CN116564521A - Chronic disease risk assessment model establishment method, medium and system - Google Patents
Chronic disease risk assessment model establishment method, medium and system Download PDFInfo
- Publication number
- CN116564521A CN116564521A CN202310531805.8A CN202310531805A CN116564521A CN 116564521 A CN116564521 A CN 116564521A CN 202310531805 A CN202310531805 A CN 202310531805A CN 116564521 A CN116564521 A CN 116564521A
- Authority
- CN
- China
- Prior art keywords
- chronic disease
- disease risk
- risk assessment
- risk
- layer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 208000017667 Chronic Disease Diseases 0.000 title claims abstract description 175
- 238000012502 risk assessment Methods 0.000 title claims abstract description 103
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000012549 training Methods 0.000 claims abstract description 104
- 238000013528 artificial neural network Methods 0.000 claims abstract description 32
- 238000005457 optimization Methods 0.000 claims abstract description 8
- 230000006870 function Effects 0.000 claims description 42
- 201000010099 disease Diseases 0.000 claims description 34
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims description 34
- 235000005911 diet Nutrition 0.000 claims description 24
- 230000037213 diet Effects 0.000 claims description 24
- 230000001684 chronic effect Effects 0.000 claims description 22
- 230000001788 irregular Effects 0.000 claims description 18
- 230000004913 activation Effects 0.000 claims description 16
- 239000013598 vector Substances 0.000 claims description 13
- HVYWMOMLDIMFJA-DPAQBDIFSA-N cholesterol Chemical compound C1C=C2C[C@@H](O)CC[C@]2(C)[C@@H]2[C@@H]1[C@@H]1CC[C@H]([C@H](C)CCCC(C)C)[C@@]1(C)CC2 HVYWMOMLDIMFJA-DPAQBDIFSA-N 0.000 claims description 12
- 230000000694 effects Effects 0.000 claims description 12
- 238000011156 evaluation Methods 0.000 claims description 12
- 230000007774 longterm Effects 0.000 claims description 12
- 238000004364 calculation method Methods 0.000 claims description 10
- 238000013135 deep learning Methods 0.000 claims description 10
- 230000037081 physical activity Effects 0.000 claims description 9
- 238000003860 storage Methods 0.000 claims description 9
- 238000006243 chemical reaction Methods 0.000 claims description 8
- 238000012795 verification Methods 0.000 claims description 8
- 230000033001 locomotion Effects 0.000 claims description 7
- 239000008280 blood Substances 0.000 claims description 6
- 210000004369 blood Anatomy 0.000 claims description 6
- 230000036772 blood pressure Effects 0.000 claims description 6
- 235000012000 cholesterol Nutrition 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 6
- 230000001186 cumulative effect Effects 0.000 claims description 4
- 238000011478 gradient descent method Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 4
- 238000012360 testing method Methods 0.000 claims description 4
- 238000001514 detection method Methods 0.000 claims description 3
- 235000013305 food Nutrition 0.000 claims description 3
- 230000001766 physiological effect Effects 0.000 claims description 3
- 230000009286 beneficial effect Effects 0.000 description 8
- 230000006872 improvement Effects 0.000 description 7
- 238000010606 normalization Methods 0.000 description 7
- 238000012545 processing Methods 0.000 description 5
- 230000036541 health Effects 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 235000009508 confectionery Nutrition 0.000 description 1
- 238000013501 data transformation Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000035622 drinking Effects 0.000 description 1
- 210000002216 heart Anatomy 0.000 description 1
- 210000003734 kidney Anatomy 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 230000001717 pathogenic effect Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 230000033764 rhythmic process Effects 0.000 description 1
- 230000000391 smoking effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
-
- 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
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/50—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Abstract
The invention provides a method, medium and system for establishing a chronic disease risk assessment model, which belong to the technical field of risk assessment, acquire relevant data of chronic disease patients and healthy people and establish an initial neural network; calculating risk grades according to the chronic disease influencing factors, wherein the risk grades are divided into four grades of super-risk, high-risk, medium-risk and low-risk according to the possibility of the chronic disease; performing risk analysis according to the risk level to obtain a risk event of chronic disease risk, and obtaining a chronic disease risk analysis result according to the risk event; performing optimization training on the initial neural network to obtain a chronic disease risk assessment model, and inputting sample parameters of an object to be analyzed into the chronic disease risk assessment model to perform chronic disease risk assessment; and obtaining a chronic disease risk assessment result of the object to be analyzed according to the output data of the chronic disease risk assessment model.
Description
Technical Field
The invention belongs to the technical field of risk assessment, and particularly relates to a method, medium and system for establishing a chronic disease risk assessment model.
Background
With the rapid development of socioeconomic performance, chronic diseases have become a major medical problem affecting the health of residents. The modern society life rhythm is accelerated, and people often neglect health, ingest excessive high calorie, high fat, high confectionery, lack motion, stay up night, smoking and drinking and other bad life habits increase the risk of chronic diseases. According to statistics of the ministry of health, the number of chronic patients diagnosed in China at present exceeds 2.6 hundred million people, and the number of chronic patients dying each year exceeds 300 ten thousand people. The damage of chronic diseases mainly causes damage to important organs such as brain, heart, kidney and the like, is easy to cause disability, affects working capacity and life quality, has extremely high medical cost, and increases the economic burden of society and families. The technology of chronic patients in China is huge, the chronic disease management difficulty is huge, and the method has important significance on how to effectively evaluate the chronic disease.
Along with the development of medical technology, the diagnosis capability and consciousness of chronic diseases are improved, so that a plurality of chronic diseases are discovered and diagnosed in early stages, and further statistical data of diseased age groups are increased, but due to the fact that the chronic diseases have a plurality of pathogenic factors and the relationship among variables is complicated, the internal rules of the lesions are difficult to discover by adopting a traditional risk assessment method, and effective implementation of accurate prevention and control measures of the chronic diseases is seriously influenced.
Disclosure of Invention
In view of the above, the invention provides a method, medium and system for establishing a chronic disease risk assessment model, which can judge whether a patient is ill according to the daily living habit of a person to be detected, solve the problem that chronic diseases are easy to be ignored in daily life, and can realize that people can independently judge whether the patient is ill with chronic diseases.
The invention is realized in the following way:
the first aspect of the present invention provides a method for establishing a chronic disease risk assessment model, wherein,
s10, acquiring sample parameters, wherein the sample parameters comprise relevant data of chronic disease patient parameters and healthy crowd parameters, the sample parameters are used for analyzing common chronic disease influencing factors, the chronic disease influencing factors comprise diet, exercise, stay up and genetics, and the relevant data comprise age, sex, body mass index, blood pressure, blood sugar and cholesterol indexes of the chronic disease patients and the healthy crowd;
s20, acquiring related data of chronic patients and healthy people, and establishing an initial neural network;
s30, calculating risk grades according to the chronic disease influence factors, wherein the risk grades are divided into four grades of super-risk, high-risk, medium-risk and low-risk according to the possibility of chronic disease;
s40, performing risk analysis according to the risk level to obtain a risk event of chronic disease risk, and obtaining a chronic disease risk analysis result according to the risk event, wherein the risk event is a factor of potential influence of the chronic disease and comprises unreasonable diet, lack of exercise, long-term stay up and inheritance;
s50, performing optimization training on the initial neural network to obtain a chronic disease risk assessment model, and inputting sample parameters of an object to be analyzed into the chronic disease risk assessment model to perform chronic disease risk assessment;
s60, obtaining a chronic disease risk assessment result of the object to be analyzed according to output data of the chronic disease risk assessment model.
The method, medium and system for establishing the chronic disease risk assessment model provided by the invention have the technical effects that: the chronic disease risk assessment model can be established through a convolutional neural network, and the specific steps are as follows:
acquiring relevant data of the chronic patient, inputting training data of the chronic patient, outputting the risk of illness judged by the doctor, establishing the initial neural network, using the initial neural network, passing through the training set and the verification set,
the method comprises the steps of verifying and testing a chronic disease risk assessment model, optimizing parameters through back propagation, updating parameters of a backbone layer in real time, converting initial neural network parameters, realizing an operation process from low precision to high precision, obtaining more accurate parameters, obtaining the chronic disease risk assessment model which is more similar to the disease risk judged by a doctor, calculating a loss function value through calculation of a cross entropy loss function, optimizing the parameters, enabling the chronic disease risk assessment model to be more accurate, carrying out risk assessment on the object to be analyzed according to sample parameters of the object to be analyzed by establishing the chronic disease risk assessment model, and obtaining a chronic disease risk assessment result of the object to be analyzed according to output data of the chronic disease risk assessment model.
On the basis of the technical scheme, the method, the medium and the system for establishing the chronic disease risk assessment model can be further improved as follows:
the method for establishing the chronic disease risk assessment model comprises the following specific steps:
acquiring relevant data of the chronic disease influencing factors and duration as training data;
the training data specifically comprises: randomly collecting 100 chronic patients, wherein the chronic patients comprise severe chronic patients and mild chronic patients, and evaluating the activity of the 100-day organism;
extracting the potential influencing factors of the chronic diseases according to the trained data and determining specific values of the related data, wherein the related data comprise age, sex, body mass index, blood pressure, blood sugar and cholesterol indexes of the chronic disease patient;
based on the training data and the disease risk judged by doctors, the deep learning network training sample is taken as the deep learning network training sample, and the deep learning network training sample is prepared according to the following steps: 4, dividing the ratio into a training set and a verification set, and constructing the initial neural network;
training the initial neural network by adopting the training set to obtain the chronic disease risk assessment model;
adopting the verification set to verify and test the chronic disease risk assessment model, and optimizing the chronic disease risk assessment model;
wherein the chronic disease potential influencing factors include: unreasonable diet, lack of exercise, long-term stay up and inheritance.
The beneficial effects of adopting above-mentioned improvement scheme are: by collecting the training data of the severe chronic disease patient and the mild chronic disease patient, risk analysis can be carried out on the chronic disease potential influence factors, and the influence degree of the chronic disease potential influence factors on chronic disease is obtained;
the training sample of the deep learning network is divided into the training set and the verification set, the chronic disease risk assessment model can be obtained through training, the initial neural network is trained and tested, and the obtained chronic disease risk assessment model is optimized.
Further, the training set is adopted to train the initial neural network to obtain the chronic disease risk assessment model, and the specific steps are as follows:
obtaining the training data, wherein the chronic disease potential influence factors correspond to a marking score according to the duration, and according to statistics, the chronic disease potential influence is defined as marking integral 1, and other marking integral is 0;
the potential influencing factors of chronic diseases comprise unreasonable diet, lack of exercise and long-term stay up;
wherein the unreasonable diet is irregular diet time accumulated for more than 40 days in the 100-day body activity evaluation, and the mark score is 1; wherein the irregular diet time comprises irregular time and irregular structure, the irregular time refers to that food is not fed on time, and the irregular structure refers to that the diet structure is single or the physiological activity requirement of a human body can not be met;
the lack of movement is that no movement activity is performed for more than 40 days in the 100-day body movement assessment, and the mark integral is 1;
the long-term stay up is that the person falls asleep after the cumulative time exceeds 40 days and the ratio of 11:00 in the 100-day body activity evaluation, and the mark score is 1;
taking the training data as a sample and the corresponding mark score as a label, and performing vector calculation on the body activity evaluation result;
acquiring the integral vector and the disease risk corresponding to the judgment of the doctor, weighting the output of the integral vector for the disease risk judged by the doctor of the chronic disease risk assessment model, and summing the weighted doctor outputs, wherein the output chronic disease risk is the weighted doctor output summation;
training the initial neural network to obtain a chronic disease risk assessment model.
The beneficial effects of adopting above-mentioned improvement scheme are: the chronic disease potential influence is respectively marked and integrated, the chronic disease influence factor can be defined, the chronic disease potential influence is defined as marked and integrated 1, other marked and integrated 0, and the doctor corresponding to the chronic disease influence factor can obtain the judging risk of the disease through the marked and integrated 1;
by weighting the output of the weighted vector of the weights judged by the doctor of the chronic disease risk assessment model as to the severity of the disease risk and summing the weighted doctor outputs, the output chronic disease risk can be obtained as an average of a plurality of results, enabling a more accurate chronic disease risk assessment of the output.
The initial neural network comprises 1 input layer, 1 backbone layer and 1 output layer, when the output result of the chronic disease risk assessment model is different from the disease risk judged by the doctor by more than 10%, parameters are optimized through back propagation, the back propagation network training is carried out through a gradient descent method through the parameter optimization, and the parameters of the backbone layer are updated;
the parameters are W and b, wherein the W and b are respectively connected with the input layer and the backbone layer, and the backbone layer and the output layer; the W represents the weight and the b represents the deviation.
The beneficial effects of adopting above-mentioned improvement scheme are: the parameters are subjected to the gradient descent methodUpdating is performed, and the learning speed of the chronic disease risk assessment model depends on two values: 1. a learning rate; 2. a bias guide value; wherein the learning rate is a set super-parameter, and the magnitude of the bias value depends on x o And [ e(s) -y]Said [ e(s) -y]The magnitude of (2) reflects the degree of error of the model, wherein x is o The y is the output of the chronic disease risk assessment model;
the larger the value is, the worse the model effect is, the faster the model learning speed is, the learning speed is fast when the model effect is bad and the learning speed is slow when the model effect is good when the cross entropy is combined with the loss function.
Further, the backbone layer is used for processing the training data, preprocessing the training data and obtaining a preprocessed data set;
the training data is converted from the input layer to the backbone layer, and the training data is converted into the backbone layer through a matrix operation formula:
H=X×W1+b1;
wherein, H is the backbone layer; the X is the input layer; the W1 and b1 connect the input layer and the backbone layer; the training data is converted from the backbone layer to the output layer, and the training data is converted into the output layer through a matrix operation formula:
Y=H×W2+b2;
wherein, Y is the output layer; the H is the backbone layer; the W2 and b2 connect the backbone layer and the output layer.
The beneficial effects of adopting above-mentioned improvement scheme are: the training data can be normalized by preprocessing the training data, including standardization and normalization, the extension of the data is controlled within a certain range, the training of the initial neural network can be accelerated, the stability is improved, and the standardized data is used for training in the initial neural network.
Further, the backbone layer includes an activation layer, and an activation function in the activation layer performs nonlinear conversion on the result after the rectangular operation, where the nonlinear conversion is performed by a ReLU function, and the ReLU function formula is:
f(V)=max(0,V);
the f (V) is the activation function ReLU function, the max is the result after the rectangular operation, and the V is the training data of the input layer; the ReLU function output result is:
R=max(0,WV+B);
and R is the result of outputting the ReLU function, max is the result after the rectangular operation, and W and b are coefficients connecting the input layer and the backbone layer.
The beneficial effects of adopting above-mentioned improvement scheme are: the nonlinear conversion is carried out on the result after rectangular operation through the activation function in the activation layer, and the nonlinear conversion is carried out through the ReLU function, wherein the specific steps of the ReLU function are as follows:
adding a ReLU activation function to the backbone layer of the deep learning network, and performing stack operation between layers in the secondary forward propagation process; between the two layers, after the upper layer passes through the ReLU, the upper layer is transferred into the lower layer, and each neuron of the lower layer contains add operation of the upper layer; the stack and add operations are mutually blended, so that the ReLU has nonlinear expression capability;
finally, the linear inseparable data of the result after the rectangular operation is changed into the linear inseparable data of the high-dimensional space, and the result after the rectangular operation is mapped to Gaussian distribution, so that variance can be stabilized and deviation can be minimized.
Further, the output layer includes a Softmax layer, and the Softmax layer normalizes the ReLU function output result in the output layer, and the specific operation formula is as follows:
wherein the S is i For the normalized output result, i represents the training function sequence number, e l Representing the exponentiation of all training data underlying e, said e j Representing all exponentiations of the exponent and the j tableThe number of the training functions is shown;
and calculating the accuracy of the normalized output result through cross entropy loss so as to reduce the cross entropy loss and improve the accuracy of the result, wherein the cross entropy loss calculation method is used for calculating the negative number of the logarithm of the normalized output result.
The beneficial effects of adopting above-mentioned improvement scheme are: the loss function value can be calculated by calculating the accuracy of the normalized output result through cross entropy loss, the cross entropy loss function can capture the difference of the prediction effect between the initial neural network and the chronic disease risk assessment model, and the accuracy of the result can be improved by reducing the cross entropy loss.
Wherein, the disease risk value, the formula of calculation is:
S=P×Q;
wherein: the S is the possibility of occurrence of the disease risk value event; the P is the occurrence frequency of the disease risk value event; the Q refers to the influence degree of the disease risk value event;
the likelihood is classified into four classes according to the occurrence frequency of the chronic disease influencing factors from low to high, wherein class one is rarely or unlikely to occur; a level two is likely to occur or not more than once during detection; level three is likely to occur or likely to occur multiple times throughout the job; grade four is a frequent occurrence.
The beneficial effects of adopting above-mentioned improvement scheme are: by calculating the disease risk value, the risk level can be evaluated according to the occurrence frequency of the disease risk value event and the influence degree of the disease risk value event, and the risk level is ranked according to the disease risk value by the evaluation of the disease risk value; sequentially sorting into four grades of super-risk, high-risk, medium-risk and low-risk.
A second aspect of the present invention provides a computer readable storage medium, wherein the computer readable storage medium has stored therein program instructions, which when executed, are configured to perform a method for establishing a chronic disease risk assessment model as described above.
A third aspect of the present invention provides a method for establishing a chronic disease risk assessment model, comprising a computer readable storage medium as described above.
Compared with the prior art, the method, medium and system for establishing the chronic disease risk assessment model provided by the invention have the beneficial effects that: acquiring relevant data of chronic patients and healthy people, and establishing an initial neural network; calculating risk grades according to the chronic disease influencing factors, wherein the risk grades are divided into four grades of super-risk, high-risk, medium-risk and low-risk according to the possibility of the chronic disease; performing risk analysis according to the risk level to obtain a risk event of chronic disease risk, and obtaining a chronic disease risk analysis result according to the risk event; and carrying out optimization training on the initial neural network to obtain a chronic disease risk assessment model, and inputting sample parameters of an object to be analyzed into the chronic disease risk assessment model to carry out chronic disease risk assessment. Whether the patient is ill or not can be judged according to the daily life habit of the person to be detected, the problem that chronic diseases are easy to ignore in daily life can be solved, and whether people suffer from chronic diseases or not can be independently judged.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the description of the embodiments of the present invention will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a method for establishing a chronic disease risk assessment model according to a first aspect of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments, based on the embodiments of the invention, which are apparent to those of ordinary skill in the art without inventive faculty, are intended to be within the scope of the invention.
Thus, the following detailed description of the embodiments of the invention, as presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, based on the embodiments of the invention, which are apparent to those of ordinary skill in the art without inventive faculty, are intended to be within the scope of the invention.
It should be noted that: like reference numerals and letters denote like items in the following figures, and thus once an item is defined in one figure, no further definition or explanation thereof is necessary in the following figures.
In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings are merely for convenience in describing the present invention and simplifying the description, and do not indicate or imply that the apparatus or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and thus should not be construed as limiting the present invention.
Furthermore, the terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature. In the description of the present invention, the meaning of "a plurality" is two or more, unless explicitly defined otherwise.
As shown in fig. 1, the present invention provides a flowchart of a method for establishing a chronic disease risk assessment model, which includes,
s10, acquiring sample parameters, wherein the sample parameters comprise relevant data of chronic disease patient parameters and healthy crowd parameters, the sample parameters are used for analyzing common chronic disease influencing factors, the chronic disease influencing factors comprise diet, exercise, stay up and genetics, and the relevant data comprise age, sex, body mass index, blood pressure, blood sugar and cholesterol indexes of the chronic disease patients and the healthy crowd;
s20, acquiring related data of chronic patients and healthy people, and establishing an initial neural network;
s30, calculating risk grades according to chronic disease influence factors, wherein the risk grades are divided into four grades of super-risk, high-risk, medium-risk and low-risk according to the possibility of chronic disease;
s40, performing risk analysis according to the risk level, obtaining a risk event of the chronic disease risk, and obtaining a chronic disease risk analysis result according to the risk event, wherein the risk event is a potential influence factor of the chronic disease and comprises unreasonable diet, lack of exercise, long-term stay up and inheritance;
s50, performing optimization training on the initial neural network to obtain a chronic disease risk assessment model, and inputting sample parameters of an object to be analyzed into the chronic disease risk assessment model to perform chronic disease risk assessment;
s60, obtaining a chronic disease risk assessment result of the object to be analyzed according to the output data of the chronic disease risk assessment model.
The risk level is calculated, and the specific steps are as follows:
establishing a risk grade for common chronic disease influencing factors, and assigning a magnitude based on a systematic method and chronic disease risk assessment experience according to the chronic disease influencing factors;
performing risk analysis to obtain main risk events of chronic disease risks;
calculating a disease risk value;
according to the risk level evaluation of the disease risk value, the risk level evaluation is ordered according to the risk value from big to small;
the risk value is divided into four grades of super-risk, high-risk, medium-risk and low-risk from large to small in sequence,
the super risk is that the risk value exceeds 80%, the high risk is that the risk value exceeds 60%, the medium risk is that the risk value exceeds 40%, and the low risk is that the risk value is lower than 40%.
In the technical scheme, a chronic disease risk assessment model is established, and the method specifically comprises the following steps:
acquiring relevant data of chronic disease influencing factors and duration as training data;
the training data is specifically: randomly collecting 100 chronic patients, wherein the chronic patients comprise severe chronic patients and mild chronic patients, and evaluating the activity of the 100-day organism;
extracting potential influencing factors of the chronic diseases according to the trained data and determining specific numerical values of related data, wherein the related data comprise age, sex, body mass index, blood pressure, blood sugar and cholesterol indexes of a patient suffering from the chronic diseases;
based on training data and the disease risk judged by doctors, the training data is taken as a deep learning network training sample, and the deep learning network training sample is prepared according to the following steps: 4, dividing the ratio into a training set and a verification set, and constructing an initial neural network;
training an initial neural network by using a training set to obtain a chronic disease risk assessment model;
adopting a verification set to verify and test the chronic disease risk assessment model, and optimizing the chronic disease risk assessment model;
among the potential contributors to chronic disease include: unreasonable diet, lack of exercise, long-term stay up and inheritance.
Further, in the above technical solution, training the initial neural network by using the training set to obtain a chronic disease risk assessment model, which specifically includes the steps of:
obtaining training data, wherein the chronic disease potential influence factors correspond to mark scores according to duration time, and according to statistics, the chronic disease potential influence is defined as mark integral 1, and other mark integral is 0;
potential contributors to chronic disease include improper diet, lack of exercise, long-term stay up;
the unreasonable diet is that diet time is irregular after accumulation for more than 40 days in 100 days of body activity evaluation, and the marking score is 1; wherein, irregular diet time includes irregular time and irregular structure, irregular time means that food is not fed on time, irregular structure means that diet structure is single or the physiological activity requirement of human body can not be met;
lack of exercise is that no exercise activity is performed in 100 days of machine activity assessment, and the mark integral is 1;
the long-term stay up is that the person falls asleep after the cumulative time exceeds 40 days and the ratio of the cumulative time to the total time is 11:00 in the 100-day body activity evaluation, and the mark score is 1;
taking training data as a sample and corresponding mark scores as labels, and performing vector calculation on the body activity evaluation result;
acquiring an integral vector and the disease risk judged by the corresponding doctor, weighting the output of the integral vector for the disease risk judged by the doctor of the chronic disease risk assessment model, and summing the weighted doctor outputs, wherein the output chronic disease risk is the weighted doctor output summation;
the initial neural network is trained to obtain a chronic risk assessment model.
In the technical scheme, the initial neural network comprises 1 input layer, 1 backbone layer and 1 output layer, when the output result of the chronic disease risk assessment model is different from the disease risk judged by a doctor by more than 10%, parameters are optimized through back propagation, the back propagation network training is carried out through a gradient descent method through the parameter optimization, and the parameters of the backbone layer are updated;
wherein, the parameters are W and b, W and b are respectively connected with the input layer and the backbone layer, and the backbone layer and the output layer; w represents the weight and b represents the deviation.
The gradient descent method comprises the following specific operation steps:
the first step: calculating the gradients of parameters W and b of the training data on the loss function, calculating the sum of the gradients of all the parameters W, and calculating the sum of the gradients of the training data of all the parameters b;
and a second step of: calculating the average value of the gradients of the parameter W and the parameter b of all training data;
and a third step of: by calculating the weights and deviations for updating the training data,
the calculation formula is as follows:
wherein a is learning rate, W t+1 W is the updated weight t The initial value of the weight, t is the serial number of the training data;
wherein a is learning rate, b t+1 B for updated bias t And t is the serial number of the training data and is the initial value of the deviation.
Further, in the above technical solution, the backbone layer is configured to process training data, and perform preprocessing on the training data to obtain a preprocessed data set;
the training data is converted from the input layer to the backbone layer, and the training data is converted into the backbone layer through a matrix operation formula:
H=X×W1+b1;
wherein H is a backbone layer; x is an input layer; w1 and b1 connect the input layer and the backbone layer; the training data is converted from the backbone layer to the output layer, and the training data is converted into the output layer through a matrix operation formula:
Y=H×W2+b2;
wherein Y is an output layer; h is a backbone layer; w2 and b2 connect the backbone layer and the output layer.
The step of preprocessing the training data to obtain a preprocessed data set specifically comprises
Acquiring training data, and performing data cleaning on the training data, namely processing a missing value and an abnormal value;
performing differential operation on the related data of the duration, performing data normalization processing on the training data, wherein the data normalization processing is performed through maximum and minimum normalization, and the maximum and minimum normalization formula is as follows:
wherein X' is training data after data normalization processing, X is training data, and X min To train data minimum value, X max Maximum value of training data;
selecting training data by a statistical method, wherein the training data is used for modeling and analysis;
and carrying out data transformation on the training data through exponential transformation to complete the preprocessing data set.
Further, in the above technical solution, the backbone layer includes an activation layer, and the activation function in the activation layer performs nonlinear conversion on the result after the rectangular operation, where the nonlinear conversion is performed by a ReLU function, and a ReLU function formula is as follows:
f(V)=max(0,V);
f (V) is an activation function ReLU function, max is a result after rectangular operation, and V is training data of an input layer; the result of the ReLU function output is:
R=max(0,WV+B);
wherein R is a result of ReLU function output, max is a result after rectangular operation, and W and b are coefficients of a connection input layer and a backbone layer.
Further, in the above technical solution, the output layer includes a Softmax layer, and the Softmax layer normalizes a ReLU function output result in the output layer, and the specific operation formula is:
wherein the S is i For the normalized output result, i represents the training function sequence number, e i Representing the exponentiation of all training data underlying e, said e j Representing all exponent powers for summation, wherein j represents the number of training functions;
the accuracy of the normalized output result is calculated through the cross entropy loss, so that the cross entropy loss is reduced, the accuracy of the result is improved, and the cross entropy loss calculation method is used for calculating the negative number of the logarithm of the normalized output result.
Wherein Softmax is an activation function, which is understood as normalization, a numerical vector can be normalized to a probability distribution vector, and the sum of the vectors is 1; the Softmax layer is used as the last layer of the convolutional neural network for output;
the cross entropy loss calculation comprises the following specific formulas:
L=-[y log y+(1-y)log(1-y)];
wherein L is a loss function result, and y is a normalized output result.
In the above technical solution, the disease risk value is calculated according to the following formula:
S=P×Q;
wherein: s is the possibility of occurrence of a disease risk value event; p is the frequency of occurrence of the disease risk value event; q refers to the influence degree of the occurrence of the disease risk value event;
the probability is divided into four grades from low to high according to the occurrence frequency of chronic disease influencing factors, wherein grade one is rarely or unlikely to occur; a level two is likely to occur or not more than once during detection; level three is likely to occur or likely to occur multiple times throughout the job; grade four is a frequent occurrence.
A second aspect of the present invention provides a computer readable storage medium, wherein the computer readable storage medium has stored therein program instructions, which when executed, are configured to perform a method for establishing a chronic disease risk assessment model as described above.
A third aspect of the present invention provides a method for establishing a chronic disease risk assessment model, comprising a computer readable storage medium as described above.
The foregoing is merely illustrative of the present invention, and the present invention is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present invention. Therefore, the protection scope of the invention is subject to the protection scope of the claims.
Claims (10)
1. A method, medium and system for establishing a chronic disease risk assessment model are characterized in that,
s10, acquiring sample parameters, wherein the sample parameters comprise relevant data of chronic disease patient parameters and healthy crowd parameters, the sample parameters are used for analyzing common chronic disease influencing factors, the chronic disease influencing factors comprise diet, exercise, stay up and genetics, and the relevant data comprise age, sex, body mass index, blood pressure, blood sugar and cholesterol indexes of the chronic disease patients and the healthy crowd;
s20, acquiring related data of chronic patients and healthy people, and establishing an initial neural network;
s30, calculating risk grades according to the chronic disease influence factors, wherein the risk grades are divided into four grades of super-risk, high-risk, medium-risk and low-risk according to the possibility of chronic disease;
s40, performing risk analysis according to the risk level to obtain a risk event of chronic disease risk, and obtaining a chronic disease risk analysis result according to the risk event, wherein the risk event is a factor of potential influence of the chronic disease and comprises unreasonable diet, lack of exercise, long-term stay up and inheritance;
s50, performing optimization training on the initial neural network to obtain a chronic disease risk assessment model, and inputting sample parameters of an object to be analyzed into the chronic disease risk assessment model to perform chronic disease risk assessment;
s60, obtaining a chronic disease risk assessment result of the object to be analyzed according to output data of the chronic disease risk assessment model.
2. The method, medium and system for establishing a chronic disease risk assessment model according to claim 1, wherein the method, medium and system for establishing the chronic disease risk assessment model comprises the following specific steps:
acquiring relevant data of the chronic disease influencing factors and duration as training data;
the training data specifically comprises: randomly collecting 100 chronic patients, wherein the chronic patients comprise severe chronic patients and mild chronic patients, and evaluating the activity of the 100-day organism;
extracting the potential influencing factors of the chronic diseases according to the trained data and determining specific values of the related data, wherein the related data comprise age, sex, body mass index, blood pressure, blood sugar and cholesterol indexes of the chronic disease patient;
based on the training data and the disease risk judged by doctors, the deep learning network training sample is taken as the deep learning network training sample, and the deep learning network training sample is prepared according to the following steps: 4, dividing the ratio into a training set and a verification set, and constructing the initial neural network;
training the initial neural network by adopting the training set to obtain the chronic disease risk assessment model;
adopting the verification set to verify and test the chronic disease risk assessment model, and optimizing the chronic disease risk assessment model;
wherein the chronic disease potential influencing factors include: unreasonable diet, lack of exercise, long-term stay up and inheritance.
3. The method, medium and system for establishing a chronic disease risk assessment model according to claim 2, wherein the training of the initial neural network by using the training set is performed to obtain the chronic disease risk assessment model, and the specific steps are as follows:
obtaining the training data, wherein the chronic disease potential influence factors correspond to a marking score according to the duration, and according to statistics, the chronic disease potential influence is defined as marking integral 1, and other marking integral is 0;
the potential influencing factors of chronic diseases comprise unreasonable diet, lack of exercise and long-term stay up;
wherein the unreasonable diet is irregular diet time accumulated for more than 40 days in the 100-day body activity evaluation, and the mark score is 1; wherein the irregular diet time comprises irregular time and irregular structure, the irregular time refers to that food is not fed on time, and the irregular structure refers to that the diet structure is single or the physiological activity requirement of a human body can not be met;
the lack of movement is that no movement activity is performed for more than 40 days in the 100-day body movement assessment, and the mark integral is 1;
the long-term stay up is that the person falls asleep after the cumulative time exceeds 40 days and the ratio of 11:00 in the 100-day body activity evaluation, and the mark score is 1;
taking the training data as a sample and the corresponding mark score as a label, and performing vector calculation on the body activity evaluation result;
acquiring the integral vector and the disease risk corresponding to the judgment of the doctor, weighting the output of the integral vector for the disease risk judged by the doctor of the chronic disease risk assessment model, and summing the weighted doctor outputs, wherein the output chronic disease risk is the weighted doctor output summation;
training the initial neural network to obtain a chronic disease risk assessment model.
4. The method, medium and system for establishing a chronic disease risk assessment model according to claim 1, wherein the initial neural network comprises 1 input layer, 1 backbone layer and 1 output layer, and when the output result of the chronic disease risk assessment model differs from the disease risk judged by the doctor by more than 10%, parameters are optimized by back propagation, the parameter optimization carries out back propagation network training by a gradient descent method, and the parameters of the backbone layer are updated;
the parameters are W and b, wherein the W and b are respectively connected with the input layer and the backbone layer, and the backbone layer and the output layer; the W represents the weight and the b represents the deviation.
5. The method, medium and system for establishing a chronic disease risk assessment model according to claim 4, wherein the backbone layer is configured to process the training data, and perform preprocessing on the training data to obtain a preprocessed data set;
the training data is converted from the input layer to the backbone layer, and the training data is converted into the backbone layer through a matrix operation formula:
H=X×W1+b1;
wherein, H is the backbone layer; the X is the input layer; the W1 and b1 connect the input layer and the backbone layer; the training data is converted from the backbone layer to the output layer, and the training data is converted into the output layer through a matrix operation formula:
Y=H×W2+b2;
wherein, Y is the output layer; the H is the backbone layer; the W2 and b2 connect the backbone layer and the output layer.
6. The method, medium and system for establishing a chronic disease risk assessment model according to claim 4, wherein the backbone layer comprises an activation layer, and an activation function in the activation layer performs nonlinear conversion on the result after the rectangular operation, where the nonlinear conversion is performed by a ReLU function, and the ReLU function formula is as follows:
f(V)=max(0,V);
the f (V) is the activation function ReLU function, the max is the result after the rectangular operation, and the V is the training data of the input layer; the ReLU function output result is:
R=max(0,WV+B);
and R is the result of outputting the ReLU function, max is the result after the rectangular operation, and W and b are coefficients connecting the input layer and the backbone layer.
7. The method, medium and system for establishing a chronic disease risk assessment model according to claim 6, wherein the output layer comprises a Softmax layer, the Softmax layer normalizes the ReLU function output result in the output layer, and the specific operation formula is:
wherein the S is i For the normalized output result, i represents the training function sequence number, e i Representing the exponentiation of all training data underlying e, said e j Representing all exponent powers for summation, wherein j represents the number of training functions;
and calculating the accuracy of the normalized output result through cross entropy loss so as to reduce the cross entropy loss and improve the accuracy of the result, wherein the cross entropy loss calculation method is used for calculating the negative number of the logarithm of the normalized output result.
8. The method, medium and system for establishing a chronic disease risk assessment model according to claim 1, wherein the disease risk value is calculated according to the following formula:
S=P×Q;
wherein: the S is the possibility of occurrence of the disease risk value event; the P is the occurrence frequency of the disease risk value event; the Q refers to the influence degree of the disease risk value event;
the likelihood is classified into four classes according to the occurrence frequency of the chronic disease influencing factors from low to high, wherein class one is rarely or unlikely to occur; a level two is likely to occur or not more than once during detection; level three is likely to occur or likely to occur multiple times throughout the job; grade four is a frequent occurrence.
9. A computer readable storage medium, characterized in that the computer readable storage medium has stored therein program instructions, which when run, are adapted to perform a chronic disease risk assessment model building method according to any of claims 1-8.
10. A method of modeling chronic disease risk assessment comprising a computer readable storage medium according to claim 9.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310531805.8A CN116564521A (en) | 2023-05-11 | 2023-05-11 | Chronic disease risk assessment model establishment method, medium and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310531805.8A CN116564521A (en) | 2023-05-11 | 2023-05-11 | Chronic disease risk assessment model establishment method, medium and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN116564521A true CN116564521A (en) | 2023-08-08 |
Family
ID=87494123
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310531805.8A Pending CN116564521A (en) | 2023-05-11 | 2023-05-11 | Chronic disease risk assessment model establishment method, medium and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116564521A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117558460A (en) * | 2024-01-11 | 2024-02-13 | 卓世未来(天津)科技有限公司 | Chronic disease management method and system based on small sample learning and large language model |
-
2023
- 2023-05-11 CN CN202310531805.8A patent/CN116564521A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117558460A (en) * | 2024-01-11 | 2024-02-13 | 卓世未来(天津)科技有限公司 | Chronic disease management method and system based on small sample learning and large language model |
CN117558460B (en) * | 2024-01-11 | 2024-04-05 | 卓世未来(天津)科技有限公司 | Chronic disease management method and system based on small sample learning and large language model |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107153774A (en) | The disease forecasting system of the structure and application of chronic disease risk assessment the hyperbolic model model | |
CN110706822B (en) | Health management method based on logistic regression model and decision tree model | |
CN107145715B (en) | Clinical medicine intelligence discriminating gear based on electing algorithm | |
Jalali et al. | Advanced analytics for outcome prediction in intensive care units | |
CN113838577B (en) | Convenient layered old people MODS early death risk assessment model, device and establishment method | |
CN114023441A (en) | Severe AKI early risk assessment model and device based on interpretable machine learning model and development method thereof | |
CN112786203A (en) | Machine learning diabetic retinopathy morbidity risk prediction method and application | |
CN115862850B (en) | Modeling method and device of hepatocellular carcinoma monitoring model based on longitudinal multidimensional data | |
CN116564521A (en) | Chronic disease risk assessment model establishment method, medium and system | |
CN114974585A (en) | Construction method of early risk prediction and evaluation model of metabolic syndrome in gestational period | |
JP2023036652A (en) | System for estimating health state of human body | |
CN111145902A (en) | Asthma diagnosis method based on improved artificial neural network | |
CN112967803A (en) | Early mortality prediction method and system for emergency patients based on integrated model | |
CN116864104A (en) | Chronic thromboembolic pulmonary artery high-pressure risk classification system based on artificial intelligence | |
TWI688371B (en) | Intelligent device for atrial fibrillation signal pattern acquisition and auxiliary diagnosis | |
CN116189866A (en) | Remote medical care analysis system based on data analysis | |
CN110610766A (en) | Apparatus and storage medium for deriving probability of disease based on symptom feature weight | |
CN109192306A (en) | A kind of judgment means of diabetes, equipment and computer readable storage medium | |
CN112086130B (en) | Method for predicting obesity risk prediction device based on sequencing and data analysis | |
CN113096127A (en) | System and method for generating brain network evolution model | |
CN113744877A (en) | Chronic disease assessment and intervention system with disease related factor extraction module | |
CN113273959A (en) | Portable diabetic retinopathy diagnosis and treatment instrument | |
Kavya et al. | Heart Disease Prediction Using Logistic Regression | |
Kour et al. | An Advance Approach for Diabetes Detection by Implementing Machine Learning Algorithms | |
CN116864062B (en) | Health physical examination report data analysis management system based on Internet |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |