, and fine-grained assessment can be given to the current public opinion risk more comprehensively and comprehensively; different statistical strategies are carried out on the negative topics and the positive topics, the method is more reasonable than the traditional method, the public opinion tendency can be better reflected, the constructed index model is beneficial to refining of risk factors, and meanwhile, the risk indexes have more objective influence degree evaluation.
Description
Technical Field
The invention relates to the technical field of risk management and fuzzy mathematics, in particular to a multi-factor public opinion risk assessment method based on fuzzy number similarity.
Background
With the rapid development of the internet, the forming and spreading speed of public opinions is increasing, and the public opinions aiming at hot events in the network media bring great risks and challenges to the work. Therefore, the demand for accurate assessment of public opinion risk in the working process is gradually growing, wherein how to construct the assessment content of public opinion risk and how to perform quantitative analysis of risk level are two core problems to be solved. The existing methods such as machine learning and fuzzy matrix are difficult to be effectively applied to actual business due to limited historical data, high risk assessment precision requirement and the like.
For this reason, the risk assessment model based on the fuzzy number similarity has strong technical advantages in terms of processing risk ambiguity and uncertainty. The method applies fuzzy set theory to risk assessment and multi-attribute decision making, converts qualitative evaluation into interval number form mathematical expression on the basis of fuzzy number continuous membership function, and performs fuzzy operation, fuzzy aggregation or fuzzy reasoning and other operations on the interval number form mathematical expression on the basis, thereby obtaining accurate risk level. Therefore, the risk assessment based on the fuzzy analysis method can effectively solve the conversion problem between qualitative evaluation and quantitative analysis. The method mainly comprises four types of models: an evaluation model based on fuzzy sorting, an evaluation model based on fuzzy number similarity, a fuzzy inference model and a fuzzy matrix model. In the evaluation model based on fuzzy sorting, a sorting benchmark needs to be established first, and then the risk level is determined according to the sorting result. In an actual working scene, a certain workflow is usually evaluated independently, and the comparison of a plurality of evaluation objects is less likely to occur, so that the sequencing model is difficult to establish. In the aspect of Fuzzy Inference, researchers usually describe membership functions and If-Then rules by using a risk matrix structure in a Fuzzy Inference System (FIS for short), and evaluate and predict risks based on event information.
Fuzzy matrix models represent a class of qualitative analysis methods that incorporate fuzzy set theory. The model replaces real numbers with fuzzy interval numbers, reflects artificial evaluation more truly, and thereby refines the matrix granularity. Although the number of risk knots (risk tie) is reduced to a certain extent, and the risk grade distinguishing capability is improved, the method still belongs to qualitative assessment in nature and is very complicated in construction process, and an effective quantitative analysis model is difficult to establish for a business containing multi-factor risks.
The fuzzy number similarity-based assessment model determines the early warning level by utilizing the similarity of the fuzzy risk measure and the risk level, a single assessment object directly corresponds to the risk level, a plurality of objects do not depend on each other, when only one work flow exists, the corresponding risk early warning level can be output, and human factors are eliminated to determine the risk level. Therefore, the evaluation requirement of the service scene is met, and meanwhile, higher evaluation accuracy is achieved.
In conclusion, the fuzzy number similarity-based evaluation model not only can effectively convert qualitative risk evaluation in a working scene, but also can establish an independent evaluation mechanism for single work. But the method still has defects in the aspects of special data processing, evaluation accuracy and the like, and becomes a bottleneck for restricting accurate public opinion risk evaluation.
One of the existing technologies is a management system, which mainly serves basic business process specification management and time management, and lacks the evaluation and management of multi-factor public opinion risks. The existing workflow risk prevention and control system is defined in a form of a table, and each risk factor entry is defined as a risk item, a risk level and a risk responsible person.
The defects of the prior art are as follows: (1) the existing mature management system is lack of assessment and management of multi-factor work public opinion risks; (2) the existing risk prevention and control system has single risk point factor and does not macroscopically consider multiple risk factors; (3) the existing risk prevention and control system has fixed risk grade, the evaluation process is not fully combined with objective data, and the evaluation result is inaccurate.
Another prior art is a network public opinion multidimensional analysis system, which comprises the following main functions: negative public opinion analysis, hotspot analysis, public opinion outbreak trend analysis, and the like. The negative public sentiment analysis is based on positive/negative sentiment value analysis of relevant texts of each large mainstream network platform to obtain an evaluation result.
The second disadvantage of the prior art is that: (1) in the system, a risk assessment algorithm for multidimensional public opinion analysis is lacked, and the analysis content has more dimensions but is independent of each other; (2) negative public sentiment in the system is simple statistics of the sentiment value of the text, the analysis of focal topics is lacked, and the accuracy of negative sentiment analysis is not high.
Disclosure of Invention
The invention aims to overcome the defects of the existing method and provides a multi-factor public opinion risk assessment method based on fuzzy number similarity. The invention solves the main problems that (1) in the existing public opinion risk index system, the risk factor is single, and the risk grade is fixed; (2) the conventional public opinion analysis platform only performs simple emotion analysis on a text and does not perform accurate analysis on a focus topic; (3) the existing public opinion risk management method cannot overcome the ambiguity and uncertainty of risks, does not comprehensively and dynamically analyze the current public opinion state, and only gives a fixed assessment result.
In order to solve the problems, the invention provides a multi-factor public opinion risk assessment method based on fuzzy number similarity, which comprises the following steps:
extracting public opinion risk factors based on historical working documents and a workflow rule base, constructing a risk factor set, constructing an expert group, expressing the influence degree of the risk factor set by using a phrase set by an expert, and collecting the data to form a risk factor influence degree data set;
the method comprises the steps that a three-layer structure of an initial public opinion risk index model is built by using an analytic hierarchy process, and each level node is defined, wherein the model consists of a target layer, a criterion layer and an index layer, the target layer determines that the main body of evaluation is public opinion risk, the criterion layer defines four standard criteria of a decision maker, a party, situation attributes and a public opinion state based on fishbone image analysis of historical working documents, and the index layer consists of risk factors and corresponding influence degrees;
constructing an evaluation matrix based on the risk factor influence degree data set, and calculating the influence degree of each risk factor by using an extended good and bad solution distance method and integrating a multi-dimensional evaluation matrix;
determining the arrangement order of each layer element in the initial public opinion risk index model by using an analytic hierarchy process based on the influence degree of each risk factor to obtain a final public opinion risk index model with a complete system structure;
collecting and analyzing relevant public opinion data of the current processed work, and evaluating risk factors in the final public opinion risk index model to obtain a fact coefficient of each risk index in a model index layer;
fusing the risk factor influence degree in the final public opinion risk index model index layer with the fact coefficient to obtain a public opinion risk comprehensive value based on multi-factor evaluation;
based on the graphic characteristics of the fuzzy number, mapping the public opinion risk comprehensive value based on the multi-factor evaluation to a risk grade and fuzzy number conversion system by calculating the fuzzy number similarity based on the gyration radius, thereby obtaining the public opinion risk grade of the current processing work;
and if the public opinion risk level or the variation trend thereof meets the triggering early warning condition, executing corresponding early warning operation according to a risk plan.
Preferably, the phrase set specifically includes:
the phrase set LT { "considerably low" (absolute low), "very low" (very low), "low" (far low), "medium)," high "(high)," very high "(very high)," considerably high "(absolute high).
Preferably, the evaluation matrix is constructed based on the risk factor influence degree data set, and the influence degrees of the risk factors are calculated by using an extended good-bad solution distance method and a multi-dimensional evaluation matrix, specifically:
constructing an evaluation matrix based on the risk factor influence degree data set;
constructing positive and negative ideal solutions for risk factor evaluation;
calculating the Euclidean distance between each evaluation scheme in the evaluation matrix and the positive ideal solution and the Euclidean distance between each evaluation scheme in the evaluation matrix and the negative ideal solution;
calculating the distance ratio of each evaluation scheme in the evaluation matrix to the positive and negative ideal solutions, and normalizing the distance ratio to obtain the weight of each evaluation scheme;
and carrying out weighted summation on the weight of each evaluation scheme and each evaluation scheme to obtain the evaluation of the influence degree of each risk factor.
The multi-factor public opinion risk assessment method based on the fuzzy number similarity can be used for fusing multi-dimensional information such as risk factors in a workflow from multiple sources, and comprehensively providing fine-grained assessment for the current public opinion risk; the method has the advantages that different statistical strategies are carried out on the negative topics and the positive topics, the method is more reasonable than the traditional method, the public opinion tendency can be better reflected, the index model constructed by the method is beneficial to refining of risk factors in the construction stage, the model is beneficial to expanding in the maintenance stage, and meanwhile, the risk indexes have more objective influence degree evaluation.
Drawings
Fig. 1 is a general flowchart of a multi-factor public opinion risk assessment method based on fuzzy number similarity according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a system for converting a linguistic description into an ambiguity number according to an embodiment of the present invention;
FIG. 3 is a diagram of a public opinion risk index model according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a fuzzy number and comparison scale conversion system of an embodiment of the present invention;
FIG. 5 is a numerical query graph of a random consistency indicator RI according to an embodiment of the present invention;
fig. 6 is a flowchart of calculating negative emotion value of public sentiment text according to an embodiment of 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.
Fig. 1 is a method for multi-factor public opinion risk assessment based on fuzzy number similarity according to an embodiment of the present invention, as shown in fig. 1, the method includes:
s1, extracting public opinion risk factors based on historical working documents and a workflow rule base, constructing a risk factor set, constructing an expert group, expressing the influence degree of the risk factor set by the expert by using a phrase set, and collecting the data to form a risk factor influence degree data set;
s2, constructing a three-layer structure of an initial public opinion risk index model by using an analytic hierarchy process and defining each level node, wherein the model consists of a target layer, a criterion layer and an index layer, the target layer determines that the main body of evaluation is public opinion risk, the criterion layer defines four standard criteria of a decision maker, a party, case attributes and a public opinion state based on fishbone image analysis of historical working documents, and the index layer consists of risk factors and corresponding influence degrees;
s3, constructing an evaluation matrix based on the risk factor influence degree data set, and calculating the influence degree of each risk factor by using an extended good-bad solution distance method and a comprehensive multidimensional evaluation matrix;
s4, determining the arrangement sequence of each layer element in the initial public opinion risk index model by using an analytic hierarchy process based on the influence degree of each risk factor to obtain a final public opinion risk index model with a complete system structure;
s5, collecting and analyzing the current processing work related public sentiment data, and evaluating the risk factors in the final public sentiment risk index model to obtain the fact coefficient of each risk index in the model index layer;
s6, fusing the influence degree of the risk factors in the final public opinion risk index model index layer and the fact coefficient to obtain a public opinion risk comprehensive value based on multi-factor evaluation;
s7, based on the graphic characteristics of the fuzzy number, mapping the public opinion risk comprehensive value based on multi-factor evaluation to a risk grade and fuzzy number conversion system by calculating the fuzzy number similarity based on the gyration radius, thereby obtaining the current processing public opinion risk grade;
and S8, if the public opinion risk level or the variation trend thereof meets the triggering early warning condition, executing corresponding early warning operation according to a risk plan.
Step S1, wherein the RISK factor set is RISK ═ { r1,r2,…,rnThe phrase set LT { "considerably low" (absolute low), "very low" (very low), "low" (low), "medium)", "high" (high), "very high" (very high), "considerably high" (absolute high).
Step S3 is specifically as follows:
s3-1, constructing an evaluation matrix based on the risk factor influence degree data set acquired in the S1 preprocessing stage:
wherein N represents the number of risk factors, M is the number of evaluations in the data set, smnAn evaluation value, s, is given for the mth item to evaluate the degree of influence of the risk factor for the nth itemmnIs a generalized fuzzy number (a) in the form of a section corresponding to the language description of the evaluation1,a2,a3,a4(ii) a w), the mapping relationship is shown in fig. 2.
S3-2, constructing a positive and negative ideal solution of risk factor evaluation:
wherein the content of the first and second substances,
and
are respectively defined as:
s3-3, calculating each evaluation scheme SmEuclidean distance from the positive ideal solution:
calculating each evaluation scenario smEuclidean distance from the negative ideal solution:
s3-4, calculating the distance ratio of each evaluation scheme to the positive and negative ideal solutions, and normalizing the distance ratios to obtain the weight of the evaluation scheme:
s3-5, weighting each evaluation scheme and each evaluation scheme SmCarrying out weighted summation to obtain the evaluation of the influence degree of each risk factor
In step S4, the public opinion risk index model includes a target layer a, a criterion layer B, and an index layer C, as shown in fig. 3, which is specifically as follows:
and S4-1, the risk factors are programmed into the index layer C according to the influence degree of the risk factors, and the risk factors are connected with corresponding criteria according to the own attributes of the risk factors. In the child nodes of B, the indexes are sorted according to the degree of influence. And assigning initial sequencing to the nodes of the B layer, and performing consistency verification of the model.
S4-2, constructing a comparison matrix delta,
obtaining the importance comparison result delta of the risk factor i and the risk factor j according to the figure 4 for the evaluation weight gap between the risk factor i and the risk factor jijThereby constructing a pair-wise comparison matrix delta, deltaijIs a scalar, having the properties:
and S4-3, verifying the consistency of the hierarchy list sorting. And calculating the eigenvector of the maximum characteristic root of the matrix delta, and normalizing to obtain the ranking weight of the relative importance of the same layer of factors to the previous layer. The consistency index CI to be verified is as follows:
wherein λ is the maximum characteristic root of the matrix C, and n is the dimension of the matrix C. If CI is 0, the ordering has complete consistency; the inconsistency increases as CI increases.
In order to further quantify the consistency index, a random consistency index RI is introduced, and the relationship between RI and the matrix order n is shown in fig. 5. The quantified consistency indicator CR is defined as:
if CR <0.1, the matrix C passes the consistency check.
And S4-4, calculating the overall hierarchical ranking of all risk factors of the index layer relative to the highest layer. The ranking weight of the 4 nodes B1, B2, B3 and B4 on the highest layer of the criterion layer is a1,a2,a3,a4The hierarchical list of n risk factors of the index layer to the BJ is ordered as b1j,b2j,…,bnj. Wherein CI1Corresponding to the contrast matrix constructed by C1, C2 and B1. The consistency ratio of the total hierarchical ordering is:
when CR <0.1, the hierarchical total ordering passes the consistency check.
And S4-5, traversing the sequencing and sequencing weight calculation of the nodes in the criterion layer until the multilevel public opinion risk index model passes consistency verification, thereby completing the construction of the public opinion risk index model. A more objective influence degree weight is obtained by integrating the weight evaluation given by a plurality of experts in the expert group to the risk factors by using a TOPSIS (TOPSIS) method. Besides, a more systematic public opinion risk index system is constructed by using an Analytic Hierarchy Process (AHP), which is beneficial to updating and maintaining a risk index model.
Step S5 is specifically as follows:
and S5-1, calculating the fact coefficient of the risk index in the risk index model according to each item of public opinion data updated in real time. The index evaluation data sources are classified into 3 types: 1) data provided by other related systems; 2) risk data actively reported by related staff of the workflow; 3) and the public opinion analysis module provides data.
In the final public opinion risk index model constructed by S5-2 and S4, C5 and C7 are actively input by a decision maker, and input data are elements of a phrase set LT.
In the final public opinion risk indicator model constructed in S5-3 and S4, the C3 historical legacy work is determined by the attribute field, True is mapped to 'quite low' and False is mapped to 'quite high'. The C4 is determined by the key job type by its attribute field. The C9 well-known decision maker is provided by a decision maker knowledge graph with True mapped to "fairly low" and False mapped to "fairly high".
In the final public opinion risk index model constructed by S5-4 and S4, C1, C2, C6 and C8 are calculated according to public opinion data collected in real time. Wherein, C2 is the ratio of the white lists of the accounts with high influence on public opinion coverage, and C6 and C8 are the values of the occurrence status of each hot search list. C1 is the negative sentiment value based on the focused topic, and the specific processing flow is shown in fig. 6.
(1) After text preprocessing and vectorization are completed, topic clustering is carried out based on a K-Means algorithm, a focus topic text is extracted according to word frequency information in the text, and an emotion value of the topic text is calculated by using a Bayes classifier on the basis.
(2) And (4) extracting topic texts from the public opinion texts to obtain residual texts, and obtaining emotion values of all the residual texts related to the topics by using a Bayes classifier.
(3) Fusing negative face values of public sentiments, which are defined as:
wherein n isiFor the number of negative emotion text in a topic, otherwise piIs the number of positive emotion text in the topic. τ is the threshold of negative emotion text triggering public sentiment risk, and M is the number of focal topics relevant to the job. And map the emotion values by 1 to 9 to elements in phrase set LT. Compared with the traditional method, the sentiment value analysis method based on the focal topic is more reasonable in judgment of public opinion tendency.
Step S6 is specifically as follows:
s6-1, the probability of the event corresponding to the risk index, which is the numerical value of the fact coefficient corresponding to the risk index obtained by the evaluation in S5, is converted into the fuzzy number according to the graph of FIG. 2. Order GFNSm＝(ltm,GFNm), ltmIs a language expression phrase, GFNmIs the corresponding generalized ambiguity number. Value v of the factual coefficient of riskiThe converted blur number is expressed as:
s6-2, fusing the multiple risk factors according to the influence degree impact and the fact coefficient evaluate to obtain the public opinion risk level appearing in the current work flow, wherein the fusing method comprises the following steps:
wherein N is the number of risk factors, and finally, the fuzzy number obtained by public opinion risk comprehensive evaluation represents risk.
Step S7 is specifically as follows:
based on the graphic characteristics of the fuzzy number, mapping the public opinion risk comprehensive evaluation risk obtained in the step S6 to a risk level and fuzzy number conversion system by calculating the fuzzy number similarity based on the gyration radius, so as to obtain the public opinion risk level of the current workflow:
the multi-factor public opinion risk assessment method based on the fuzzy number similarity provided by the embodiment of the invention can be used for fusing multi-dimensional information such as risk factors in a workflow from multiple sources, and comprehensively providing fine-grained assessment for the current public opinion risk; the method has the advantages that different statistical strategies are carried out on the negative topics and the positive topics, the method is more reasonable than the traditional method, the public opinion tendency can be better reflected, the index model constructed by the method is beneficial to refining of risk factors in the construction stage, the model is beneficial to expanding in the maintenance stage, and meanwhile, the risk indexes have more objective influence degree evaluation.
Those skilled in the art will appreciate that all or part of the steps in the methods of the above embodiments may be implemented by associated hardware instructed by a program, which may be stored in a computer-readable storage medium, and the storage medium may include: read Only Memory (ROM), Random Access Memory (RAM), magnetic or optical disks, and the like.
In addition, the multi-factor public opinion risk assessment method based on fuzzy number similarity provided by the embodiment of the invention is described in detail, a specific example is applied in the text to explain the principle and the implementation mode of the invention, and the description of the embodiment is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
Claims (3)
1. A multi-factor public opinion risk assessment method based on fuzzy number similarity is characterized by comprising the following steps:
extracting public opinion risk factors based on historical working documents and a workflow rule base, constructing a risk factor set, constructing an expert group, expressing the influence degree of the risk factor set by using a phrase set by an expert, and collecting the data to form a risk factor influence degree data set;
the method comprises the steps that a three-layer structure of an initial public opinion risk index model is built by using an analytic hierarchy process, and each level node is defined, wherein the model consists of a target layer, a criterion layer and an index layer, the target layer determines that the main body of evaluation is public opinion risk, the criterion layer defines four standard criteria of a decision maker, a party, case attribute and public opinion state based on fishbone image analysis of historical working documents, and the index layer consists of risk factors and corresponding influence degrees;
constructing an evaluation matrix based on the risk factor influence degree data set, and calculating the influence degree of each risk factor by using an extended good and bad solution distance method and integrating a multi-dimensional evaluation matrix;
determining the arrangement order of each layer element in the initial public opinion risk index model by using an analytic hierarchy process based on the influence degree of each risk factor to obtain a final public opinion risk index model with a complete system structure;
collecting and analyzing relevant public opinion data of the current processed work, and evaluating risk factors in the final public opinion risk index model to obtain a fact coefficient of each risk index in a model index layer;
fusing the risk factor influence degree in the final public opinion risk index model index layer with the fact coefficient to obtain a public opinion risk comprehensive value based on multi-factor evaluation;
based on the graphic characteristics of the fuzzy number, mapping the public opinion risk comprehensive value based on the multi-factor evaluation to a risk grade and fuzzy number conversion system by calculating the fuzzy number similarity based on the gyration radius, thereby obtaining the public opinion risk grade of the current processing work;
and if the public opinion risk level or the variation trend thereof meets the triggering early warning condition, executing corresponding early warning operation according to a risk plan.
2. The method for multi-factor public opinion risk assessment based on fuzzy number similarity as claimed in claim 1, wherein the phrase set specifically comprises:
the phrase set LT { "considerably low" (absolute low), "very low" (very low), "low" (far low), "medium)," high "(high)," very high "(very high)," considerably high "(absolute high).
3. The method for multi-factor public opinion risk assessment based on fuzzy number similarity as claimed in claim 1, wherein the evaluation matrix is constructed based on the risk factor influence degree data set, and the influence degree of each risk factor is calculated by using an extended good and bad solution distance method to synthesize a multi-dimensional evaluation matrix, specifically:
constructing an evaluation matrix based on the risk factor influence degree data set;
constructing positive and negative ideal solutions for risk factor evaluation;
calculating the Euclidean distance between each evaluation scheme in the evaluation matrix and the positive ideal solution and the Euclidean distance between each evaluation scheme in the evaluation matrix and the negative ideal solution;
calculating the distance ratio of each evaluation scheme in the evaluation matrix to the positive and negative ideal solutions, and normalizing the distance ratio to obtain the weight of each evaluation scheme;
and carrying out weighted summation on the weight of each evaluation scheme and each evaluation scheme to obtain the evaluation of the influence degree of each risk factor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011279436.0A CN112734154B (en) | 2020-11-16 | 2020-11-16 | Multi-factor public opinion risk assessment method based on fuzzy number similarity |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011279436.0A CN112734154B (en) | 2020-11-16 | 2020-11-16 | Multi-factor public opinion risk assessment method based on fuzzy number similarity |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112734154A true CN112734154A (en) | 2021-04-30 |
| CN112734154B CN112734154B (en) | 2023-08-01 |
Family
ID=75597516
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011279436.0A Active CN112734154B (en) | 2020-11-16 | 2020-11-16 | Multi-factor public opinion risk assessment method based on fuzzy number similarity |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112734154B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113360843A (en) * | 2021-08-11 | 2021-09-07 | 上海建工集团股份有限公司 | Multi-factor coupling-based security risk management and control method |
| CN113570182A (en) * | 2021-05-06 | 2021-10-29 | 深圳怀新企业投资顾问股份有限公司 | Reputation risk management capability assessment method, reputation risk management capability assessment device, reputation risk management equipment and storage medium |
| CN113779195A (en) * | 2021-08-31 | 2021-12-10 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Hot event state evaluation method |
| CN114565408A (en) * | 2022-03-01 | 2022-05-31 | 北京派瑞威行互联技术有限公司 | Bidding prediction method and system for advertisement putting |
| CN115221416A (en) * | 2022-08-18 | 2022-10-21 | 上海金仕达软件科技有限公司 | Reputation risk positioning analysis method and system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110738399A (en) * | 2019-12-16 | 2020-01-31 | 中山大学 | judicial judgment flow deviation early warning method based on fuzzy set theory |
| CN111914087A (en) * | 2020-07-30 | 2020-11-10 | 广州城市信息研究所有限公司 | Public opinion analysis method |
-
2020
- 2020-11-16 CN CN202011279436.0A patent/CN112734154B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110738399A (en) * | 2019-12-16 | 2020-01-31 | 中山大学 | judicial judgment flow deviation early warning method based on fuzzy set theory |
| CN111914087A (en) * | 2020-07-30 | 2020-11-10 | 广州城市信息研究所有限公司 | Public opinion analysis method |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113570182A (en) * | 2021-05-06 | 2021-10-29 | 深圳怀新企业投资顾问股份有限公司 | Reputation risk management capability assessment method, reputation risk management capability assessment device, reputation risk management equipment and storage medium |
| CN113360843A (en) * | 2021-08-11 | 2021-09-07 | 上海建工集团股份有限公司 | Multi-factor coupling-based security risk management and control method |
| CN113360843B (en) * | 2021-08-11 | 2021-11-05 | 上海建工集团股份有限公司 | Multi-factor coupling-based security risk management and control method |
| CN113779195A (en) * | 2021-08-31 | 2021-12-10 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Hot event state evaluation method |
| CN113779195B (en) * | 2021-08-31 | 2023-12-22 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Hot event state evaluation method |
| CN114565408A (en) * | 2022-03-01 | 2022-05-31 | 北京派瑞威行互联技术有限公司 | Bidding prediction method and system for advertisement putting |
| CN115221416A (en) * | 2022-08-18 | 2022-10-21 | 上海金仕达软件科技有限公司 | Reputation risk positioning analysis method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112734154B (en) | 2023-08-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112734154B (en) | Multi-factor public opinion risk assessment method based on fuzzy number similarity | |
| CN111428054B (en) | Construction and storage method of knowledge graph in network space security field | |
| US6321217B1 (en) | Data analyzing method for generating rules | |
| Pratt et al. | Employee attrition estimation using random forest algorithm | |
| CN110046250A (en) | Three embedded convolutional neural networks model and its more classification methods of text | |
| Mezei et al. | Credit risk evaluation in peer-to-peer lending with linguistic data transformation and supervised learning | |
| CN116127084A (en) | Knowledge graph-based micro-grid scheduling strategy intelligent retrieval system and method | |
| CN116090757A (en) | Method for evaluating capability demand satisfaction of information guarantee system | |
| Baherifard et al. | Improving the effect of electric vehicle charging on imbalance index in the unbalanced distribution network using demand response considering data mining techniques | |
| Zhang et al. | Research on borrower's credit classification of P2P network loan based on LightGBM algorithm | |
| CN114528405A (en) | Public opinion monitoring method based on network burst hotspot | |
| CN113516189A (en) | Website malicious user prediction method based on two-stage random forest algorithm | |
| CN113129188A (en) | Provincial education teaching evaluation system based on artificial intelligence big data | |
| Lee et al. | Wind power pattern forecasting based on projected clustering and classification methods | |
| CN115293827A (en) | Novel model interpretability analysis method for assisting fine operation of enterprise | |
| CN114741515A (en) | Social network user attribute prediction method and system based on graph generation | |
| CN113987536A (en) | Method and device for determining security level of field in data table, electronic equipment and medium | |
| You et al. | Credit Grade Prediction Based on Decision Tree Model | |
| Liu et al. | A trend-based stock index forecasting model with gated recurrent neural network | |
| Zheng | Individualized Recommendation Method of Multimedia Network Teaching Resources Based on Classification Algorithm in a Smart University | |
| CN114282875A (en) | Flow approval certainty rule and semantic self-learning combined judgment method and device | |
| Li et al. | Research on listed companies’ credit ratings, considering classification performance and interpretability | |
| Salih et al. | A Type-2 Fuzzy Logic Based System for Decision Support to Minimize Financial Default in the Sudanese Banking Sector | |
| KR102663632B1 (en) | Device and method for artwork trend data prediction using artificial intelligence | |
| CN113407727B (en) | Qualitative measure and era recommendation method based on legal knowledge graph and related equipment |
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 |