KR101859410B1 - Method, Apparatus for Regional Food Safety Factor Computing, And a Computer-readable Storage Medium for executing the Method - Google Patents

Abstract

The present invention relates to a computer-readable recording medium storing a program for executing a method for calculating a food safety index of a region, a device thereof, and a method for calculating the safety index of a region, And deriving a safety diagnostic model by constructing a nonlinear regression model on the correlation between the information on the weather and the occurrence of food poisoning; And calculating a regional food safety index based on the safety diagnosis model.
Thus, by using the information on the cultivation environment of the region of the foodstuffs expected to increase the food poisoning risk of the foodstuffs as the basic data of the foodstuff safety index, a more accurate food safety index can be calculated. In addition, by applying the nonlinear regression model, it is easy to apply even when the input and output patterns are nonlinear, and it is possible to test easily by adding input variables. By applying a new sampling method to solve the data unbalance problem, Can be improved.

Description

TECHNICAL FIELD [0001] The present invention relates to a method for calculating a food-material safety index for a region, a method for calculating the food-material safety index for each region, and a computer-readable storage medium storing a program for executing the method,

The present invention relates to a method for calculating a food safety index of a region, a device therefor, and a computer-readable recording medium recording a program for executing the method, and more particularly, to a method for calculating a food safety index of a region to which a nonlinear regression model is applied, The present invention relates to a computer-readable recording medium storing a program for causing a computer to execute the program.

Recently, as food safety accidents are frequent, public concern about food safety is increasing. As the causes and behavior of food safety accidents become complicated due to the increase of food additives, the emergence of new harmful substances, the development of distribution network and the increase of international trade, the need for systematic and efficient management of food safety is emerging.

Conventionally, the food poisoning index and the food management system of the food and the like are based on only temperature and humidity information, and the safety index is calculated to predict the food poisoning index. For example, based on the table of food poisoning indices based on temperature and humidity information, the food poisoning risk index is calculated by applying regional temperature and humidity to the table.

However, food poisoning accidents are caused by hygiene control of group foodservice and restaurants, but there are also a few cases of contamination problems in the production and distribution of raw materials.

According to data from 2002 to 2013, it is known that accidents caused by harmful microorganisms account for about 60% of domestic food safety accidents. Currently, food poisoning accidents due to freshly consumed agricultural products are increasing in the world, and it is pointed out that contamination of agricultural products due to environmental pollution is one of food safety risk factors due to the characteristic of ingestion without heating.

Food pollution caused by harmful microorganisms is mainly caused by feces of people and animals, and the possibility of soil and water pollution increases according to weather conditions such as flood, flood, temperature rise, and the risk of food contamination increases.

However, the conventional method of calculating food poisoning index does not take into consideration the possibility of food poisoning due to food poisoning virus or food poisoning causative bacteria which may be introduced during the distribution process or cooking process of such food ingredients.

In the past, time series regression models such as Poisson regression model, Auto Regressive (AR), and Auto Regressive Integrated Moving Average (ARIMA) have been used predominantly in food poisoning prediction models.

However, the conventional time series regression model has a disadvantage that it can not be applied when the pattern is nonlinear. In addition, data on risk prediction, such as food poisoning, are generally problematic due to data imbalance due to very few occurrence samples.

The present invention aims at providing a food material safety index calculating device and method thereof, in which the information on the cultivation environment of a region of a foodstuff expected to increase the risk of food poisoning of food ingredients is used as a basis for calculating the food safety index. Also, it is an object of the present invention to provide a food material safety index calculating apparatus and a method thereof that employ a nonlinear regression model. Another object of the present invention is to provide a new sampling method of food poisoning occurrence data.

The above object is achieved by a non-linear regression model for the correlation between food-related information and food-related-environment-related information on the cultivation environment and weather of a food material, And calculating a regional food safety index based on the safety diagnosis model.

Here, the non-linear regression model may include a support vector regression model (SVR) or a least squares support vector machine (LS-SVM). .

The information on the cultivation environment for each region includes the flood information for each region and the information on the livestock farming, and the information about the weather for each region may include temperature information, humidity information, and precipitation amount information.

In addition, the food safety index may include a food poisoning index.

Further, the step of deriving the safety diagnosis model may include converting the cultivation environment, meteorological conditions and food poisoning occurrence information for the local ingredients into data for learning; Selecting samples for building the nonlinear regression model; And learning the nonlinear regression model using the samples.

Here, the step of selecting the samples may include sampling the food poisoning occurrence information of each city or district unit by taking an average value in units of a trial.

Further, the information on the cultivation environment by region may further include information on agricultural products by region.

The above object can also be achieved by a computer-readable recording medium recording a program for executing the method of calculating the food safety index.

The above object is achieved by a nonlinear regression modeling unit for deriving a safety diagnosis model by constructing a nonlinear regression model on the correlation between information on the cultivation environment and weather of the region and the occurrence of food poisoning, ; And a safety index calculating unit that receives data on the cultivation environment and weather of each region and calculates a safety index for each region based on the safety diagnosis model.

Here, the nonlinear regression modeling unit may include a database unit that constructs the collected information as learning data for model learning and test data for model testing; A nonlinear regression model learning unit for learning the nonlinear regression model using the learning data; And a model verification unit for verifying the nonlinear regression model using test data.

As described above, the apparatus and method for calculating the safety index of a food according to the present invention can provide information on the cultivation environment of a foodstuff, which is expected to increase the risk of food poisoning of foodstuffs, The safety index can be calculated. In addition, the application of the nonlinear regression model makes it easy to apply even when the input and output patterns are non-linear, and it is advantageous in that the input variables can be easily added and tested. In addition, modeling performance can be improved by applying a new sampling method to solve the problem of data imbalance.

1 is a schematic view of a food ingredient safety index calculating device according to a first embodiment of the present invention.
2 is a schematic diagram of a nonlinear regression modeling unit 20 according to an embodiment of the present invention.
FIG. 3 shows the model fitting results for the learning set of the safety diagnosis model according to the present invention learned from the data from May 1, 2006 to April 30, 2012.
Fig. 4 shows the result of prediction of the model learned from the data of 2006/05/01 to 2012/04/30 with the data of the next four months.
FIG. 5 is a flowchart illustrating a method of calculating the safe index of foodstuffs according to a second embodiment of the present invention using the regional food safety index calculating device of FIG. 1, and FIG. 6 is a detailed view of step S10 of FIG.

Hereinafter, specific embodiments of the present invention will be described with reference to the drawings.

According to the present invention, the food safety index calculating device and the calculating method thereof can predict the food safety by region in consideration of the weather conditions and environment factors in which food poisoning can be expressed and various regional conditions that may affect the cultivation environment. The present invention proposes a safety diagnosis model based on a nonlinear regression model that predicts the risk of food poisoning at a given date and region by analyzing input data patterns at the time of food poisoning, using past information on actual food poisoning.

1 is a schematic view of a food ingredient safety index calculating device according to a first embodiment of the present invention.

Referring to FIG. 1, the regional food safety index calculating apparatus according to the first embodiment of the present invention includes a nonlinear regression modeling unit 20 and a safety index calculating unit 30.

The nonlinear regression modeling unit 20 constructs a safety diagnosis model for the correlation between the information on the environment and the climate of the area and the occurrence of food poisoning in each region by using a nonlinear regression model. Unlike the food poisoning index calculation model, it is a model that considers the characteristics of the cultivation environment in each region in order to diagnose food safety according to the cultivation environment, and the support vector regression (SVR) as a nonlinear regression model in deriving the safety diagnosis model. There is a big feature in that it utilizes.

According to the present invention, by using the bottom-up approach, it is possible to predict future food poisoning risks by analyzing patterns of cultivation environment, regions, and weather conditions at the time of food poisoning using past information that actual food poisoning occurred Develop a safety diagnosis model. Time series regression models such as the Poisson regression model, the Auto Regressive (AR), and the Auto Regressive Integrated Moving Average (ARIMA) used in the general food poisoning prediction model are disadvantages that can not be applied when the pattern is nonlinear . The safety diagnosis model of the present invention uses a nonlinear regression model. Among them, a support vector regression model or a least squares support vector machine (Least Squares), which is a regression analysis version of a support vector machine (SVM) Support Vector Machines (LS-SVM). The application of the support vector regression model to the food poisoning prediction model is first attempted in the present invention. The superiority over the existing regression model is that it can be easily applied even when the pattern is nonlinear, There are advantages in.

Data necessary for constructing the safety diagnosis model according to the present invention can be collected through the data collection unit 10. [ The data collecting unit 10 may be included in the regional food safety index calculating apparatus or may be implemented as a separate data collection server.

1, the information on the safety of the foodstuffs collected by the data collecting unit 10 includes information on the cultivation environment of the foodstuffs, weather information and food poisoning information, information on the cultivation environment by region includes information on flood information by region, , And livestock farming information, and the weather information includes past, present, and current weather information including temperature, humidity, and precipitation information by region. Food poisoning information may include information on the status of food poisoning and on-site investigation results, possible cause foods related to food poisoning, place of consumption, time, and patient statistics. In addition, the flood information may include information on the flooded area, the source of flooding, the type of flood, the use of the floodplain, and the time of flood. The data collecting unit 10 collects specific information in a specific site targeted in advance or collects information related to food ingredients and food safety contained in news or SNS data. The information collected by the data collecting unit 10 includes past and present information on the cultivation environment of the foodstuff, weather information, and food poisoning information.

2 is a schematic diagram of a nonlinear regression modeling unit 20 according to an embodiment of the present invention. 2, the nonlinear regression modeling unit 20 includes a database unit 21, a nonlinear regression model learning unit 23, a model verification unit 25, and a safety criterion calculation unit 27. [ The database unit 21 constructs the information collected by the data collecting unit 10 as learning data for model learning and test data for model testing. The nonlinear regression modeling unit 20 learns the model with the learning data and then verifies the prediction accuracy of the learned model with the test data.

The nonlinear regression modeling unit 20 according to the present invention has a great feature in that it takes into consideration various local conditions that may affect the cultivation environment and various weather environment factors in which food poisoning can be manifested, Factors influencing the cultivation environment and food poisoning can be influenced by the size of the regional farmland, past flood occurrence, daily maximum temperature, daily minimum temperature, daily average temperature, humidity, and precipitation. In addition, as an input variable, agricultural products information, agricultural distribution information, or food information can be additionally considered in the area. It should be noted that the above-described input variables are only examples, and other input variables may be considered in the cultivation environment. The data to be built in the database is region-specific data, for example, data of units of city or district. This is to calculate the safety diagnosis index by region.

When creating a predictive model using a machine learning model, it is important to define the input samples for the purpose of the prediction and to define the output values to be predicted. It is also important to select and define the features that make up the input value. In the present invention, in order to faithfully reflect the difference of the cultivation environment in each region, the safety diagnosis index can be predicted every day in 226 city and county districts nationwide. For this purpose, each input sample should be generated for each town / village, and since the safety diagnosis index should be calculated daily, one input sample is generated for each town / village by date.

Table 1 below shows input sample qualification definitions of the safety diagnostic model of the present invention.

Input qualities meaning highest temperature Date, local maximum temperature Lowest temperature The date, the lowest temperature in the region Average temperature The date, the average temperature of the area Precipitation Average date of precipitation Average humidity Average date, average humidity Number of livestock farms Number of livestock farms in the area Frequency of immersion for 5 years Number of inundations in the region during the last five years That month Month of the date That week Week of the day

The output value corresponding to each input sample is defined as the number of food poisoning occurrences in the area that day. When the input sample is the learning data, the value is the number of food poisoning occurred in the past and extracted from past food poisoning statistical data. When the input sample is test data, the value corresponds to the value to be predicted. The database unit 21 defines input samples and output values as described above, and then builds data for model learning and testing using the collected data.

characteristic value Collected data duration 2002/01/01 - 2016/06/30 Collection area 226 municipalities nationwide Number of input samples 1,079,577 Number of positive samples of food poisoning 3,419 Number of negative samples without food poisoning 1,076,158

Table 2 shows an example of the characteristics of the data constructed in the database unit 21. The data collection period starts from January 1, 2002, when food poisoning statistics are available, until June 30, 2016 And investigated nationwide city and county districts. On the other hand, food poisoning data has a problem of Imbalanced Data Learning, in which the number of samples of food poisoning is much smaller than that of samples that do not have food poisoning. The present invention proposes a new sampling method to solve this problem. Conventionally, in order to solve the sampling imbalance method, if it was the main method to adopt sampling method considering the distance between the samples on the input coordinates, in the present invention, an optimized sampling method of taking a representative value by grouping input given by region Solve the problem. That is, a negative sample without food poisoning is not used as it is, but a preprocessing process is performed. For example, a sample of a city or municipality unit is tied to a unit of trial to take an average value of each property. As a result, 226 samples of a specific date are reduced to 17, and the thus-processed negative samples can be reduced from 1,076,158 to 76,561.

The nonlinear regression model learning unit 23 is for learning a model with learning data for model learning. As described above, the safety diagnosis model utilizes a support vector regression model as a nonlinear regression model. In the case of the conventional linear relationship, logarithmic linear relation, or autoregressive model, the correlation between the input and the output must be predetermined and modeled, while the nonlinear regression model according to the present invention finds correlation patterns through learning without such constraints . That is, unlike the conventional method, input variables can be directly added without assuming a correlation, so that modeling of a new input is easy.

Equation (1) below expresses a support vector regression model (SVR) as a safety diagnosis model of the present invention.

,

st

,

The nonlinear regression model learning unit 23 according to the present invention learns the support vector regression model using the scikit-learn SVM library, which is a python machine learning toolkit. When the model learning is finished with the given learning data, the test samples are put into the learned model formula and the result is calculated. The resulting value is the number of daily food poisonings as defined above, and the higher the value, the higher the likelihood of food poisoning.

As described above, the conventional food poisoning prediction model can not be directly modeled in city or municipal district due to the problem that food poisoning occurrence samples are very insufficient. Therefore, the conventional food poisoning prediction model has a problem that the input value of the local weather information is not used as it is, but it is used as an average value and it is difficult to accurately reflect the weather condition at the time of food poisoning occurrence. However, It is possible to construct a nonlinear model in which the conditions of the time are accurately reflected.

The model verification unit 25 is used for verifying the predicted accuracy of the learned model with the test data. The model verification unit 25 verifies the learning data itself in order to determine how well the support vector regression model, which is the learned nonlinear regression model, And verification of the learned model. Ideally, a value of 1 or more should be given if the actual sample of the food poisoning occurred, or a value of 0 if the sample is not included. Therefore, it is assumed that food poisoning occurs if the value is 0.5 or more based on 0.5, which is the midpoint between 0 and 1, and not occurs when the value is 0.5 or less.

 FIG. 3 shows the model fitting results for the learning set of the safety diagnosis model according to the present invention learned from the data from May 1, 2006 to April 30, 2012. The model calculations for each sample are divided into categories and counted, and are shown as normalized histograms. The blue histogram on the left is the result of the negative samples and the green histogram on the right is the result of the positive samples. The precision of the negative samples is 90.8% and the precision of the positive samples is 79% Able to know.

Fig. 4 shows the result of prediction of the model learned from the data of 2006/05/01 to 2012/04/30 with the data of the next four months. There is no obvious difference as shown in FIG. 3, but there is a significant difference in distribution. For example, 26.26% of positive samples have a result of 0.35 or more, and 19.48% of negative samples have a result of 0.35 or more.

The safety criterion estimating section 27 calculates the safety diagnosis level according to the safety diagnosis index based on the verification result of the model verifying section 25. [ Here, the safety diagnosis index and the safety diagnosis level of the ingredients are indexes indicating how safe the food ingredients are in the area, including the food poisoning index and the like.

Table 3 below illustrates the level of safety diagnosis according to the safety diagnosis index according to an embodiment of the present invention.

step Judgment Criteria warning [Safety Diagnostic Index]> = 0.5 AND [Rate of sub-region where the exponent value is 0.5 or more]> 0 caution [Safety Diagnosis Index]> = 0.25 AND [Safety Diagnosis Index] <0.5 AND [Rate of Sub Area with Index Value of 0.5 or More]> 0 Attention (Safety Diagnostic Index) <0.25 OR [Percentage of subregion where index value is 0.5 or more] <= 0) AND [Safety Diagnostic Index]> = 0.175 normal [Safety Diagnostic Index] <0.175

It should be noted that the safety diagnostic level in Table 3 is based on the optimized safety diagnostic level derived through repeated experiments and can be changed according to the learning result and the verification result.

The safety index calculating unit 30 receives data on the cultivation environment and weather of each region and calculates the safety index of the foodstuffs according to the region based on the safety diagnosis model derived from the nonlinear regression modeling unit 20, To determine the diagnostic level. For example, the food safety index is calculated by inputting the weather information (daily maximum temperature, daily minimum temperature, daily average temperature, humidity, precipitation), information on livestock, and flood information in the safety diagnosis model of the present invention. If the value is greater than 0.5, the safety diagnostic level is determined to be 'Warning' for 'Warning' 0.25 to 0.5, 'Interest' for 0.175 to 0.25, and 'Normal' if less than 0.175.

Since the safety diagnosis model according to the present invention is calculated on a regional basis, for example, in a city or municipal district, when a food safety index is calculated in units of a trial, the result of a city or municipal unit is collectively grouped into trial units.

Table 4 below shows the results of the food safety diagnosis from June 1, 2016 to June 30, 2016, which is the result of verifying whether food poisoning actually occurred among the areas judged to be abnormal.

Verification item value Recall (positive) 0.5 Precision (positive) 0.06779661016949153 Recall (negative) 0.6381578947368421 Precision (negative) 0.9603960396039604 Micro precision / recall 0.63125

As a result of the above verification, it can be confirmed that the determination result according to the safety diagnosis level constructed in the present invention is very accurate.

As described above, according to another embodiment of the present invention, it is possible to further consider the agricultural product information and the agricultural product distribution information corresponding to the occurrence of the flooding, and apply the risk of contamination of agricultural products consumed in the area to the model, The diagnostic index can be calculated. In addition, it is possible to diagnose the risk of food ingredients in local foodservice stations by taking into account the information on agricultural products and the information on food provided in the foodservice stations.

FIG. 5 is a flowchart illustrating a method of calculating the safe index of foodstuffs according to a second embodiment of the present invention using the regional food safety index calculating device of FIG. 1, and FIG. 6 is a detailed view of step S10 of FIG. Hereinafter, a description overlapping with the above-described embodiment will be omitted.

Referring to FIG. 5, a safety diagnostic model is derived by constructing a nonlinear regression model on the correlation between the information about the cultivation environment and the weather of the foodstuff and the food poisoning occurrence information (S10).

Referring to FIG. 6, the nonlinear regression modeling unit 20 according to the present invention collects and stores region-specific cultivation environment information, weather information, and / or food poisoning information on the ingredients in the data collection unit 10 (S11). Here, the cultivating environment information includes the local flood information, the agricultural product information, and the livestock farming information, and the regional weather information includes the temperature information, the humidity information, and the precipitation information. The information source may be a website or SNS or news where relevant information is disclosed.

The data collected by the database unit 21 is constructed as learning data for learning and test data for testing (S21). At this time, in order to solve the problem of data imbalance in the food poisoning occurrence data, Sampling is performed by taking a representative value.

The nonlinear regression model learning unit 23 learns a nonlinear regression model using the thus-sampled learning data (S23).

Then, the nonlinear regression model, that is, the safety diagnosis model, which has been learned through the model verifying unit 25, is verified (S25). Then, the safety criterion estimating section 27 selects a criterion for classifying the local food safety index derived by the learned nonlinear regression model into the safety diagnosis level (S27).

The safety index calculating unit 30 calculates the food safety index of the corresponding date by using the safety diagnosis model according to the present invention derived by the procedure of FIG. 6, wherein the weather information and the cultivation environment information corresponding to the input variables And applied to the safety diagnosis model to calculate the food safety index of the corresponding date (S11).

Thus, although the conventional food poisoning risk index calculation model only considers temperature and humidity, the present invention calculates a more accurate food safety index by considering various data that may affect the contamination of foodstuffs such as immersion status and livestock farming information .

Meanwhile, the method for calculating the food material safety index according to the present invention can also be achieved by a computer-readable recording medium having recorded thereon a program therefor.

While the present invention has been described in connection with what is presently considered to be the most practical and preferred embodiment, it is to be understood that the invention is not limited to the disclosed embodiments. That is, within the scope of the present invention, all of the components may be selectively coupled to one or more of them. In addition, although all of the components may be implemented as one independent hardware, some or all of the components may be selectively combined to perform a part or all of the functions in one or a plurality of hardware. As shown in FIG. The codes and code segments constituting the computer program may be easily deduced by those skilled in the art. Such a computer program can be stored in a computer-readable storage medium, readable and executed by a computer, thereby realizing an embodiment of the present invention. As the storage medium of the computer program, a magnetic recording medium, an optical recording medium, a carrier wave medium, or the like may be included.

Furthermore, the terms "comprises", "comprising", or "having" described above mean that a component can be implanted unless otherwise specifically stated, But should be construed as including other elements. All terms, including technical and scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs, unless otherwise defined. Commonly used terms, such as predefined terms, should be interpreted to be consistent with the contextual meanings of the related art, and are not to be construed as ideal or overly formal, unless expressly defined to the contrary.

The foregoing description is merely illustrative of the technical idea of the present invention, and various changes and modifications may be made by those skilled in the art without departing from the essential characteristics of the present invention. Therefore, the embodiments disclosed in the present invention are intended to illustrate rather than limit the scope of the present invention, and the scope of the technical idea of the present invention is not limited by these embodiments. The scope of protection of the present invention should be construed according to the following claims, and all technical ideas within the scope of equivalents should be construed as falling within the scope of the present invention.

10: data collecting unit 20: nonlinear regression modeling unit
21: Database part 23: Nonlinear regression model learning part
25: Model verification part 27: Safety standard Acquisition part
30: Safety Index Calculation Section

Claims (18)

In calculating the regional food safety index,
Establishing a nonlinear regression model on the correlation between information on food environment and regional climate and information on the occurrence of food poisoning, and deriving a safety diagnosis model; And
Calculating a regional food safety index based on the safety diagnostic model;
The safety diagnosis model is constructed by using a nonlinear regression model so that the possibility of food contamination in the cultivation environment is reflected in the calculation of the food safety index, using the scale of the local farm farmers and the presence or absence of past flooding as input variables. Way.
The method according to claim 1,
Wherein the nonlinear regression model includes a support vector regression model (SVR) or a least squares support vector machine (LS-SVM).
The method according to claim 1,
Wherein the regional weather information includes temperature information, humidity information, and precipitation information.
The method according to claim 1,
Wherein the food safety index includes a food poisoning index.
The method according to claim 1,
The step of deriving the safety diagnosis model includes the steps of converting the cultivation environment, weather and food poisoning occurrence information for each food material into data for learning;
Selecting samples for building the nonlinear regression model; And
And learning the nonlinear regression model using the samples. &Lt; RTI ID = 0.0 &gt; 8. &lt; / RTI &gt;
6. The method of claim 5,
Wherein the step of selecting the samples includes a step of selecting a sample by taking an average value of food poisoning occurrence information per city unit as a unit of trial.
The method according to claim 1,
Wherein the information on the cultivation environment of each region further includes information on agricultural products by region.
A computer-readable recording medium recording a program for executing the method of calculating the food material safety index according to any one of claims 1 to 7.
In the regional food safety index calculating device,
Nonlinear regression modeling unit to derive a safety diagnosis model by constructing a nonlinear regression model on the correlation between information on food - related environmental conditions and regional climate and information on the occurrence of food poisoning; And
A safety index calculation unit for receiving data on a cultivation environment and weather of each region and calculating a safety index for each region based on the safety diagnosis model;
The safety diagnosis model is constructed by using a nonlinear regression model so that the possibility of food contamination in the cultivation environment is reflected in the calculation of the food safety index, using the scale of the local farm farmers and the presence or absence of past flooding as input variables. Device.
10. The method of claim 9,
Wherein the nonlinear regression model is based on a support vector regression model (SVR) or a least squares support vector machine (LS-SVM).
11. The method according to claim 9 or 10,
And a data collecting unit for collecting information on the cultivation environment, weather information, and food poisoning occurrence information about the local ingredients of the food.
12. The method of claim 11,
Wherein the nonlinear regression modeling unit comprises: a database unit for constructing information collected by the data collecting unit as learning data for model learning and test data for model testing;
A nonlinear regression model learning unit for learning the nonlinear regression model using the learning data; And
And a model verifying unit for verifying the nonlinear regression model using test data.
13. The method of claim 12,
Wherein the database unit constructs the learning data and the test data by taking an average value of the food poisoning occurrence information per city unit in units of a trial.
13. The method of claim 12,
Wherein the nonlinear regression modeling unit includes a safety criterion calculating unit for calculating a criterion of a safety diagnosis level according to the food safety index.
11. The method according to claim 9 or 10,
Wherein the regional weather information includes temperature information, humidity information, and rainfall information.
13. The method of claim 12,
Wherein the food poisoning occurrence information includes information on a food poisoning occurrence area, cause food, causative bacteria, occurrence place, number of incidents, number of patients, and number of consumed persons.
10. The method of claim 9,
Wherein the local flood information includes information on a flooded area, a source of flooding, a kind of flooded water, a use of a flooded field, and a flood time.
11. The method according to claim 9 or 10,
Wherein the food safety index includes a food poisoning index.

Images