CN111461576A - Fuzzy comprehensive evaluation method for safety risk of chemical hazards in food - Google Patents
Fuzzy comprehensive evaluation method for safety risk of chemical hazards in food Download PDFInfo
- Publication number
- CN111461576A CN111461576A CN202010343593.7A CN202010343593A CN111461576A CN 111461576 A CN111461576 A CN 111461576A CN 202010343593 A CN202010343593 A CN 202010343593A CN 111461576 A CN111461576 A CN 111461576A
- Authority
- CN
- China
- Prior art keywords
- risk
- evaluation
- food
- level
- value
- 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
- 238000011156 evaluation Methods 0.000 title claims abstract description 243
- 235000013305 food Nutrition 0.000 title claims abstract description 121
- 239000000126 substance Substances 0.000 title claims abstract description 61
- 238000001514 detection method Methods 0.000 claims abstract description 62
- 238000004364 calculation method Methods 0.000 claims abstract description 56
- 238000000034 method Methods 0.000 claims abstract description 49
- 238000010606 normalization Methods 0.000 claims abstract description 32
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 23
- 239000011159 matrix material Substances 0.000 claims description 21
- 230000006870 function Effects 0.000 claims description 20
- 238000004458 analytical method Methods 0.000 claims description 19
- 230000015572 biosynthetic process Effects 0.000 claims description 19
- 238000007689 inspection Methods 0.000 claims description 19
- 238000003786 synthesis reaction Methods 0.000 claims description 19
- 239000013598 vector Substances 0.000 claims description 18
- 238000012545 processing Methods 0.000 claims description 16
- 238000006243 chemical reaction Methods 0.000 claims description 8
- 238000009826 distribution Methods 0.000 claims description 7
- 230000000694 effects Effects 0.000 claims description 6
- 238000010801 machine learning Methods 0.000 claims description 6
- 238000012502 risk assessment Methods 0.000 claims description 5
- 239000002131 composite material Substances 0.000 claims description 4
- 238000000556 factor analysis Methods 0.000 claims description 4
- 230000005484 gravity Effects 0.000 claims description 4
- 238000012935 Averaging Methods 0.000 claims description 3
- 238000004891 communication Methods 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 238000012800 visualization Methods 0.000 claims description 3
- 238000012549 training Methods 0.000 claims description 2
- 238000012795 verification Methods 0.000 claims 1
- 239000000447 pesticide residue Substances 0.000 description 14
- 235000013311 vegetables Nutrition 0.000 description 10
- SNICXCGAKADSCV-UHFFFAOYSA-N nicotine Natural products CN1CCCC1C1=CC=CN=C1 SNICXCGAKADSCV-UHFFFAOYSA-N 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 238000011160 research Methods 0.000 description 5
- 238000005070 sampling Methods 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 4
- 235000013399 edible fruits Nutrition 0.000 description 4
- 238000007781 pre-processing Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- SNICXCGAKADSCV-JTQLQIEISA-N (-)-Nicotine Chemical compound CN1CCC[C@H]1C1=CC=CN=C1 SNICXCGAKADSCV-JTQLQIEISA-N 0.000 description 3
- 241000427649 Melongena Species 0.000 description 3
- 230000019771 cognition Effects 0.000 description 3
- 230000007547 defect Effects 0.000 description 3
- 238000013178 mathematical model Methods 0.000 description 3
- 229960002715 nicotine Drugs 0.000 description 3
- 238000012797 qualification Methods 0.000 description 3
- 230000004044 response Effects 0.000 description 3
- 238000007619 statistical method Methods 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- PVNIIMVLHYAWGP-UHFFFAOYSA-N Niacin Chemical compound OC(=O)C1=CC=CN=C1 PVNIIMVLHYAWGP-UHFFFAOYSA-N 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000013210 evaluation model Methods 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 210000001015 abdomen Anatomy 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 230000001149 cognitive effect Effects 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000013278 delphi method Methods 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 239000003344 environmental pollutant Substances 0.000 description 1
- 238000012854 evaluation process Methods 0.000 description 1
- 235000012631 food intake Nutrition 0.000 description 1
- 235000021393 food security Nutrition 0.000 description 1
- 239000000383 hazardous chemical Substances 0.000 description 1
- 239000013056 hazardous product Substances 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 231100000206 health hazard Toxicity 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 229960003512 nicotinic acid Drugs 0.000 description 1
- 235000001968 nicotinic acid Nutrition 0.000 description 1
- 239000011664 nicotinic acid Substances 0.000 description 1
- 235000016709 nutrition Nutrition 0.000 description 1
- 230000035764 nutrition Effects 0.000 description 1
- 231100000719 pollutant Toxicity 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000007711 solidification Methods 0.000 description 1
- 230000008023 solidification Effects 0.000 description 1
- 230000002194 synthesizing 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
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- 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
- G06Q50/26—Government or public services
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Educational Administration (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Primary Health Care (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a first purpose of providing a fuzzy comprehensive evaluation method for safety risks of chemical hazards in food, which comprises the following steps of S1, establishing an evaluation factor set, S2, establishing a fuzzy comprehensive evaluation comment set, using a national food safety limit standard MR L as an evaluation standard, and adopting a 'none/micro, low, medium, high and ultrahigh' 5-grade scaling method to evaluate the safety risks of the food, S3, using the evaluation value of each chemical hazard of S1 as a calculation basis, and adopting an entropy algorithm to determine multi-level risk factor weights, wherein a fixed coefficient method is adopted to revise an S1 original detection data normalization method, and S4, adopting the fuzzy comprehensive evaluation method to calculate comprehensive safety risk comprehensive evaluation indexes FSI of chemical hazards at all levels in the food.
Description
Technical Field
The invention relates to a fuzzy comprehensive evaluation method for safety risk of chemical hazards in food, and belongs to the field of food safety.
Background
Since ancient times, "people eat as a day". With the continuous development of modern society and economy and the continuous improvement of the living science and technology level, the demand of people on food is no longer only satisfied with 'fruit belly', the nutrition, delicacy, health and diversity become the food pursuit of the current society masses, and in the food, the safety is the most fundamental requirement on the food quality, and the food quality has 'one-vote-negative power' for measuring the food quality. Food is taken safely as the first, food quality supervision and spot check and risk assessment are scientifically developed under the current market economic operation and management mode, risk exchange is actively developed, the current food safety situation is actively, timely and comprehensively mastered, and the method is one of basic guarantee means for continuously improving the food safety level of the whole society.
The food safety relates to a plurality of factors of each link of food production, circulation, consumption and the like, the food safety is essentially food quality, and the most direct objective reflection is food quality detection data. With the long-term wide development of the work of food safety supervision spot check and the like, massive food quality safety detection data are gradually accumulated, and the large data have extremely important research and application values for researching food safety characteristics and harm factor distribution growth and degradation rules. But the food quality detection data is huge in quantity and complex in composition, and extremely high requirements are put forward on data mining and analysis technologies. The traditional descriptive statistical method is low in working efficiency, weak in comprehensive analysis capability of data and limited in usability and practicability under the background of big data, and the actual demand promotes the deep exploration of the comprehensive evaluation method based on various mathematical models.
Currently, in the fields of comprehensive evaluation research and application of food safety risks, evaluation methods such as a delphire method, an analytic hierarchy process, machine learning and the like are mainly adopted. The Delphi method is one of the main modes of expert survey methods, and solves the key problems in the comprehensive evaluation process, namely risk factor weight setting and risk classification index system determination, by adopting multiple rounds of anonymous feedback function inquiry and collecting and analyzing professional opinions of experts in related fields. The analytic hierarchy process is a hierarchical weight decision analysis process, which decomposes complex decision problems of multiple targets, multiple criteria or no structural features step by step, constructs a multilevel analysis model with reasonable hierarchical relation, determines the most basic scheme, measure and other relatively important weights relative to decision through methods such as expert evaluation, and finally realizes the decision target. However, in the specific implementation process, the two methods depend on expert opinions, the participation degree of subjective factors is high, objective and general standard rules are difficult to form, meanwhile, the requirement of the hierarchical analysis method on evaluation elements is high, the judgment matrix needs to pass consistency inspection, and the practical effect is influenced. The machine learning method has extremely high-efficiency deep analysis capability on complex and diverse data, and is one of the most popular application tools in the big data era, wherein an artificial neural network is commonly used for risk assessment and prediction. The artificial neural network is an algorithm with nonlinear adaptive information processing capability, and the implicit rule in the system is acquired through large sample training, so that the state prediction based on the existing quantitative and qualitative knowledge is realized, but the 'black box' model is the 'black box' model, so that the reasonability and the accuracy of a data prediction model cannot be objectively distinguished, and the reliability and the usability of a prediction result are finally influenced.
In addition, because of a plurality of food safety influence factors, in the past research reports, in order to realize the comprehensiveness of risk evaluation, evaluation indexes are often complicated in structure and are various, such as various indexes including food quality level, food consumption level, risk factor exposure level, food safety public opinion and the like, and food quality inspection is often only one of the indexes. Under the evaluation index structure, the entropy weight method is mostly based on indirect indexes such as qualification rate, standard exceeding degree and the like, so as to obtain the weight of a single index related to the food quality. Due to the lack of comparability among different types of evaluation indexes, subjective weighting methods such as AHP are still mostly adopted for the weights of other evaluation indexes in the comprehensive evaluation indexes.
Disclosure of Invention
In order to overcome the defects of subjectivity, one-sidedness, low efficiency and the like of the conventional comprehensive evaluation method, the invention establishes an index system with clear structure and quantification and comparability according to the high representativeness of food quality supervision big data, the scientific connotation of national limit standards of food safety and the practical application requirements of food safety risk evaluation, takes food quality inspection detection data as a direct research object, revises weights for all research indexes by observing the dispersion degree of information entropy and applying an entropy value calculation method, and meanwhile, based on the principle that entropy intensity can be added, in a multi-stage evaluation index system, revises the weights for all layers of evaluation indexes layer by adding and comprehensively comparing entropy values. Therefore, in a multi-level evaluation index system, the whole evaluation system is regarded as a whole by the entropy weight algorithm, the calculation method is objective and uniform, and subjective factors are effectively avoided.
The invention embodies real 'mirror image' of the overall parameters according to the statistical characteristics of large sample data, applies an improved entropy weight calculation method to determine the weight of multi-level risk factors, establishes a universal 5-level comment system which accords with food safety standards based on objective cognition of national limit standards of food safety and standard detection method parameters, then comprehensively evaluates the risk level of chemical hazards in food from the view point of risk probability theory essence by adopting a fuzzy mathematical model, realizes the visual output and application of analysis data by a large data sharing platform, and comprehensively guarantees the scientificity, objectivity, high efficiency and practicability of the evaluation method system.
The invention aims to provide a fuzzy comprehensive evaluation method for the safety risk of chemical hazards in food, which comprises the following steps:
s1: detecting the content of each chemical hazard in the food, determining the evaluation value of each chemical hazard, and constructing an evaluation factor set;
s2, establishing a comment set of fuzzy comprehensive evaluation, namely evaluating the food safety risk by using a 'none (micro), low, medium, high and ultrahigh' 5-grade scaling method by taking a national food safety limit standard (Maximum Residue L imits, MR L) as an evaluation standard;
s3: determining the weight of the multilevel risk factors by using an entropy algorithm based on the evaluation value of each chemical hazard of S1; wherein, the S1 original detection data normalization method is revised by a fixed coefficient method;
s4: and calculating the comprehensive evaluation index (FSI) of the safety risk of chemical hazards at all levels in the Food by adopting a fuzzy comprehensive evaluation method.
In one embodiment of the present invention, the evaluation value in S1 is obtained by performing dimensionless processing on the original detection result with reference to the highest limit value MR L in the corresponding national food safety standard:
xij＝Tij/MRL(1)
in the formula, xijThe ith raw test result (T) of the jth risk factorij) And (4) evaluation value after dimensionless processing.
In one embodiment of the present invention, the set of evaluation factors is constructed in S1 as follows: in a first-order fuzzy evaluation system, m types of chemical risk factors influencing food safety form a first-order evaluation factor set (U) which is expressed as: u ═ U1，U2，U3，…，UmAnd (f) forming a first-level evaluation index matrix X (X) by each risk factor and n detection results contained in the risk factorsij)n×mWherein i ═ 1, 2, 3, …, n; j is 1, 2, 3, …, m.
In one embodiment of the present invention, the fixed coefficient revised in S3 is 1/1000; adopting formula (2) to evaluate the ith evaluation value x of the jth risk factorijAnd (3) carrying out normalization treatment:
in the formula: vijRepresents the calculation result of the ith evaluation value of the jth risk factor after normalization processing, xj(max)And xj(min)Respectively representing the maximum value and the minimum value of the n evaluation values of the j-th risk factor,
is a revision coefficient to the normalization method.
In one embodiment of the present invention, S3 calculates the entropy weight ω of each chemical hazard using the following steps:
adopting formula (2) to evaluate the ith evaluation value x of the jth risk factorijAnd (3) carrying out normalization treatment:
in the formula: vijRepresents the calculation result of the ith evaluation value of the jth risk factor after normalization processing, xj(max)And xj(min)Respectively representing the maximum value and the minimum value of the n evaluation values of the j-th risk factor,
is a coefficient revision to the conventional normalization method;
the characteristic specific gravity (P) of each evaluation value in the index is calculated by using the formula (3)ij)：
Then, the j-th type risk factor entropy value (e)j) Can be calculated by equation (4):
among the m classes of risk factors, the entropy weight ω of the jth class of risk factorsjCalculated by equation (5) and satisfies
In one embodiment of the present invention, if there are multiple levels of evaluation indexes above the one-level index in S2For an evaluation system containing k types of secondary evaluation indexes, applying entropy-adding principle and algorithm, and obtaining difference coefficient (D) of l (l ═ 1, 2, 3, …, k) type secondary indexesl) I.e. the difference coefficient (d) of m primary indexesj) The sum (formula 6, 7):
wherein: dj＝1-ej(7)
Then, in the second-level index, the entropy weight (ω) of the first-term indexl) Calculated by equation (8) and satisfies
And (4) carrying out iterative calculation on more than two levels of evaluation indexes by using the method.
In one embodiment of the invention, S4 adopts a fuzzy comprehensive evaluation method to calculate the comprehensive evaluation index FSI of the chemical hazard safety risk in food, and comprises the following steps:
(a) constructing a membership function: chemical hazards belong to negative effect indexes, and a membership function adopts a half-reduced gradient distribution model; in the first-level evaluation index system, the ith evaluation value (x) of the jth risk factorij) Degree of membership (h) thereofijq) The calculation method is as follows:
when q is equal to 1, the reaction is carried out,
when q is 2, 3, 4,
when q is equal to 5, the reaction is carried out,
in the formula, LqA representative value of a general type set for the 5-scale evaluation scale, q being 1, 2, 3, 4, 5, i.e., { L }1，L2，L3，L4，L5The evaluation results are (1), (0), (0.25), (0.50), (0.75), (1.00), (2.00), (V) and (V) are respectively corresponding to the risk levels in a comment set of fuzzy comprehensive evaluation, wherein (V) is no (micro), low, medium, high and ultrahigh risk;
(b) calculating the comprehensive risk index of a first-level evaluation system: based on single factor analysis, the evaluation values of the jth risk factor are subjected to membership function calculation to obtain the fuzzy judgment subset (r) of the risk factorj)：
To rjRespectively averaging the membership degrees of all levels in the matrix to obtain a composite membership degree vector R of the jth risk factorjq，
Risk membership degree vector R based on m-class primary indexesjqConstructing a fuzzy matrix R ═ (R)jq)m×5Wherein j is 1, 2, 3, …, m, q is 1, 2, 3, 4, 5; and then carrying out weighted W synthesis calculation on the fuzzy matrix R to obtain a first-level evaluation index comprehensive evaluation vector B:
in the formula: omegamFor each of the risk factor entropy weights is,
is a fuzzy synthesis operator. In order to highlight the entropy weight and take the influence of each risk factor into consideration, and improve the comprehensive evaluation capability of the model, the comprehensive evaluation method selects
(product, sum operator) synthesizing the fuzzy matrix;
by adopting a fuzzy comprehensive index method (formula 12), a first-level risk factor comprehensive evaluation index (FSI) can be calculated, and the corresponding risk grade is judged according to the maximum membership principle:
in the formula:
for fuzzy synthesis of operators
sqFor each rank of the panel, q is 1, 2, 3, 4, 5, i.e. { s }1，s2，s3，s4， s 50, 1, 2, 3, 4. The FSI correspondence comment set V ═ i, { i, ii, iii, iv, V } ═ zero (micro) risk, low risk, medium risk, high risk, ultra high risk }, and its interval representative value is {0, 1, 2, 3, 4 }. And directly judging the risk grade of the FSI according to the maximum membership rule and the boundary value interval of each risk grade. In one embodiment of the present invention, the method specifically includes the steps of:
1) acquiring, standardizing and arranging food quality inspection and detection big data, establishing a risk factor evaluation factor set (U), and applying an improved entropy weight method to give weight to evaluation indexes; the evaluation factor set can be a first-level risk factor or a multi-level risk evaluation index comprising different levels and dimensions.
A1. The method comprises the following steps of raw data preprocessing, namely acquiring and preprocessing raw detection data of chemical hazards in food, and mainly comprises the steps of cleaning and regularizing data, assigning non-numerical detection results such as character types and the like based on rules, and carrying out dimensionless processing on the raw detection results (formula 1) by taking the maximum limit value MR L in the corresponding national food safety standard as a reference according to the chemical hazard types:
xij＝Tij/MRL(1)
in the formula, xijThe ith raw test result (T) of the jth risk factorij) And (4) evaluation value after dimensionless processing.
A2. Constructing an evaluation factor set: in a first-level fuzzy evaluation (or single-factor fuzzy evaluation) system, m types of chemical risk factors influencing food safety form a first-level evaluation factor set (U) which is expressed as: u ═ U1，U2，U3，…，UmAnd (f) forming a first-level evaluation index matrix X (X) by each risk factor and n detection results contained in the risk factorsij)n×mWherein i ═ 1, 2, 3, …, n; j is 1, 2, 3, …, m.
A3. Calculating the entropy weight of the first-level risk factor: for the ith evaluation value (x) of the jth risk factorij) Normalization treatment (formula 2):
in the formula: vijRepresents the calculation result of the ith evaluation value of the jth risk factor after normalization processing, xj(max)And xj(min)Respectively representing the maximum value and the minimum value of the n evaluation values of the j-th risk factor.
Is a coefficient revision to the conventional normalization method. The revision significance lies in that: the method avoids the problem that the subsequent operation steps cannot be automatically carried out under the condition of big data when the calculation result of the step is 0 or meaningless (namely the denominator is 0). By revising the coefficient, the original calculation result is not obviously influenced, and the automatic data calculation and analysis capability under the condition of big data can be improved.
The characteristic specific gravity (P) of each evaluation value in the index is calculated by using the formula (3)ij)：
Then, the j-th type risk factor entropy value (e)j) Can be calculated by equation (4):
among the m-class risk factors, the entropy weight of the jth class risk factor is calculated by equation (5) and satisfies
A4. And (3) carrying out multilevel risk factor entropy weight calculation: if a plurality of levels of evaluation indexes exist above the first level index, applying entropy intensification principle and algorithm to an evaluation system containing k types of second level evaluation indexes, wherein the first type (l is 1, 2, 3, …, k) of second level indexes has difference coefficient (D)l) I.e. the difference coefficient (d) of m primary indexesj) The sum (formula 6, 7):
wherein: dj＝1-ej(7)
Then, in the second-level index, the entropy weight (ω) of the first-term indexl) Calculated by equation (8) and satisfies
And (4) carrying out iterative calculation on more than two levels of evaluation indexes by using the method.
2) And calculating the comprehensive evaluation vector of each risk factor by adopting a fuzzy comprehensive evaluation model, and then calculating the comprehensive evaluation index FSI of the chemical hazard safety risk in the food based on the entropy weight and the synthesis operator.
B1. Establishing a comment set of fuzzy comprehensive evaluation, namely evaluating the food safety risk by adopting a 'no (micro), low, medium, high and ultrahigh' 5-level scale method, setting a comment set V as { I, II, III, IV, V }, as { no (micro) risk, low risk, medium risk, high risk and ultrahigh risk }, setting a universal representative value L for each evaluation level based on the highest residual limit value (MR L) of hazard factors in national food safety standards and technical parameters of a standard detection methodq(q-1, 2, 3, 4, 5), i.e. { L1，L2，L3，L4，L5}＝{0.25，0.50，0.75，1.00，2.00}；
B2. Constructing a membership function: chemical hazards belong to negative effect indexes, and a half-reduced gradient distribution model is adopted as a membership function. In the first-level evaluation index system, the ith evaluation value (x) of the jth risk factorij) Degree of membership (h) thereofijq) The calculation method is (formula 9-11):
when q is equal to 1, the reaction is carried out,
when q is 2, 3, 4,
when q is equal to 5, the reaction is carried out,
B3. calculating the comprehensive risk index of a first-level evaluation system: based on single factor analysis, the evaluation values of the jth risk factor are subjected to membership function calculation to obtain the fuzzy judgment subset (r) of the risk factorj)：
To rjRespectively averaging the membership degrees of all levels in the matrix to obtain the j-th riskFactor complex membership vector Rjq，
Risk membership degree vector R based on m-class primary indexesjqConstructing a fuzzy matrix R ═ (R)jq)m×5Wherein j is 1, 2, 3, …, m, q is 1, 2, 3, 4, 5; and then carrying out weighted W synthesis calculation on the fuzzy matrix R to obtain a first-level evaluation index comprehensive evaluation vector B:
in the formula: omegamFor each of the risk factor entropy weights is,
a fuzzy synthesis operator; in order to highlight the entropy weight and take the influence of each risk factor into consideration, and improve the comprehensive evaluation capability of the model, the comprehensive evaluation method selects
The (product, sum operator) performs a synthesis operation on the fuzzy matrix.
By adopting a fuzzy comprehensive index algorithm (formula 12), a first-level risk factor comprehensive evaluation index (FSI) can be calculated, and the corresponding risk grade is judged according to the maximum membership principle:
in the formula:
for fuzzy synthesis of operators
sqRank (q is 1, 2, 3, 4, 5) of each rank of the panel of comments, i.e., { s {(s) }1，s2，s3，s4， s 50, 1, 2, 3, 4. FSI corresponds to the comment set V ═ { i, ii, iii, iv, V } ═ no (micro) riskLow risk, medium risk, high risk, ultra high risk, with interval representative value of {0, 1, 2, 3, 4 }. And directly judging the risk grade of the FSI according to the maximum membership rule and the boundary value interval of each risk grade.
3) In a multi-level risk factor evaluation system, a lower-level evaluation factor comprehensive evaluation vector forms a fuzzy judgment subset of a previous-level evaluation index, and then a matrix synthesis algorithm can be iteratively applied to calculate the comprehensive evaluation index of each level of risk factors step by step.
A second object of the present invention is to provide a food safety risk assessment device, comprising:
the detection module is used for detecting the content of each chemical hazard in the food, determining the evaluation value of each chemical hazard and constructing an evaluation factor set;
a calculation module: the entropy weight calculation module is used for calculating the entropy weight of each chemical hazard, calculating the entropy weight of each chemical hazard according to the detection data and evaluation value of each chemical hazard content obtained by the detection module, and determining the weight of the multilevel risk factor;
the analysis module is used for calculating the comprehensive evaluation index FSI of the chemical hazard safety risk in the food, calculating the entropy weight of each chemical hazard according to the calculation module, then calculating the comprehensive evaluation index FSI of the chemical hazard safety risk at each level in the food by adopting a fuzzy comprehensive evaluation method, and determining the risk level, wherein the risk level is determined by adopting a 'no (micro), low, medium, high and ultrahigh' 5-level scaling method by taking the national limit standard MR L of the food safety as an evaluation standard.
In one embodiment of the present invention, the computing unit includes a revision module, configured to revise the original detection data normalization method, wherein the revision is a fixed coefficient method, and the fixed coefficient is 1/1000; adopting formula (2) to evaluate the ith evaluation value x of the jth risk factorijAnd (3) carrying out normalization treatment:
in the formula: vijRepresenting class j risk factorsThe calculated result x of the sub ith evaluation value after normalization processingj(max)And xj(min)Respectively representing the maximum value and the minimum value of the n evaluation values of the j-th risk factor,
is a coefficient revision to the conventional normalization method.
In one embodiment of the invention, the device further comprises a big data sharing platform, wherein the big data sharing platform is connected with the analysis module, the obtained analysis module information is integrated into the Internet communication-based food safety inspection detection big data sharing platform, and the multidimensional comprehensive risk evaluation result is actively displayed in a circulating manner through conventional statistical charts, hot spot maps, multi-graph linkage and other visualization manners; in the related area of the webpage, an operator can select the concerned target item by adopting a mouse or touch screen clicking mode, so that food safety sampling inspection and evaluation result conditional output are realized.
In one embodiment of the invention, the big data sharing platform can also realize a risk prediction and early warning function, and can adopt machine learning methods such as a BP neural network and the like to train and verify by taking historical detection data based on a time sequence as a sample, and preferentially establish a prediction model; and (3) loading the detection data of the current time node as a prediction set into a network model, wherein the obtained output value is a risk trend prediction value, and in practical application, a response rule is set for the food safety risk level, so that a corresponding scheme can be provided according to the prediction value.
The third purpose of the invention is to provide the application of the method in food quality safety evaluation, the formed food safety index based on food inspection and detection big data has comprehensiveness and comparability, and multi-level and multi-dimensional risk evaluation and comparison can be directly carried out; meanwhile, the fuzzy grade evaluation result of the food safety risk accords with the subjective cognitive mode of the public on the risk, has wider acceptability and operability than evaluation indexes such as single qualification rate statistics and the like, and is suitable for practical application of different subjects such as food safety supervision departments, industrial enterprises, social public, media and the like.
The invention has the beneficial effects that:
the invention provides a food chemical hazard safety risk fuzzy comprehensive evaluation method based on entropy weight, which is based on food quality safety inspection and detection information data from the view point of overall safety of food, fully utilizes authenticity and representativeness of big data, firstly determines objective weight of multi-level risk evaluation factors by applying an improved entropy weight calculation method, and establishes a general risk evaluation grade according with food safety based on objective cognition of national limit standard and standard detection method parameters of food safety; and then, starting from objective understanding of the nature of the risk, a fuzzy mathematical model is adopted to comprehensively evaluate the risk level of the chemical hazards in the food, and visual display and interactive information service of the evaluation result are realized by virtue of a food safety big data sharing platform based on the Internet. The method is based on high-quality food safety inspection detection data, has sufficient scientific basis, is low in subjective factor participation degree of an evaluation system, is objective and efficient in analysis module, is high in adaptability, is suitable for different information application main bodies to flexibly construct an evaluation index system according to self requirements, quickly completes data acquisition, analysis and output, is high in evaluation result comprehensiveness and high in matching degree with subjective risk cognition, overcomes the defects of single information, lack of integrity and comprehensiveness of the traditional statistical method, overcomes the defect of structural mode solidification of a common comprehensive evaluation method depending on an expert system, and is suitable for popularization and application in the fields of food safety supervision and risk exchange practice with increasingly diversified information requirements.
Drawings
Fig. 1 is a flowchart of a risk assessment method adopted in embodiment 1 of the present invention.
FIG. 2 is a multi-level evaluation index system for agricultural product agricultural residue risk constructed in example 1 of the present invention.
Fig. 3 is a webpage screenshot of a big data sharing platform for food security in embodiment 2 of the present invention.
Fig. 4 is a view showing the food safety risk comprehensive evaluation result in a visualized manner in embodiment 2 of the present invention.
Fig. 5 is a graph of agricultural residue risk time trend pre-warning for agricultural products in embodiment 2 of the present invention.
Detailed Description
The following description of the preferred embodiments of the present invention is provided for the purpose of better illustrating the invention and is not intended to limit the invention thereto.
Example 1: fuzzy comprehensive evaluation method for safety risk of chemical hazards in food
A safety risk fuzzy comprehensive evaluation method of chemical hazards in food based on entropy weight is disclosed, which is implemented according to food inspection and detection big data, and the implementation flow is shown in figure 1: firstly, acquiring food detection data, establishing an evaluation factor set and an index system according to a risk evaluation purpose, and determining weights for risk factors by adopting an improved entropy weight method; then, setting a risk level comment set and assigning values, and establishing a membership function by adopting a halving ladder distribution model; secondly, calculating the membership degree of the hazardous material detection values in the food, constructing an evaluation fuzzy subset, and calculating comprehensive risk indexes of each layer step by adopting a weight-based matrix merging algorithm; and finally, carrying out visual display and interactive information service on the comprehensive evaluation result by using a big data sharing platform.
The specific implementation steps are as follows:
1) risk evaluation index system construction and entropy weight calculation
A1. Preprocessing raw data: in order to comprehensively evaluate the pesticide residue risk level of agricultural products sold in the current market, agricultural product supervision and spot inspection data between 2017 and 2019 in 7 months in a certain market are selected in a simulation mode and are shown in table 1.
Table 1 agricultural product pesticide residue supervision sampling inspection original data
Note: in addition to the above information, the supervised sampling inspection information includes report number, sample specification and model, sampling inspection information, inspection institution and inspection personnel information, detailed information of production and sales subject, etc., which are exemplified in the table and not listed in detail.
Classifying the original data according to evaluation indexes and carrying out standard processing, which mainly comprises the steps of assigning half values of the detection limit (L OD) of character type detection results of characters such as 'undetected', 'lower than the detection limit' and the like by a corresponding standard detection method so as to carry out quantitative calculation, carrying out non-dimensionalization processing on the original detection results by taking a corresponding national standard limit value (MR L) as a reference, and carrying out (formula 1):
xij＝Tij/MRL(1)
in the formula, xijThe ith raw test result (T) of the jth risk factorij) And (4) evaluation value after dimensionless processing. Obtaining evaluation value x according to equation 1ijAnd the differences in dimensions between different indexes are eliminated, and the table 2 shows.
TABLE 2 raw test data normalization and Pre-processing
A2. Constructing an evaluation factor set: establishing a three-level evaluation index system (see figure 2) according to the detection object and the risk factor category, wherein in a first-level fuzzy evaluation (or single-factor fuzzy evaluation) system, m types of chemical risk factors influencing food safety form a first-level evaluation factor set (U) which is expressed as: u ═ U1，U2，U3，…，UmAnd (f) forming a first-level evaluation index matrix X (X) by each risk factor and n detection results contained in the risk factorsij)n×mWherein i ═ 1, 2, 3, …, n; j is 1, 2, 3, …, m.
A3. Calculating the entropy weight of the first-level risk factor: for the ith evaluation value (x) of the jth risk factorij) Normalization treatment (formula 2):
in the formula: vijRepresents the calculation result of the ith evaluation value of the jth risk factor after normalization processing, xj(max)And xj(min)Respectively representing the maximum value and the minimum value of the n evaluation values of the j-th risk factor,
is a revision coefficient to the normalization method. The revision significance lies in that: (1) due to the wide distribution of the big data, a meaningless calculation result also appears when the original detection value is subjected to normalization processing, namely the denominator in the formula (2) is zero, so that the problem that under the condition of big data, when the calculation result in the step is 0 or meaningless (namely the denominator is 0), the subsequent operation step cannot be automatically performed is avoided.
(2) Since coefficient revision is adopted instead of constant revision, there is no significant difference between the revision result and the calculation result for any original data. The inventor has conducted comparative studies on the revision coefficients of 1/100, 1/1000 and 1/10000, and from the actual analysis results, 1/1000 has satisfied the accuracy requirement of data analysis, and by means of the revision of the coefficients, the automatic calculation and analysis capability of data under the condition of large data can be improved while the original calculation results are not significantly affected.
The characteristic specific gravity (P) of each evaluation value in the index is calculated by using the formula (3)ij)：
Then, the j-th type risk factor entropy value (e)j) Can be calculated by equation (4):
among the m-class risk factors, the entropy weight of the jth class risk factor is calculated by equation (5) and satisfies
A4. And (3) carrying out multilevel risk factor entropy weight calculation: if at the first levelOn the basis of the existing multi-level evaluation indexes, for an evaluation system comprising k types of secondary evaluation indexes, applying an entropy-adding principle and an algorithm, wherein the difference coefficient (D) of the l type (l is 1, 2, 3, …, k) secondary indexesl) I.e. the difference coefficient (d) of m primary indexesj) The sum (formula 6, 7):
wherein: dj＝1-ej(7)
Then, in the second-level index, the entropy weight (ω) of the first-term indexl) Calculated by equation (8) and satisfies
And (4) carrying out iterative calculation on more than two levels of evaluation indexes by using the method.
Evaluating the value x by various first-level indexesijRespectively calculating the entropy values (e) by using an improved entropy algorithm for calculation basisj) And coefficient of difference (d)j) The entropy calculation process of the nicotine pesticide residue level of solanaceous vegetables in agricultural products is shown in table 3. And then, based on the entropy strong adding principle and algorithm, the iterative summation and normalization calculation are carried out on the risk factor difference coefficients to obtain the evaluation index weights of two levels or more (see table 4).
TABLE 3 calculation of entropy values of nicotine pesticide residues in solanaceous vegetables
Table 4 weight table of safety risk evaluation index system for pesticide residue of edible agricultural products
2) Fuzzy function construction and first-level index comprehensive risk index calculation
B1. Establishing a comment set of fuzzy comprehensive evaluation, namely evaluating the food safety risk by adopting a 'no (micro), low, medium, high and ultrahigh' 5-level scale method, setting a comment set V (micro) risk, low risk, medium risk, high risk and ultrahigh risk), and setting a universal representative value L for evaluation levels of various risk factors based on the highest residual limit value (MR L) of the hazard factors in the national standard of food safety and the technical parameters of a standard detection methodqWhere q is 1, 2, 3, 4, 5, i.e. { L1，L2，L3，L4，L5The terms "0.25," 0.50, "0.75," 1.00, ", and" 2.00 "correspond to the fuzzy evaluation set" V, "(no (micro) risk, low risk, medium risk, high risk, and ultra high risk", respectively, as shown in table 5.
TABLE 5 evaluation level and representative value of food safety risk
The national food safety limit standard (MR L) is established on the basis of comprehensive evaluation of health hazards of hazard factors and is a universal index for judging food quality safety and marketing permission, thus MR L is set as a key representative value for judging high risk of food safety, the detection result of the hazard factor of 'undetected' is equivalent to the level of 'no (micro) risk' of the risk factor, in food quality detection and evaluation, according to 'credible evaluation recommendation of low-level pollutants in food' and relevant rules, the detection result is regularly assigned to L OD/2 in quantitative analysis, on the premise that the quantitative limit (L OQ) of a standard detection method at least meets the preset value of MR L, the ratio relation between the detection limit (365638 OD) and the quantitative limit (7 OQ) is calculated to be 1:2 by taking the detection method S/N (signal-to-noise ratio) as a quantitative basis, the upper limit of the grade of 'no (micro) risk' is 0.25, between the 'no (micro) risk representative value and the quantitative limit (7 OQ) is calculated to be 1:2, the intermediate grade of the' no (352) is calculated to be equal and the average risk factor of the average risk factor of the grade of the risk factor of the risk.
B2. Constructing a membership function: chemical hazards belong to negative effect indexes, and a half-reduced gradient distribution model is adopted as a membership function. In the first-level evaluation index system, the ith evaluation value (x) of the jth risk factorij) Degree of membership (h) thereofijq) The calculation method is as follows:
when q is equal to 1, the reaction is carried out,
when q is 2, 3, 4,
when q is equal to 5, the reaction is carried out,
the risk factors in the food belong to negative effect indexes, a half-reduced gradient function is selected for fuzzy evaluation, and a general type comment set is combined to evaluate the value x of each indexijFor the calculation basis, a general membership function is established:
for the class i rating scale, there are:
for the class ii rating scale, there are:
grade III and IV and so on.
For a rating scale of v, there are:
B3. calculating the comprehensive risk index of a first-level evaluation system: based on single factor analysis, the evaluation values of the jth risk factor are subjected to membership function calculation to obtain the fuzzy judgment subset (r) of the risk factorj)：
And (3) substituting the evaluation values of the first-level risk factors in the table 2 into the membership functions respectively to obtain evaluation matrixes, and then performing subsequent synthesis calculation. Substituting evaluation values of nicotine pesticide residue in solanaceous vegetables into membership functions to obtain fuzzy judgment subsets (r)Nicotinic acid):
Based on the mean value algorithm, the composite membership degree vector (R) of the mean value algorithm is obtainedNicotine)：
RNicotine＝[0.9048 0.0390 0.0087 0.0199 0.0277]
Risk membership degree vector R based on m-class primary indexesj(j ═ 1, 2, 3, …, m) to construct fuzzy subset R, and perform weighting (W) synthesis calculation on the fuzzy matrix, so as to obtain the comprehensive evaluation vector of the primary evaluation index:
in the formula: omegamFor each of the risk factor entropy weights is,
is a fuzzy synthesis operator. In order to highlight the entropy weight and take the influence of each risk factor into consideration, and improve the comprehensive evaluation capability of the model, the comprehensive evaluation method selects
(product, sum operator) is used for carrying out synthesis operation on the fuzzy matrix, the algorithm is used for respectively solving the composite membership degrees of various pesticide residues of the solanaceous vegetables, and then the comprehensive evaluation vector B of the solanaceous vegetables is calculatedSolanum melongena fruit：
And (3) calculating a first-level risk factor comprehensive evaluation index (FSI) by adopting a fuzzy comprehensive index algorithm (formula 12), and judging the grade of the FSI according to the maximum membership degree of the FSI in the comment set.
In the formula: sqRank (q is 1, 2, 3, 4, 5) of each rank of the panel of comments, i.e., { s {(s) }1，s2，s3，s4，s5}＝{0，1，2，3，4}；
A fuzzy synthesis operator;
comprehensive risk index (FSI) of pesticide residues of solanaceous vegetablesSolanum melongena fruit) Comprises the following steps:
according to the standard of the comment set and the maximum membership rule, FSISolanum melongena fruitAnd (3) the comprehensive risk of the solanaceous fruit vegetable pesticide residues is judged to be I-level no (micro) risk within the range of I-level risk grade interval (0-0.5).
3) In a multi-level risk factor evaluation system, a secondary evaluation index fuzzy judgment subset is formed by a subordinate primary evaluation factor comprehensive evaluation vector, and then a matrix synthesis algorithm can be iteratively applied to calculate the risk factor comprehensive evaluation index of each layer step by step. Respectively calculating the membership degree vectors of pesticide residue risks of various vegetables by using the same algorithm, and calculating the comprehensive evaluation vector (B) of the vegetables by using a fuzzy synthesis algorithmVegetable product)：
The comprehensive risk index of the pesticide residues of the vegetables is as follows:
and judging the comprehensive risk of the vegetable pesticide residues to be I-level risk-free (micro) risk according to the risk degree interval standard of the evaluation set.
4) And (3) iteratively calculating the agricultural product pesticide residue risk comprehensive index (see table 6) by using the algorithm and the index entropy weights of all levels, and judging the current agricultural product pesticide residue comprehensive risk to be I-level no (micro) risk according to the evaluation set standard and the maximum membership judgment principle.
TABLE 6 comprehensive risk evaluation of agricultural residues in edible agricultural products
Example 2: food safety risk evaluation device
The device comprises: the detection module is used for detecting the content of each chemical hazard in the food, determining the evaluation value of each chemical hazard and constructing an evaluation factor set;
a calculation module: the entropy weight calculation module is used for calculating the entropy weight of each chemical hazard, calculating the entropy weight of each chemical hazard according to the detection data and evaluation value of each chemical hazard content obtained by the detection module, and determining the weight of the multilevel risk factor;
the analysis module is used for calculating the comprehensive evaluation index FSI of the chemical hazard safety risk in the food, calculating the entropy weight of each chemical hazard according to the calculation module, then calculating the comprehensive evaluation index FSI of the chemical hazard safety risk at each level in the food by adopting a fuzzy comprehensive evaluation method, and determining the risk level, wherein the risk level is determined by adopting a 5-level scaling method of 'none or little, low, medium, high and ultrahigh' by taking the national limit standard MR L of the food safety as an evaluation standard.
Calculating a comprehensive evaluation index FSI of the chemical hazard safety risk in the food according to the method of the embodiment 1, integrating the established comprehensive evaluation model and an analysis module into a food safety inspection detection big data sharing platform based on internet communication, and actively displaying a multi-dimensional comprehensive risk evaluation result in a circulating manner through visualization manners such as a conventional statistical chart, a hot spot map, multi-map linkage and the like; in a related area in a webpage, an operator can select a concerned target item by adopting a mouse or touch screen clicking mode, so that food safety sampling inspection and evaluation result conditional output are realized; the method can visually display the conventional statistical analysis indexes such as food quality safety detection rate, qualification rate, standard exceeding rate and the like, provide real-time information and a trend chart for comprehensive evaluation of food safety risks, train and verify historical detection data based on time series by adopting machine learning methods such as a BP neural network and the like as samples, and preferentially establish a prediction model. The detection data of the current time node is used as a prediction set, a network model is loaded, the obtained output value is a risk trend prediction value, in practical application, a response rule is set for the food safety risk level, a corresponding scheme can be provided according to the prediction value, a prediction model based on machine learning and a food safety risk response rule are established, and the food safety risk prediction early warning service requirements are met (see fig. 3, 4 and 5).
Although the present invention has been described with reference to the preferred embodiments, it should be understood that various changes and modifications can be made therein by those skilled in the art without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (10)
1. A method for fuzzy comprehensive evaluation of chemical hazard safety risk in food is characterized by comprising the following steps:
s1: detecting the content of each chemical hazard in the food, determining the evaluation value of each chemical hazard, and constructing an evaluation factor set;
s2, establishing a comment set of fuzzy comprehensive evaluation, namely evaluating the food safety risk by using a 'none/micro, low, medium, high and ultrahigh' 5-grade scaling method by taking the national food safety limit standard MR L as an evaluation standard;
s3: determining the weight of the multilevel risk factors by using an entropy algorithm based on the evaluation value of each chemical hazard of S1; wherein, the S1 original detection data normalization method is revised by a fixed coefficient method;
s4: and calculating the comprehensive evaluation index FSI of the safety risk of chemical hazards at all levels in the food by adopting a fuzzy comprehensive evaluation method, and directly judging whether the FSI corresponds to the risk level of no/little risk, low risk, medium risk, high risk or ultrahigh risk according to the maximum membership principle and the boundary value interval of each level of the comment set.
2. The method according to claim 1, wherein the evaluation value in S1 is obtained by performing non-dimensionalization on the original detection result with reference to the highest limit value MR L in the corresponding national standard for food safety:
xij＝Tij/MRI (1)
in the formula, xijThe ith original detection result T of the jth risk factorijEvaluation value after dimensionless treatment;
the construction evaluation factor set is as follows: in a first-level fuzzy evaluation system, m types of chemical risk factors influencing food safety form a first-level evaluation factor set U, which is expressed as: u ═ U1，U2，U3，…，UmAnd (f) forming a first-level evaluation index matrix X (X) by each risk factor and n detection results contained in the risk factorsij)n×mWherein i ═ 1, 2, 3, …, n; j is 1, 2, 3, …, m.
3. The method of claim 1, wherein the fixed coefficient revised in S3 is 1/1000; adopting formula (2) to evaluate the ith evaluation value x of the jth risk factorijAnd (3) carrying out normalization treatment:
in the formula: vijRepresents the calculation result of the ith evaluation value of the jth risk factor after normalization processing, xj(max)And xj(min)Respectively representing the maximum value and the minimum value of the n evaluation values of the j-th risk factor,
is a revision coefficient to the normalization method.
4. The method of claim 1, wherein S3 calculates the entropy weight ω of each chemical hazard by:
adopting formula (2) to evaluate the ith evaluation value x of the jth risk factorijAnd (3) carrying out normalization treatment:
in the formula: vijRepresents the calculation result of the ith evaluation value of the jth risk factor after normalization processing, xj(max)And xj(min)Respectively representing the maximum value and the minimum value of the n evaluation values of the j-th risk factor,
is a revision coefficient to the normalization method;
calculating the characteristic specific gravity P of each evaluation value in the index by adopting the formula (3)ij：
Then, the j-th type risk factor entropy value ejCan be calculated by equation (4):
among the m classes of risk factors, the entropy weight ω of the jth class of risk factorsjCalculated by equation (5) and satisfies
Wherein j is 1, 2, 3, …, m.
5. The method according to claim 4, wherein if there are multiple levels of evaluation indexes above the first level index in S3, applying entropy-enhancing principle and algorithm to the evaluation system containing k types of second level evaluation indexes, and the difference coefficient D of the l type of second level indexlI.e. the difference coefficient d of m primary indexesjAnd, formula 6, 7, where l ═ 1, 2, 3, …, k:
wherein: dj＝1-ej(7)
Then, in the second-level index, the entropy weight ω of the first-term indexlCalculated by equation (8) and satisfies
And (4) carrying out iterative calculation on more than two levels of evaluation indexes by using the method.
6. The method of claim 1, wherein the step of S4 calculating the chemical hazard safety risk comprehensive assessment index FSI in the food by using the fuzzy comprehensive assessment method comprises the following steps:
(a) constructing a membership function: chemical hazards belong to negative effect indexes and membership functionsA halving ladder distribution model is adopted; in a first-level evaluation index system, for the ith evaluation value x of the jth risk factorijDegree of membership h thereofijqThe calculation method is as follows:
when q is equal to 1, the reaction is carried out,
when q is 2, 3, 4,
when q is equal to 5, the reaction is carried out,
in the formula, LqA representative value of a general type set for the 5-scale evaluation scale, q being 1, 2, 3, 4, 5, i.e., { L }1，L2，L3，L4，L5The evaluation results are (1, 0, 0.25, 0.50, 0.75, 1.00, 2.00), which correspond to a comment set V of fuzzy comprehensive evaluation, i, ii, iii, iv, V, i, V;
(b) calculating the comprehensive risk index of a first-level evaluation system: on the basis of single-factor analysis, performing membership function calculation on each evaluation value of the jth risk factor to obtain a fuzzy judgment subset r of the risk factorj：
To rjRespectively averaging the membership degrees of all levels in the matrix to obtain a composite membership degree vector R of the jth risk factorjq，
q＝1，2，3，4，5；
Risk membership degree vector R based on m-class primary indexesjqConstructing a fuzzy matrix R ═ (R)jq)m×5Wherein j is 1, 2, 3, …, m, q is 1, 2, 3, 4, 5; and then carrying out weighted W synthesis calculation on the fuzzy matrix R to obtain a first-level evaluation index comprehensive evaluation vector B:
in the formula: omegamFor each of the risk factor entropy weights is,
a fuzzy synthesis operator;
by adopting a fuzzy comprehensive index algorithm, as shown in formula (12), a first-level risk factor comprehensive evaluation index FSI can be calculated, and the corresponding risk grade is judged according to a maximum membership principle:
in the formula:
for fuzzy synthesis of operators, sqRank for the panel of comments, where q is 1, 2, 3, 4, 5, i.e., { s }1，s2，s3，s4，s50, 1, 2, 3, 4; the assessment set V ═ i, { i, ii, iii, iv, V } ═ i, { no/low risk, medium risk, high risk, ultra-high risk }, and the interval representative value is {0, 1, 2, 3, 4 }; and directly judging the risk grade of the FSI according to the maximum membership rule and the boundary value interval of each risk grade.
7. A food safety risk assessment device, characterized in that the device comprises:
the detection module is used for detecting the content of each chemical hazard in the food, determining the evaluation value of each chemical hazard and constructing an evaluation factor set;
a calculation module: the entropy weight calculation module is used for calculating the entropy weight of each chemical hazard, calculating the entropy weight of each chemical hazard according to the detection data and evaluation value of each chemical hazard content obtained by the detection module, and determining the weight of the multilevel risk factor;
the analysis module is used for calculating the comprehensive evaluation index FSI of the chemical hazard safety risk in the food, calculating the entropy weight of each chemical hazard according to the calculation module, then calculating the comprehensive evaluation index FSI of the chemical hazard safety risk at each level in the food by adopting a fuzzy comprehensive evaluation method, and determining the risk level, wherein the risk level is determined by adopting a 5-level scaling method of 'none/micro, low, medium, high and ultrahigh' by taking the national limit standard MR L of the food safety as an evaluation standard.
8. The apparatus of claim 7, wherein the computing unit comprises a revision module for revising the original detection data normalization method by using a fixed coefficient method, wherein the fixed coefficient method is 1/1000; adopting formula (2) to evaluate the ith evaluation value x of the jth risk factorijAnd (3) carrying out normalization treatment:
in the formula: vijRepresents the calculation result of the ith evaluation value of the jth risk factor after normalization processing, xj(max)And xj(min)Respectively representing the maximum value and the minimum value of the n evaluation values of the j-th risk factor,
is a revision coefficient to the normalization method.
9. The device according to claim 7, further comprising a big data sharing platform, wherein the big data sharing platform is connected with the analysis module, risk level data acquired by the analysis module are integrated into the big data sharing platform for food safety inspection detection based on internet communication, and then the multi-dimensional comprehensive risk evaluation result is actively displayed in a circulating manner through a conventional statistical chart, a hot spot map and a multi-graph linkage visualization manner in the big data sharing platform; the big data sharing platform can also realize a risk prediction and early warning function, a machine learning method is adopted, historical detection data based on a time sequence are used as samples for training and verification, and a prediction model is preferentially established; and (4) loading the network model by taking the detection data of the current time node as a prediction set, wherein the obtained output value is the risk trend prediction value.
10. Use of the method of any one of claims 1 to 6 for the assessment of the safety of food quality.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010343593.7A CN111461576A (en) | 2020-04-27 | 2020-04-27 | Fuzzy comprehensive evaluation method for safety risk of chemical hazards in food |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010343593.7A CN111461576A (en) | 2020-04-27 | 2020-04-27 | Fuzzy comprehensive evaluation method for safety risk of chemical hazards in food |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111461576A true CN111461576A (en) | 2020-07-28 |
Family
ID=71678664
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010343593.7A Pending CN111461576A (en) | 2020-04-27 | 2020-04-27 | Fuzzy comprehensive evaluation method for safety risk of chemical hazards in food |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111461576A (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112258069A (en) * | 2020-10-30 | 2021-01-22 | 中国农业科学院农业质量标准与检测技术研究所 | Agricultural product safety evaluation method and system based on risk entropy |
| CN112381364A (en) * | 2020-10-30 | 2021-02-19 | 浪潮云信息技术股份公司 | Comprehensive evaluation method for food quality spot check |
| CN112465264A (en) * | 2020-12-07 | 2021-03-09 | 湖北省食品质量安全监督检验研究院 | Food safety risk grade prediction method and device and electronic equipment |
| CN113011796A (en) * | 2021-05-06 | 2021-06-22 | 北京工商大学 | Edible oil safety early warning method based on hierarchical analysis-neural network |
| CN113360898A (en) * | 2021-06-03 | 2021-09-07 | 中国电子信息产业集团有限公司第六研究所 | Index weight determination method, network attack evaluation method and electronic equipment |
| CN113393127A (en) * | 2021-06-17 | 2021-09-14 | 江苏省环境科学研究院 | Screening optimization determination method for regional environment control chemical substances |
| CN114520030A (en) * | 2022-02-18 | 2022-05-20 | 广州机械科学研究院有限公司 | Oil product performance evaluation method and device, electronic equipment and readable storage medium |
| CN114764682A (en) * | 2022-03-25 | 2022-07-19 | 北京工商大学 | Rice safety risk assessment method based on multi-machine learning algorithm fusion |
| CN115758888A (en) * | 2022-11-17 | 2023-03-07 | 厦门智康力奇数字科技有限公司 | Agricultural product safety risk assessment method based on fusion of multi-machine learning algorithm |
| CN116227982A (en) * | 2022-12-30 | 2023-06-06 | 中国矿业大学(北京) | Quantification method and device for pollution degree of coal dust |
| CN116862231A (en) * | 2023-07-07 | 2023-10-10 | 上海福智宠物食品有限公司 | Risk assessment system and method for pet production full link |
| CN117251492A (en) * | 2023-09-20 | 2023-12-19 | 怀宁县政通中小企业服务中心 | Intelligent park industry cluster data comparison analysis system and method |
| CN117726489A (en) * | 2024-02-18 | 2024-03-19 | 昆明理工大学 | Method for evaluating disaster risk of smooth stone slope |
| CN118536808A (en) * | 2024-05-15 | 2024-08-23 | 国贸食品科学研究院有限公司 | Food safety risk evaluation method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103606031A (en) * | 2013-11-04 | 2014-02-26 | 清华大学 | Evaluation method of kitchen waste feed product safety |
| CN104915793A (en) * | 2015-06-30 | 2015-09-16 | 北京西塔网络科技股份有限公司 | Public information intelligent analysis platform based on big data analysis and mining |
| US20160086123A1 (en) * | 2014-09-24 | 2016-03-24 | Research Center For Justice Standards Ltd. | Methods and systems for evaluating judicial system |
| CN106920044A (en) * | 2017-03-07 | 2017-07-04 | 北京工商大学 | Residues of pesticides methods of risk assessment based on model of fuzzy synthetic evaluation |
-
2020
- 2020-04-27 CN CN202010343593.7A patent/CN111461576A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103606031A (en) * | 2013-11-04 | 2014-02-26 | 清华大学 | Evaluation method of kitchen waste feed product safety |
| US20160086123A1 (en) * | 2014-09-24 | 2016-03-24 | Research Center For Justice Standards Ltd. | Methods and systems for evaluating judicial system |
| CN104915793A (en) * | 2015-06-30 | 2015-09-16 | 北京西塔网络科技股份有限公司 | Public information intelligent analysis platform based on big data analysis and mining |
| CN106920044A (en) * | 2017-03-07 | 2017-07-04 | 北京工商大学 | Residues of pesticides methods of risk assessment based on model of fuzzy synthetic evaluation |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112258069A (en) * | 2020-10-30 | 2021-01-22 | 中国农业科学院农业质量标准与检测技术研究所 | Agricultural product safety evaluation method and system based on risk entropy |
| CN112381364A (en) * | 2020-10-30 | 2021-02-19 | 浪潮云信息技术股份公司 | Comprehensive evaluation method for food quality spot check |
| CN112258069B (en) * | 2020-10-30 | 2021-07-30 | 中国农业科学院农业质量标准与检测技术研究所 | Agricultural product safety evaluation method and system based on risk entropy |
| CN112465264A (en) * | 2020-12-07 | 2021-03-09 | 湖北省食品质量安全监督检验研究院 | Food safety risk grade prediction method and device and electronic equipment |
| WO2022121312A1 (en) * | 2020-12-07 | 2022-06-16 | 湖北省食品质量安全监督检验研究院 | Food safety risk level prediction method and apparatus, and electronic device |
| CN113011796A (en) * | 2021-05-06 | 2021-06-22 | 北京工商大学 | Edible oil safety early warning method based on hierarchical analysis-neural network |
| CN113011796B (en) * | 2021-05-06 | 2024-02-02 | 北京工商大学 | Edible oil safety early warning method based on' analytic hierarchy process-neural network |
| CN113360898A (en) * | 2021-06-03 | 2021-09-07 | 中国电子信息产业集团有限公司第六研究所 | Index weight determination method, network attack evaluation method and electronic equipment |
| CN113360898B (en) * | 2021-06-03 | 2024-05-24 | 中国电子信息产业集团有限公司第六研究所 | Index weight determining method, network attack evaluating method and electronic equipment |
| CN113393127A (en) * | 2021-06-17 | 2021-09-14 | 江苏省环境科学研究院 | Screening optimization determination method for regional environment control chemical substances |
| CN114520030A (en) * | 2022-02-18 | 2022-05-20 | 广州机械科学研究院有限公司 | Oil product performance evaluation method and device, electronic equipment and readable storage medium |
| CN114764682A (en) * | 2022-03-25 | 2022-07-19 | 北京工商大学 | Rice safety risk assessment method based on multi-machine learning algorithm fusion |
| CN114764682B (en) * | 2022-03-25 | 2023-04-07 | 北京工商大学 | Rice safety risk assessment method based on multi-machine learning algorithm fusion |
| CN115758888A (en) * | 2022-11-17 | 2023-03-07 | 厦门智康力奇数字科技有限公司 | Agricultural product safety risk assessment method based on fusion of multi-machine learning algorithm |
| CN115758888B (en) * | 2022-11-17 | 2024-04-23 | 厦门智康力奇数字科技有限公司 | Agricultural product security risk assessment method based on multi-machine learning algorithm fusion |
| CN116227982B (en) * | 2022-12-30 | 2023-10-31 | 中国矿业大学(北京) | Quantification method and device for pollution degree of coal dust |
| CN116227982A (en) * | 2022-12-30 | 2023-06-06 | 中国矿业大学(北京) | Quantification method and device for pollution degree of coal dust |
| CN116862231A (en) * | 2023-07-07 | 2023-10-10 | 上海福智宠物食品有限公司 | Risk assessment system and method for pet production full link |
| CN117251492A (en) * | 2023-09-20 | 2023-12-19 | 怀宁县政通中小企业服务中心 | Intelligent park industry cluster data comparison analysis system and method |
| CN117726489A (en) * | 2024-02-18 | 2024-03-19 | 昆明理工大学 | Method for evaluating disaster risk of smooth stone slope |
| CN117726489B (en) * | 2024-02-18 | 2024-04-26 | 昆明理工大学 | Method for evaluating disaster risk of smooth stone slope |
| CN118536808A (en) * | 2024-05-15 | 2024-08-23 | 国贸食品科学研究院有限公司 | Food safety risk evaluation method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111461576A (en) | Fuzzy comprehensive evaluation method for safety risk of chemical hazards in food | |
| De la Rosa et al. | Qualitative and quantitative land evaluations | |
| CN113011796B (en) | Edible oil safety early warning method based on' analytic hierarchy process-neural network | |
| CN106920044B (en) | Pesticide residue risk assessment method based on fuzzy comprehensive evaluation model | |
| CN112700319A (en) | Enterprise credit line determination method and device based on government affair data | |
| CN110020762A (en) | A kind of comprehensive monitoring and evaluation system and method for school | |
| Saleh et al. | Palm oil classification using deep learning | |
| Bąk et al. | Fuzzy cognitive maps and their application in the economic sciences | |
| CN108459997A (en) | High skewness data value probability forecasting method based on deep learning and neural network | |
| Чіков et al. | Modeling of the synthetic indicator of competitiveness of agricultural enterprises: a methodological approach to the use of neural network tools | |
| Yuan et al. | Research on AHP-fuzzy comprehensive evaluation method and application | |
| Tian et al. | A two stage risk assessment model based on interval-valued fuzzy numbers and risk attitudes | |
| CN116933928A (en) | TabNet-GRA-based food safety risk prediction method and visual analysis system | |
| CN112766739A (en) | BWM-E model-based evaluation method for heavy metal pollution in meat products | |
| CN117132383A (en) | Credit data processing method, device, equipment and readable storage medium | |
| CN114764682B (en) | Rice safety risk assessment method based on multi-machine learning algorithm fusion | |
| Zhang et al. | EVALUATION SYSTEM FOR LEAN KNOWLEDGE MANAGEMENT ABILITY BASED ON IMPROVED GRAY CORRELATION ANALYSIS. | |
| Abubakar et al. | Performance management antecedents and public sector organizational performance: data screening and preliminary analysis | |
| CN114331053A (en) | Shallow lake eutrophication evaluation method based on fuzzy hierarchical evaluation model | |
| KR101918359B1 (en) | A Data Processing System Using AI | |
| Indrasari et al. | ELECTRE III for Human Resource Management: A Study of Recruitment and Retention Strategies | |
| Sarveswaran et al. | MilkSafe: A Hardware-Enabled Milk Quality Prediction using Machine Learning | |
| Bokaei et al. | Portfolio Selection with Fuzzy Synthetic Evaluation and Genetic Algorithm based on Performance Perspectives | |
| CN116976694A (en) | Comprehensive evaluation method for beach ecosystem | |
| CN117436689A (en) | Safety diagnosis method, platform, electronic equipment and storage medium for coal chemical industry enterprises |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200728 |
|
| RJ01 | Rejection of invention patent application after publication |