WO2021232588A1 - 食品安全风险评估方法、装置、设备及存储介质 - Google Patents

食品安全风险评估方法、装置、设备及存储介质 Download PDF

Info

Publication number
WO2021232588A1
WO2021232588A1 PCT/CN2020/105061 CN2020105061W WO2021232588A1 WO 2021232588 A1 WO2021232588 A1 WO 2021232588A1 CN 2020105061 W CN2020105061 W CN 2020105061W WO 2021232588 A1 WO2021232588 A1 WO 2021232588A1
Authority
WO
WIPO (PCT)
Prior art keywords
food safety
risk
characteristic data
preset
risk assessment
Prior art date
Application number
PCT/CN2020/105061
Other languages
English (en)
French (fr)
Inventor
冷新波
刘佳
Original Assignee
平安国际智慧城市科技股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 平安国际智慧城市科技股份有限公司 filed Critical 平安国际智慧城市科技股份有限公司
Publication of WO2021232588A1 publication Critical patent/WO2021232588A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0635—Risk analysis of enterprise or organisation activities
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/24323—Tree-organised classifiers
    • 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
    • G06Q50/12—Hotels or restaurants

Definitions

  • This application relates to the field of machine learning technology, in particular to food safety risk assessment methods, devices, equipment and storage media.
  • the main purpose of this application is to propose a food safety risk assessment method, device, equipment and storage medium, aiming to improve the efficiency and accuracy of food safety risk assessment.
  • the first aspect of this application provides a food safety risk assessment method.
  • the food safety risk assessment method includes:
  • Generate a training set according to the positive sample and the negative sample input the training set into a preset decision tree model for training, and obtain a target decision tree;
  • the second aspect of the present application provides a food safety risk assessment device, the food safety risk assessment device comprising: a memory, a processor, and a food safety risk assessment program stored in the memory and running on the processor When the processor executes the food safety risk assessment program, the following steps are implemented:
  • Generate a training set according to the positive sample and the negative sample input the training set into a preset decision tree model for training, and obtain a target decision tree;
  • the third aspect of the present application provides a storage medium, a computer-readable storage medium in which computer instructions are stored in the computer-readable storage medium, and when the computer instructions are executed on a computer, the computer executes the following steps:
  • Generate a training set according to the positive sample and the negative sample input the training set into a preset decision tree model for training, and obtain a target decision tree;
  • the fourth aspect of the present application provides a food safety risk assessment device, the food safety risk assessment device comprising:
  • the first acquisition module is used to acquire risk characteristic data of multiple catering companies according to preset food safety risk characteristic items
  • the judgment module is used to judge whether the multiple catering companies have had food safety accidents within a preset period of time, use the risk characteristic data of the catering companies that have had food safety accidents among the multiple catering companies as a positive sample, and use all catering companies as a positive sample. State the risk characteristic data of catering companies that have not had a food safety accident among multiple catering companies as a negative sample;
  • a training module configured to generate a training set according to the positive sample and the negative sample, and input the training set into a preset decision tree model for training to obtain a target decision tree;
  • the second acquisition module is configured to respectively acquire the path from the root node of the target decision tree to each leaf node, and respectively generate a corresponding food safety risk feature combination according to each acquired path;
  • the calculation module is used to calculate the risk assessment value corresponding to each food safety risk feature combination, and mark the risk assessment value on the corresponding leaf node of the target decision tree to obtain a food safety risk assessment model;
  • the risk assessment module is used to obtain the risk characteristic data of the catering company to be assessed, input the risk characteristic data of the catering company to be assessed into the food safety risk assessment model, and use the food safety risk assessment model to analyze the risk characteristics of the catering company to be assessed. Assess catering companies to conduct food safety risk assessments.
  • the food safety risk assessment method proposed in this application first obtains the risk characteristic data of multiple catering companies, generates a training set based on the risk characteristic data and inputs it into the decision tree model for training to obtain the target decision tree, and then obtains the target decision tree according to the target decision tree. Calculate the risk assessment value corresponding to each food safety risk feature combination, and mark the risk assessment value on the corresponding leaf node of the target decision tree to obtain the food safety risk assessment model. Finally, pass the food safety risk assessment model.
  • the risk assessment model conducts food safety risk assessment for the catering companies to be assessed.
  • This method of establishing a food safety risk assessment model based on a decision tree to assess food safety risks avoids problems such as untimely assessment, single assessment data items, and multiple subjective factors. Improve the efficiency and accuracy of food safety risk assessment.
  • Figure 1 is a schematic flow chart of an embodiment of the food safety risk assessment method of this application.
  • FIG. 2 is a schematic diagram of modules of an embodiment of the food safety risk assessment device of the present application.
  • Figure 3 is a schematic structural diagram of a food safety risk assessment device provided by an embodiment of the application.
  • the embodiments of the present application provide a food safety risk assessment method, device, equipment and storage medium.
  • the method of building a food safety risk assessment model based on a decision tree to assess food safety risks is compared with the artificial methods in the prior art. Evaluation avoids problems such as untimely evaluation, single evaluation data items, and multiple subjective factors, and improves the efficiency and accuracy of food safety risk evaluation.
  • Fig. 1 is a schematic flowchart of an embodiment of the food safety risk assessment method of this application, and the method includes:
  • Step 101 Obtain risk characteristic data of multiple catering companies according to preset food safety risk characteristic items
  • the execution subject of this application may be a food safety risk assessment device, or a terminal or a server, which is not specifically limited here.
  • the embodiment of the present application takes the server as the execution subject as an example for description.
  • Qualification permission category such as company name, company type, business address, legal representative information, unified social credit code, business license information, annual turnover, number of employees, maximum number of diners, business area, whether to sell self-made wine, whether to sell Seafood, whether they sell cold meat and cold dishes, whether they are restaurants around the school, etc.;
  • Administrative supervision categories such as the number of abnormal video inspections in the past six months, the number of overdue rectifications in the past six months of smart patrol inspections, the number of failed smart patrols and rectifications in the past six months, the number of abnormal work clothes and hats, the number of abnormal inspections of potentially contaminated food, trash cans The number of abnormal inspections that are not stamped or overflowed, the number of rats found in the kitchen in the past six months, the number of non-compliance inspections, disinfection and disinfection, the most recent quantitative inspection conclusion, the number of abnormal inspections for unlicensed operations, etc.;
  • Operational risks such as whether it is currently a blacklist of commercial entities, whether it is currently a list of abnormal operations, whether it is currently in arrears with utility bills, the total number of asset mortgages, and the cumulative number of government debts in arrears;
  • Social responsibility categories such as the cumulative number of tax payments, whether to invest abroad, whether to guarantee externally, whether employees pay social insurance, the proportion of employees paying social insurance, etc.
  • the assessor can flexibly select multiple food safety risk feature items from the above-mentioned influencing factors in advance, and the server obtains risk feature data of multiple catering companies according to the preset food safety risk feature items. For example, when the pre-set food safety risk feature items include the operating area and the last quantitative inspection conclusion, it is necessary to obtain the operating area data of multiple catering companies and the last quantitative inspection conclusion data. These acquired data are the catering company Risk characteristic data.
  • the above step 101 may specifically include: extracting historical operating data of multiple catering companies from a preset database, and performing data cleaning on the historical operating data to obtain a basic corporate profile data set; for each catering company in the corporate basic profile data set, from The risk feature data corresponding to the preset food safety risk feature items are extracted from the historical business data after data cleaning.
  • the server may first extract historical operating data of multiple catering companies from a preset database, where the preset database may include a market supervision bureau registration license database, an administrative law enforcement database, a comprehensive supervision database, an enterprise annual report database, Commercial credit database and random inspection database, etc.
  • historical operating data can include qualification license, administrative supervision, business risk, and social responsibility data generated by catering companies in the historical operation process; after extracting the history of multiple catering companies
  • the server performs data cleaning on these historical operating data to obtain the basic image data set of the enterprise.
  • the data cleaning includes deleting incomplete data, incomplete data, and erroneous data. Through data cleaning, data quality can be improved;
  • the server extracts the risk feature data corresponding to the preset food safety risk feature items from the historical business data after data cleaning, so as to obtain the risk feature data of each catering company.
  • the risk characteristic data of the catering enterprise is extracted from multiple dimensions, and the food safety risk assessment model is established through the risk characteristic data, which is beneficial to improve the accuracy of the food safety risk assessment.
  • the blockchain referred to in this application is a new application mode of computer technology such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm.
  • Blockchain essentially a decentralized database, is a series of data blocks associated with cryptographic methods. Each data block contains a batch of network transaction information for verification. The validity of the information (anti-counterfeiting) and the generation of the next block.
  • the blockchain can include the underlying platform of the blockchain, the platform product service layer, and the application service layer.
  • Step 102 Determine whether multiple catering companies have had food safety incidents within a preset period of time, take the risk characteristic data of catering companies that have had food safety incidents in multiple catering companies as a positive sample, and take the risk characteristics of catering companies that have not occurred in multiple catering companies The risk characteristic data of catering companies that have experienced food safety incidents are used as negative samples;
  • the server judges whether food safety incidents have occurred in the above-mentioned multiple catering companies within a preset time period.
  • the preset time period can be flexibly set, such as 1 year, half a year, etc.; the server will determine whether food safety incidents have occurred in multiple catering companies.
  • the risk characteristic data of catering companies are taken as positive samples, and the risk characteristic data of catering companies that have not had food safety accidents among multiple catering companies are taken as negative samples.
  • the step of judging whether multiple catering companies have had food safety incidents within a preset time period may specifically include: obtaining food safety report information of multiple catering companies within a preset time period from a preset media database; Food safety report information determines whether multiple catering companies have experienced food poisoning incidents, whether there are judicial or administrative punishment records, and whether they are included in the list of business abnormalities; when a food poisoning incident has occurred in a catering company, or there is a judicial penalty or When the administrative penalty record is included in the list of business abnormalities, it is determined that the catering company has had a food safety incident within a preset period of time.
  • the server may first obtain food safety report information of multiple catering companies within a preset time period from a preset media database, where the preset media database includes, but is not limited to, databases such as news websites, Weibo, corporate websites, and catering official accounts; Then, the server can perform keyword identification on the obtained food safety report information to determine whether food poisoning incidents have occurred in various catering companies, whether there are judicial or administrative punishment records, and whether they are included in the list of business abnormalities, such as Dang When the “food poisoning” keyword exists in the food safety report information of a catering company, it can be determined that the catering company has had a food poisoning incident.
  • the preset media database includes, but is not limited to, databases such as news websites, Weibo, corporate websites, and catering official accounts.
  • the server determines that the catering company Food safety incidents have occurred within a period of time.
  • Step 103 Generate a training set according to the positive sample and the negative sample, and input the training set into a preset decision tree model for training to obtain a target decision tree;
  • the server After determining the positive sample and the negative sample, the server generates a training set based on the positive sample and the negative sample. Specifically, in one embodiment, the server may directly merge the positive sample and the negative sample determined above and use it as the training set input In another embodiment, in order to improve the accuracy and efficiency of model training, the server can also further filter the risk characteristic data contained in the positive and negative samples to obtain Risk characteristic data with predictive value.
  • the risk characteristic data with predictive value refers to the risk characteristic data that is strongly related to the risk assessment result, that is, it can determine to a large extent whether a food safety incident occurs in a catering company, for example, nearly half a year later The number of times the kitchen found rats, the conclusion of the last quantitative inspection, etc.
  • the server After obtaining the training set, the server inputs the training set into the preset decision tree model for training, and obtains the target decision tree.
  • Decision tree is a method of machine learning.
  • each node represents a judgment on an attribute
  • each branch represents the output of a judgment result
  • each leaf node represents a classification result.
  • the algorithm for generating the target decision tree can use the existing C4.5 or C5.0 algorithm, which will not be repeated here.
  • Step 104 Obtain the path from the root node of the target decision tree to each leaf node, and generate a corresponding food safety risk feature combination according to each obtained path;
  • the server separately obtains the path from the root node of the target decision tree to each leaf node, and generates a corresponding food safety risk feature combination according to the risk feature data corresponding to the node on each path.
  • the path from the root node of the target decision tree to a certain leaf node is: the number of rats found in the kitchen in the past six months is greater than or equal to 2 ⁇ the last quantitative inspection concluded that it is qualified ⁇ the operating area is greater than 200 square meters ⁇ the rectification passed in the past month ,
  • the food safety risk feature combination corresponding to this path is “the number of rats found in the kitchen after the past six months is greater than or equal to 2, the last quantitative inspection concluded that it is qualified, the operating area is greater than 200 square meters, and the rectification passed in the past month”.
  • Step 105 Calculate the risk assessment value corresponding to each food safety risk feature combination, and mark the risk assessment value on the corresponding leaf node of the target decision tree to obtain a food safety risk assessment model;
  • the server further calculates the risk assessment value corresponding to each food safety risk feature combination, and marks the risk assessment value on the corresponding leaf node of the target decision tree to obtain a food safety risk assessment model.
  • the above step of calculating the risk assessment value corresponding to each food safety risk feature combination may specifically include: counting the number of positive samples and the number of negative samples corresponding to each food safety risk feature combination, which will correspond to each food safety risk feature combination Add the number of positive samples and the number of negative samples to get the total number of samples; calculate the ratio of the number of positive samples corresponding to each combination of food safety risk characteristics to the total number of samples, and use the ratio as the risk assessment corresponding to each combination of food safety risk characteristics value.
  • server statistics meet the food safety risk feature combination "The number of times the kitchen has found rats greater than or equal to 2, the last quantitative inspection concluded that it is qualified, the business area is greater than 200 square meters, and the total number of samples passed the rectification in the past month" is 100. Where the number of positive samples is 10 and the number of negative samples is 90, the risk assessment value corresponding to the food safety risk feature combination is 0.1.
  • the above method uses the ratio of the number of positive samples in the food safety risk feature combination to the total number of samples as the risk assessment value corresponding to the food safety risk feature combination, and realizes the objective and accurate evaluation of a food safety risk feature combination based on the number of samples. risk.
  • Step 106 Obtain the risk characteristic data of the catering company to be assessed, input the risk characteristic data of the catering company to be assessed into a food safety risk assessment model, and conduct a food safety risk assessment for the catering company to be assessed through the food safety risk assessment model.
  • the server can extract the risk characteristic data of the catering company to be assessed from the preset database according to the preset food safety risk feature items for the catering company to be assessed for the subsequent food safety risk assessment, where
  • the preset database includes, but is not limited to, the registration and license database of the Market Supervision Bureau, the administrative law enforcement database, the comprehensive supervision database, the enterprise annual report database, the commercial credit database, and the sampling database.
  • the server inputs the risk characteristic data of the catering company to be assessed into the In the food safety risk assessment model, the food safety risk assessment model outputs the combination of food safety risk characteristics and the corresponding risk value to which the catering company to be assessed belongs.
  • the server can also send the food safety risk feature combination and corresponding risk value of the catering company to be assessed to a preset comprehensive supervision system, so that supervisors can take corresponding measures in time for high-risk catering companies, such as On-site inspection or remote inspection processing, etc.; in addition, the inspection results can also be fed back to the server's decision tree training model as a sample to further optimize and improve the accuracy of the model.
  • the food safety risk assessment method proposed in this embodiment first obtains the risk characteristic data of multiple catering companies, generates a training set according to the risk characteristic data and inputs it to the decision tree model for training to obtain the target decision tree, and then obtains the target decision tree according to the target decision tree. Multiple food safety risk feature combinations, calculate the risk assessment value corresponding to each food safety risk feature combination, and mark the risk assessment value on the corresponding leaf node of the target decision tree to obtain a food safety risk assessment model. Finally, pass the food The safety risk assessment model conducts food safety risk assessment for the catering companies to be assessed.
  • This method of establishing a food safety risk assessment model based on a decision tree to assess food safety risks avoids problems such as untimely assessment, single assessment data items, and multiple subjective factors. Improve the efficiency and accuracy of food safety risk assessment.
  • This application can be applied to the field of smart government affairs, thereby promoting the construction of smart cities.
  • the above step 103 may specifically include: screening the target discrete risk feature data with risk prediction value from positive samples and negative samples according to the chi-square test algorithm; according to the weight of evidence algorithm and the information value algorithm, from Target continuous risk feature data with risk prediction value is selected from the positive samples and negative samples; the target discrete risk feature data and target continuous risk feature data are input as the training set into the preset decision tree model for training. Target decision tree.
  • the chi-square test is a commonly used feature selection algorithm.
  • the chi-square test is the degree of deviation between the actual observation value of the statistical sample and the theoretical inferred value.
  • the degree of deviation between the actual observation value and the theoretical inferred value determines the chi-square
  • the server can use the chi-square test algorithm to filter out the target discrete risk characteristic data with risk prediction value.
  • Weight of evidence algorithm (WOE) and information value algorithm (information value, IV) are algorithms used to measure the predictive ability of a variable. The larger the value of WOE and IV, the stronger the predictive ability of this variable.
  • WOE weight of evidence algorithm
  • information value algorithm information value, IV
  • the server can select the target continuous risk characteristic data with risk prediction value through the evidence weight algorithm and the information value algorithm.
  • the step of screening the target discrete risk characteristic data with risk prediction value from the positive sample and the negative sample may include: for the discrete risk characteristic data in the positive sample and the negative sample, statistics and The number of positive samples and the number of negative samples corresponding to the discrete risk feature data; the number of positive samples, the number of negative samples corresponding to the discrete risk feature data, and the theoretical inferred value of the sample number corresponding to the discrete risk feature data are substituted into the chi-square test algorithm for calculation, The chi-square value of the discrete risk feature data is obtained; when the chi-square value is less than or equal to the first preset threshold, the discrete risk feature data is used as the target discrete risk feature data with risk prediction value.
  • the server first counts the number of positive samples and the number of negative samples corresponding to the discrete risk characteristic data, and then calculates the number of positive samples corresponding to the discrete risk characteristic data.
  • the number of negative samples and the theoretical inferred value of the preset sample number are substituted into the chi-square test algorithm for calculation to obtain the chi-square value of the discrete risk characteristic data.
  • the formula of the chi-square test algorithm is:
  • X 2 is the chi-square value
  • observed is the observed value of the number of samples
  • expected is the theoretical inferred value of the number of samples.
  • the server determines whether the calculated chi-square value is less than or equal to the first preset threshold, and if so, uses the discrete risk characteristic data as target discrete risk characteristic data with risk prediction value.
  • the step of screening the target continuous risk characteristic data with risk prediction value from the positive sample and the negative sample may include: for the continuous risk characteristic data in the positive sample and the negative sample , Divide the continuous risk characteristic data into segments; calculate the evidence weight corresponding to each segment according to the evidence weight algorithm, and calculate the information value corresponding to each segment according to the calculated evidence weight and information value algorithm; The information value corresponding to each segment is summed to obtain the information value corresponding to the continuous risk characteristic data; when the information value corresponding to the continuous risk characteristic data is greater than or equal to the second preset threshold, the continuous risk Feature data is used as target continuous risk feature data with risk prediction value.
  • the server inputs the target discrete risk characteristic data and target continuous risk characteristic data as a training set into the preset decision tree model for training, and obtains the target decision tree, thereby improving the accuracy and efficiency of model training.
  • the embodiment of the application also provides a food safety risk assessment device.
  • Fig. 2 is a schematic diagram of a module of an embodiment of the food safety risk assessment device of the present application.
  • the food safety risk assessment device includes:
  • the first acquisition module 201 is configured to acquire risk characteristic data of multiple catering companies according to preset food safety risk characteristic items;
  • the judging module 202 is used to judge whether the multiple catering companies have had food safety accidents within a preset time period, taking the risk characteristic data of the catering companies that have had food safety accidents among the multiple catering companies as a positive sample, and taking The risk characteristic data of the catering companies that have not had a food safety accident among the plurality of catering companies are taken as a negative sample;
  • the training module 203 is configured to generate a training set according to the positive sample and the negative sample, and input the training set into a preset decision tree model for training, to obtain a target decision tree;
  • the second obtaining module 204 is configured to obtain the path from the root node of the target decision tree to each leaf node, and respectively generate a corresponding food safety risk feature combination according to each obtained path;
  • the calculation module 205 is configured to calculate a risk assessment value corresponding to each food safety risk feature combination, and mark the risk assessment value on the corresponding leaf node of the target decision tree to obtain a food safety risk assessment model;
  • the risk assessment module 206 is configured to obtain the risk characteristic data of the catering company to be assessed, input the risk characteristic data of the catering company to be assessed into the food safety risk assessment model, and compare the risk characteristics of the catering company to the food safety risk assessment model.
  • the catering companies to be assessed conduct food safety risk assessments.
  • the first obtaining module 201 is further configured to:
  • the risk characteristic data corresponding to the preset food safety risk characteristic items are extracted from the historical operating data after data cleaning.
  • the judgment module 202 is further configured to:
  • the food safety report information determine whether the multiple catering companies have experienced food poisoning incidents, whether there are judicial or administrative punishment records, and whether they are included in the list of business abnormalities;
  • the training module 203 is also used to:
  • the target discrete risk characteristic data and the target continuous risk characteristic data are input into a preset decision tree model as a training set for training, to obtain a target decision tree.
  • the training module 203 is also used to:
  • the discrete risk characteristic data is used as target discrete risk characteristic data with risk prediction value.
  • the training module 203 is also used to:
  • the evidence weight algorithm calculate the evidence weight corresponding to each segment, and calculate the information value corresponding to each segment according to the calculated evidence weight and information value algorithm;
  • the continuous risk characteristic data is used as target continuous risk characteristic data with risk prediction value.
  • calculation module 205 is further configured to:
  • the food safety risk assessment device in the embodiment of the present application is described in detail above from the perspective of modular functional entity, and the food safety risk assessment device in the embodiment of the present application is described in detail from the perspective of hardware processing.
  • Fig. 3 is a schematic structural diagram of a food safety risk assessment device provided by an embodiment of the application.
  • the food safety risk assessment device 300 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 310 (for example, one or more processors) and a memory 320.
  • processors central processing units, CPU
  • One or more storage media 330 for storing application programs 333 or data 332 for example, one or one storage device with a large amount of storage).
  • the memory 320 and the storage medium 330 may be short-term storage or persistent storage.
  • the program stored in the storage medium 330 may include one or more modules (not shown in the figure), and each module may include a series of command operations in the food safety risk assessment device 300.
  • the processor 310 may be configured to communicate with the storage medium 330 and execute a series of instruction operations in the storage medium 330 on the food safety risk assessment device 300.
  • the food safety risk assessment device 300 may also include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input and output interfaces 360, and/or one or more operating systems 331, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc.
  • FIG. 3 does not constitute a limitation on the food safety risk assessment equipment, and may include more or less components than shown in the figure, or combine certain components, or Different component arrangements.
  • the present application also provides a storage medium, which may be a non-volatile storage medium or a volatile storage medium, the storage medium stores a food safety risk assessment program, and the food safety risk assessment program When executed by the processor, the steps of the food safety risk assessment method as described above are realized.
  • the computer-readable storage medium may mainly include a storage program area and a storage data area, where the storage program area may store an operating system, an application program required by at least one function, etc.; the storage data area may store a block chain node Use the created data, etc.
  • the blockchain referred to in this application is a new application mode of computer technology such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm.
  • Blockchain essentially a decentralized database, is a series of data blocks associated with cryptographic methods. Each data block contains a batch of network transaction information for verification. The validity of the information (anti-counterfeiting) and the generation of the next block.
  • the blockchain can include the underlying platform of the blockchain, the platform product service layer, and the application service layer.
  • the technical solution of the present application essentially or the part that contributes to the existing technology or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium , Including several instructions to make a computer device (which may be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the methods described in the various embodiments of the present application.
  • the aforementioned storage media include: U disk, mobile hard disk, read-only memory (read-only memory, ROM), random access memory (random access memory, RAM), magnetic disks or optical disks and other media that can store program codes. .

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Tourism & Hospitality (AREA)
  • General Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Business, Economics & Management (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Biology (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Development Economics (AREA)
  • Educational Administration (AREA)
  • Primary Health Care (AREA)
  • Game Theory and Decision Science (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Medical Treatment And Welfare Office Work (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种食品安全风险评估方法、装置、设备及存储介质，用于提高食品安全风险评估的效率和准确性。该方法包括：获取多个餐饮企业的风险特征数据（101）；根据风险特征数据确定正样本和负样本（102）；根据正样本和负样本生成训练集，将训练集输入至预设的决策树模型中进行训练，得到目标决策树（103）；分别获取目标决策树的根节点到每个叶子节点的路径，根据每一条路径分别生成一个对应的食品安全风险特征组合（104），并计算对应的风险评估值，将风险评估值标记在目标决策树的对应叶子节点上，得到食品安全风险评估模型（105），对待评估餐饮企业进行食品安全风险评估（106）。

Description

食品安全风险评估方法、装置、设备及存储介质
本申请要求于2020年5月21日提交中国专利局、申请号为202010433430.8、发明名称为“食品安全风险评估方法、装置、设备及存储介质”的中国专利申请的优先权，其全部内容通过引用结合在申请中。
技术领域
本申请涉及机器学习技术领域，尤其涉及食品安全风险评估方法、装置、设备及存储介质。
背景技术
食品安全问题一直是社会高度重视的问题。近年来，餐饮企业的数量多、地域分散、更新迭代快与食品***人员少的矛盾日益突出，在各区域风险程度不明确的前提下，依旧只能凭借经验或其他随机条件选择部分监管对象或区域进行监管，比如对每个餐饮企业每年至少检查一次，这种方式无法及时发现风险，很容易造成风险失控和有限监管力量的无效使用。
随着食品***理念从事后查处向事前预防转变，食品安全风险评估工作的重要性越发突显出来。现有的食品安全风险评估方法，一般是由监管人员现场评估，发明人意识到这种评估方式效率低下，且各项评估指标对食品安全风险的影响程度依赖于人为判断，无法保证风险评估的客观性和准确性。
发明内容
本申请的主要目的在于提出一种食品安全风险评估方法、装置、设备及存储介质，旨在提高食品安全风险评估的效率和准确性。
本申请第一方面提供了一种食品安全风险评估方法，所述食品安全风险评估方法包括：
根据预设的食品安全风险特征项，获取多个餐饮企业的风险特征数据；
判断所述多个餐饮企业在预设时长内是否发生过食品安全事故，将所述多个餐饮企业中发生过食品安全事故的餐饮企业的风险特征数据作为正样本，将所述多个餐饮企业中未发生过食品安全事故的餐饮企业的风险特征数据作为负样本；
根据所述正样本和所述负样本生成训练集，将所述训练集输入至预设的决策树模型中进行训练，得到目标决策树；
分别获取所述目标决策树的根节点到每个叶子节点的路径，根据获取到的每一条路径分别生成一个对应的食品安全风险特征组合；
计算与每个食品安全风险特征组合对应的风险评估值，将所述风险评估值标记在所述目标决策树的对应叶子节点上，得到食品安全风险评估模型；
获取待评估餐饮企业的风险特征数据，将所述待评估餐饮企业的风险特征数据输入至所述食品安全风险评估模型中，通过所述食品安全风险评估模型对所述待评估餐饮企业进行食品安全风险评估。
本申请第二方面提供了一种食品安全风险评估设备，所述食品安全风险评估设备包括：存储器、处理器以及存储在所述存储器上并可在所述处理器上运行的食品安全风险评估程序，所述处理器执行所述食品安全风险评估程序时实现如下步骤：
根据预设的食品安全风险特征项，获取多个餐饮企业的风险特征数据；
判断所述多个餐饮企业在预设时长内是否发生过食品安全事故，将所述多个餐饮企业中发生过食品安全事故的餐饮企业的风险特征数据作为正样本，将所述多个餐饮企业中未发生过食品安全事故的餐饮企业的风险特征数据作为负样本；
根据所述正样本和所述负样本生成训练集，将所述训练集输入至预设的决策树模型中进行训练，得到目标决策树；
分别获取所述目标决策树的根节点到每个叶子节点的路径，根据获取到的每一条路径分别生成一个对应的食品安全风险特征组合；
计算与每个食品安全风险特征组合对应的风险评估值，将所述风险评估值标记在所述目标决策树的对应叶子节点上，得到食品安全风险评估模型；
获取待评估餐饮企业的风险特征数据，将所述待评估餐饮企业的风险特征数据输入至所述食品安全风险评估模型中，通过所述食品安全风险评估模型对所述待评估餐饮企业进行食品安全风险评估。
本申请的第三方面提供了一种存储介质，一种计算机可读存储介质，所述计算机可读存储介质中存储计算机指令，当所述计算机指令在计算机上运行时，使得计算机执行如下步骤：
根据预设的食品安全风险特征项，获取多个餐饮企业的风险特征数据；
判断所述多个餐饮企业在预设时长内是否发生过食品安全事故，将所述多个餐饮企业中发生过食品安全事故的餐饮企业的风险特征数据作为正样本，将所述多个餐饮企业中未发生过食品安全事故的餐饮企业的风险特征数据作为负样本；
根据所述正样本和所述负样本生成训练集，将所述训练集输入至预设的决策树模型中进行训练，得到目标决策树；
分别获取所述目标决策树的根节点到每个叶子节点的路径，根据获取到的每一条路径分别生成一个对应的食品安全风险特征组合；
计算与每个食品安全风险特征组合对应的风险评估值，将所述风险评估值标记在所述目标决策树的对应叶子节点上，得到食品安全风险评估模型；
获取待评估餐饮企业的风险特征数据，将所述待评估餐饮企业的风险特征数据输入至所述食品安全风险评估模型中，通过所述食品安全风险评估模型对所述待评估餐饮企业进行食品安全风险评估。
本申请第四方面提供了一种食品安全风险评估装置，所述食品安全风险评估装置包括：
第一获取模块，用于根据预设的食品安全风险特征项，获取多个餐饮企业的风险特征数据；
判断模块，用于判断所述多个餐饮企业在预设时长内是否发生过食品安全事故，将所述多个餐饮企业中发生过食品安全事故的餐饮企业的风险特征数据作为正样本，将所述多个餐饮企业中未发生过食品安全事故的餐饮企业的风险特征数据作为负样本；
训练模块，用于根据所述正样本和所述负样本生成训练集，将所述训练集输入至预设的决策树模型中进行训练，得到目标决策树；
第二获取模块，用于分别获取所述目标决策树的根节点到每个叶子节点的路径，根据获取到的每一条路径分别生成一个对应的食品安全风险特征组合；
计算模块，用于计算与每个食品安全风险特征组合对应的风险评估值，将所述风险评估值标记在所述目标决策树的对应叶子节点上，得到食品安全风险评估模型；
风险评估模块，用于获取待评估餐饮企业的风险特征数据，将所述待评估餐饮企业的风险特征数据输入至所述食品安全风险评估模型中，通过所述食品安全风险评估模型对所述待评估餐饮企业进行食品安全风险评估。
本申请提出的食品安全风险评估方法，首先获取多个餐饮企业的风险特征数据，根据风险特征数据生成训练集输入至决策树模型中进行训练，得到目标决策树，然后，根据目标决策树获取多个食品安全风险特征组合，计算与每个食品安全风险特征组合对应的风险评估值，将风险评估值标记在目标决策树的对应叶子节点上，得到食品安全风险评估模型，最后，通过该食品安全风险评估模型对待评估餐饮企业进行食品安全风险评估。这种基于 决策树建立食品安全风险评估模型以对食品安全风险进行评估的方式，相比于现有技术中的人为评估，避免了评估不及时、评估数据项单一、主观性因素多等问题，提高了食品安全风险评估的效率和准确性。
附图说明
图1为本申请食品安全风险评估方法的一个实施例的流程示意图；
图2为本申请食品安全风险评估装置的一个实施例的模块示意图；
图3为本申请实施例提供的食品安全风险评估设备的结构示意图。
具体实施方式
本申请实施例提供了一种食品安全风险评估方法、装置、设备及存储介质，通过基于决策树建立食品安全风险评估模型以对食品安全风险进行评估的方式，相比于现有技术中的人为评估，避免了评估不及时、评估数据项单一、主观性因素多等问题，提高了食品安全风险评估的效率和准确性。
本申请的说明书和权利要求书及上述附图中的术语“第一”、“第二”、“第三”、“第四”等(如果存在)是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。应该理解这样使用的数据在适当情况下可以互换，以便这里描述的实施例能够以除了在这里图示或描述的内容以外的顺序实施。此外，术语“包括”或“具有”及其任何变形，意图在于覆盖不排他的包含，例如，包含了一系列步骤或单元的过程、方法、***、产品或设备不必限于清楚地列出的那些步骤或单元，而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备固有的其它步骤或单元。
为便于理解，下面对本申请食品安全风险评估方法实施例的具体流程进行描述。
参照图1，图1为本申请食品安全风险评估方法的一个实施例的流程示意图，该方法包括：
步骤101，根据预设的食品安全风险特征项，获取多个餐饮企业的风险特征数据；
可以理解的是，本申请的执行主体可以为食品安全风险评估装置，还可以是终端或者服务器，具体此处不做限定。本申请实施例以服务器为执行主体为例进行说明。
现实中餐饮企业的食品安全影响因素有很多，具体可以划分为以下四种类型的影响因素：
资质许可类：如企业名称、企业类型、经营地址、法定代表人信息、统一社会信用代码、营业执照信息、年营业额、员工人数、最大就餐人数、经营面积、是否售卖自泡酒、是否售卖海鲜类、是否售卖冷荤凉菜、是否学校周边餐企等；
行政监管类：如近半年视频巡检异常次数、近半年智能巡检逾期未整改次数、近半年智能巡检整改未通过次数、工作衣帽检查异常次数、可能污染食品行为检查异常次数、垃圾桶未加盖或溢出检查异常次数、近半年后厨发现老鼠次数、检查消洗消毒不合规次数、最近一次量化检查结论、无证经营检查异常次数等；
经营风险类：如当前是否为商事主体黑名单、当前是否为经营异常名录、当前是否拖欠水电费、资产抵押总次数、累计拖欠政府债务次数等；
社会责任类：如累计税务缴纳次数、是否对外投资、是否对外担保、员工是否缴纳社保、员工缴纳社保比例等。
在本实施例中，可由评估人员预先从上述影响因素中灵活选取多个作为食品安全风险特征项，服务器根据预先设置的食品安全风险特征项，获取多个餐饮企业的风险特征数据。比如，当预先设置的食品安全风险特征项包括经营面积和最近一次量化检查结论时，需要对应获取多个餐饮企业的经营面积数据和最近一次量化检查结论数据，这些获取到的数据即为餐饮企业的风险特征数据。
上述步骤101具体可以包括：从预设数据库中抽取多个餐饮企业的历史经营数据，对历史经营数据进行数据清洗，得到企业基础画像数据集；对于企业基础画像数据集中的每个餐饮企业，从进行数据清洗后的历史经营数据中提取出与预设的食品安全风险特征项对应的风险特征数据。
在本实施例中，服务器可以首先从预设数据库中抽取多个餐饮企业的历史经营数据，其中，预设数据库可以包括市场监管局登记许可数据库、行政执法数据库、综合监督数据库、企业年报数据库、商事信用数据库和抽检数据库等，历史经营数据可以包括餐饮企业在历史经营过程中所产生的资质许可类、行政监管类、经营风险类和社会责任类数据等；在抽取到多个餐饮企业的历史经营数据后，服务器对这些历史经营数据进行数据清洗，得到企业基础画像数据集，其中数据清洗包括删除不完整数据、残缺数据和错误数据等，通过数据清洗，可以提高数据质量；之后，对于企业基础画像数据集中的每个餐饮企业，服务器从进行数据清洗后的历史经营数据中提取出与预设的食品安全风险特征项对应的风险特征数据，从而得到每个餐饮企业的风险特征数据。
通过上述方式，实现了从多维度提取餐饮企业的风险特征数据，通过这些风险特征数据建立食品安全风险评估模型，有利于提高食品安全风险评估的准确性。
其中，为了避免数据被篡改，可以将餐饮企业的风险特征数据存储于区块链中。本申请所指区块链是分布式数据存储、点对点传输、共识机制、加密算法等计算机技术的新型应用模式。区块链(Blockchain)，本质上是一个去中心化的数据库，是一串使用密码学方法相关联产生的数据块，每一个数据块中包含了一批次网络交易的信息，用于验证其信息的有效性(防伪)和生成下一个区块。区块链可以包括区块链底层平台、平台产品服务层以及应用服务层等。
步骤102，判断多个餐饮企业在预设时长内是否发生过食品安全事故，将多个餐饮企业中发生过食品安全事故的餐饮企业的风险特征数据作为正样本，将多个餐饮企业中未发生过食品安全事故的餐饮企业的风险特征数据作为负样本；
该步骤中，服务器判断上述多个餐饮企业在预设时长内是否发生过食品安全事件，其中预设时长可以灵活设置，比如1年、半年等；服务器将多个餐饮企业中发生过食品安全事故的餐饮企业的风险特征数据作为正样本，将多个餐饮企业中未发生过食品安全事故的餐饮企业的风险特征数据作为负样本。
在一实施方式中，上述判断多个餐饮企业在预设时长内是否发生过食品安全事故的步骤具体可以包括：从预设媒体数据库获取预设时长内多个餐饮企业的食品安全报导信息；根据食品安全报导信息判断多个餐饮企业是否发生过食品中毒事件、是否存在司法处罚或行政处罚记录，以及是否被列入经营异常名录；当某个餐饮企业发生过食品中毒事件，或存在司法处罚或行政处罚记录，或被列入经营异常名录时，判定餐饮企业在预设时长内发生过食品安全事故。
具体地，服务器可以首先从预设媒体数据库获取预设时长内多个餐饮企业的食品安全报导信息，其中预设媒体数据库包括但不限于新闻网站、微博、企业网站、餐饮公众号等数据库；然后，服务器可以对获取到的食品安全报导信息进行关键字识别，以判断各餐饮企业是否发生过食品中毒事件、是否存在司法处罚或行政处罚记录，以及是否被列入经营异常名录，比如当某个餐饮企业的食品安全报导信息中存在“食品中毒”关键字时，可以判定该餐饮企业发生过食品中毒事件，当某个餐饮企业的食品安全报导信息中存在“责令整改”关键字时，可以判定该餐饮企业被给予过司法及行政处罚；当某个餐饮企业发生过食品中毒事件，或存在司法处罚或行政处罚记录，或被列入经营异常名录时，服务器即判定该餐饮企业在预设时长内发生过食品安全事故。
通过上述方式，实现了对餐饮企业在预设时长内是否发生过食品安全事故进行准确、可靠地判断。
步骤103，根据正样本和负样本生成训练集，将训练集输入至预设的决策树模型中进行训练，得到目标决策树；
在确定了正样本和负样本后，服务器根据正样本和负样本生成训练集，具体地，在一实施方式中，服务器可以直接将上述确定的正样本和负样本进行合并后，作为训练集输入至预设的决策树模型中进行训练，在另一实施方式中，为提高模型训练的准确性和效率，服务器也可以对正样本和负样本中包含的风险特征数据做进一步筛选，以获得具有预测价值的风险特征数据，该具有预测价值的风险特征数据指的是与风险评估结果强相关的风险特征数据，即能够在较大程度上决定一个餐饮企业是否发生食品安全事故，比如近半年后厨发现老鼠次数、最近一次量化检查结论等。
在获得训练集后，服务器将训练集输入至预设的决策树模型中进行训练，得到目标决策树。决策树是一种机器学习的方法，在生成的目标决策树中，每个节点表示一个属性上的判断，每个分支代表一个判断结果的输出，最后每个叶子节点代表一种分类结果，在本实施例中，目标决策树的生成算法可以采用现有的C4.5或C5.0算法，此处不做赘述。
步骤104，分别获取目标决策树的根节点到每个叶子节点的路径，根据获取到的每一条路径分别生成一个对应的食品安全风险特征组合；
该步骤中，服务器分别获取目标决策树的根节点到每个叶子节点的路径，根据每条路径上的节点对应的风险特征数据分别生成一个对应的食品安全风险特征组合。
比如，当目标决策树的根节点到某个叶子节点的路径为：近半年后厨发现老鼠次数大于或等于2→最近一次量化检查结论为合格→经营面积大于200平→近一个月整改通过时，则将“近半年后厨发现老鼠次数大于或等于2、最近一次量化检查结论为合格、经营面积大于200平，且近一个月整改通过”作为与该路径对应的食品安全风险特征组合。
步骤105，计算与每个食品安全风险特征组合对应的风险评估值，将风险评估值标记在目标决策树的对应叶子节点上，得到食品安全风险评估模型；
该步骤中，服务器进一步计算与每个食品安全风险特征组合对应的风险评估值，并将风险评估值标记在目标决策树的对应叶子节点上，得到食品安全风险评估模型。
上述计算与每个食品安全风险特征组合对应的风险评估值的步骤具体可以包括：统计与每个食品安全风险特征组合对应的正样本数和负样本数，将与每个食品安全风险特征组合对应的正样本数和负样本数相加，得到样本总数；计算与每个食品安全风险特征组合对应的正样本数占样本总数的比例，将比例作为与每个食品安全风险特征组合对应的风险评估值。
比如，服务器统计符合食品安全风险特征组合“近半年后厨发现老鼠次数大于或等于2、最近一次量化检查结论为合格、经营面积大于200平，且近一个月整改通过”的样本总数为100，其中正样本数为10，负样本数为90，则与该食品安全风险特征组合对应的风险评估值为0.1。
上述方式通过将食品安全风险特征组合中的正样本数占样本总数的比例作为与该食品安全风险特征组合对应的风险评估值，实现了根据样本数客观而准确地评价一个食品安全风险特征组合的风险。
步骤106，获取待评估餐饮企业的风险特征数据，将待评估餐饮企业的风险特征数据输入至食品安全风险评估模型中，通过食品安全风险评估模型对待评估餐饮企业进行食品安全风险评估。
在得到食品安全风险评估模型后，对于后续待进行食品安全风险评估的餐饮企业，服 务器可以根据预设的食品安全风险特征项，从预设数据库中提取该待评估餐饮企业的风险特征数据，其中预设数据库包括但不限于市场监管局登记许可数据库、行政执法数据库、综合监督数据库、企业年报数据库、商事信用数据库和抽检数据库等，然后，服务器将该待评估餐饮企业的风险特征数据输入至该食品安全风险评估模型中，以使食品安全风险评估模型输出该待评估餐饮企业所属的食品安全风险特征组合及对应的风险值。
进一步地，服务器还可以将该待评估餐饮企业所属的食品安全风险特征组合及对应的风险值发送至预设的综合监管***，以使监管人员能够对高风险的餐饮企业及时采取相应措施，如现场检查或远程巡查处理等；此外，还可以将检查结果作为样本反馈到服务器的决策树训练模型中，以进一步优化和提升模型的精度。
本实施例提出的食品安全风险评估方法，首先获取多个餐饮企业的风险特征数据，根据风险特征数据生成训练集输入至决策树模型中进行训练，得到目标决策树，然后，根据目标决策树获取多个食品安全风险特征组合，计算与每个食品安全风险特征组合对应的风险评估值，将风险评估值标记在目标决策树的对应叶子节点上，得到食品安全风险评估模型，最后，通过该食品安全风险评估模型对待评估餐饮企业进行食品安全风险评估。这种基于决策树建立食品安全风险评估模型以对食品安全风险进行评估的方式，相比于现有技术中的人为评估，避免了评估不及时、评估数据项单一、主观性因素多等问题，提高了食品安全风险评估的效率和准确性。
本申请可应用于智慧政务领域，从而推动智慧城市的建设。
进一步地，基于本申请中食品安全风险评估方法的第一实施例，提出本申请中食品安全风险评估方法的第二实施例。
在本实施例中，上述步骤103具体可以包括：根据卡方检验算法，从正样本和负样本中筛选出具有风险预测价值的目标离散型风险特征数据；根据证据权重算法和信息价值算法，从正样本和负样本中筛选出具有风险预测价值的目标连续型风险特征数据；将目标离散型风险特征数据和目标连续型风险特征数据作为训练集输入至预设的决策树模型中进行训练，得到目标决策树。
其中，卡方检验是一种常用的特征选择算法，卡方检验就是统计样本的实际观测值与理论推断值之间的偏离程度，实际观测值与理论推断值之间的偏离程度就决定卡方值的大小，如果卡方值越大，二者偏差程度越大；反之，二者偏差越小；若两个值完全相等时，卡方值就为0，表明理论值完全符合。对于正样本和负样本中的离散型风险特征数据，如：是否对外投资、是否学校周边餐企等，服务器可以通过卡方检验算法，从中筛选出具有风险预测价值的目标离散型风险特征数据。证据权重算法(weight of evidence，WOE)和信息价值算法(information value，IV)是一种用来衡量变量的预测能力的算法，WOE和IV的值越大，表示此变量的预测能力越强。对于正样本和负样本中的连续型风险特征数据，如：经营面积、员工人数等，服务器可以通过证据权重算法和信息价值算法，从中筛选出具有风险预测价值的目标连续型风险特征数据。
进一步地，根据卡方检验算法，从正样本和负样本中筛选出具有风险预测价值的目标离散型风险特征数据的步骤可以包括：对于正样本和负样本中的离散型风险特征数据，统计与离散型风险特征数据对应的正样本数和负样本数；将与离散型风险特征数据对应的正样本数、负样本数，以及预设的样本数理论推断值代入卡方检验算法中进行计算，得到离散型风险特征数据的卡方值；当卡方值小于或等于第一预设阈值时，将离散型风险特征数据作为具有风险预测价值的目标离散型风险特征数据。
具体地，对于正样本和负样本中的离散型风险特征数据，服务器首先统计与该离散型风险特征数据对应的正样本数和负样本数，然后将与离散型风险特征数据对应的正样本数、 负样本数，以及预设的样本数理论推断值代入卡方检验算法中进行计算，得到离散型风险特征数据的卡方值，其中，卡方检验算法的公式为：
Figure PCTCN2020105061-appb-000001
其中，X 2为卡方值，observed为样本数的观测值，expected为样本数理论推断值。
服务器判断计算的卡方值是否小于或等于第一预设阈值，若是，则将该离散型风险特征数据作为具有风险预测价值的目标离散型风险特征数据。
进一步地，根据证据权重算法和信息价值算法，从正样本和负样本中筛选出具有风险预测价值的目标连续型风险特征数据的步骤可以包括：对于正样本和负样本中的连续型风险特征数据，将连续型风险特征数据进行分段；根据证据权重算法，计算与每个分段对应的证据权重，根据计算得到的证据权重和信息价值算法，计算与每个分段对应的信息价值；将与每个分段对应的信息价值进行求和，得到与连续型风险特征数据对应的信息价值；当与连续型风险特征数据对应的信息价值大于或等于第二预设阈值时，将连续型风险特征数据作为具有风险预测价值的目标连续型风险特征数据。
具体地，对于正样本和负样本中的连续型风险特征数据，服务器首先将该连续型风险特征数据进行分段，其中分段数及每个分段的大小可以灵活设置，然后根据证据权重算法，计算与每个分段对应的证据权重WOE，其中WOE＝ln(P1/P2)，P1为当前分段中正样本占所有正样本的比例，P2为当前分段中负样本占所有负样本的比例；之后，再将WOE值代入信息价值算法中进行计算，得到与每个分段对应的信息价值IV，其中IV＝(P1-P2)*WOE；将与每个分段对应的信息价值进行求和，即得到与该连续型风险特征数据对应的信息价值，服务器判断该信息价值是否大于或等于第二预设阈值，若是，则将该连续型风险特征数据作为具有风险预测价值的目标连续型风险特征数据。
之后，服务器将上述目标离散型风险特征数据和目标连续型风险特征数据作为训练集输入至预设的决策树模型中进行训练，得到目标决策树，由此能够提高模型训练的准确性和效率。
本申请实施例还提供一种食品安全风险评估装置。
参照图2，图2为本申请食品安全风险评估装置的一个实施例的模块示意图。本实施例中，所述食品安全风险评估装置包括：
第一获取模块201，用于根据预设的食品安全风险特征项，获取多个餐饮企业的风险特征数据；
判断模块202，用于判断所述多个餐饮企业在预设时长内是否发生过食品安全事故，将所述多个餐饮企业中发生过食品安全事故的餐饮企业的风险特征数据作为正样本，将所述多个餐饮企业中未发生过食品安全事故的餐饮企业的风险特征数据作为负样本；
训练模块203，用于根据所述正样本和所述负样本生成训练集，将所述训练集输入至预设的决策树模型中进行训练，得到目标决策树；
第二获取模块204，用于分别获取所述目标决策树的根节点到每个叶子节点的路径，根据获取到的每一条路径分别生成一个对应的食品安全风险特征组合；
计算模块205，用于计算与每个食品安全风险特征组合对应的风险评估值，将所述风险评估值标记在所述目标决策树的对应叶子节点上，得到食品安全风险评估模型；
风险评估模块206，用于获取待评估餐饮企业的风险特征数据，将所述待评估餐饮企业的风险特征数据输入至所述食品安全风险评估模型中，通过所述食品安全风险评估模型 对所述待评估餐饮企业进行食品安全风险评估。
可选的，所述第一获取模块201还用于：
从预设数据库中抽取多个餐饮企业的历史经营数据，对所述历史经营数据进行数据清洗，得到企业基础画像数据集；
对于企业基础画像数据集中的每个餐饮企业，从进行数据清洗后的历史经营数据中提取出与预设的食品安全风险特征项对应的风险特征数据。
可选的，所述判断模块202还用于：
从预设媒体数据库获取预设时长内所述多个餐饮企业的食品安全报导信息；
根据所述食品安全报导信息判断所述多个餐饮企业是否发生过食品中毒事件、是否存在司法处罚或行政处罚记录，以及是否被列入经营异常名录；
当某个餐饮企业发生过食品中毒事件，或存在司法处罚或行政处罚记录，或被列入经营异常名录时，判定所述餐饮企业在预设时长内发生过食品安全事故。
可选的，所述训练模块203还用于：
根据卡方检验算法，从所述正样本和所述负样本中筛选出具有风险预测价值的目标离散型风险特征数据；
根据证据权重算法和信息价值算法，从所述正样本和所述负样本中筛选出具有风险预测价值的目标连续型风险特征数据；
将所述目标离散型风险特征数据和所述目标连续型风险特征数据作为训练集输入至预设的决策树模型中进行训练，得到目标决策树。
可选的，所述训练模块203还用于：
对于所述正样本和所述负样本中的离散型风险特征数据，统计与所述离散型风险特征数据对应的正样本数和负样本数；
将与所述离散型风险特征数据对应的正样本数、负样本数，以及预设的样本数理论推断值代入卡方检验算法中进行计算，得到所述离散型风险特征数据的卡方值；
当所述卡方值小于或等于第一预设阈值时，将所述离散型风险特征数据作为具有风险预测价值的目标离散型风险特征数据。
可选的，所述训练模块203还用于：
对于所述正样本和所述负样本中的连续型风险特征数据，将所述连续型风险特征数据进行分段；
根据证据权重算法，计算与每个分段对应的证据权重，根据计算得到的所述证据权重和信息价值算法，计算与每个分段对应的信息价值；
将与每个分段对应的信息价值进行求和，得到与所述连续型风险特征数据对应的信息价值；
当与所述连续型风险特征数据对应的信息价值大于或等于第二预设阈值时，将所述连续型风险特征数据作为具有风险预测价值的目标连续型风险特征数据。
可选的，所述计算模块205还用于：
统计与每个食品安全风险特征组合对应的正样本数和负样本数，将与每个食品安全风险特征组合对应的正样本数和负样本数相加，得到样本总数；
计算与每个食品安全风险特征组合对应的正样本数占所述样本总数的比例，将所述比例作为与每个食品安全风险特征组合对应的风险评估值。
上述食品安全风险评估装置中各个模块的功能实现及有益效果与上述食品安全风险评估方法实施例中各步骤相对应，此处不再赘述。
上面从模块化功能实体的角度对本申请实施例中的食品安全风险评估装置进行了详细 描述，下面从硬件处理的角度对本申请实施例中食品安全风险评估设备进行详细描述。
参照图3，图3为本申请实施例提供的食品安全风险评估设备的结构示意图。该食品安全风险评估设备300可因配置或性能不同而产生比较大的差异，可以包括一个或一个以上处理器(central processing units，CPU)310(例如，一个或一个以上处理器)和存储器320，一个或一个以上存储应用程序333或数据332的存储介质330(例如一个或一个以上海量存储设备)。其中，存储器320和存储介质330可以是短暂存储或持久存储。存储在存储介质330的程序可以包括一个或一个以上模块(图示没标出)，每个模块可以包括对食品安全风险评估设备300中的一系列指令操作。更进一步地，处理器310可以设置为与存储介质330通信，在食品安全风险评估设备300上执行存储介质330中的一系列指令操作。
食品安全风险评估设备300还可以包括一个或一个以上电源340，一个或一个以上有线或无线网络接口350，一个或一个以上输入输出接口360，和/或，一个或一个以上操作***331，例如Windows Serve，Mac OS X，Unix，Linux，FreeBSD等等。本领域技术人员可以理解，图3示出的食品安全风险评估设备结构并不构成对食品安全风险评估设备的限定，可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件布置。
本申请还提供一种存储介质，该存储介质可以为非易失性存储介质，也可以为易失性存储介质，所述存储介质中存储有食品安全风险评估程序，所述食品安全风险评估程序被处理器执行时实现如上所述的食品安全风险评估方法的步骤。
其中，在所述处理器上运行的食品安全风险评估程序被执行时所实现的方法及有益效果可参照本申请食品安全风险评估方法的各个实施例，此处不再赘述。
进一步地，该计算机可读存储介质可主要包括存储程序区和存储数据区，其中，存储程序区可存储操作***、至少一个功能所需的应用程序等；存储数据区可存储根据区块链节点的使用所创建的数据等。
本申请所指区块链是分布式数据存储、点对点传输、共识机制、加密算法等计算机技术的新型应用模式。区块链(Blockchain)，本质上是一个去中心化的数据库，是一串使用密码学方法相关联产生的数据块，每一个数据块中包含了一批次网络交易的信息，用于验证其信息的有效性(防伪)和生成下一个区块。区块链可以包括区块链底层平台、平台产品服务层以及应用服务层等。
本领域技术人员可以理解，上述集成的模块或单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个取存储介质中。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的全部或部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(read-only memory，ROM)、随机存取存储器(random access memory，RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
以上实施例仅用以说明本申请的技术方案，而非对其限制；尽管参照前述实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的精神和范围。

Claims (20)

  1. 一种食品安全风险评估方法，其中，所述食品安全风险评估方法包括如下步骤：
    根据预设的食品安全风险特征项，获取多个餐饮企业的风险特征数据；
    判断所述多个餐饮企业在预设时长内是否发生过食品安全事故，将所述多个餐饮企业中发生过食品安全事故的餐饮企业的风险特征数据作为正样本，将所述多个餐饮企业中未发生过食品安全事故的餐饮企业的风险特征数据作为负样本；
    根据所述正样本和所述负样本生成训练集，将所述训练集输入至预设的决策树模型中进行训练，得到目标决策树；
    分别获取所述目标决策树的根节点到每个叶子节点的路径，根据获取到的每一条路径分别生成一个对应的食品安全风险特征组合；
    计算与每个食品安全风险特征组合对应的风险评估值，将所述风险评估值标记在所述目标决策树的对应叶子节点上，得到食品安全风险评估模型；
    获取待评估餐饮企业的风险特征数据，将所述待评估餐饮企业的风险特征数据输入至所述食品安全风险评估模型中，通过所述食品安全风险评估模型对所述待评估餐饮企业进行食品安全风险评估。
  2. 如权利要求1所述的食品安全风险评估方法，其中，所述根据预设的食品安全风险特征项，获取多个餐饮企业的风险特征数据的步骤包括：
    从预设数据库中抽取多个餐饮企业的历史经营数据，对所述历史经营数据进行数据清洗，得到企业基础画像数据集；
    对于企业基础画像数据集中的每个餐饮企业，从进行数据清洗后的历史经营数据中提取出与预设的食品安全风险特征项对应的风险特征数据。
  3. 如权利要求1所述的食品安全风险评估方法，其中，所述判断所述多个餐饮企业在预设时长内是否发生过食品安全事故的步骤包括：
    从预设媒体数据库获取预设时长内所述多个餐饮企业的食品安全报导信息；
    根据所述食品安全报导信息判断所述多个餐饮企业是否发生过食品中毒事件、是否存在司法处罚或行政处罚记录，以及是否被列入经营异常名录；
    当某个餐饮企业发生过食品中毒事件，或存在司法处罚或行政处罚记录，或被列入经营异常名录时，判定所述餐饮企业在预设时长内发生过食品安全事故。
  4. 如权利要求1-3中任一项所述的食品安全风险评估方法，其中，所述根据所述正样本和所述负样本生成训练集，将所述训练集输入至预设的决策树模型中进行训练，得到目标决策树的步骤包括：
    根据卡方检验算法，从所述正样本和所述负样本中筛选出具有风险预测价值的目标离散型风险特征数据；
    根据证据权重算法和信息价值算法，从所述正样本和所述负样本中筛选出具有风险预测价值的目标连续型风险特征数据；
    将所述目标离散型风险特征数据和所述目标连续型风险特征数据作为训练集输入至预设的决策树模型中进行训练，得到目标决策树。
  5. 如权利要求4所述的食品安全风险评估方法，其中，所述根据卡方检验算法，从所述正样本和所述负样本中筛选出具有风险预测价值的目标离散型风险特征数据的步骤包括：
    对于所述正样本和所述负样本中的离散型风险特征数据，统计与所述离散型风险特征数据对应的正样本数和负样本数；
    将与所述离散型风险特征数据对应的正样本数、负样本数，以及预设的样本数理论推 断值代入卡方检验算法中进行计算，得到所述离散型风险特征数据的卡方值；
    当所述卡方值小于或等于第一预设阈值时，将所述离散型风险特征数据作为具有风险预测价值的目标离散型风险特征数据。
  6. 如权利要求4所述的食品安全风险评估方法，其中，所述根据证据权重算法和信息价值算法，从所述正样本和所述负样本中筛选出具有风险预测价值的目标连续型风险特征数据的步骤包括：
    对于所述正样本和所述负样本中的连续型风险特征数据，将所述连续型风险特征数据进行分段；
    根据证据权重算法，计算与每个分段对应的证据权重，根据计算得到的所述证据权重和信息价值算法，计算与每个分段对应的信息价值；
    将与每个分段对应的信息价值进行求和，得到与所述连续型风险特征数据对应的信息价值；
    当与所述连续型风险特征数据对应的信息价值大于或等于第二预设阈值时，将所述连续型风险特征数据作为具有风险预测价值的目标连续型风险特征数据。
  7. 如权利要求1所述的食品安全风险评估方法，其中，所述计算与每个食品安全风险特征组合对应的风险评估值的步骤包括：
    统计与每个食品安全风险特征组合对应的正样本数和负样本数，将与每个食品安全风险特征组合对应的正样本数和负样本数相加，得到样本总数；
    计算与每个食品安全风险特征组合对应的正样本数占所述样本总数的比例，将所述比例作为与每个食品安全风险特征组合对应的风险评估值。
  8. 一种食品安全风险评估设备，其中，所述食品安全风险评估设备包括：存储器、处理器以及存储在所述存储器上并可在所述处理器上运行的食品安全风险评估程序，所述处理器执行所述食品安全风险评估程序时实现如下步骤：
    根据预设的食品安全风险特征项，获取多个餐饮企业的风险特征数据；
    判断所述多个餐饮企业在预设时长内是否发生过食品安全事故，将所述多个餐饮企业中发生过食品安全事故的餐饮企业的风险特征数据作为正样本，将所述多个餐饮企业中未发生过食品安全事故的餐饮企业的风险特征数据作为负样本；
    根据所述正样本和所述负样本生成训练集，将所述训练集输入至预设的决策树模型中进行训练，得到目标决策树；
    分别获取所述目标决策树的根节点到每个叶子节点的路径，根据获取到的每一条路径分别生成一个对应的食品安全风险特征组合；
    计算与每个食品安全风险特征组合对应的风险评估值，将所述风险评估值标记在所述目标决策树的对应叶子节点上，得到食品安全风险评估模型；
    获取待评估餐饮企业的风险特征数据，将所述待评估餐饮企业的风险特征数据输入至所述食品安全风险评估模型中，通过所述食品安全风险评估模型对所述待评估餐饮企业进行食品安全风险评估。
  9. 如权利要求8所述的食品安全风险评估设备，其中，所述处理器执行所述食品安全风险评估程序实现所述根据预设的食品安全风险特征项，获取多个餐饮企业的风险特征数据时，包括以下步骤：
    从预设数据库中抽取多个餐饮企业的历史经营数据，对所述历史经营数据进行数据清洗，得到企业基础画像数据集；
    对于企业基础画像数据集中的每个餐饮企业，从进行数据清洗后的历史经营数据中提取出与预设的食品安全风险特征项对应的风险特征数据。
  10. 如权利要求8所述的食品安全风险评估设备，其中，所述处理器执行所述食品安全风险评估程序实现所述判断所述多个餐饮企业在预设时长内是否发生过食品安全事故时，包括以下步骤：
    从预设媒体数据库获取预设时长内所述多个餐饮企业的食品安全报导信息；
    根据所述食品安全报导信息判断所述多个餐饮企业是否发生过食品中毒事件、是否存在司法处罚或行政处罚记录，以及是否被列入经营异常名录；
    当某个餐饮企业发生过食品中毒事件，或存在司法处罚或行政处罚记录，或被列入经营异常名录时，判定所述餐饮企业在预设时长内发生过食品安全事故。
  11. 如权利要求8-10中任一项所述的食品安全风险评估设备，其中，所述处理器执行所述食品安全风险评估程序实现所述根据所述正样本和所述负样本生成训练集，将所述训练集输入至预设的决策树模型中进行训练，得到目标决策树时，包括以下步骤：
    根据卡方检验算法，从所述正样本和所述负样本中筛选出具有风险预测价值的目标离散型风险特征数据；
    根据证据权重算法和信息价值算法，从所述正样本和所述负样本中筛选出具有风险预测价值的目标连续型风险特征数据；
    将所述目标离散型风险特征数据和所述目标连续型风险特征数据作为训练集输入至预设的决策树模型中进行训练，得到目标决策树。
  12. 如权利要求11所述的食品安全风险评估设备，其中，所述处理器执行所述食品安全风险评估程序实现所述根据卡方检验算法，从所述正样本和所述负样本中筛选出具有风险预测价值的目标离散型风险特征数据时，包括以下步骤：
    对于所述正样本和所述负样本中的离散型风险特征数据，统计与所述离散型风险特征数据对应的正样本数和负样本数；
    将与所述离散型风险特征数据对应的正样本数、负样本数，以及预设的样本数理论推断值代入卡方检验算法中进行计算，得到所述离散型风险特征数据的卡方值；
    当所述卡方值小于或等于第一预设阈值时，将所述离散型风险特征数据作为具有风险预测价值的目标离散型风险特征数据。
  13. 如权利要求11所述的食品安全风险评估设备，其中，所述处理器执行所述食品安全风险评估程序实现所述根据证据权重算法和信息价值算法，从所述正样本和所述负样本中筛选出具有风险预测价值的目标连续型风险特征数据时，包括以下步骤：
    对于所述正样本和所述负样本中的连续型风险特征数据，将所述连续型风险特征数据进行分段；
    根据证据权重算法，计算与每个分段对应的证据权重，根据计算得到的所述证据权重和信息价值算法，计算与每个分段对应的信息价值；
    将与每个分段对应的信息价值进行求和，得到与所述连续型风险特征数据对应的信息价值；
    当与所述连续型风险特征数据对应的信息价值大于或等于第二预设阈值时，将所述连续型风险特征数据作为具有风险预测价值的目标连续型风险特征数据。
  14. 如权利要求8所述的食品安全风险评估设备，其中，所述处理器执行所述食品安全风险评估程序实现所述计算与每个食品安全风险特征组合对应的风险评估值时，包括以下步骤：
    统计与每个食品安全风险特征组合对应的正样本数和负样本数，将与每个食品安全风险特征组合对应的正样本数和负样本数相加，得到样本总数；
    计算与每个食品安全风险特征组合对应的正样本数占所述样本总数的比例，将所述比 例作为与每个食品安全风险特征组合对应的风险评估值。
  15. 一种计算机可读存储介质，所述计算机可读存储介质中存储计算机指令，当所述计算机指令在计算机上运行时，使得计算机执行如下步骤：
    根据预设的食品安全风险特征项，获取多个餐饮企业的风险特征数据；
    判断所述多个餐饮企业在预设时长内是否发生过食品安全事故，将所述多个餐饮企业中发生过食品安全事故的餐饮企业的风险特征数据作为正样本，将所述多个餐饮企业中未发生过食品安全事故的餐饮企业的风险特征数据作为负样本；
    根据所述正样本和所述负样本生成训练集，将所述训练集输入至预设的决策树模型中进行训练，得到目标决策树；
    分别获取所述目标决策树的根节点到每个叶子节点的路径，根据获取到的每一条路径分别生成一个对应的食品安全风险特征组合；
    计算与每个食品安全风险特征组合对应的风险评估值，将所述风险评估值标记在所述目标决策树的对应叶子节点上，得到食品安全风险评估模型；
    获取待评估餐饮企业的风险特征数据，将所述待评估餐饮企业的风险特征数据输入至所述食品安全风险评估模型中，通过所述食品安全风险评估模型对所述待评估餐饮企业进行食品安全风险评估。
  16. 如权利要求15所述的计算机可读存储介质，所述计算机可读存储介质执行所述计算机指令实现所述根据预设的食品安全风险特征项，获取多个餐饮企业的风险特征数据时，包括以下步骤：
    从预设数据库中抽取多个餐饮企业的历史经营数据，对所述历史经营数据进行数据清洗，得到企业基础画像数据集；
    对于企业基础画像数据集中的每个餐饮企业，从进行数据清洗后的历史经营数据中提取出与预设的食品安全风险特征项对应的风险特征数据。
  17. 如权利要求15所述的计算机可读存储介质，所述计算机可读存储介质执行所述计算机指令实现所述判断所述多个餐饮企业在预设时长内是否发生过食品安全事故时，包括以下步骤：
    从预设媒体数据库获取预设时长内所述多个餐饮企业的食品安全报导信息；
    根据所述食品安全报导信息判断所述多个餐饮企业是否发生过食品中毒事件、是否存在司法处罚或行政处罚记录，以及是否被列入经营异常名录；
    当某个餐饮企业发生过食品中毒事件，或存在司法处罚或行政处罚记录，或被列入经营异常名录时，判定所述餐饮企业在预设时长内发生过食品安全事故。
  18. 如权利要求15-17中任一项所述的计算机可读存储介质，所述计算机可读存储介质执行所述计算机指令实现所述根据所述正样本和所述负样本生成训练集，将所述训练集输入至预设的决策树模型中进行训练，得到目标决策树时，包括以下步骤：
    根据卡方检验算法，从所述正样本和所述负样本中筛选出具有风险预测价值的目标离散型风险特征数据；
    根据证据权重算法和信息价值算法，从所述正样本和所述负样本中筛选出具有风险预测价值的目标连续型风险特征数据；
    将所述目标离散型风险特征数据和所述目标连续型风险特征数据作为训练集输入至预设的决策树模型中进行训练，得到目标决策树。
  19. 如权利要求18所述的计算机可读存储介质，所述计算机可读存储介质执行所述计算机指令实现所述根据卡方检验算法，从所述正样本和所述负样本中筛选出具有风险预测价值的目标离散型风险特征数据时，包括以下步骤：
    对于所述正样本和所述负样本中的离散型风险特征数据，统计与所述离散型风险特征数据对应的正样本数和负样本数；
    将与所述离散型风险特征数据对应的正样本数、负样本数，以及预设的样本数理论推断值代入卡方检验算法中进行计算，得到所述离散型风险特征数据的卡方值；
    当所述卡方值小于或等于第一预设阈值时，将所述离散型风险特征数据作为具有风险预测价值的目标离散型风险特征数据。
  20. 一种食品安全风险评估装置，其中，所述食品安全风险评估装置包括：
    第一获取模块，用于根据预设的食品安全风险特征项，获取多个餐饮企业的风险特征数据；
    判断模块，用于判断所述多个餐饮企业在预设时长内是否发生过食品安全事故，将所述多个餐饮企业中发生过食品安全事故的餐饮企业的风险特征数据作为正样本，将所述多个餐饮企业中未发生过食品安全事故的餐饮企业的风险特征数据作为负样本；
    训练模块，用于根据所述正样本和所述负样本生成训练集，将所述训练集输入至预设的决策树模型中进行训练，得到目标决策树；
    第二获取模块，用于分别获取所述目标决策树的根节点到每个叶子节点的路径，根据获取到的每一条路径分别生成一个对应的食品安全风险特征组合；
    计算模块，用于计算与每个食品安全风险特征组合对应的风险评估值，将所述风险评估值标记在所述目标决策树的对应叶子节点上，得到食品安全风险评估模型；
    风险评估模块，用于获取待评估餐饮企业的风险特征数据，将所述待评估餐饮企业的风险特征数据输入至所述食品安全风险评估模型中，通过所述食品安全风险评估模型对所述待评估餐饮企业进行食品安全风险评估。
PCT/CN2020/105061 2020-05-21 2020-07-28 食品安全风险评估方法、装置、设备及存储介质 WO2021232588A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010433430.8 2020-05-21
CN202010433430.8A CN111738549A (zh) 2020-05-21 2020-05-21 食品安全风险评估方法、装置、设备及存储介质

Publications (1)

Publication Number Publication Date
WO2021232588A1 true WO2021232588A1 (zh) 2021-11-25

Family

ID=72647526

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/105061 WO2021232588A1 (zh) 2020-05-21 2020-07-28 食品安全风险评估方法、装置、设备及存储介质

Country Status (2)

Country Link
CN (1) CN111738549A (zh)
WO (1) WO2021232588A1 (zh)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114492214A (zh) * 2022-04-18 2022-05-13 支付宝(杭州)信息技术有限公司 利用机器学习的选择算子确定、策略组合优化方法及装置
CN115375194A (zh) * 2022-10-24 2022-11-22 浙江省标准化研究院(金砖国家标准化(浙江)研究中心、浙江省物品编码中心) 风险预警方法、装置、电子设备及存储介质
CN116596300A (zh) * 2023-04-19 2023-08-15 浙江物芯数科信息产业有限公司 一种食品生产企业的智慧监管方法、***、设备和介质
CN116934359A (zh) * 2023-09-15 2023-10-24 深圳市金谷园实业发展有限公司 一种基于互联网的食品安全全流程监管***
CN118037141A (zh) * 2024-04-15 2024-05-14 中国标准化研究院 一种餐饮企业数字化综合评估方法及***

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112101819A (zh) * 2020-10-28 2020-12-18 平安国际智慧城市科技股份有限公司 食品风险预测方法、装置、设备及存储介质
CN113298438A (zh) * 2021-06-22 2021-08-24 中国平安财产保险股份有限公司 区域风险等级的评估方法、装置、计算机设备及存储介质
CN113762764B (zh) * 2021-09-02 2024-04-12 南京大学 一种进口食品安全风险自动分级与预警***及方法
CN113886716B (zh) * 2021-11-30 2022-04-05 国家食品安全风险评估中心 食品安全突发事件的应急处置推荐方法及***
CN114140095A (zh) * 2021-12-13 2022-03-04 成都市食品检验研究院 一种智能生成食品生产经营企业关键控制措施的***
CN115496372B (zh) * 2022-09-26 2023-10-24 广东顺峰智慧能源研究院有限公司 充电***安全风险预测方法、装置、设备及可读存储介质
CN115829321B (zh) * 2022-11-22 2023-08-15 成都市食品检验研究院 一种食品企业加工过程风险智能化处置***
CN117314149A (zh) * 2023-05-17 2023-12-29 南京汉卫公共卫生研究院有限公司 一种食品安全监督管理***

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107563645A (zh) * 2017-09-04 2018-01-09 杭州云算信达数据技术有限公司 一种基于大数据的金融风险分析方法
CN108269012A (zh) * 2018-01-12 2018-07-10 中国平安人寿保险股份有限公司 风险评分模型的构建方法、装置、存储介质及终端
US20180232805A1 (en) * 2016-06-12 2018-08-16 Tencent Technology (Shenzhen) Company Limited User credit rating method and apparatus, and storage medium
CN108764273A (zh) * 2018-04-09 2018-11-06 中国平安人寿保险股份有限公司 一种数据处理的方法、装置、终端设备及存储介质
CN109657894A (zh) * 2018-09-27 2019-04-19 深圳壹账通智能科技有限公司 企业信用风险预警方法、装置、设备及存储介质
CN109858762A (zh) * 2018-12-29 2019-06-07 重庆誉存大数据科技有限公司 一种企业风险预测方法及装置
CN111062736A (zh) * 2018-10-17 2020-04-24 百度在线网络技术（北京）有限公司 模型训练、线索排序方法、装置及设备

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20180232805A1 (en) * 2016-06-12 2018-08-16 Tencent Technology (Shenzhen) Company Limited User credit rating method and apparatus, and storage medium
CN107563645A (zh) * 2017-09-04 2018-01-09 杭州云算信达数据技术有限公司 一种基于大数据的金融风险分析方法
CN108269012A (zh) * 2018-01-12 2018-07-10 中国平安人寿保险股份有限公司 风险评分模型的构建方法、装置、存储介质及终端
CN108764273A (zh) * 2018-04-09 2018-11-06 中国平安人寿保险股份有限公司 一种数据处理的方法、装置、终端设备及存储介质
CN109657894A (zh) * 2018-09-27 2019-04-19 深圳壹账通智能科技有限公司 企业信用风险预警方法、装置、设备及存储介质
CN111062736A (zh) * 2018-10-17 2020-04-24 百度在线网络技术（北京）有限公司 模型训练、线索排序方法、装置及设备
CN109858762A (zh) * 2018-12-29 2019-06-07 重庆誉存大数据科技有限公司 一种企业风险预测方法及装置

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114492214A (zh) * 2022-04-18 2022-05-13 支付宝(杭州)信息技术有限公司 利用机器学习的选择算子确定、策略组合优化方法及装置
CN115375194A (zh) * 2022-10-24 2022-11-22 浙江省标准化研究院(金砖国家标准化(浙江)研究中心、浙江省物品编码中心) 风险预警方法、装置、电子设备及存储介质
CN116596300A (zh) * 2023-04-19 2023-08-15 浙江物芯数科信息产业有限公司 一种食品生产企业的智慧监管方法、***、设备和介质
CN116934359A (zh) * 2023-09-15 2023-10-24 深圳市金谷园实业发展有限公司 一种基于互联网的食品安全全流程监管***
CN116934359B (zh) * 2023-09-15 2023-12-29 深圳市金谷园实业发展有限公司 一种基于互联网的食品安全全流程监管***
CN118037141A (zh) * 2024-04-15 2024-05-14 中国标准化研究院 一种餐饮企业数字化综合评估方法及***

Also Published As

Publication number Publication date
CN111738549A (zh) 2020-10-02

Similar Documents

Publication Publication Date Title
WO2021232588A1 (zh) 食品安全风险评估方法、装置、设备及存储介质
Dai et al. Utilizing blockchain and smart contracts to enable audit 4.0: From the perspective of accountability audit of air pollution control in China
CN114066242A (zh) 一种企业风险的预警方法及装置
US20110191138A1 (en) Risk scorecard
CN112102076B (zh) 一种平台综合风险预警***
CN101236638A (zh) 一种基于Web的银行卡风险监测方法及***
CN111291076A (zh) 基于大数据的异常用水监测报警***及其构建方法
CN112419030B (zh) 财务舞弊风险评估的方法、***及设备
CN111583033A (zh) 基于上市公司和股东之间联系的关联分析方法及装置
CN111222955A (zh) 一种基于区块链的供应商监管方法及***
CN109118411B (zh) 基于智能辅助平台和移动终端的刑事执行检察***和方法
CN111062597A (zh) 一种上市公司财务报表粉饰嫌疑检测方法及装置
CN113449964A (zh) 一种企业金融风险监测预警***及监测预警方法
CN112330474A (zh) 核保风控监控方法、装置、设备及存储介质
CN113283806A (zh) 企业信息评估方法、装置、计算机设备及存储介质
KR101666824B1 (ko) 기업 연쇄부도위험 관리 시스템 및 이를 이용한 연쇄부도위험 관리 방법
CN116777140A (zh) 一种企业业务管理方法、装置、设备及介质
CN115271514A (zh) 通讯企业的监控方法、装置、电子设备及存储介质
CN114169686A (zh) 一种上市公司esg评价方法
CN110738570A (zh) 基于多通道异质数据的信息型操纵自动识别方法
CN111899112A (zh) 一种不动产信托项目管理方法与装置
CN111709669A (zh) 一种企业信用评分***及方法
Sari The Successful Implementation of Accounting Information Systems and the Quality of Government Financial Statements at Regencies
CN113052415B (zh) 一种基于案件分析的法官审判预警监控方法、装置、电子设备和存储介质
LU503268B1 (en) A business environment assessment and analysis method and system

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20936800

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20936800

Country of ref document: EP

Kind code of ref document: A1