CN112700882B - Self-adaptive early warning method and robot for infectious diseases based on big data artificial intelligence - Google Patents

Abstract

An adaptive early warning method and a robot for infectious diseases based on big data artificial intelligence comprise the following steps: a target infectious disease acquisition step; a target infectious disease first condition acquisition step; a first adjusting step of early warning threshold value; and a second adjusting step of early warning threshold value. According to the method, the system and the robot, when early warning is carried out, the early warning threshold value is set according to the accuracy of target infectious disease monitoring and the severity of target infectious disease development, so that early warning can be adaptively adjusted according to the accuracy of target infectious disease monitoring and the severity of target infectious disease development, the early warning effect is improved, and the early warning can control and inhibit the infectious disease development to the greatest extent.

Description

Adaptive early warning method and robot for infectious diseases based on big data artificial intelligence

technical field

The invention relates to the technical field of artificial intelligence, in particular to an adaptive early warning method and robot for infectious diseases based on big data artificial intelligence.

Background technique

In the process of realizing the present invention, the inventors found that there are at least the following problems in the prior art: under the prior art, because of the peak period of the development of new major infectious diseases, medical resources are bound to be scarce, which will cause medical problems due to the fixed warning threshold. Suffering from conflicts and even social conflicts.

Therefore, the prior art still needs to be improved and developed.

Contents of the invention

Based on this, it is necessary to address the defects or insufficiencies of the existing technology, and provide an adaptive early warning method and robot for infectious diseases based on big data artificial intelligence, so as to solve the problem of conflicts between doctors and patients and even social conflicts caused by fixed warning thresholds in the prior art .

In a first aspect, an embodiment of the present invention provides an artificial intelligence method, the method comprising:

Target infectious disease acquisition step: acquire the infectious disease that needs early warning as the target infectious disease;

Steps to obtain the first status of the target infectious disease: obtain the accuracy of target infectious disease monitoring and the severity of the development of the target infectious disease;

The first adjustment step of the early warning threshold: adopt the corresponding early warning threshold in the early warning system according to the accuracy of target infectious disease monitoring;

The second adjustment step of the early warning threshold: adopt the corresponding early warning threshold in the early warning system according to the severity of the development of the target infectious disease.

Preferably, the method also includes:

Reinforcement learning step: Obtain user feedback after early warning. If the user feedback early warning is correct, use the user's data and required early warning as input and expected output to carry out supervised training on the early warning model, and use the user's data and confirmed infection Infectious diseases are used as input and expected output to carry out supervised training on the monitoring model; if the user feedbacks that the early warning is incorrect, the user's data and no early warning are used as input and expected output to carry out supervised training on the early warning model, and the user The data and the confirmed non-infected infectious disease are used as input and expected output to supervise the training of the monitoring model; the early warning model includes a deep learning neural network model or a convolutional neural network model or other neural network models or robot learning models; the monitoring Models include deep learning neural network models or convolutional neural network models or other neural network models or robot learning models.

Preferably, the method also includes:

Second adjustment step: Obtain the change of monitoring accuracy before and after the adjustment of the warning threshold, and the change of warning accuracy. If the monitoring accuracy or warning accuracy both decrease, the warning threshold is called back at a preset ratio.

Preferably,

The secondary adjustment step specifically includes: the preset ratio is a fraction less than 1, if the warning threshold has recently increased by the first preset value, multiplying the first preset value by the preset ratio as the second preset value, and then Decrease the warning threshold by the second preset value; if the warning threshold has recently decreased by the first preset value, multiply the first preset value by the preset ratio as the second preset value, and then increase the warning threshold by the second preset value value.

Preferably, the method also includes:

The step of obtaining the second status of the target infectious disease: obtaining the tightness of medical resources related to the target infectious disease and the priority of the user; the user includes an individual or a family or a department;

The third adjustment step of the early warning threshold: adopt the corresponding early warning threshold in the early warning system according to the tightness of medical resources related to the target infectious disease;

The fourth adjustment step of the warning threshold: adopting the corresponding warning threshold in the warning system according to the user's priority.

Preferably, the method also includes:

The first adjustment step of the early warning threshold by region: for each region, adopt the corresponding early warning threshold for each region in the early warning system according to the accuracy of the target infectious disease monitoring in each region;

The second adjustment step of the subregional early warning threshold: for each area, adopt the corresponding early warning threshold for each area in the early warning system according to the severity of the development of the target infectious disease in each area;

The third adjustment step of the regional early warning threshold: for each region, adopt the corresponding early warning threshold for each region in the early warning system according to the degree of strain on the target infectious disease-related medical resources in each region;

The fourth adjustment step of the early warning threshold by area: for each area, adopt the corresponding early warning threshold for each area in the early warning system according to the priority of users in each area.

Preferably, it is characterized in that,

The first adjustment step of the early warning threshold specifically includes: if the accuracy rate of target infectious disease monitoring decreases, then lower the early warning threshold; if the target infectious disease monitoring accuracy rate increases, then increase the early warning threshold; if the target infectious disease monitoring accuracy rate is lower than the accurate rate preset threshold, lower the early warning threshold to the sensitive early warning range; if the target infectious disease monitoring accuracy is higher than the accuracy preset threshold, increase the early warning threshold to the insensitive early warning range;

The second adjustment step of the early warning threshold specifically includes: if the severity of the development of the target infectious disease decreases, then increase the early warning threshold; if the severity of the development of the target infectious disease increases, then reduce the early warning threshold; If the severity preset threshold is lowered, the early warning threshold is lowered to the sensitive early warning range; if the severity of the development of the target infectious disease is lower than the severity preset threshold, the early warning threshold is raised to the insensitive early warning range.

Preferably,

The third adjustment step of the early warning threshold specifically includes: if the tightness of medical resources related to the target infectious disease decreases, lower the early warning threshold; if the tightness of medical resources related to the target infectious disease increases, increase the early warning threshold; if the tightness of medical resources related to the target infectious disease If the tension is higher than the preset threshold, the early warning threshold is increased to the insensitive early warning range; if the target infectious disease-related medical resource tension is lower than the tense preset threshold, the early warning threshold is lowered to the sensitive early warning range;

The fourth adjustment step of the early warning threshold specifically includes: if the user's priority decreases, then increase the early warning threshold; if the user's priority increases, then reduce the early warning threshold; if the user's priority is higher than the priority preset threshold, then reduce the early warning threshold to the sensitive warning range; if the user's priority is lower than the priority preset threshold, the warning threshold is increased to the insensitive warning range;

The first adjustment step of the early warning threshold, the second adjustment step of the early warning threshold, the third adjustment step of the early warning threshold and the fourth adjustment step of the early warning threshold also specifically include: if the early warning threshold is in the sensitive early warning range, the credibility of the suspected infection of the target infectious disease When the threshold is low, it will also be warned; if the warning threshold is in the insensitive warning range, it will also be warned when the credibility of the suspected infection of the target infectious disease is high; if the credibility of the suspected infection of the target infectious disease is higher than the credible If the degree preset threshold, the confidence of suspected infection of the target infectious disease is high, otherwise it is low.

In a second aspect, an embodiment of the present invention provides an artificial intelligence system, the system comprising:

Target infectious disease acquisition module: acquire infectious diseases that require early warning as target infectious diseases;

Obtaining the first status of the target infectious disease module: Acquiring the accuracy of target infectious disease monitoring and the severity of the target infectious disease development;

The first adjustment module of the early warning threshold: adopt the corresponding early warning threshold in the early warning system according to the accuracy of target infectious disease monitoring;

The second adjustment module of the early warning threshold: according to the severity of the development of the target infectious disease, the corresponding early warning threshold is adopted in the early warning system.

Preferably, the system also includes:

Reinforcement learning module: Obtain user feedback after early warning. If the user feedback early warning is correct, the user's data and required early warning will be used as input and expected output to carry out supervised training on the early warning model, and the user's data and confirmed infection Infectious diseases are used as input and expected output to carry out supervised training on the monitoring model; if the user feedbacks that the early warning is incorrect, the user's data and no early warning are used as input and expected output to carry out supervised training on the early warning model, and the user The data and the confirmed non-infected infectious disease are used as input and expected output to supervise the training of the monitoring model; the early warning model includes a deep learning neural network model or a convolutional neural network model or other neural network models or robot learning models; the monitoring Models include deep learning neural network models or convolutional neural network models or other neural network models or robot learning models.

Preferably, the system also includes:

Secondary adjustment module: Obtain the changes in monitoring accuracy and early warning accuracy before and after the adjustment of the early warning threshold. If the monitoring accuracy or early warning accuracy both decrease, the early warning threshold will be called back at a preset ratio.

Preferably,

The secondary adjustment module specifically includes: the preset ratio is a fraction less than 1, if the early warning threshold value has recently increased the first preset value, the first preset value is multiplied by the preset ratio as the second preset value, and then Decrease the warning threshold by the second preset value; if the warning threshold has recently decreased by the first preset value, multiply the first preset value by the preset ratio as the second preset value, and then increase the warning threshold by the second preset value value.

Preferably, the system also includes:

The second status acquisition module of the target infectious disease: acquire the tightness of medical resources related to the target infectious disease and the priority of the user; the user includes an individual or a family or a department;

The third adjustment module of early warning threshold: adopt the corresponding early warning threshold in the early warning system according to the tightness of medical resources related to the target infectious disease;

The fourth adjustment module of the warning threshold: according to the priority of the user, the corresponding warning threshold is adopted in the warning system.

Preferably, the system also includes:

Subregional early warning threshold first adjustment module: for each area, adopt the corresponding early warning threshold for each area in the early warning system according to the accuracy of target infectious disease monitoring in each area;

Subregional early warning threshold second adjustment module: for each area, adopt the corresponding early warning threshold for each area in the early warning system according to the severity of the development of the target infectious disease in each area;

The third adjustment module of the regional early warning threshold: for each region, adopt the corresponding early warning threshold for each region in the early warning system according to the degree of strain on the target infectious disease-related medical resources in each region;

The fourth adjustment module of the early warning threshold by area: for each area, adopt the corresponding early warning threshold for each area in the early warning system according to the priority of users in each area.

Preferably,

The first adjustment module of the early warning threshold specifically includes: if the accuracy of target infectious disease monitoring decreases, then reduce the early warning threshold; if the accuracy of target infectious disease monitoring increases, then increase the early warning threshold; if the target infectious disease monitoring accuracy is lower than the accurate rate preset threshold, lower the early warning threshold to the sensitive early warning range; if the target infectious disease monitoring accuracy is higher than the accuracy preset threshold, increase the early warning threshold to the insensitive early warning range;

The second adjustment module of the early warning threshold specifically includes: if the severity of the development of the target infectious disease decreases, then increase the early warning threshold; if the severity of the development of the target infectious disease increases, then reduce the early warning threshold; If the severity preset threshold is lowered, the early warning threshold is lowered to the sensitive early warning range; if the severity of the development of the target infectious disease is lower than the severity preset threshold, the early warning threshold is raised to the insensitive early warning range.

Preferably,

The third adjustment module of the early warning threshold specifically includes: if the tightness of medical resources related to the target infectious disease decreases, then reduce the early warning threshold; if the tightness of medical resources related to the target infectious disease increases, increase the early warning threshold; if the tightness of medical resources related to the target infectious disease If the tension is higher than the preset threshold, the early warning threshold is increased to the insensitive early warning range; if the target infectious disease-related medical resource tension is lower than the tense preset threshold, the early warning threshold is lowered to the sensitive early warning range;

The fourth adjustment module of the early warning threshold specifically includes: if the user's priority decreases, then increase the early warning threshold; if the user's priority increases, then reduce the early warning threshold; if the user's priority is higher than the priority preset threshold, then reduce the early warning threshold to the sensitive warning range; if the user's priority is lower than the priority preset threshold, the warning threshold is increased to the insensitive warning range;

The first warning threshold adjustment module, the second warning threshold adjustment module, the third warning threshold adjustment module, and the fourth warning threshold adjustment module also specifically include: if the warning threshold is in the sensitive warning range, the credibility of the suspected infection of the target infectious disease When the threshold is low, it will also be warned; if the warning threshold is in the insensitive warning range, it will also be warned when the credibility of the suspected infection of the target infectious disease is high; if the credibility of the suspected infection of the target infectious disease is higher than the credible If the degree preset threshold, the confidence of suspected infection of the target infectious disease is high, otherwise it is low.

In a third aspect, an embodiment of the present invention provides an artificial intelligence device, and the device includes modules of the system described in any one of the embodiments of the second aspect.

In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of any one of the methods described in the embodiments of the first aspect are implemented.

In the fifth aspect, an embodiment of the present invention provides a robot, including a memory, a processor, and an artificial intelligence robot program stored in the memory and operable on the processor, and the processor implements the first aspect when executing the program. Examples of the steps of any one of the described methods.

The self-adaptive early warning method and robot for infectious diseases based on big data artificial intelligence provided in this embodiment include: the step of obtaining the target infectious disease; the step of obtaining the first condition of the target infectious disease; the first adjustment step of the early warning threshold; the second adjustment of the early warning threshold step. The above-mentioned method, system and robot, when performing early warning, also set the early warning threshold according to the accuracy of target infectious disease monitoring and the severity of the development of the target infectious disease, so that the early warning can be adaptively based on the accuracy of target infectious disease monitoring , Adjust the severity of the development of the target infectious disease, so as to improve the effect of early warning, and enable the early warning to control and curb the development of infectious diseases to the greatest extent.

Description of drawings

Fig. 1 is the flowchart of the artificial intelligence method that the embodiment of the present invention provides;

Fig. 2 is the flowchart that the artificial intelligence method that the embodiment of the present invention provides includes;

Fig. 3 is the flow chart that the artificial intelligence method provided by the embodiment of the present invention includes;

Fig. 4 is a schematic diagram of the early warning adjustment of infectious diseases based on big data provided by the embodiment of the present invention.

Detailed ways

The technical solutions in the embodiments of the present invention will be described in detail below in conjunction with the embodiments of the present invention.

Basic embodiment of the invention

An artificial intelligence method provided by an embodiment of the present invention, as shown in Figure 1, said method includes: the step of acquiring the target infectious disease; the step of acquiring the first condition of the target infectious disease; the first adjustment step of the early warning threshold; the second early warning threshold Adjustment steps. Technical effect: the existing method only considers the monitoring results of the target infectious disease during the early warning, while the method described in this embodiment also sets the early warning threshold according to the accuracy of the target infectious disease monitoring and the severity of the development of the target infectious disease when performing the early warning Therefore, the early warning can be adaptively adjusted according to the accuracy of target infectious disease monitoring and the severity of target infectious disease development, thereby improving the effect of early warning and enabling early warning to control and curb the development of infectious diseases to the greatest extent.

Preferably, the method further includes: a reinforcement learning step. Technical effect: In the existing method, early warning and monitoring are two different links, each of which uses a different model, and has no relationship with each other. The method described in this embodiment continuously improves the prediction effect of the monitoring model through the feedback of the early warning, so that the early warning and the monitoring promote each other, and the effects of the monitoring and the early warning are continuously improved.

Preferably, the method further includes: a secondary adjustment step. Technical effect: the method avoids problems caused by over-regulation through a callback mechanism, so that the adjustment effect of the warning threshold is better.

Preferably, as shown in FIG. 2 , the method further includes: a second condition acquiring step of the target infectious disease; a third adjusting step of the warning threshold; and a fourth adjusting step of the warning threshold. Technical effect: The method further considers the degree of tension of medical resources related to the target infectious disease and the priority of users to adjust the warning threshold to further improve the adaptability of the warning, so that the warning can adapt to various situations and achieve the best results. Good warning effect.

Preferably, as shown in FIG. 3 , the method further includes: a first adjustment step of the early warning threshold by region; a second adjustment step of the early warning threshold by region; a third adjustment step of the early warning threshold by region; a fourth adjustment step of the early warning threshold by region . Technical effect: the method further improves the accuracy of early warning by carrying out early warning for each area, because different areas have different conditions, so early warning for different areas can achieve better early warning effects, and can make the early warning adapt to the situation of different areas . .

Preferred Embodiments of the Invention

Target infectious disease acquisition step: acquire the infectious disease that needs early warning as the target infectious disease;

Steps to obtain the first status of the target infectious disease: obtain the accuracy of target infectious disease monitoring and the severity of the development of the target infectious disease;

The first adjustment step of the early warning threshold: adopt the corresponding early warning threshold in the early warning system according to the accuracy of target infectious disease monitoring;

The second adjustment step of the early warning threshold: adopt the corresponding early warning threshold in the early warning system according to the severity of the development of the target infectious disease;

Reinforcement learning step: Obtain user feedback after early warning. If the user feedback early warning is correct, use the user's data and required early warning as input and expected output to carry out supervised training on the early warning model, and use the user's data and confirmed infection Infectious diseases are used as input and expected output to carry out supervised training on the monitoring model; if the user feedbacks that the early warning is incorrect, the user's data and no early warning are used as input and expected output to carry out supervised training on the early warning model, and the user The data and the confirmed non-infected infectious disease are used as input and expected output to supervise the training of the monitoring model; the early warning model includes a deep learning neural network model or a convolutional neural network model or other neural network models or robot learning models; the monitoring Models include deep learning neural network models or convolutional neural network models or other neural network models or robotic learning models;

Secondary adjustment step: Obtain the change of monitoring accuracy before and after the adjustment of the warning threshold, and the change of warning accuracy. If the monitoring accuracy or warning accuracy both decrease, the warning threshold is called back at a preset ratio;

The secondary adjustment step specifically includes: the preset ratio is a fraction less than 1, if the warning threshold has recently increased by the first preset value, multiplying the first preset value by the preset ratio as the second preset value, and then Decrease the warning threshold by the second preset value; if the warning threshold has recently decreased by the first preset value, multiply the first preset value by the preset ratio as the second preset value, and then increase the warning threshold by the second preset value value;

The step of obtaining the second status of the target infectious disease includes: obtaining the degree of strain on medical resources related to the target infectious disease and the priority of the user; the user includes an individual or a family or a department;

The third adjustment step of the early warning threshold: adopt the corresponding early warning threshold in the early warning system according to the tightness of medical resources related to the target infectious disease;

The fourth adjustment step of the early warning threshold: adopt the corresponding early warning threshold in the early warning system according to the priority of the user;

The first adjustment step of the early warning threshold by region: for each region, adopt the corresponding early warning threshold for each region in the early warning system according to the accuracy of the target infectious disease monitoring in each region;

The second adjustment step of the subregional early warning threshold: for each area, adopt the corresponding early warning threshold for each area in the early warning system according to the severity of the development of the target infectious disease in each area;

The third adjustment step of the regional early warning threshold: for each region, adopt the corresponding early warning threshold for each region in the early warning system according to the degree of strain on the target infectious disease-related medical resources in each region;

The fourth adjustment step of the early warning threshold by area: for each area, adopt the corresponding early warning threshold for each area in the early warning system according to the priority of users in each area;

The first adjustment step of the early warning threshold specifically includes: if the accuracy rate of target infectious disease monitoring decreases, then lower the early warning threshold; if the target infectious disease monitoring accuracy rate increases, then increase the early warning threshold; if the target infectious disease monitoring accuracy rate is lower than the accurate rate preset threshold, lower the early warning threshold to the sensitive early warning range; if the target infectious disease monitoring accuracy is higher than the accuracy preset threshold, increase the early warning threshold to the insensitive early warning range;

The second adjustment step of the early warning threshold specifically includes: if the severity of the development of the target infectious disease decreases, then increase the early warning threshold; if the severity of the development of the target infectious disease increases, then reduce the early warning threshold; If the severity preset threshold is lower than the threshold, the early warning threshold is lowered to the sensitive early warning range; if the severity of the development of the target infectious disease is lower than the severity preset threshold, the early warning threshold is increased to the insensitive early warning range;

The third adjustment step of the early warning threshold specifically includes: if the tightness of medical resources related to the target infectious disease decreases, lower the early warning threshold; if the tightness of medical resources related to the target infectious disease increases, increase the early warning threshold; if the tightness of medical resources related to the target infectious disease If the tension is higher than the preset threshold, the early warning threshold is increased to the insensitive early warning range; if the target infectious disease-related medical resource tension is lower than the tense preset threshold, the early warning threshold is lowered to the sensitive early warning range;

The fourth adjustment step of the early warning threshold specifically includes: if the user's priority decreases, then increase the early warning threshold; if the user's priority increases, then reduce the early warning threshold; if the user's priority is higher than the priority preset threshold, then reduce the early warning threshold to the sensitive warning range; if the user's priority is lower than the priority preset threshold, the warning threshold is increased to the insensitive warning range;

The first adjustment step of the early warning threshold, the second adjustment step of the early warning threshold, the third adjustment step of the early warning threshold and the fourth adjustment step of the early warning threshold also specifically include: if the early warning threshold is in the sensitive early warning range, the credibility of the suspected infection of the target infectious disease When the threshold is low, it will also be warned; if the warning threshold is in the insensitive warning range, it will also be warned when the credibility of the suspected infection of the target infectious disease is high; if the credibility of the suspected infection of the target infectious disease is higher than the credible If the degree preset threshold, the confidence of suspected infection of the target infectious disease is high, otherwise it is low.

Other embodiments of the invention

Because of the peak period of the development of new major infectious diseases, medical resources are bound to be scarce. How to adjust the early warning threshold to alleviate the contradiction between doctors and patients and even social conflicts caused by the lack of medical resources? The method to be adopted in this embodiment is to improve the accuracy of system monitoring and early warning on the one hand, so that people who are not sick will not be misjudged and go to the hospital for examination to occupy medical resources. The level of epidemic development, user attributes, etc. are used to dynamically adjust the early warning threshold to balance medical resources and diagnosis and treatment needs, so that they can be intelligently matched.

Adaptive early warning of emerging major infectious diseases based on big data: Different early warning thresholds are used at different development stages of the system and emerging major infectious diseases. First of all, at the system level, when the system monitoring accuracy is not high, the threshold should be lowered as much as possible, so as not to delay the early warning of new major infectious diseases; when the system monitoring accuracy is gradually improved, the threshold can be gradually increased to reduce the follow-up medical examination caused by the early warning cost and the cost of causing panic to society. Secondly, at the level of the development process of new major infectious diseases, when the new major infectious diseases are not serious, the risk of infection is low at this time, and the threshold can be raised to avoid interference with people's life and work; During the period, the threshold needs to be lowered for early detection, early isolation and treatment. At the same time, the adjustment of the warning threshold should also take into account the epidemic situation in each region, the current resource shortage in each region, and the different attributes of each user. Figure 4 shows a schematic diagram of the early warning adjustment of new major infectious diseases based on big data:

Feedback adjustment of early warning threshold for new major infectious diseases based on big data: using reinforcement learning method based on big data, verify the accuracy of monitoring according to the feedback after response, and adjust the early warning threshold according to the accuracy. When the accuracy rate is high, the early warning threshold is increased, because the monitoring is more accurate at this time, so the early warning is also more accurate, and the probability of missing infected users is small. When the accuracy rate is not high, set the early warning threshold lower, because the monitoring is not very accurate at this time, so the early warning is not very accurate. In order to avoid omissions, I would rather make some false positives (some false negatives that are not infected are infected), Try to ensure that infected users are not missed. Therefore, in the initial stage of system use, the early warning threshold should be set lower, and then with the use of the system, more response feedback can be used to correct the monitoring model, so that the monitoring becomes more and more accurate, and then the early warning threshold can be gradually increased. When the early warning threshold is increased, the accuracy rate of monitoring and early warning does not decrease, which indicates that the threshold is raised appropriately. The threshold will be slightly lowered until the accuracy of monitoring and early warning will not decrease due to the increase of the threshold. The reason why the increase in the threshold will lead to a decline in the accuracy of monitoring and early warning is because if the threshold is increased too much, more users will inevitably be missed, and these users will be found to be infected later, which will lead to the accuracy of monitoring and early warning. Of course, there will be a certain delay in the feedback process, so the adjustment of the monitoring and early warning threshold is a process of gradual optimization from coarse to fine.

Dynamic adjustment of the early warning threshold of new major infectious diseases based on big data: New major infectious diseases generally have a development cycle, which is divided into occurrence, development, peak, low peak, and extinction (if prevention is relaxed during the extinction stage, it may lead to resurgence , and re-enter the development stage) several stages, then through big data analysis can determine which stage it is currently in, and dynamically adjust the early warning threshold according to the stage it is in. In the occurrence stage, the early warning threshold should be set lower, so that mild symptoms can be screened out for medical treatment, so that the signs of new major infectious diseases can be stopped. If the prevention is good, then it will not enter the later stage. . If it is not stopped, it will enter the development or even peak period. In these two stages, the early warning threshold needs to be set properly, because if the early warning threshold is set too low, many infected people will be missed, which will cause even more infection. Large-scale transmission, and if the setting is too high, it will cause a shortage of medical resources, because more people have become sick during the development and peak stages, so medical resources must be very tight, if the early warning threshold is set too low, There will be many people with very low risk of illness who will also occupy medical resources, which will inevitably lead to a shortage of medical resources, and the inability to treat those who are really sick, and even lead to social chaos. At the same time, the adjustment of the warning threshold should also take into account the epidemic situation in each region, the current resource shortage in each region, and the different attributes of each user. The dynamic adjustment of early warning thresholds for emerging major infectious diseases based on big data can be realized through a three-level early warning and response rule base based on intelligent classification for emerging major infectious diseases, and a three-level early warning based on intelligent classification for emerging major infectious diseases The response rule base is given in the research plan for the response to major emerging infectious diseases based on big data.

The above-mentioned embodiments only express several implementation modes of the present invention, and the description thereof is relatively specific and detailed, but should not be construed as limiting the patent scope of the present invention. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent for the present invention should be based on the appended claims.

Claims (8)

1. an artificial intelligence method, is characterized in that, described method comprises: Target infectious disease acquisition step: acquire the infectious disease that needs early warning as the target infectious disease; Steps to obtain the first status of the target infectious disease: obtain the accuracy of target infectious disease monitoring and the severity of the development of the target infectious disease; The first adjustment step of the early warning threshold: adopt the corresponding early warning threshold in the early warning system according to the accuracy of target infectious disease monitoring; The second adjustment step of the early warning threshold: adopt the corresponding early warning threshold in the early warning system according to the severity of the development of the target infectious disease; The first adjustment step of the early warning threshold specifically includes: if the accuracy rate of target infectious disease monitoring decreases, then lower the early warning threshold; if the target infectious disease monitoring accuracy rate increases, then increase the early warning threshold; if the target infectious disease monitoring accuracy rate is lower than the accurate rate preset threshold, lower the early warning threshold to the sensitive early warning range; if the target infectious disease monitoring accuracy is higher than the accuracy preset threshold, increase the early warning threshold to the insensitive early warning range; The second adjustment step of the early warning threshold specifically includes: if the severity of the development of the target infectious disease decreases, then increase the early warning threshold; if the severity of the development of the target infectious disease increases, then reduce the early warning threshold; If the severity preset threshold is lower than the threshold, the early warning threshold is lowered to the sensitive early warning range; if the severity of the development of the target infectious disease is lower than the severity preset threshold, the early warning threshold is increased to the insensitive early warning range; At the system level, lower the threshold when the system monitoring accuracy is not high; increase the threshold when the system monitoring accuracy is improved; At the level of the development process of emerging major infectious diseases, increase the threshold during the period when the emerging major infectious disease is not serious; lower the threshold during the period when the emerging major infectious disease is severe; Using the method of reinforcement learning based on big data, the accuracy of monitoring is verified according to the feedback after the response, and the threshold of early warning is adjusted according to the accuracy; when the accuracy is high, the threshold of early warning is increased; The threshold is set low. With the use of the system, the early warning threshold is increased; when the early warning threshold is increased, the accuracy of monitoring and early warning decreases, and the threshold of monitoring and early warning needs to be lowered until the accuracy of monitoring and early warning will not be affected by the threshold. up and down; Reinforcement learning step: Obtain user feedback after early warning. If the user feedback early warning is correct, use the user's data and required early warning as input and expected output to carry out supervised training on the early warning model, and use the user's data and confirmed infection Infectious diseases are used as input and expected output to carry out supervised training on the monitoring model; if the user feedbacks that the early warning is incorrect, the user's data and no early warning are used as input and expected output to carry out supervised training on the early warning model, and the user The data and the confirmed non-infected infectious disease are used as input and expected output to supervise the training of the monitoring model; the early warning model includes a deep learning neural network model or a convolutional neural network model or other neural network models or robot learning models; the monitoring Models include deep learning neural network models or convolutional neural network models or other neural network models or robot learning models. 2. the artificial intelligence method according to claim 1, is characterized in that, described method also comprises: Second adjustment step: Obtain the change of monitoring accuracy before and after the adjustment of the warning threshold, and the change of warning accuracy. If the monitoring accuracy or warning accuracy both decrease, the warning threshold is called back at a preset ratio. 3. artificial intelligence method according to claim 1, is characterized in that, The secondary adjustment step specifically includes: the preset ratio is a fraction less than 1, if the warning threshold has recently increased by the first preset value, the first preset value is multiplied by the preset ratio as the second preset value, and then the warning threshold The threshold is decreased by a second preset value; if the warning threshold has recently decreased by the first preset value, the first preset value is multiplied by a preset ratio as the second preset value, and then the warning threshold is increased by the second preset value. 4. artificial intelligence method according to claim 1, is characterized in that, described method also comprises: The step of obtaining the second status of the target infectious disease: obtaining the tightness of medical resources related to the target infectious disease and the priority of the user; the user includes an individual or a family or a department; The third adjustment step of the early warning threshold: adopt the corresponding early warning threshold in the early warning system according to the tightness of medical resources related to the target infectious disease; The fourth adjustment step of the warning threshold: adopting the corresponding warning threshold in the warning system according to the user's priority. 5. artificial intelligence method according to claim 1, is characterized in that, described method also comprises: The first adjustment step of the early warning threshold by region: for each region, adopt the corresponding early warning threshold for each region in the early warning system according to the accuracy of the target infectious disease monitoring in each region; The second adjustment step of the subregional early warning threshold: for each area, adopt the corresponding early warning threshold for each area in the early warning system according to the severity of the development of the target infectious disease in each area; The third adjustment step of the regional early warning threshold: for each region, adopt the corresponding early warning threshold for each region in the early warning system according to the degree of strain on the target infectious disease-related medical resources in each region; The fourth adjustment step of the early warning threshold by area: for each area, adopt the corresponding early warning threshold for each area in the early warning system according to the priority of users in each area. 6. artificial intelligence method according to claim 4, is characterized in that, The third adjustment step of the early warning threshold specifically includes: if the tightness of medical resources related to the target infectious disease decreases, lower the early warning threshold; if the tightness of medical resources related to the target infectious disease increases, increase the early warning threshold; if the tightness of medical resources related to the target infectious disease If the tension is higher than the preset threshold, the early warning threshold is increased to the insensitive early warning range; if the target infectious disease-related medical resource tension is lower than the tense preset threshold, the early warning threshold is lowered to the sensitive early warning range; The fourth adjustment step of the early warning threshold specifically includes: if the user's priority decreases, then increase the early warning threshold; if the user's priority increases, then reduce the early warning threshold; if the user's priority is higher than the priority preset threshold, then reduce the early warning threshold to the sensitive warning range; if the user's priority is lower than the priority preset threshold, the warning threshold is increased to the insensitive warning range; The first adjustment step of the early warning threshold, the second adjustment step of the early warning threshold, the third adjustment step of the early warning threshold and the fourth adjustment step of the early warning threshold also specifically include: if the early warning threshold is in the sensitive early warning range, the credibility of the suspected infection of the target infectious disease When the threshold is low, it will also be warned; if the warning threshold is in the insensitive warning range, it will also be warned when the credibility of the suspected infection of the target infectious disease is high; if the credibility of the suspected infection of the target infectious disease is higher than the credible If the degree preset threshold, the confidence of suspected infection of the target infectious disease is high, otherwise it is low. 7. A robot comprising a memory, a processor and an artificial intelligence robot program stored on the memory and operable on the processor, wherein the processor implements any one of claims 1-6 when executing the program. steps of the method described in the item. 8. A computer-readable storage medium, on which a computer program is stored, wherein the program implements the steps of any one of claims 1-6 when executed by a processor.

Images