CN116825383A - Training method, identification method, terminal and storage medium for drug resistance identification model - Google Patents
Training method, identification method, terminal and storage medium for drug resistance identification model Download PDFInfo
- Publication number
- CN116825383A CN116825383A CN202310770822.7A CN202310770822A CN116825383A CN 116825383 A CN116825383 A CN 116825383A CN 202310770822 A CN202310770822 A CN 202310770822A CN 116825383 A CN116825383 A CN 116825383A
- Authority
- CN
- China
- Prior art keywords
- vector
- drug resistance
- coding
- user
- anomaly
- 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
- 206010059866 Drug resistance Diseases 0.000 title claims abstract description 120
- 238000000034 method Methods 0.000 title claims abstract description 65
- 238000012549 training Methods 0.000 title claims abstract description 63
- 239000013598 vector Substances 0.000 claims abstract description 190
- 238000001514 detection method Methods 0.000 claims abstract description 24
- 230000002159 abnormal effect Effects 0.000 claims abstract description 20
- 238000005070 sampling Methods 0.000 claims description 24
- 230000006870 function Effects 0.000 claims description 21
- 238000012544 monitoring process Methods 0.000 claims description 14
- 238000013527 convolutional neural network Methods 0.000 claims description 13
- 230000009466 transformation Effects 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 7
- 238000011282 treatment Methods 0.000 abstract description 4
- 239000003242 anti bacterial agent Substances 0.000 description 8
- 229940088710 antibiotic agent Drugs 0.000 description 8
- 230000036772 blood pressure Effects 0.000 description 8
- 230000008569 process Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 229940079593 drug Drugs 0.000 description 6
- 239000003814 drug Substances 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 6
- 230000036457 multidrug resistance Effects 0.000 description 6
- 238000013145 classification model Methods 0.000 description 5
- 238000013473 artificial intelligence Methods 0.000 description 4
- 238000003745 diagnosis Methods 0.000 description 4
- 230000002924 anti-infective effect Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 238000011269 treatment regimen Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 206010061822 Drug intolerance Diseases 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 1
- 230000003115 biocidal effect Effects 0.000 description 1
- 239000008280 blood Substances 0.000 description 1
- 210000004369 blood Anatomy 0.000 description 1
- 230000036760 body temperature Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 229910052760 oxygen Inorganic materials 0.000 description 1
- 239000001301 oxygen Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000007723 transport mechanism Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Landscapes
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The embodiment of the application relates to the technical field of digital medical treatment, and particularly provides a training method, a recognition method, a terminal and a storage medium of a drug resistance recognition model. The training method comprises the following steps: obtaining a training sample, wherein the training sample comprises relevant information of a user and a drug resistance label of the user; coding the related information by using a coding layer of the drug resistance identification model to obtain a coding vector; performing anomaly detection on the coded vector according to an anomaly identification layer of the drug resistance identification model to obtain an anomaly tag corresponding to the coded vector; determining a loss function of the drug resistance identification model according to the drug resistance label and the abnormal label; and carrying out iterative updating on the drug resistance identification model based on the training sample and the loss function to obtain a target drug resistance identification model. According to the application, the recognition result of the abnormal recognition layer for recognizing the drug resistance is used as an abnormal event, and the abnormal event is used for prediction training, so that the complexity of the model is reduced and the accuracy of the model is improved.
Description
Technical Field
The application relates to the technical field of digital medical treatment, in particular to a training method, a recognition method, a terminal and a storage medium of a drug resistance recognition model.
Background
Along with the development of artificial intelligence technology, the development of the artificial intelligence technology is involved in various aspects of life of people, such as medical treatment, and the artificial intelligence technology can solve the problems of auxiliary diagnosis, health management, remote consultation and the like of diseases in medical treatment, so that the timeliness of patient's diagnosis and the high efficiency of doctor diagnosis are greatly improved.
Resistance to antibiotics is a very serious problem in the medical field that needs to be faced, and predicting resistance to antibiotics helps doctors to better formulate anti-infective treatment strategies for patients. Multidrug resistance is a recent research hotspot in the medical community and refers to the development of multiple antibiotics by patients. At present, the problem of multi-drug resistance is getting more and more attention because of the problems of abuse and evolution of antibiotics and the like. In the prior art, the artificial intelligence technology can be adopted to predict the multi-drug resistance of the antibiotics, but the multi-drug resistance of the antibiotics is an event with a relatively small occurrence ratio, so that the prediction result is poor and the prediction precision is not high when the multi-drug resistance of the antibiotics is predicted.
Disclosure of Invention
The embodiment of the application mainly aims to provide a training method, a recognition method, a terminal and a storage medium for a drug resistance recognition model, and aims to solve the problem that in the prior art, when multi-drug resistance of antibiotics is predicted, because fewer drug resistance events cause fewer training samples, prediction accuracy is lower when model prediction is performed.
In a first aspect, an embodiment of the present application provides a training method for a drug resistance identification model, including:
obtaining a training sample, wherein the training sample comprises relevant information of a user and a drug resistance label of the user, and the relevant information is used for representing information related to drug resistance identification of the user;
coding the related information by using a coding layer of the drug resistance identification model to obtain a coding vector;
performing anomaly detection on the coded vector according to an anomaly identification layer of the drug resistance identification model to obtain an anomaly tag corresponding to the coded vector;
determining a loss function of the drug resistance identification model according to the drug resistance tag and the abnormal tag;
and carrying out iterative updating on the drug resistance identification model based on the training sample and the loss function to obtain a target drug resistance identification model.
In a second aspect, an embodiment of the present application further provides a method for identifying drug resistance, including:
and inputting relevant information of the user into a drug resistance recognition model to obtain a drug resistance recognition result, wherein the drug resistance recognition model is obtained according to the training step of any drug resistance recognition model provided by the specification of the application.
In a third aspect, an embodiment of the present application further provides a terminal device, where the terminal device includes a processor, a memory, a computer program stored on the memory and executable by the processor, and a data bus for implementing a connection communication between the processor and the memory, where the computer program, when executed by the processor, implements the training method steps and the drug resistance identification method steps of any one of the drug resistance identification models provided in the specification of the present application.
In a fourth aspect, an embodiment of the present application further provides a storage medium, where the storage medium stores one or more programs, and the one or more programs are executable by one or more processors to implement the steps of the training method and the steps of the drug resistance identification method of any one of the drug resistance identification models provided in the specification of the present application.
The embodiment of the application provides a training method, a recognition method, a terminal and a storage medium of a drug resistance recognition model, wherein the training method obtains a coding vector by coding related information of a user in a training sample, and further carries out anomaly detection on the coding vector by utilizing an anomaly recognition layer, so as to obtain an anomaly label corresponding to the coding vector; the method comprises the steps of comparing an abnormal label with a drug resistance label corresponding to related information of a user, obtaining a target drug resistance identification model when the abnormal label is consistent with the drug resistance label or reaches a preset range, and further detecting drug resistance of the user to be identified by using the target drug resistance identification model, so that the drug resistance can be predicted by using the drug resistance as an abnormal event, the problem that a high-quality identification model cannot be trained when training data are fewer is solved, the time complexity of the drug resistance identification model is reduced, the accuracy of model identification is improved, and a doctor can be better helped to formulate an anti-infection treatment strategy of the patient.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flow chart of a training method of a drug resistance recognition model according to an embodiment of the present application;
fig. 2 is a schematic block diagram of a structure corresponding to a drug resistance identification model according to an embodiment of the present application;
FIG. 3 is a flowchart corresponding to one embodiment of step S2 in FIG. 1;
FIG. 4 is a flowchart corresponding to one embodiment of step S3 in FIG. 1;
fig. 5 is a schematic block diagram of a structure of a terminal device according to an embodiment of the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
The flow diagrams depicted in the figures are merely illustrative and not necessarily all of the elements and operations/steps are included or performed in the order described. For example, some operations/steps may be further divided, combined, or partially combined, so that the order of actual execution may be changed according to actual situations.
It is to be understood that the terminology used in the description of the application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
The embodiment of the application provides a training method, an identification method, a terminal and a storage medium of a drug resistance identification model.
According to the training method of the drug resistance recognition model, the coding vector is obtained by coding the related information of the user in the training sample, and then the coding vector is subjected to anomaly detection by utilizing the anomaly recognition layer, so that an anomaly label corresponding to the coding vector is obtained; the method comprises the steps of comparing an abnormal label with a drug resistance label corresponding to related information of a user, obtaining a target drug resistance identification model when the abnormal label is consistent with the drug resistance label or reaches a preset range, and further detecting drug resistance of the user to be identified by using the target drug resistance identification model, so that the drug resistance can be predicted by using the drug resistance as an abnormal event, the problem that a high-quality identification model cannot be trained when training data are fewer is solved, the time complexity of the drug resistance identification model is reduced, the accuracy of model identification is improved, and a doctor can be better helped to formulate an anti-infection treatment strategy of the patient.
Some embodiments of the application are described in detail below with reference to the accompanying drawings. The following embodiments and features of the embodiments may be combined with each other without conflict.
Referring to fig. 1, fig. 1 is a flowchart of a training method of a drug resistance recognition model according to an embodiment of the application.
As shown in fig. 1, the training method of the drug resistance recognition model includes steps S1 to S5. Fig. 2 is a schematic block diagram of a structure corresponding to the text classification model.
Step S1: a training sample is obtained, the training sample comprising relevant information of a user and a drug resistance tag of the user, the relevant information being used to characterize information associated with a drug resistance identification of the user.
Illustratively, data of the user is obtained as input for model training, wherein each of the training samples comprises relevant information of the user and a drug resistance label corresponding to the user. The relevant information of the user includes, but is not limited to, basic information of the user, diagnosis information of a doctor, physiological detection information, medication information of the doctor, operation intervention information, and the like. Information of the user information associated with drug resistance identification of the user can be incorporated into the training sample.
For example, after the multi-dimensional information of the user is collected, the multi-dimensional information of the user can be analyzed by utilizing the correlation degree, so that information related to drug resistance identification in the multi-dimensional information of the user is obtained, and the related information is further used as a training sample.
For example, the information obtained from the user a includes information such as age, height, heart rate, blood pressure, etc., and after correlation analysis, it is found that the height in the information of the user a is irrelevant to drug resistance identification, and further the remaining information can be used as relevant information of the user, that is, information such as age, heart rate, blood pressure, etc.
Step S2: and coding the related information by using a coding layer of the drug resistance identification model to obtain a coding vector.
Illustratively, after obtaining the relevant information, the relevant information needs to be converted into a language that can be recognized by a machine, and then the relevant information can be encoded by an encoding layer to obtain an encoding vector.
For example, when the gender of the user in the related information is encoded, the encoding method can be used for encoding by using a one-hot encoding method, or can be used for encoding by using a word2vec or an ebedding method, and the encoding method of the related information is not particularly limited in the application, and can be selected according to the self requirement.
As shown in fig. 3, in some embodiments, the coding layer is a convolutional neural network, the related information includes basic information and physiological monitoring information of the user, and step S2 includes steps S21 to S23.
Step S21: encoding the basic information by utilizing the convolutional neural network to obtain a first convolutional vector corresponding to the basic information;
for example, after obtaining the relevant information of the user, the relevant information may be divided into basic information of the user and physiological monitoring information of the user according to information types.
For example, the basic information of the user comprises information such as age, gender, height, weight and the like of the user, and the physiological monitoring information of the user is heart rate, blood pressure, body temperature, blood oxygen saturation, medication information and operation intervention information.
Illustratively, the coding layer is a convolutional neural network with 5 layers, and the basic information of the user passes through the convolutional layer to obtain a first convolutional vector corresponding to the basic information.
For example, if the basic information of the user includes age and gender, the age and gender are respectively input into a 5-layer convolutional neural network to respectively obtain an age convolutional vector corresponding to the age and a gender convolutional vector corresponding to the gender, and then the age convolutional vector and the gender convolutional vector are spliced to obtain a spliced vector, the spliced vector can be used as a first convolutional vector, feature extraction can be performed on the spliced vector, and the extracted vector is used as the first convolutional vector.
Step S22: the convolutional neural network is utilized to encode the physiological monitoring information to obtain a second convolutional vector corresponding to the physiological monitoring information;
the coding layer is a convolutional neural network of 5 layers, and physiological monitoring information of a user passes through the convolutional layer to obtain a second convolutional vector corresponding to the physiological monitoring information.
For example, if the physiological monitoring information of the user includes heart rate and blood pressure, the heart rate and the blood pressure are respectively input into a convolutional neural network of 5 layers to respectively obtain a heart rate convolutional vector corresponding to the heart rate and a blood pressure convolutional vector corresponding to the blood pressure, and then the heart rate convolutional vector and the blood pressure convolutional vector are spliced to obtain a spliced vector, the spliced vector can be used as a second convolutional vector, feature extraction can also be performed on the spliced vector, and the extracted vector is used as the second convolutional vector.
And S23, determining the coding vector of the user according to the first convolution vector and the second convolution vector.
Illustratively, after the first convolution vector and the second convolution vector are obtained, the first convolution vector and the second convolution vector are fused, and the fusion result is used as a coding vector corresponding to the relevant information of the user.
For example, the first convolution vector and the second convolution vector may be added bit by bit, and the vector obtained by adding bit by bit may be used as the encoded vector corresponding to the correlation information.
In some implementations, the determining the user's encoding vector from the first convolution vector and the second convolution vector includes: and performing nonlinear transformation according to the first convolution vector and the second convolution vector, so as to obtain the coding vector of the user after nonlinear transformation processing.
Illustratively, the first convolution vector and the second convolution vector may be fused by using a nonlinear transformation manner, so as to obtain a code vector after the nonlinear transformation processing.
For example, the first convolution vector and the second convolution vector are transformed non-linearly by bits. If the first convolution vector is [1,2,3], the second convolution vector is [4,5,6], and the nonlinear transformation is 3x+2y, where x is a vector value in the first convolution vector and y is a vector value in the second convolution vector, the code vector after nonlinear transformation is [3×1+2×4,3×2+2×5,3×3+2×6], that is, the code vector is [11,16,21].
And S3, carrying out anomaly detection on the coded vector according to an anomaly identification layer of the drug resistance identification model to obtain an anomaly tag corresponding to the coded vector.
For example, after obtaining the encoded vector of the related information, the encoded vector is subjected to anomaly detection by using an anomaly identification layer, where the anomaly identification layer may be a classification model, so as to obtain an anomaly tag corresponding to the encoded vector.
For example, the anomaly identification layer is a classification model layer, the label categories classified by the classification model layer include two categories of drug resistance and drug intolerance, and if the user A and the user B exist, the first relevant information corresponding to the user A and the second relevant information corresponding to the user B are respectively input into the coding layer to obtain a first coding vector corresponding to the first relevant information and a second coding vector corresponding to the second relevant information. And then the first code vector and the second code vector are input into the classification model layer, so that a first classification label corresponding to the first code vector and a second classification label corresponding to the second code vector can be obtained, and if the first classification label is drug-resistant and the second classification label is non-drug-resistant, a first abnormal label of the user A can be obtained to be drug-resistant, and a second abnormal label of the user B can be obtained to be non-drug-resistant.
As shown in fig. 4, in some embodiments, the anomaly identification layer is an isolated forest, and step S3 includes steps S31 to S32.
Step S31: setting parameters of trees in the isolated forest to obtain a target isolated forest;
the basic principle of the isolated forest is that an abnormal sample can be isolated through random feature segmentation for a small number of times compared with a common sample, so that abnormal data in sample data is obtained, and abnormal detection is realized.
For example, when the isolated forest is utilized for anomaly identification, the height of the tree of the isolated forest is limited to ensure that data with shorter paths are obtained as anomaly data, and normal data with longer paths are ignored.
For example, the height of the tree in the convergence state can be obtained according to the judgment of the real data. In the present application, the height of the tree may be set to 100, and further the height of the tree in the isolated forest may be set to 100 as the target isolated forest.
Step S32: and carrying out anomaly detection on the coding vector according to the target isolated forest, and further obtaining an anomaly tag corresponding to the coding vector.
By way of example, the anomaly detection is performed on the coded vector through the target isolated forest, and then the anomaly tag corresponding to the coded vector is obtained.
For example, there are a user a, a user B, a user C and a user D, where the user a, the user B and the user C are all intolerant, and the user D is resistant, and if the user D is an outlier in the data, it is necessary to detect the anomaly from the user a, the user B, the user C and the user D through the target isolated forest.
For example, the training data may be divided into a plurality of groups, where the number of samples in each group may be set according to the requirement, for example, the number of training data is 120, and the training data is now randomly grouped to obtain 30 groups, where the number of samples in each group is 4, and taking one group as an example, including sample 1, sample 2, sample 3 and sample 4, the code vectors corresponding to sample 1, sample 2, sample 3 and sample 4 are first obtained respectively as code vector 1, code vector 2, code vector 3 and code vector 4, and then the target isolated forest is used to perform anomaly detection on the code vector 1, code vector 2, code vector 3 and code vector 4, and then any one of the code vector 1, code vector 2, code vector 3 and code vector 4 may be obtained as anomaly data, where the anomaly data indicates that an anomaly value is detected from the obtained code vector 1, code vector 2, code vector 3 and code vector 4, and the symptom of the antibiotic is represented by the user corresponding to the anomaly value.
In some embodiments, the anomaly detection is performed on the encoded vector according to the target isolated forest, so as to obtain an anomaly tag corresponding to the encoded vector, including: sampling the coding vector to obtain a sampling vector corresponding to the coding vector; and carrying out anomaly detection on the sampling vector according to the target isolated forest to obtain an anomaly tag corresponding to the coding vector.
For example, in order to further improve the effect of anomaly detection, the encoding vector may be sampled to obtain a characteristic representative of the encoding vector, so that normal data and abnormal data may be further distinguished more obviously, and the dimension of the vector is reduced, thereby reducing the waste of calculation time and space.
In some embodiments, the sampling the encoded vector to obtain a sampling vector corresponding to the encoded vector includes: obtaining Gaussian probability distribution parameters of the coding vector under the hidden variable space; and sampling the coding vector according to the Gaussian probability distribution parameters, so as to obtain a sampling vector.
Illustratively, a gaussian probability distribution parameter of the code vector in the hidden variable space is obtained, wherein the gaussian probability distribution parameter comprises a mean value and a variance, and the code vector is sampled by using the mean value and the variance, so that a sampling vector is obtained.
For example, if python is used for sampling, after the mean value and the variance in the gaussian probability distribution parameters corresponding to the coding vector are obtained, the dimension, the mean value and the variance of the coding vector and the sampling vector can be input by using a Lambda function in pyhon, and then the corresponding sampling vector can be obtained. Shape sample vector = Lambda (code vector, dimension of sample vector, [ mean, variance ]).
And S4, determining a loss function of the drug resistance identification model according to the drug resistance label and the abnormal label.
For example, after the anomaly identification layer, an anomaly tag corresponding to the input data can be obtained, and the anomaly tag and a drug resistance tag truly corresponding to the input data together form a loss function of the drug resistance identification model.
In some embodiments, the loss function is a cross entropy loss function.
Illustratively, the cross entropy loss function may be tanh, sigmoid, softmax, reLU or the like, and specifically may be set as required.
For example, when the input data passes through the anomaly identification layer to obtain the drug resistance data corresponding to the anomaly data, the probability value corresponding to the anomaly data is obtained, and the more the prediction categories corresponding to the input data are the same as the drug resistance labels, the better the effect of the model is indicated. The loss function may be as shown in equation 1.
Wherein,,m is the number of input samples, p i Is the probability that sample i belongs to the drug resistance tag.
And S5, carrying out iterative updating on the drug resistance identification model based on the training sample and the loss function to obtain a target drug resistance identification model.
For example, when the loss value corresponding to the loss function is higher than the threshold value, training needs to be continued according to the training sample until the loss value is lower than the threshold value, and then the target drug resistance identification model is obtained. Or setting a training round of the training sample, and taking the drug resistance identification model corresponding to the loss value in the training round as the target drug resistance identification model. Or setting a training round of the training sample, taking the corresponding drug resistance recognition model as a drug resistance recognition model to be determined when the loss value in the training round is lower than the set value, further testing the drug resistance recognition model to be determined, and taking the corresponding drug resistance recognition model to be determined as a target drug resistance recognition model when the accuracy is highest.
The application also provides a drug resistance identification method, which comprises the following steps: and inputting relevant information of the user into a drug resistance recognition model to obtain a drug resistance recognition result, wherein the drug resistance recognition model is obtained according to a training method of any drug resistance recognition model. For detailed descriptions of the related contents, please refer to the training method section of the drug resistance recognition model, and the detailed description is omitted herein.
By way of example, the training data is trained to obtain a target drug resistance recognition model, and relevant information of the user to be recognized is input into the target drug resistance recognition model, so that a drug resistance recognition result of the user to be recognized is obtained.
Referring to fig. 5, fig. 5 is a schematic block diagram of a structure of a terminal device according to an embodiment of the present application.
As shown in fig. 5, the terminal device 300 includes a processor 301 and a memory 302, the processor 301 and the memory 302 being connected by a bus 303, such as an I2C (Inter-integrated Circuit) bus.
In particular, the processor 301 is used to provide computing and control capabilities, supporting the operation of the entire server. The processor 301 may be a central processing unit (Central Processing Unit, CPU), the processor 301 may also be other general purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), field-programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. Wherein the general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
Specifically, the Memory 302 may be a Flash chip, a Read-Only Memory (ROM) disk, an optical disk, a U-disk, a removable hard disk, or the like.
It will be appreciated by those skilled in the art that the structure shown in fig. 5 is merely a block diagram of a portion of the structure associated with an embodiment of the present application and is not intended to limit the terminal device to which an embodiment of the present application is applied, and that a particular terminal device may include more or fewer components than shown, or may combine certain components, or may have a different arrangement of components.
The processor 301 is configured to run a computer program stored in the memory, and implement the training method of the drug resistance identification model provided in any embodiment of the present application when the computer program is executed.
In some embodiments, the processor 301 is configured to run a computer program stored in a memory, apply to a terminal device, and implement the following steps when executing the computer program:
obtaining a training sample, wherein the training sample comprises relevant information of a user and a drug resistance label of the user, and the relevant information is used for representing information related to drug resistance identification of the user;
coding the related information by using a coding layer of the drug resistance identification model to obtain a coding vector;
performing anomaly detection on the coded vector according to an anomaly identification layer of the drug resistance identification model to obtain an anomaly tag corresponding to the coded vector;
determining a loss function of the drug resistance identification model according to the drug resistance tag and the abnormal tag;
and carrying out iterative updating on the drug resistance identification model based on the training sample and the loss function to obtain a target drug resistance identification model.
In some embodiments, the coding layer is a convolutional neural network, the relevant information includes basic information and physiological monitoring information of the user, and the processor 301 performs, in the process of coding the relevant information by using the coding layer of the drug resistance identification model to obtain a coding vector:
encoding the basic information by utilizing the convolutional neural network to obtain a first convolutional vector corresponding to the basic information;
the convolutional neural network is utilized to encode the physiological monitoring information to obtain a second convolutional vector corresponding to the physiological monitoring information;
and determining the coding vector of the user according to the first convolution vector and the second convolution vector.
In some implementations, the processor 301 performs, in determining the coding vector for the user from the first convolution vector and the second convolution vector:
and performing nonlinear transformation according to the first convolution vector and the second convolution vector, so as to obtain the coding vector of the user after nonlinear transformation processing.
In some embodiments, the anomaly recognition layer is an isolated forest, and the processor 301 performs, in the process of obtaining the anomaly label corresponding to the encoded vector by performing anomaly detection on the encoded vector according to the anomaly recognition layer of the drug resistance recognition model:
setting parameters of trees in the isolated forest to obtain a target isolated forest;
and carrying out anomaly detection on the coding vector according to the target isolated forest, and further obtaining an anomaly tag corresponding to the coding vector.
In some embodiments, the processor 301 performs, in performing anomaly detection on the encoded vector according to the target isolated forest, to obtain an anomaly tag corresponding to the encoded vector, the following steps:
sampling the coding vector to obtain a sampling vector corresponding to the coding vector;
and carrying out anomaly detection on the sampling vector according to the target isolated forest to obtain an anomaly tag corresponding to the coding vector.
In some embodiments, the processor 301 performs, in sampling the encoded vector to obtain a sampling vector corresponding to the encoded vector:
obtaining Gaussian probability distribution parameters of the coding vector under the hidden variable space;
and sampling the coding vector according to the Gaussian probability distribution parameters, so as to obtain a sampling vector.
In some embodiments, the loss function is a cross entropy loss function.
It should be noted that, for convenience and brevity of description, a specific working process of the above-described terminal device may refer to a corresponding process in the foregoing embodiment of the training method of the drug resistance identification model, which is not described herein again.
The embodiment of the application also provides a storage medium for computer readable storage, the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of the training method of any drug resistance identification model provided by the embodiment of the specification of the application.
The storage medium may be an internal storage unit of the terminal device of the foregoing embodiment, for example, a memory of the terminal device. The storage medium may also be an external storage device of the terminal device, such as a plug-in hard disk provided on the terminal device, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), or the like.
Those of ordinary skill in the art will appreciate that all or some of the steps of the methods, functional modules/units in the apparatus disclosed above may be implemented as software, firmware, hardware, and suitable combinations thereof. In a hardware embodiment, the division between the functional modules/units mentioned in the above description does not necessarily correspond to the division of physical components; for example, one physical component may have multiple functions, or one function or step may be performed cooperatively by several physical components. Some or all of the physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on computer readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media). The term computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data, as known to those skilled in the art. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital Versatile Disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can be accessed by a computer. Furthermore, as is well known to those of ordinary skill in the art, communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media.
It should be understood that the term "and/or" as used in the present specification and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes such combinations. It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or system that comprises the element.
The foregoing embodiment numbers of the present application are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments. The present application is not limited to the above embodiments, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the scope of the present application, and these modifications and substitutions are intended to be included in the scope of the present application. Therefore, the protection scope of the application is subject to the protection scope of the claims.
Claims (10)
1. A method of training a drug resistance recognition model, the method comprising:
obtaining a training sample, wherein the training sample comprises relevant information of a user and a drug resistance label of the user, and the relevant information is used for representing information related to drug resistance identification of the user;
coding the related information by using a coding layer of the drug resistance identification model to obtain a coding vector;
performing anomaly detection on the coded vector according to an anomaly identification layer of the drug resistance identification model to obtain an anomaly tag corresponding to the coded vector;
determining a loss function of the drug resistance identification model according to the drug resistance tag and the abnormal tag;
and carrying out iterative updating on the drug resistance identification model based on the training sample and the loss function to obtain a target drug resistance identification model.
2. The method of claim 1, wherein the coding layer is a convolutional neural network, the related information includes basic information and physiological monitoring information of the user, and the coding layer using the drug resistance identification model encodes the related information to obtain a coding vector, comprising:
encoding the basic information by utilizing the convolutional neural network to obtain a first convolutional vector corresponding to the basic information;
the convolutional neural network is utilized to encode the physiological monitoring information to obtain a second convolutional vector corresponding to the physiological monitoring information;
and determining the coding vector of the user according to the first convolution vector and the second convolution vector.
3. The method of claim 2, wherein said determining the user's code vector from the first and second convolution vectors comprises:
and performing nonlinear transformation according to the first convolution vector and the second convolution vector, so as to obtain the coding vector of the user after nonlinear transformation processing.
4. The method according to claim 1, wherein the anomaly recognition layer is an isolated forest, the anomaly detection is performed on the encoded vector by the anomaly recognition layer according to the drug resistance recognition model, and an anomaly tag corresponding to the encoded vector is obtained, including:
setting parameters of trees in the isolated forest to obtain a target isolated forest;
and carrying out anomaly detection on the coding vector according to the target isolated forest, and further obtaining an anomaly tag corresponding to the coding vector.
5. The method of claim 4, wherein the performing anomaly detection on the encoded vector according to the target isolated forest to obtain an anomaly tag corresponding to the encoded vector comprises:
sampling the coding vector to obtain a sampling vector corresponding to the coding vector;
and carrying out anomaly detection on the sampling vector according to the target isolated forest to obtain an anomaly tag corresponding to the coding vector.
6. The method of claim 5, wherein sampling the encoded vector to obtain a sampled vector corresponding to the encoded vector comprises:
obtaining Gaussian probability distribution parameters of the coding vector under the hidden variable space;
and sampling the coding vector according to the Gaussian probability distribution parameters, so as to obtain a sampling vector.
7. The method of claim 1, wherein the loss function is a cross entropy loss function.
8. A method of identifying resistance, the method comprising:
inputting relevant information of a user into a drug resistance recognition model to obtain a drug resistance recognition result, wherein the drug resistance recognition model is obtained according to the training method of the drug resistance recognition model of any one of claims 1-7.
9. A terminal device, characterized in that the terminal device comprises a processor and a memory;
the memory is used for storing a computer program;
the processor is configured to execute the computer program and to implement the training method of the resistance identification model according to any one of claims 1 to 7 and the resistance identification method according to claim 8 when the computer program is executed.
10. A computer-readable storage medium, which when executed by one or more processors causes the one or more processors to perform the steps of the method of training the drug resistance identification model of any one of claims 1 to 7 and the steps of the method of drug resistance identification of claim 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310770822.7A CN116825383A (en) | 2023-06-27 | 2023-06-27 | Training method, identification method, terminal and storage medium for drug resistance identification model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310770822.7A CN116825383A (en) | 2023-06-27 | 2023-06-27 | Training method, identification method, terminal and storage medium for drug resistance identification model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116825383A true CN116825383A (en) | 2023-09-29 |
Family
ID=88112177
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310770822.7A Pending CN116825383A (en) | 2023-06-27 | 2023-06-27 | Training method, identification method, terminal and storage medium for drug resistance identification model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116825383A (en) |
-
2023
- 2023-06-27 CN CN202310770822.7A patent/CN116825383A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111767707B (en) | Method, device, equipment and storage medium for detecting Leideogue cases | |
| CN112365987B (en) | Diagnostic data abnormality detection method, diagnostic data abnormality detection device, computer device, and storage medium | |
| Adeel et al. | Entropy‐controlled deep features selection framework for grape leaf diseases recognition | |
| CN110245685B (en) | Method, system and storage medium for predicting pathogenicity of genome single-site variation | |
| Wang et al. | Research on Healthy Anomaly Detection Model Based on Deep Learning from Multiple Time‐Series Physiological Signals | |
| CN112132624A (en) | Medical claims data prediction system | |
| CN114358169B (en) | Colorectal cancer detection system based on XGBoost | |
| Li et al. | Predicting clinical outcomes with patient stratification via deep mixture neural networks | |
| WO2019180310A1 (en) | A method, an apparatus and a computer program product for an interpretable neural network representation | |
| CN110929877B (en) | Model building method, device, equipment and storage medium based on transfer learning | |
| CN113658655B (en) | Physical examination recommendation method, device, storage medium and equipment | |
| CN114428860A (en) | Pre-hospital emergency case text recognition method and device, terminal and storage medium | |
| CN116797817A (en) | Autism disease prediction technology based on self-supervision graph convolution model | |
| Yousif | Classification of mental disorders figures based on soft computing methods | |
| CN116825383A (en) | Training method, identification method, terminal and storage medium for drug resistance identification model | |
| CN115346084A (en) | Sample processing method, sample processing apparatus, electronic device, storage medium, and program product | |
| Sandhya et al. | An optimized elman neural network for contactless palm-vein recognition framework | |
| CN112447292B (en) | Human body index-stroke relation analysis system based on machine learning interpretability | |
| CN111476775B (en) | DR symptom identification device and method | |
| Mandala et al. | A Study on the Development of Machine Learing in Health Analysis | |
| Che et al. | Time series feature learning with applications to health care | |
| Padole et al. | Early Detection of Alzheimer’s Disease Using Graph Signal Processing and Deep Learning. | |
| CN113688854A (en) | Data processing method and device and computing equipment | |
| CN110580955A (en) | Disease information identification method and device, computer equipment and storage medium | |
| CN115620802B (en) | Gene data processing method and system |
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 |