CN110428121A - Hidden Markov model food quality appraisal procedure based on grey correlation analysis - Google Patents
Hidden Markov model food quality appraisal procedure based on grey correlation analysis Download PDFInfo
- Publication number
- CN110428121A CN110428121A CN201910330705.2A CN201910330705A CN110428121A CN 110428121 A CN110428121 A CN 110428121A CN 201910330705 A CN201910330705 A CN 201910330705A CN 110428121 A CN110428121 A CN 110428121A
- Authority
- CN
- China
- Prior art keywords
- index
- markov model
- hidden markov
- follows
- food quality
- 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
- 235000013305 food Nutrition 0.000 title claims abstract description 46
- 238000010219 correlation analysis Methods 0.000 title claims abstract description 28
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000012549 training Methods 0.000 claims abstract description 32
- 239000011159 matrix material Substances 0.000 claims description 19
- 238000012545 processing Methods 0.000 claims description 5
- 238000012502 risk assessment Methods 0.000 abstract description 15
- 238000004519 manufacturing process Methods 0.000 abstract description 5
- 238000001303 quality assessment method Methods 0.000 abstract description 4
- 238000012954 risk control Methods 0.000 abstract description 4
- 235000013336 milk Nutrition 0.000 description 23
- 239000008267 milk Substances 0.000 description 23
- 210000004080 milk Anatomy 0.000 description 23
- 238000010586 diagram Methods 0.000 description 17
- 238000001514 detection method Methods 0.000 description 11
- 238000011156 evaluation Methods 0.000 description 11
- 238000012360 testing method Methods 0.000 description 11
- 238000004458 analytical method Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 6
- 241001269238 Data Species 0.000 description 5
- 230000004044 response Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 229930195730 Aflatoxin Natural products 0.000 description 2
- VYZAMTAEIAYCRO-UHFFFAOYSA-N Chromium Chemical compound [Cr] VYZAMTAEIAYCRO-UHFFFAOYSA-N 0.000 description 2
- 239000005409 aflatoxin Substances 0.000 description 2
- 229910052785 arsenic Inorganic materials 0.000 description 2
- RQNWIZPPADIBDY-UHFFFAOYSA-N arsenic atom Chemical compound [As] RQNWIZPPADIBDY-UHFFFAOYSA-N 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 229910052804 chromium Inorganic materials 0.000 description 2
- 239000011651 chromium Substances 0.000 description 2
- 238000013480 data collection Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 235000013861 fat-free Nutrition 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- QSHDDOUJBYECFT-UHFFFAOYSA-N mercury Chemical compound [Hg] QSHDDOUJBYECFT-UHFFFAOYSA-N 0.000 description 2
- 229910052753 mercury Inorganic materials 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 102000004169 proteins and genes Human genes 0.000 description 2
- 108090000623 proteins and genes Proteins 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 231100000279 safety data Toxicity 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 230000002123 temporal effect Effects 0.000 description 2
- 238000012800 visualization Methods 0.000 description 2
- XWIYFDMXXLINPU-UHFFFAOYSA-N Aflatoxin G Chemical compound O=C1OCCC2=C1C(=O)OC1=C2C(OC)=CC2=C1C1C=COC1O2 XWIYFDMXXLINPU-UHFFFAOYSA-N 0.000 description 1
- PEDCQBHIVMGVHV-UHFFFAOYSA-N Glycerine Chemical compound OCC(O)CO PEDCQBHIVMGVHV-UHFFFAOYSA-N 0.000 description 1
- 231100000678 Mycotoxin Toxicity 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 239000003344 environmental pollutant Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000004615 ingredient Substances 0.000 description 1
- 235000020191 long-life milk Nutrition 0.000 description 1
- 230000000813 microbial effect Effects 0.000 description 1
- 230000002906 microbiologic effect Effects 0.000 description 1
- 244000005700 microbiome Species 0.000 description 1
- 239000002636 mycotoxin Substances 0.000 description 1
- 230000007935 neutral effect Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 231100000719 pollutant Toxicity 0.000 description 1
- 238000003908 quality control method Methods 0.000 description 1
- 238000010948 quality risk assessment Methods 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 235000020185 raw untreated milk Nutrition 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000001954 sterilising effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- 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
- 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
- G06Q10/06395—Quality analysis or management
-
- 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
- G06Q50/265—Personal security, identity or safety
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Tourism & Hospitality (AREA)
- Educational Administration (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Computer Security & Cryptography (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The hidden Markov model food quality appraisal procedure based on grey correlation analysis that the invention discloses a kind of, it include: that the corresponding index weights of the influence factor are obtained according to grey correlation analysis algorithm and each influence factor, the corresponding risk index of the index is obtained according to each index and the index weights, the hidden Markov model is trained using the risk index as the input data of hidden Markov model, food quality is assessed and predicted using hidden Markov model after training.Technical solution provided by the invention realizes the risk assessment to food quality more accurately, to help to improve the accuracy and efficiency of food quality assessment.Therefore, technical solution provided by the invention realizes assessment and prediction to food quality, can be analyzed with seriousness the reason of generation, formulates or adjust risk control measure in advance, to food production, examines and administrative department advises.
Description
Technical field
The present invention relates to technical field of food safety more particularly to a kind of hidden Markovs based on grey correlation analysis
Model food method for evaluating quality.
Background technique
The risk assessment of Safety of Food Quality and early warning system need perfect.
Summary of the invention
A kind of hidden horse based on grey correlation analysis is provided to solve limitation and defect, the present invention of the existing technology
Er Kefu model food method for evaluating quality, comprising:
It obtains reference vector and compares vector, the reference vector x0With the relatively vector xiIt is respectively as follows:
x0={ x0(1), x0(2) ..., x0(n)}；
xi={ xi(1), xi(2) ..., xi(n)}；
Wherein, n is number of samples, and i=0,1,2 ..., m-1, m are the number of all indexs；
Each index is normalized, normalized processing formula is as follows:
Obtain yi(k) and y0(k) in the grey incidence coefficient at k moment, the grey incidence coefficient are as follows:
Wherein, (0,1) ρ ∈；
Sequences y is obtained according to the grey incidence coefficient0With sequences yiBetween related coefficient, the related coefficient are as follows:
The correlation matrix of all indexs, the correlation matrix are obtained according to the related coefficient are as follows:
The corresponding index weights of each index are obtained according to the correlation matrix；
The corresponding risk index of the index is obtained according to each index and the index weights；
The hidden Markov model is instructed using the risk index as the input data of hidden Markov model
Practice；
Food quality is assessed and predicted using hidden Markov model after training.
Optionally, the step that the corresponding risk index of the index is obtained according to each index and the index weights
Suddenly include:
Local state is initialized according to viterbi algorithm are as follows:
δ1(i)=πibi(o1), i=1,2 ..., N
Dynamic Programming recursion moment t=2,3 is carried out ..., the local state at T moment are as follows:
Maximum value, hidden state are obtained at the T moment are as follows:
Corresponding risk index is obtained according to the hidden state and the index weights.
Optionally, described the step of obtaining corresponding risk index according to the hidden state and the index weights it
Before include:
δT(i) maximum value, the probability that hidden state sequence occurs are obtained at the T moment are as follows:
Optionally, further includes:
Use local stateRecalled, for t=T-1, T-2 ..., 1, corresponding hidden state sequence are as follows:
Optimal path, the optimal path are obtained according to the hidden state sequence are as follows:
I*={ i1*, i2..., i *T*}。
Optionally, the step of index weights corresponding according to each index of correlation matrix acquisition include:
The average value of each index is obtained according to each index among the related coefficient square；
Using the average value of each index as the corresponding index weights of each index.
The present invention have it is following the utility model has the advantages that
Hidden Markov model food quality appraisal procedure provided by the invention based on grey correlation analysis includes: root
The corresponding index weights of the influence factor are obtained according to grey correlation analysis algorithm and each influence factor, according to each index
Risk index corresponding with the index weights acquisition index, using the risk index as hidden Markov model
Input data is trained the hidden Markov model, using hidden Markov model after training to food quality
It is assessed and is predicted.Technical solution provided by the invention realizes the risk assessment to food quality more accurately, thus
Help to improve the accuracy and efficiency of food quality assessment.Therefore, technical solution provided by the invention is realized to food matter
The assessment and prediction of amount can be analyzed with seriousness the reason of generation, formulate or adjust risk control in advance
Measure is examined to food production and administrative department advises.
Detailed description of the invention
Fig. 1 is the hidden Markov model schematic diagram based on grey correlation analysis that the embodiment of the present invention one provides.
Fig. 2 is the risk class criteria for classifying schematic diagram that the embodiment of the present invention one provides.
Fig. 3 is the sterile milk risk assessment hidden Markov model schematic diagram that the embodiment of the present invention one provides.
Fig. 4 is accuracy schematic diagram of the model that provides of the embodiment of the present invention one in training set.
Fig. 5 is accuracy schematic diagram of the model that provides of the embodiment of the present invention one in training set and test set.
Fig. 6 is accuracy schematic diagram of the model that provides of the embodiment of the present invention one in various training sets and test set.
Fig. 7 is the viterbi algorithm visualization process schematic diagram that the embodiment of the present invention one provides.
Fig. 8 is regular schematic diagram of the sterile milk quality that provides of the embodiment of the present invention one with time change.
Fig. 9 is the hidden Markov model risk evaluation result schematic diagram that the embodiment of the present invention one provides.
Specific embodiment
To make those skilled in the art more fully understand technical solution of the present invention, with reference to the accompanying drawing to the present invention
The hidden Markov model food quality appraisal procedure based on grey correlation analysis provided is described in detail.
Embodiment one
The present embodiment carries out risk assessment using hidden Markov model, and the training model first has to obtain each evaluation and refers to
Weight is marked, the prior art seeks index weights using expert system and the method for step analysis.However, the method for expert system
The shortcomings that there are biggish subjectivity and inconsistencies, sensitivity is low and big by subjective impact.Although analytic hierarchy process (AHP) uses
Method that is qualitative and quantitatively combining, but level building is complicated, relies on the judgement of expert's subjectivity.Due to influencing food safety
Factor is complicated, and the influence factor and detection data of food safety all have very strong non-linear, therefore nonlinear correlation analysis is more
Suitable for solving food data.
Grey correlation analysis (Grey Relational Analysis, GRA) is important point in gray system theory
Branch, basic thought is to judge whether the connection between different sequences is close according to sequence curve geometry, is suitable for analysis
With processing nonlinear data.For the non-linear behavior of food inspection data, grey correlation analysis algorithm building food can be used
Incidence coefficient matrix in product safety between each index, and then obtain each index weights.
Hidden Markov model (Hidden Markov Model, HMM) is developed on the basis of Markov chain
Come, the presence of perception state and its model of process are gone by a random process.Hidden Markov model is applied to risk
Assessment, so that assessment result has better real-time and specific aim, can realize risk with the temporal characteristics of combined data
Dynamic evaluation.Hidden Markov model is a kind of Statistic analysis models, is usually indicated with five members, including 2 state sets
With 3 probability matrixs, respectively hidden state Q, Observable state O, initial state probabilities matrix Π, hidden state transfer is generally
Rate matrix A and Observable state transition probability matrix B.Π, A and B are initialized first, are combined using Baum-Welch algorithm defeated
Enter data training hidden Markov model, obtain trained hidden Markov model, reuses viterbi algorithm in conjunction with considerable
Survey state carries out risk assessment and prediction.Using hidden Markov model, can be adjusted in advance with dynamic evaluation Safety of Food Quality
Whole risk control measure reduces the loss of quality security problem bring.
Fig. 1 is the hidden Markov model schematic diagram based on grey correlation analysis that the embodiment of the present invention one provides.Such as figure
Shown in 1, the weight of each index is calculated first with grey correlation analysis for the present embodiment, and comprehensive each influence factor obtains
Hidden Markov model is trained to the risk index of index, then using risk index as the input data of hidden Markov model,
Obtain trained model dynamic evaluation and prediction Safety of Food Quality.The more accurate ground of technical solution provided by the invention is real
The risk assessment to food quality is showed, to help to improve the accuracy and efficiency of food quality assessment.
Diversity for food safety data and non-linear, traditional hidden Markov model cannot be effectively to non-thread
Property data analyzed, the present embodiment utilize grey correlation analysis algorithm improvement hidden Markov model, solve each index
Weight reuses Baum-Welch algorithm and the training of Viterbi algorithm implementation model and risk assessment, to realize risk
Dynamic evaluation.Grey Incidence Analysis is for data non-linear, finds out the weight of each index, realizes more accurately
The analysis of food-safe data.Therefore, hidden Markov model can pacify quality in conjunction with the time response of food data
It is complete to carry out dynamic risk assessment.
The present embodiment passes through grey correlation analysis first, solves each index weights, refers in conjunction with weight calculation risk
Number.Each index respectively serves as a reference sequences, remaining index, which is used as, compares sequence.Setting reference vector is x0={ x0(1),
x0(2),…,x0(n) }, n is number of samples, and comparing vector is xi={ xi(1),xi(2),…,xi(n) }, i=0,1,2 ...,
M-1, m are all index numbers, establish correlation matrix further according to grey correlation analysis.
M group index is normalized in the present embodiment, eliminates the influence of dimension, and processing formula is as follows:
The present embodiment calculates grey incidence coefficient, the y at k momenti(k) and y0(k) grey incidence coefficient is as follows:
Wherein, ξiIt (k) is grey incidence coefficient, ρ ∈ (0,1), adjustment parameter ρ can make the otherness of each coefficient
Enhancing.Sequences y0And sequences yiBetween incidence coefficient can indicate are as follows:
Each index respectively serves as a reference sequences, and the correlation matrix of all indexs can be obtained by formula (3):
The present embodiment takes mean value to index each in related coefficient, the corresponding weight of as each index.In addition, hidden horse
The mathematical notation of five elements of Er Kefu model is as follows:
Q={ q1, q2..., qN} (5)
O={ o1, o2..., oM} (6)
Π={ [π (i)]N, π (i)=P (qi)} (7)
A={ [aij]N×N, aij=P (qj|qi)} (8)
B={ [bj(k)]N×M, bj(k)=P (ok}qj}} (9)
The present embodiment solves the model parameter λ of hidden Markov model, random initializtion institute using Baum-Welch algorithm
Some πi, aij, bj(k).Model parameter formula provided in this embodiment is as follows:
πi=γ1(i) (10)
If πi, aij, bj(k) value has restrained, then obtains final result, otherwise, continues iteration.
Under the conditions of the present embodiment solves given observation sequence using Viterbi algorithm, most probable occurs corresponding hiding
Status switch.Firstly, the present embodiment initializes local state, formula is as follows:
δ1(i)=πibi(o1), i=1,2 ..., N (13)
The present embodiment carries out Dynamic Programming recursion moment t=2, the local state of 3 ..., T, and formula indicates as follows:
The maximum δ of moment TT(i) probability occurred for most probable hidden state sequence, the moment, T was maximumFor when
The most probable hidden state of T is carved, calculation formula is as follows:
The present embodiment uses local stateStart to recall, hiding shape corresponding for moment t=T-1, T-2 ..., 1
State sequence, formula are as follows:
The optimal path that the present embodiment finally obtains are as follows:
I*={ i1*, i2..., i *T*} (20)
Above-mentioned optimal path is the hidden state sequence that most probable occurs.
Having for Safety of Food Quality risk assessment is carried out in order to verify the hidden Markov model based on grey correlation analysis
Effect property and accuracy, the present embodiment, using sterile milk data as research object, are used based on certain province's food safety data
Risk assessment is carried out to sterile milk detection data based on the hidden Markov model of grey correlation analysis, is dynamically assessed and pre-
Risk is surveyed with the changing rule of time.
Sterile milk detection data collection includes in August, 2013 to 2 months 2016 970 detection datas, and the present embodiment uses
375 datas in August, 2013 in October, 2015 carry out model training, and remaining data is used for model prediction and assessment.It is all
The processing technology of batch sterile milk is ultra-high-temperature sterilized milk.According to national standard it is found that the technical standard of sterile milk includes
Ingredient requirement, organoleptic requirements, physical and chemical index, pollutant limitation, mycotoxin limitation and Microbiological requirements.The present embodiment uses
Testing index include: protein, fat, non-fat solid, acidity, mercury, lead, arsenic, chromium and Aflatoxins M1.2016 6
The part detection data in the moon in October, 2014 is as shown in table 1, and national standard is as shown in table 2.
1 part detection data of table
2 national standard of table
Data are normalized using formula (1) for the present embodiment, and obtained partial results are as shown in table 3.
Part detection data after the standardization of table 3
The present embodiment calculates grey incidence coefficient, each finger to the data after normalization using formula (2) and formula (3)
Mark serves as a reference sequences, obtains correlation matrix as shown in table 4.
4 correlation matrix of table
The present embodiment takes mean value to each index weights respectively, and it is as shown in table 5 to obtain the corresponding weight of each index.
5 index weights of table
The present embodiment is carried out using the difference of pattern detection data and national standard come the quality safety risk to the sample
Assessment, by national standard it is found that the content of the fat of qualified sample, protein and non-fat solid need to meet more than or equal to state
Family's standard, is defined as positive index.Acidity need to meet the section of requirements of the national standard, and defining acidity is neutral index.Lead,
Mercury, arsenic, chromium and aflatoxin, which meet, is less than or equal to national standard, is defined as negative sense index.In conjunction with the classification of index, inspection
Measured data and national standard ask difference value equation as shown in table 6.
Table 6 seeks difference value equation
With the data instance of table 1, detection data and national standard make the difference the result that value obtains calculate it is as follows, to mass number
According to carrying out dividing available five risk class according to numerical values recited.The results are shown in Table 7 for calculated result and grade classification.
Fig. 2 is the risk class criteria for classifying schematic diagram that the embodiment of the present invention one provides.As shown in Fig. 2, according to 5 scaling laws, connexus
It measures data and divides risk class.
7 difference calculated result of table
Fig. 3 is the sterile milk risk assessment hidden Markov model schematic diagram that the embodiment of the present invention one provides.Such as Fig. 3 institute
Show, the present embodiment carries out model training using Baum-Welch algorithm, and defining this hidden Markov model, there are five implicit shapes
State, state 1-5 represent the quality safety risk of the sample, and number is bigger, and representation quality security risk is higher.By risk index
As the input data of model, as Observable state.Initialization probability matrix is as follows:
πi=[0.2 0.2 0.2 0.2 0.2]T (32)
The present embodiment carries out model training by taking in August, 2013 to 375 datas in October, 2015 as an example.By above-mentioned first
Beginningization result is as follows using the hidden Markov model result that the training completion of Baum-Welch algorithm obtains.
πi=[0 100 0]T (35)
The present embodiment uses the reliability of confusion matrix judgment models, and TP is represented really, the positive sample being positive by model prediction
This, TN, which is represented, really to be born, and the negative sample being negative by model prediction, FP is representing vacation just, the negative sample being positive by model prediction, FN generation
Table vacation is negative, the positive sample being negative by model prediction.Experiment provided in this embodiment is more classification problems, is using confusion matrix
When accuracy of judgement is spent, each classification is individually considered as " just ", all other type is considered as " negative ".Risk threshold value is set as 0 to open
Begin, every wheel increases by 0.2, until reaching 1.Fig. 4 is that the model that the embodiment of the present invention one provides is illustrated in the accuracy of training set
Figure.As shown in figure 4,1,2 and 3 respectively represent the TPR and TNR when training set size is 350,250 and 450.As shown in Figure 4,
With the increase of threshold value, the rate of descent of TPR and TNR are sharply increased, while under the premise of guaranteeing to assess accuracy, choosing threshold
When value is 0.6, TPR and TNR are maximum.
Fig. 5 is accuracy schematic diagram of the model that provides of the embodiment of the present invention one in training set and test set.Such as Fig. 5 institute
Show, for when training set size is 350, is tested using an equal amount of test set data.Accuracy on training set
Accuracy slightly above on test set.It is 0.6 citing with risk threshold value, risk evaluating system is on training set and test set
TPR is respectively 83.14% and 80.48%, and TNR is respectively 95.79% and 95.12%, and the accuracy on training set is higher than test
Accuracy on collection.
Using 0.6 as risk threshold value, the present embodiment has carried out another experiment and has carried out research and training collection size to TPR and TNR
Influence.The present embodiment has used different size of training set as model training data, remaining data collection as test set,
Accuracy of the model on training set and test set is tested respectively.Fig. 6 is the model that provides of the embodiment of the present invention one various
The accuracy schematic diagram of training set and test set.As shown in fig. 6, training set size, since 250, every wheel increases by 25.Work as instruction
When white silk collection reaches 375, the accuracy of model will not become more preferably because of training dataset is increased, i.e., 375 datas are for mould
For type training enough.
Fig. 7 is the viterbi algorithm visualization process schematic diagram that the embodiment of the present invention one provides.As shown in fig. 7, this implementation
Example carries out risk profile and assessment using the good model of Viterbi algorithm combined training, it is assumed that the state of corresponding maximum probability is
State 5 is obtained given the corresponding optimal path of observation state (2,4 ..., 1,5) using retrogressive method, and solid line is shown in figure
Path is from the optimal path for starting state to end-state.
Fig. 8 is regular schematic diagram of the sterile milk quality that provides of the embodiment of the present invention one with time change.Such as Fig. 8 institute
Show, hidden Markov model is dynamic evaluation algorithm, can be assessed with the temporal characteristics of combined data data set, is analyzed
The time response of sterile milk data, high quality batch products proportion is with Seasonal fluctuation.Comprehensive fourth quarter in 2013
Data to fourth quarter in 2015 show, the second quarter in 2015, the third season in 2015, the second quarter in 2014 institute
The reason of accounting example is significantly lower than other seasons, generates this phenomenon may be high for summer temp, is unfavorable for the milk product that sterilizes
Storage, while being easy to produce the problems such as raw milk is rotten.The quality of sterilizing milk product indicated above changes with time with bright
Aobvious difference, using hidden Markov model, this dynamic evaluation algorithm can be very good binding time characteristic to sterile milk
Quality is assessed and is predicted.
Trained hidden Markov model is applied to whole sterile milk detection datas to its quality safety by the present embodiment
It is assessed, in conjunction with the time response of sterile milk, the quality safety risk evaluation result for obtaining sterile milk changes over time rule
Rule.Fig. 9 is the hidden Markov model risk evaluation result schematic diagram that the embodiment of the present invention one provides.As shown in figure 9, cylindricality
Figure be hidden Markov model be applied to the risk assessment of sterile milk detection data as a result, line chart is the input number of the model
According to, i.e., sterile milk quality change over time as a result, two groups of comparison as a result, obtained risk trend is almost the same, 2014 years the
For the second quarter, the quality safety greatest risk of the second quarter in 2015, the third season in 2015.
Pass through the risk assessment to sterile milk quality safety, in conjunction with the time response of sterile milk quality, available raising
The method of sterile milk quality.First, strict control sterilize milk product production link quality safety, summer due to temperature compared with
Height is easy foster microbial, generates threat to the quality safety of product.Therefore, manufacturing enterprise need to be stringent since being checked and accepted former milk
It checks on, guarantees the whole process quality control that raw material dispatches from the factory into factory to finished product, guarantee Product quality and safety.Second, it is transporting and is selling
Link popularization is sold, the danger of microorganism can be reduced by control measure such as temperature, time and working specifications to links
Evil determines that a critical limitation index as control standard, guarantees each critical control point limitation to each critical control point
In safe range.
By above-mentioned experiment it is found that can effectively handle nonlinear data by grey correlation analysis, each finger is obtained
Weight is marked, and then carries out model training and risk dynamic evaluation and prediction using hidden Markov model, analysis generates quality and asks
The reason of topic, while obtaining improving the direction of food quality.
Hidden Markov model food quality appraisal procedure provided in this embodiment based on grey correlation analysis includes:
The corresponding index weights of the influence factor are obtained according to grey correlation analysis algorithm and each influence factor, according to each finger
Mark risk index corresponding with the index weights acquisition index, using the risk index as hidden Markov model
Input data the hidden Markov model is trained, using hidden Markov model after training to food matter
Amount is assessed and is predicted.Technical solution provided in this embodiment realizes the risk assessment to food quality more accurately,
To help to improve the accuracy and efficiency of food quality assessment.Technical solution provided in this embodiment is realized to food
The assessment and prediction of quality can be analyzed with seriousness the reason of generation, formulate or adjust risk control in advance
Measure processed is examined to food production and administrative department advises.
It is understood that the principle that embodiment of above is intended to be merely illustrative of the present and the exemplary reality that uses
Mode is applied, however the present invention is not limited thereto.For those skilled in the art, the present invention is not being departed from
Spirit and essence in the case where, various changes and modifications can be made therein, these variations and modifications are also considered as protection of the invention
Range.
Claims (5)
1. a kind of hidden Markov model food quality appraisal procedure based on grey correlation analysis characterized by comprising
It obtains reference vector and compares vector, the reference vector x0With the relatively vector xiIt is respectively as follows:
x0={ x0(1), x0(2) ..., x0(n)}；
xi={ xi(1), xi(2) ..., xi(n)}；
Wherein, n is number of samples, and i=0,1,2 ..., m-1, m are the number of all indexs；
Each index is normalized, normalized processing formula is as follows:
Obtain yi(k) and y0(k) in the grey incidence coefficient at k moment, the grey incidence coefficient are as follows:
Wherein, (0,1) ρ ∈；
Sequences y is obtained according to the grey incidence coefficient0With sequences yiBetween related coefficient, the related coefficient are as follows:
The correlation matrix of all indexs, the correlation matrix are obtained according to the related coefficient are as follows:
The corresponding index weights of each index are obtained according to the correlation matrix；
The corresponding risk index of the index is obtained according to each index and the index weights；
The hidden Markov model is trained using the risk index as the input data of hidden Markov model；
Food quality is assessed and predicted using hidden Markov model after training.
2. the hidden Markov model food quality appraisal procedure according to claim 1 based on grey correlation analysis,
It is characterized in that, described the step of obtaining the index corresponding risk index according to each index and the index weights includes:
Local state is initialized according to viterbi algorithm are as follows:
δ1(i)=πibi(o1), i=1,2 ..., N
Dynamic Programming recursion moment t=2,3 is carried out ..., the local state at T moment are as follows:
Maximum value, hidden state are obtained at the T moment are as follows:
Corresponding risk index is obtained according to the hidden state and the index weights.
3. the hidden Markov model food quality appraisal procedure according to claim 2 based on grey correlation analysis,
It is characterized in that, includes: before described the step of obtaining corresponding risk index according to the hidden state and the index weights
δT(i) maximum value, the probability that hidden state sequence occurs are obtained at the T moment are as follows:
4. the hidden Markov model food quality appraisal procedure according to claim 2 based on grey correlation analysis,
It is characterized in that, further includes:
Use local stateRecalled, for t=T-1, T-2 ..., 1, corresponding hidden state sequence are as follows:
Optimal path, the optimal path are obtained according to the hidden state sequence are as follows:
I*=(i1 *, i2 *..., iT *}。
5. the hidden Markov model food quality appraisal procedure according to claim 1 based on grey correlation analysis,
It is characterized in that, described the step of obtaining each index corresponding index weights according to the correlation matrix includes:
The average value of each index is obtained according to each index among the related coefficient square；
Using the average value of each index as the corresponding index weights of each index.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910330705.2A CN110428121B (en) | 2019-04-23 | 2019-04-23 | Hidden Markov model food quality assessment method based on gray correlation analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910330705.2A CN110428121B (en) | 2019-04-23 | 2019-04-23 | Hidden Markov model food quality assessment method based on gray correlation analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110428121A true CN110428121A (en) | 2019-11-08 |
| CN110428121B CN110428121B (en) | 2024-02-23 |
Family
ID=68408403
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910330705.2A Active CN110428121B (en) | 2019-04-23 | 2019-04-23 | Hidden Markov model food quality assessment method based on gray correlation analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110428121B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111291967A (en) * | 2020-01-13 | 2020-06-16 | 北京化工大学 | Hidden Markov model food safety risk assessment method based on cuckoo search |
| CN111626749A (en) * | 2020-05-15 | 2020-09-04 | 贵州理工学院 | Food quality safety tracing and safety level evaluation system |
| CN111915206A (en) * | 2020-08-11 | 2020-11-10 | 成都市食品药品检验研究院 | Method for recognizing food risk conduction |
| CN114087940A (en) * | 2021-11-18 | 2022-02-25 | 得力集团有限公司 | Use method of multifunctional vernier caliper |
| CN117084683A (en) * | 2023-08-07 | 2023-11-21 | 中国人民解放军32302部队 | Method and system for monitoring and evaluating psychological states of technicians |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050050101A1 (en) * | 2003-01-23 | 2005-03-03 | Vockley Joseph G. | Identification and use of informative sequences |
| CN104063599A (en) * | 2014-06-26 | 2014-09-24 | 江苏中烟工业有限责任公司 | Index screening and processing method for evaluating quality of tobacco leaves |
| CN105608536A (en) * | 2015-12-22 | 2016-05-25 | 天津科技大学 | Food safety risk prediction method based on hidden Markov model |
| CN106372799A (en) * | 2016-08-31 | 2017-02-01 | 全球能源互联网研究院 | Power grid security risk prediction method |
| CN108329989A (en) * | 2017-11-08 | 2018-07-27 | 内蒙古师范大学 | A kind of application of desert shrub mongolian amygdalus seed grease as raw-food material |
| CN108876100A (en) * | 2018-04-28 | 2018-11-23 | 北京化工大学 | Neural network food safety risk prediction model based on ISM and AHP |
| CN109377077A (en) * | 2018-11-09 | 2019-02-22 | 贵州希望泥腿信息技术有限公司 | A kind of agricultural product quality and safety evaluation method |
| CN109447459A (en) * | 2018-10-26 | 2019-03-08 | 国网四川省电力公司达州供电公司 | A kind of energy quality comprehensive assessment method considering index's quality weighing factor |
-
2019
- 2019-04-23 CN CN201910330705.2A patent/CN110428121B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050050101A1 (en) * | 2003-01-23 | 2005-03-03 | Vockley Joseph G. | Identification and use of informative sequences |
| CN104063599A (en) * | 2014-06-26 | 2014-09-24 | 江苏中烟工业有限责任公司 | Index screening and processing method for evaluating quality of tobacco leaves |
| CN105608536A (en) * | 2015-12-22 | 2016-05-25 | 天津科技大学 | Food safety risk prediction method based on hidden Markov model |
| CN106372799A (en) * | 2016-08-31 | 2017-02-01 | 全球能源互联网研究院 | Power grid security risk prediction method |
| CN108329989A (en) * | 2017-11-08 | 2018-07-27 | 内蒙古师范大学 | A kind of application of desert shrub mongolian amygdalus seed grease as raw-food material |
| CN108876100A (en) * | 2018-04-28 | 2018-11-23 | 北京化工大学 | Neural network food safety risk prediction model based on ISM and AHP |
| CN109447459A (en) * | 2018-10-26 | 2019-03-08 | 国网四川省电力公司达州供电公司 | A kind of energy quality comprehensive assessment method considering index's quality weighing factor |
| CN109377077A (en) * | 2018-11-09 | 2019-02-22 | 贵州希望泥腿信息技术有限公司 | A kind of agricultural product quality and safety evaluation method |
Non-Patent Citations (10)
| Title |
|---|
| LAWRENCE R. RABINER: "A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition", 《PROCEEDINGS OF THE IEEE》 * |
| LAWRENCE R. RABINER: "A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition", 《PROCEEDINGS OF THE IEEE》, vol. 77, no. 2, 2 February 1989 (1989-02-02), pages 257 - 286, XP002550447, DOI: 10.1109/5.18626 * |
| 刘宇姗: "食品安全风险评估与预警方法研究", 《中国优秀硕士学位论文全文数据库 (工程科技Ⅰ辑)》 * |
| 刘宇姗: "食品安全风险评估与预警方法研究", 《中国优秀硕士学位论文全文数据库 (工程科技Ⅰ辑)》, no. 7, 15 July 2017 (2017-07-15), pages 024 - 108 * |
| 周荣喜等: "基于ARMA与BP神经网络模型的产品质量安全风险预测", 《北京化工大学学报(自然科学版)》 * |
| 周荣喜等: "基于ARMA与BP神经网络模型的产品质量安全风险预测", 《北京化工大学学报(自然科学版)》, no. 06, 20 November 2015 (2015-11-20), pages 115 - 118 * |
| 张桦: "配电系统可靠性预测模型及其算法研究", 《中国博士学位论文全文数据库 (工程科技Ⅱ辑)》 * |
| 张桦: "配电系统可靠性预测模型及其算法研究", 《中国博士学位论文全文数据库 (工程科技Ⅱ辑)》, no. 3, 15 March 2017 (2017-03-15), pages 042 - 32 * |
| 马永军等: "基于HMM的食品安全风险预测方法研究", 《天津科技大学学报》 * |
| 马永军等: "基于HMM的食品安全风险预测方法研究", 《天津科技大学学报》, no. 05, 25 October 2016 (2016-10-25), pages 69 - 73 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111291967A (en) * | 2020-01-13 | 2020-06-16 | 北京化工大学 | Hidden Markov model food safety risk assessment method based on cuckoo search |
| CN111626749A (en) * | 2020-05-15 | 2020-09-04 | 贵州理工学院 | Food quality safety tracing and safety level evaluation system |
| CN111915206A (en) * | 2020-08-11 | 2020-11-10 | 成都市食品药品检验研究院 | Method for recognizing food risk conduction |
| CN111915206B (en) * | 2020-08-11 | 2024-02-27 | 成都市食品药品检验研究院 | Method for identifying food risk conduction |
| CN114087940A (en) * | 2021-11-18 | 2022-02-25 | 得力集团有限公司 | Use method of multifunctional vernier caliper |
| CN114087940B (en) * | 2021-11-18 | 2024-05-31 | 得力集团有限公司 | Use method of multifunctional vernier caliper |
| CN117084683A (en) * | 2023-08-07 | 2023-11-21 | 中国人民解放军32302部队 | Method and system for monitoring and evaluating psychological states of technicians |
| CN117084683B (en) * | 2023-08-07 | 2024-04-19 | 中国人民解放军32302部队 | Method and system for monitoring and evaluating psychological states of technicians |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110428121B (en) | 2024-02-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110428121A (en) | Hidden Markov model food quality appraisal procedure based on grey correlation analysis | |
| Wang et al. | Comparison of variable selection methods for PLS-based soft sensor modeling | |
| Bhatt et al. | Automatic apple grading model development based on back propagation neural network and machine vision, and its performance evaluation | |
| Al-Amery et al. | Near-infrared spectroscopy used to predict soybean seed germination and vigour | |
| Singh et al. | Prediction of sensory quality of UHT milk–a comparison of kinetic and neural network approaches | |
| CN104298868A (en) | Chilled meat shelf life forecasting method and system for cold-chain logistics | |
| CN113567439A (en) | Pork freshness detection method based on color and smell data fusion | |
| WO2021114320A1 (en) | Wastewater treatment process fault monitoring method using oica-rnn fusion model | |
| CN105608536A (en) | Food safety risk prediction method based on hidden Markov model | |
| CN108876100A (en) | Neural network food safety risk prediction model based on ISM and AHP | |
| Maliepaard et al. | Bayesian versus frequentist analysis of multiple quantitative trait loci with an application to an outbred apple cross | |
| CN115034303A (en) | Directional detection method and system for harmful substances in food | |
| US11340160B2 (en) | Method for near-infrared spectral wavelength selection based on improved team progress algorithm | |
| WO2023077569A1 (en) | Deep learning-based method for updating spectral analysis model for fruit | |
| CN108830492A (en) | A method of sampling observation businessman is determined based on big data | |
| CN110705856A (en) | Nuclear power equipment forging material manufacturing quality evaluation method based on analytic hierarchy process | |
| Cheng et al. | Evaluation of the impact of skewness, clustering, and probe sampling plan on aflatoxin detection in corn | |
| Garland‐Campbell | Errors in statistical decision making | |
| Sezey et al. | Validation of mohr titration method to determine salt in olive and olive brine | |
| CN116258087B (en) | Matte grade soft measurement method and device, electronic equipment and storage medium | |
| CN112651173A (en) | Agricultural product quality nondestructive testing method based on cross-domain spectral information and generalizable system | |
| CN109255724B (en) | Pesticide residue pollution evaluation method based on multifactor and AHP-E model | |
| Gao et al. | Log transformation and the effect on estimation, implication, and interpretation of mean and measurement uncertainty in microbial enumeration | |
| CN116776252A (en) | Industrial process soft measurement method and system for improving Mallow's Cp variable selection | |
| Wang et al. | Comparison of different variable selection methods for partial least squares soft sensor development |
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 |