Method, apparatus, and device for classifying drug on the basis of drugs prohibited in clinical trial

Abstract

The present invention belongs to the technical field of drug classification. Disclosed are a method, apparatus, and device for classifying a drug on the basis of drugs prohibited in a clinical trial. The method comprises: on the basis of inputted drug information of a drug to be queried, determining a corresponding disease type and mechanism of action; by means of a Gaussian kernel function, mapping drug information obtained after processing historical drug information to a high-dimensional feature space, obtaining, by means of kernel principal component analysis, a drug information data set that has undergone spatial transformation, on the basis of the drug information data set and by using a Gaussian Bayes algorithm, constructing a drug classification model, and finally determining, by means of the drug classification model, a drug classification corresponding to the drug to be queried; and when a prohibited drug is detected, outputting a corresponding alarm prompt. By using the method, drug classification is more targeted, and the accuracy of drug classification is improved, thereby better conforming to a clinical trial research scheme, which can effectively reduce the workload of a research physician, reduce the risk during a clinical trial, and improve the quality of clinical research.

Description

Drug classification methods, devices, and equipment based on drugs prohibited in clinical trials Technical Field

[0001] This invention relates to the field of drug classification technology, and in particular to a drug classification method, apparatus and equipment based on drugs prohibited in clinical trials. Background Technology

[0002] During clinical trials, most clinical research protocols restrict concomitant drug use, listing drugs that could interfere with clinical trial results as prohibited drugs. Prohibited drugs typically include those with similar pharmacological effects to the investigational drug, drugs that affect the pharmacokinetics of the investigational drug, and drugs that enhance or mask adverse reactions. However, each clinical trial usually has multiple categories of prohibited drugs, covering a large number of medications. It is difficult for research physicians to check each drug individually when administering medications, posing a certain risk to the accuracy of clinical trial results.

[0003] Most existing drug classification methods are based on pharmacological effects, structural characteristics, and medical insurance coverage. Such classification methods lack specificity and cannot guarantee the accuracy of classification.

[0004] The above content is only used to help understand the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main objective of this invention is to provide a drug classification method, apparatus, and equipment based on drugs prohibited in clinical trials. This aims to address the technical problem that most existing drug classification methods rely on pharmacological effects, structural characteristics, and medical insurance coverage, which lack specificity and cannot guarantee the accuracy of classification.

[0006] To achieve the above objectives, the present invention provides a drug classification method based on drugs prohibited in clinical trials, the method comprising the following steps:

[0007] Receive the drug information to be queried;

[0008] Based on the drug information, determine the disease type and mechanism of action of the drug to be queried;

[0009] Historical drug information is acquired, and the historical drug information is processed to obtain processed drug information;

[0010] The processed drug information is mapped to a high-dimensional feature space through a Gaussian kernel function, and feature information is extracted through kernel principal component analysis to obtain a drug information dataset after spatial transformation.

[0011] The spatially transformed drug information dataset is divided into a training dataset and a test dataset.

[0012] The model is trained using the Gaussian Bayes algorithm on the training dataset, and the trained model is evaluated and validated using the test dataset to complete the construction of the drug classification model.

[0013] The disease type and the mechanism of action are input into the drug classification model to determine the drug classification corresponding to the drug to be queried.

[0014] If the drug to be queried is classified as a prohibited drug, a corresponding alarm message will be output.

[0015] In some embodiments, processing the historical drug information to obtain processed drug information includes:

[0016] The historical drug information was divided into several initial datasets based on disease type and mechanism of action;

[0017] Arbitrarily select an initial dataset as the current dataset, and select several samples from the other initial datasets to add to the current dataset until several reference datasets are obtained, and the number of samples among the reference datasets is balanced.

[0018] Standardized automatic scaling is used to scale the features of samples in any reference data so that the features of samples in each reference data are of the same magnitude. Based on several reference datasets that have undergone feature scaling, the processed drug information is obtained.

[0019] In some embodiments, the expression corresponding to scaling the features of samples in any reference data using standardized autoscaling is: S j '=(S j -avg) / d, where S j 'S represents the scaled feature value of feature j, where S is the feature value after scaling. j is the original feature value of feature j before scaling, avg is the average feature value of any reference dataset, and d is the variance of the feature value of any reference dataset.

[0020] In some embodiments, the step of training the model using the Gaussian Bayes algorithm based on the training dataset and evaluating and validating the trained model using the test dataset to complete the construction of the drug classification model includes:

[0021] Determine the prior probability of each category in the training set;

[0022] Determine the conditional probability of each attribute under each category;

[0023] The distance between all samples of one category and all samples of another category is determined based on the prior probability of each category and the conditional probability of each attribute under each category.

[0024] The weighted posterior probability is determined using the class-specific attribute weight matrix and the class conditional probability, and the corresponding classification accuracy is calculated.

[0025] The NSGA-II algorithm was used for iterative optimization to determine the spacing and classification accuracy between different categories, and a set of Pareto solutions was recorded.

[0026] When the termination condition of model optimization is reached, the Pareto solution set is output to obtain multiple sets of class-specific attribute weight matrices;

[0027] Cross-validation was used to evaluate and select several optimal class-specific attribute weight matrices from the multiple sets of class-specific attribute weight matrices;

[0028] The posterior probability of the test dataset is obtained by using the test dataset based on the weight matrices of the several optimal class-specific attributes.

[0029] Based on the posterior probability, the class with the highest posterior probability is selected as the predicted class of the test sample, and the predicted class label is returned.

[0030] The classification accuracy is determined based on the category labels, and the training of the drug classification model is considered complete when the classification accuracy reaches the preset accuracy.

[0031] In some embodiments, the expression corresponding to determining the distance between all samples of one category and all samples of another category is: Where p(x) j |Y0) and p(x j |Y1) represents x under category Y0. j Attributes and Y1 category x j The conditional probabilities of the attributes, P(Y0) and P(Y1) are the prior probabilities of categories Y0 and Y1, respectively, m is the number of samples, and n is the number of features.

[0032] In some embodiments, the classification accuracy is expressed as: Where δ(·) is a bisection function used to determine the number of samples correctly classified by the model, y i y represents the predicted category label, and y represents the true label.

[0033] Furthermore, to achieve the above objectives, the present invention also proposes a drug classification device based on drugs prohibited in clinical trials, the drug classification device based on drugs prohibited in clinical trials comprising:

[0034] The receiving module is used to receive the drug information of the drug to be queried.

[0035] The acquisition module is used to determine the disease type and mechanism of action of the drug to be queried based on the drug information.

[0036] The module is used to acquire historical drug information, process the historical drug information, and obtain processed drug information.

[0037] The construction module is used to map the processed drug information to a high-dimensional feature space through a Gaussian kernel function, and extract feature information through kernel principal component analysis to obtain a drug information dataset after spatial transformation.

[0038] The construction module is used to divide the spatially transformed drug information dataset to obtain a training dataset and a test dataset.

[0039] The construction module is used to train the model based on the training dataset using the Gaussian Bayes algorithm, and to evaluate and verify the trained model using the test dataset, so as to complete the construction of the drug classification model.

[0040] The classification module is used to input the disease type and the mechanism of action into the drug classification model to determine the drug classification corresponding to the drug to be queried;

[0041] The prompt module is used to output a corresponding alarm prompt if the drug to be queried is classified as a prohibited drug.

[0042] In some embodiments, the construction module is used to divide the historical drug information into several initial datasets based on disease type and mechanism of action;

[0043] Arbitrarily select an initial dataset as the current dataset, and select several samples from the other initial datasets to add to the current dataset until several reference datasets are obtained, and the number of samples among the reference datasets is balanced.

[0044] Standardized automatic scaling is used to scale the features of samples in any reference data so that the features of samples in each reference data are of the same magnitude. Based on several reference datasets that have undergone feature scaling, the processed drug information is obtained.

[0045] In some embodiments, the expression corresponding to scaling the features of samples in any reference data using standardized autoscaling is: S j '=(S j -avg) / d, where S j 'S represents the scaled feature value of feature j, where S is the feature value after scaling. jis the original feature value of feature j before scaling, avg is the average feature value of any reference dataset, and d is the variance of the feature value of any reference dataset.

[0046] Furthermore, to achieve the above objectives, the present invention also proposes a drug classification device based on drugs prohibited in clinical trials. The drug classification device based on drugs prohibited in clinical trials includes: a memory, a processor, and a drug classification program based on drugs prohibited in clinical trials stored in the memory and executable on the processor. The drug classification program based on drugs prohibited in clinical trials is configured to implement the steps of the drug classification method based on drugs prohibited in clinical trials as described above.

[0047] This invention determines the corresponding disease type and mechanism of action based on the input drug information of the drug to be queried; the processed drug information is mapped to a high-dimensional feature space using a Gaussian kernel function, and a spatially transformed drug information dataset is obtained through kernel principal component analysis. Based on the drug information dataset, a drug classification model is constructed using the Gaussian Bayes algorithm, and finally, the drug classification model determines the drug category corresponding to the queried drug; and a corresponding alarm is output when a prohibited drug is detected. This approach makes drug classification more targeted and improves its accuracy, thus better aligning with clinical trial research protocols, effectively reducing the workload of research physicians, minimizing risks during clinical trials, and improving the quality of clinical research. Attached Figure Description

[0048] Figure 1 is a flowchart illustrating the first embodiment of the drug classification method based on drugs prohibited in clinical trials of the present invention;

[0049] Figure 2 is a schematic diagram of the class-specific attribute weight matrix in the drug classification method based on drugs prohibited in clinical trials of the present invention;

[0050] Figure 3 is a structural block diagram of the first embodiment of the drug classification device based on drugs prohibited in clinical trials of the present invention.

[0051] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0052] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0053] This invention provides a drug classification method based on drugs prohibited in clinical trials. Referring to Figure 1, Figure 1 is a flowchart illustrating the first embodiment of this invention's drug classification method based on drugs prohibited in clinical trials.

[0054] In this embodiment, the drug classification method based on drugs prohibited in clinical trials includes the following steps:

[0055] Step S10: Receive the input drug information for the drug to be queried.

[0056] In this embodiment, the executing entity is a drug classification device based on drugs prohibited in clinical trials. This drug classification device has functions such as data processing, data communication, and program execution. The drug classification device based on drugs prohibited in clinical trials can be a computer terminal device or other network device, or other devices with similar functions. This embodiment does not limit the scope of the application.

[0057] It should be noted that most existing drug classification methods are based on pharmacological effects, structural characteristics, and medical insurance coverage. Such classification methods lack specificity and cannot guarantee the accuracy of the classification.

[0058] To address the aforementioned technical issues, this embodiment receives input drug information for a drug to be queried; determines the disease type and mechanism of action corresponding to the drug based on the drug information; constructs a drug classification model based on historical drug information; inputs the disease type and mechanism of action into the drug classification model to determine the drug classification corresponding to the drug to be queried; if the drug classification of the drug to be queried is a prohibited drug, a corresponding alarm is output. This method makes drug classification more targeted and improves its accuracy, thus better aligning with clinical trial research protocols. It can effectively reduce the workload of research physicians, decrease risks during clinical trials, and improve the quality of clinical research. Specifically, it can be implemented as follows.

[0059] In this specific implementation, the drug information to be queried needs to be received first. When the doctor conducts the drug query, he will input the relevant drug information, which includes, but is not limited to, the name and use of the drug.

[0060] Step S20: Determine the disease type and mechanism of action of the drug to be queried based on the drug information.

[0061] It should be noted that, unlike current drug classification, the drug classification in this embodiment is based on the disease type and mechanism of action of the drug. Based on the drug information obtained above, the disease type and mechanism of action of the drug to be queried can be determined.

[0062] Step S30: Obtain historical drug information, process the historical drug information, and obtain processed drug information.

[0063] Step S40: The processed drug information is mapped to a high-dimensional feature space through a Gaussian kernel function, and feature information is extracted through kernel principal component analysis to obtain a drug information dataset after spatial transformation.

[0064] Step S50: Divide the spatially transformed drug information dataset into a training dataset and a test dataset.

[0065] Step S60: Train the model using the Gaussian Bayes algorithm based on the training dataset, and evaluate and verify the trained model using the test dataset to complete the construction of the drug classification model.

[0066] It should be noted that the Gaussian kernel function has the following form: K(x,y)=exp(-γ||xy|| 2 ), where and y are two points in the original space, and γ is a parameter, ||xy|| 2 This represents the Euclidean distance between x and y. x and y can be obtained from the processed drug information, and the mapping to the high-dimensional feature space is achieved through the aforementioned functional relationship. The kernel principal component analysis used in this embodiment is an extension of principal component analysis (PCA) for handling dimensionality reduction problems of nonlinear data. Unlike standard PCA, kernel principal component analysis allows for the processing of nonlinear data, mapping the data to a high-dimensional space for better dimensionality reduction operations and enabling the extraction of nonlinear information from the data.

[0067] The process of processing the historical drug information to obtain processed drug information specifically includes: dividing the historical drug information into several initial datasets based on disease type and mechanism of action; arbitrarily selecting one initial dataset as the current dataset, and adding several samples from the remaining initial datasets to the current dataset until several reference datasets are obtained, with the number of samples in each reference dataset being balanced; scaling the features of the samples in any reference dataset using standardized automatic scaling to make the features of the samples in each reference dataset at the same magnitude, and obtaining the processed drug information based on the several reference datasets with completed feature scaling.

[0068] It should be noted that, to ensure sample balance, in this embodiment, for any dataset, several samples are added from other datasets. For example, the sample sizes of datasets 1 to 4 are 100, 66, 110, and 90 respectively. The adjusted dataset 1 consists of all samples from dataset 1 plus 34 randomly selected samples from each of the other datasets 2 to 4. This process is repeated for each dataset to achieve sample balance. The expression for scaling the features of samples in any reference data using standardized automatic scaling is: S j '=(S j -avg) / d, where S j 'S represents the scaled feature value of feature j, where S is the feature value after scaling. j is the original feature value of feature j before scaling, avg is the average feature value of any reference dataset, and d is the variance of the feature value of any reference dataset.

[0069] Furthermore, in this embodiment, the Gaussian Bayes algorithm is used to train the model based on the training dataset, and the trained model is evaluated and validated using the test dataset. The process of constructing the drug classification model specifically involves: determining the prior probability of each category in the training set; determining the conditional probability of each attribute under each category; determining the distance between all samples of one category and all samples of another category based on the prior probability of each category and the conditional probability of each attribute under each category; determining the weighted posterior probability using the class-specific attribute weight matrix and the class conditional probability, and calculating the corresponding classification accuracy; and using the NSGA-II algorithm for iterative optimization to determine different... The model calculates the inter-class spacing and classification accuracy, and records a set of Pareto solutions. When the model optimization termination condition is met, the Pareto solution set is output to obtain multiple sets of class-specific attribute weight matrices. Cross-validation is used to evaluate and select several optimal class-specific attribute weight matrices from the multiple sets of class-specific attribute weight matrices. The posterior probability of the test dataset is obtained based on the several optimal class-specific attribute weight matrices. Based on the posterior probability, the class with the highest posterior probability is selected as the predicted class of the test sample, and the predicted class label is returned. The classification accuracy is determined based on the class label, and the training of the drug classification model is considered complete when the classification accuracy reaches a preset accuracy.

[0070] It should be noted that the expression for determining the distance between all samples of one class and all samples of another class is: Where p(x) j |Y0) and p(x j |Y1) represents x under category Y0. j Attributes and Y1 category x jThe conditional probabilities of the attributes, P(Y0) and P(Y1), are the prior probabilities of class Y0 and class Y1, respectively, where m is the number of samples and n is the number of features. The expression for classification accuracy is: Where δ(·) is a bisection function used to determine the number of samples correctly classified by the model, y i y represents the predicted category label, and y represents the true label. The resulting class-specific attribute weight matrix is ​​shown in Figure 2. Finally, using the class-specific attribute weight matrix, according to the formula: To calculate the posterior probability of each instance, y represents the class to which the instance belongs, Y represents the set of all classes, m represents the number of all attributes in the dataset, and W y,i This represents the weight assigned to the i-th attribute under category y.

[0071] Step S70: Input the disease type and the mechanism of action into the drug classification model to determine the drug classification corresponding to the drug to be queried.

[0072] In the specific implementation, after the drug classification model is constructed, the obtained disease type and mechanism of action are input into the drug classification model to obtain the specific classification of the drug.

[0073] Step S80: If the drug to be queried is classified as a prohibited drug, then output the corresponding alarm prompt.

[0074] It should be understood that prohibited drugs can affect the results of clinical trials. When the drug to be queried is identified as a prohibited drug, this embodiment will output a corresponding alarm prompt to avoid prescribing the wrong drug.

[0075] Furthermore, in this embodiment, when searching for drugs prohibited in clinical trials, users can use the following method: Step 1: Enter drug information into the drug storage database according to the above drug classification method; Step 2: Obtain drug data from the database, and analyze and preprocess the drug information to generate an XML file; Step 3: Serialize the generated drug information XML file into JSON format data; Step 4: Transmit the serialized JSON format data to the WEB server via the network; Step 5: When the client browser sends a drug query request, the server responds to the browser's request and returns the drug information in JSON format to the browser; Step 6: The browser parses the returned JSON data; Step 7: Display the query results.

[0076] Furthermore, the drug storage information database is built within a drug management system employing the MVC pattern, separating the input, processing, and output of drug information into model, view, and control layers. The drug management system is programmed using Java technology, utilizing the user-friendly and powerful Dreamweaver as the development tool. The server-side is written in Java, while the client-side scripts are written in JavaScript. The database is Oracle 19c, and the server middleware is WebSphere Application Server 8.5. Struts session management, filters, and database integration technology are used to build the web application, and Vue.js is used to write the web pages. Specifically, step 3 involves: Step 3.1: Loading the drug information XML file into memory; Step 3.2: Parsing the XML file to form a tree structure; Step 3.3: Traversing the tree, finding the nodes (Node), and obtaining the basic element (Element); Step 3.4: Parsing into JSON data format; Step 6 involves: Step 6.1: The browser receives the JSON data from the server; Step 6.2: The browser parses the JSON data and obtains the corresponding JSONObject data; Step 6.3: Proceed to step 7.

[0077] In this embodiment, the corresponding disease type and mechanism of action are determined based on the input drug information of the drug to be queried. The processed historical drug information is mapped to a high-dimensional feature space using a Gaussian kernel function, and a spatially transformed drug information dataset is obtained through kernel principal component analysis. Based on this dataset, a drug classification model is constructed using the Gaussian Bayes algorithm. Finally, the drug classification model determines the drug category corresponding to the queried drug. When a prohibited drug is detected, a corresponding alarm is output. This approach makes drug classification more targeted and improves its accuracy, thus better aligning with clinical trial research protocols. It effectively reduces the workload of research physicians, minimizes risks during clinical trials, and improves the quality of clinical research.

[0078] Referring to Figure 3, which is a structural block diagram of the first embodiment of the drug classification device based on drugs prohibited in clinical trials of the present invention.

[0079] As shown in Figure 3, the drug classification device based on drugs prohibited in clinical trials proposed in this embodiment of the invention includes:

[0080] The receiving module 10 is used to receive the drug information of the drug to be queried.

[0081] The acquisition module 20 is used to determine the disease type and mechanism of action of the drug to be queried based on the drug information.

[0082] The construction module 30 is used to acquire historical drug information, process the historical drug information, and obtain processed drug information.

[0083] The construction module 30 is used to map the processed drug information to a high-dimensional feature space through a Gaussian kernel function, and extract feature information through kernel principal component analysis to obtain a drug information dataset after spatial transformation.

[0084] The construction module 30 is used to divide the spatially transformed drug information dataset to obtain a training dataset and a test dataset.

[0085] The construction module 30 is used to train the model based on the training dataset using the Gaussian Bayes algorithm, and to evaluate and verify the trained model using the test dataset, so as to complete the construction of the drug classification model.

[0086] The classification module 40 is used to input the disease type and the mechanism of action into the drug classification model to determine the drug classification corresponding to the drug to be queried;

[0087] The prompt module 50 is used to output a corresponding alarm prompt if the drug to be queried is classified as a prohibited drug.

[0088] In this embodiment, the corresponding disease type and mechanism of action are determined based on the input drug information of the drug to be queried. The processed historical drug information is mapped to a high-dimensional feature space using a Gaussian kernel function, and a spatially transformed drug information dataset is obtained through kernel principal component analysis. Based on this dataset, a drug classification model is constructed using the Gaussian Bayes algorithm. Finally, the drug classification model determines the drug category corresponding to the queried drug. When a prohibited drug is detected, a corresponding alarm is output. This approach makes drug classification more targeted and improves its accuracy, thus better aligning with clinical trial research protocols. It effectively reduces the workload of research physicians, minimizes risks during clinical trials, and improves the quality of clinical research.

[0089] In some embodiments, the construction module 30 is used to divide the historical drug information into several initial datasets based on disease type and mechanism of action;

[0090] Arbitrarily select an initial dataset as the current dataset, and select several samples from the other initial datasets to add to the current dataset until several reference datasets are obtained, and the number of samples among the reference datasets is balanced.

[0091] Standardized automatic scaling is used to scale the features of samples in any reference data so that the features of samples in each reference data are of the same magnitude. Based on several reference datasets that have undergone feature scaling, the processed drug information is obtained.

[0092] In some embodiments, the expression corresponding to scaling the features of samples in any reference data using standardized autoscaling is: S j '=(S j -avg) / d, where S j 'S represents the scaled feature value of feature j, where S is the feature value after scaling. j is the original feature value of feature j before scaling, avg is the average feature value of any reference dataset, and d is the variance of the feature value of any reference dataset.

[0093] In some embodiments, the construction module 30 is configured to determine the prior probability of each category in the training set;

[0094] Determine the conditional probability of each attribute under each category;

[0095] The distance between all samples of one category and all samples of another category is determined based on the prior probability of each category and the conditional probability of each attribute under each category.

[0096] The weighted posterior probability is determined using the class-specific attribute weight matrix and the class conditional probability, and the corresponding classification accuracy is calculated.

[0097] The NSGA-II algorithm was used for iterative optimization to determine the spacing and classification accuracy between different categories, and a set of Pareto solutions was recorded.

[0098] When the termination condition of model optimization is reached, the Pareto solution set is output to obtain multiple sets of class-specific attribute weight matrices;

[0099] Cross-validation was used to evaluate and select several optimal class-specific attribute weight matrices from the multiple sets of class-specific attribute weight matrices;

[0100] The posterior probability of the test dataset is obtained by using the test dataset based on the weight matrices of the several optimal class-specific attributes.

[0101] Based on the posterior probability, the class with the highest posterior probability is selected as the predicted class of the test sample, and the predicted class label is returned.

[0102] The classification accuracy is determined based on the category labels, and the training of the drug classification model is considered complete when the classification accuracy reaches the preset accuracy.

[0103] In some embodiments, the expression corresponding to determining the distance between all samples of one category and all samples of another category is: Where p(x) j |Y0) and p(x j |Y1) represents x under category Y0. j Attributes and Y1 category x j The conditional probabilities of the attributes, P(Y0) and P(Y1) are the prior probabilities of categories Y0 and Y1, respectively, m is the number of samples, and n is the number of features.

[0104] In some embodiments, the classification accuracy is expressed as: Where δ(·) is a bisection function used to determine the number of samples correctly classified by the model, y i y represents the predicted category label, and y represents the true label.

[0105] This application also provides a drug classification device based on drugs prohibited in clinical trials, including a processor, a communication interface, a memory, and a communication bus. The processor, communication interface, and memory communicate with each other through the communication bus. The memory is used to store a drug classification program based on drugs prohibited in clinical trials. When the processor executes the program stored in the memory, it implements the above-mentioned drug classification method based on drugs prohibited in clinical trials.

[0106] The communication bus mentioned in the aforementioned drug classification equipment based on drugs prohibited in clinical trials can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc.

[0107] The communication interface is used for communication between the aforementioned drug classification device based on drugs banned in clinical trials and other devices.

[0108] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0109] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0110] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid state disk (SSD)).

[0111] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0112] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0113] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

[0114] It should be understood that the above are merely illustrative examples and do not constitute any limitation on the technical solutions of the present invention. In specific applications, those skilled in the art can make settings as needed, and the present invention does not impose any restrictions on this.

[0115] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of this invention. In practical applications, those skilled in the art can select some or all of the workflow to achieve the purpose of this embodiment according to actual needs, and no restrictions are imposed here.

[0116] In addition, for technical details not described in detail in this embodiment, please refer to the drug classification method based on clinical trial prohibited drugs provided in any embodiment of the present invention, which will not be repeated here.

[0117] Furthermore, it should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0118] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0119] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory (ROM) / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0120] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

[0121] It is understood that the system provided in the embodiments of the present invention corresponds to the method provided in the embodiments of the present invention, and the explanation, examples and beneficial effects of the relevant content can be referred to the corresponding parts of the above methods.

Claims

1. A drug classification method based on drugs prohibited in clinical trials, characterized in that, The drug classification method based on drugs prohibited in clinical trials includes: Receive the drug information to be queried; Based on the drug information, determine the disease type and mechanism of action of the drug to be queried; Historical drug information is acquired, and the historical drug information is processed to obtain processed drug information; The processed drug information is mapped to a high-dimensional feature space through a Gaussian kernel function, and feature information is extracted through kernel principal component analysis to obtain a drug information dataset after spatial transformation. The spatially transformed drug information dataset is divided into a training dataset and a test dataset. The model is trained using the Gaussian Bayes algorithm on the training dataset, and the trained model is evaluated and validated using the test dataset to complete the construction of the drug classification model. The disease type and the mechanism of action are input into the drug classification model to determine the drug classification corresponding to the drug to be queried. If the drug to be queried is classified as a prohibited drug, a corresponding alarm message will be output.

2. The drug classification method based on drugs prohibited in clinical trials as described in claim 1, characterized in that, The process of processing the historical drug information to obtain processed drug information includes: The historical drug information was divided into several initial datasets based on disease type and mechanism of action; Arbitrarily select an initial dataset as the current dataset, and select several samples from the other initial datasets to add to the current dataset until several reference datasets are obtained, and the number of samples among the reference datasets is balanced. Standardized automatic scaling is used to scale the features of samples in any reference data so that the features of samples in each reference data are of the same magnitude. Based on several reference datasets that have undergone feature scaling, the processed drug information is obtained.

3. The drug classification method based on drugs prohibited in clinical trials as described in claim 2, characterized in that, The expression corresponding to scaling the features of samples in any reference data using standardized automatic scaling is: S j '=(S j -avg) / d, where S j 'S represents the scaled feature value of feature j, where S is the feature value after scaling. j is the original feature value of feature j before scaling, avg is the average feature value of any reference dataset, and d is the variance of the feature value of any reference dataset.

4. The drug classification method based on drugs prohibited in clinical trials as described in claim 1, characterized in that, The process of training a model using the Gaussian Bayes algorithm on the training dataset and evaluating and validating the trained model using the test dataset to complete the construction of a drug classification model includes: Determine the prior probability of each category in the training set; Determine the conditional probability of each attribute under each category; The distance between all samples of one category and all samples of another category is determined based on the prior probability of each category and the conditional probability of each attribute under each category. The weighted posterior probability is determined using the class-specific attribute weight matrix and the class conditional probability, and the corresponding classification accuracy is calculated. The NSGA-II algorithm was used for iterative optimization to determine the spacing and classification accuracy between different categories, and a set of Pareto solutions was recorded. When the termination condition of model optimization is reached, the Pareto solution set is output to obtain multiple sets of class-specific attribute weight matrices; Cross-validation was used to evaluate and select several optimal class-specific attribute weight matrices from the multiple sets of class-specific attribute weight matrices; The posterior probability of the test dataset is obtained by using the test dataset based on the weight matrices of the several optimal class-specific attributes. Based on the posterior probability, the class with the highest posterior probability is selected as the predicted class of the test sample, and the predicted class label is returned. The classification accuracy is determined based on the category labels, and the training of the drug classification model is considered complete when the classification accuracy reaches the preset accuracy.

5. The drug classification method based on drugs prohibited in clinical trials as described in claim 4, characterized in that, The expression corresponding to determining the distance between all samples of one category and all samples of another category is: Where p(x) j |Y0) and p(x j |Y1) represents x under category Y0. j Attributes and Y1 category x j The conditional probabilities of the attributes, P(Y0) and P(Y1) are the prior probabilities of categories Y0 and Y1, respectively, m is the number of samples, and n is the number of features.

6. The drug classification method based on drugs prohibited in clinical trials as described in claim 4, characterized in that, The expression for the classification accuracy is: Where δ(·) is a bisection function used to determine the number of samples correctly classified by the model, y i y represents the predicted category label, and y represents the true label.

7. A drug classification device based on drugs banned in clinical trials, characterized in that, The drug classification device based on drugs prohibited in clinical trials includes: The receiving module is used to receive the drug information of the drug to be queried. The acquisition module is used to determine the disease type and mechanism of action of the drug to be queried based on the drug information. The module is used to acquire historical drug information, process the historical drug information, and obtain processed drug information. The construction module is used to map the processed drug information to a high-dimensional feature space through a Gaussian kernel function, and extract feature information through kernel principal component analysis to obtain a drug information dataset after spatial transformation. The construction module is used to divide the spatially transformed drug information dataset to obtain a training dataset and a test dataset. The construction module is used to train the model based on the training dataset using the Gaussian Bayes algorithm, and to evaluate and verify the trained model using the test dataset, so as to complete the construction of the drug classification model. The classification module is used to input the disease type and the mechanism of action into the drug classification model to determine the drug classification corresponding to the drug to be queried; The prompt module is used to output a corresponding alarm prompt if the drug to be queried is classified as a prohibited drug.

8. The drug classification device based on drugs prohibited in clinical trials as described in claim 7, characterized in that, The construction module is used to divide the historical drug information into several initial datasets based on disease type and mechanism of action; Arbitrarily select an initial dataset as the current dataset, and select several samples from the other initial datasets to add to the current dataset until several reference datasets are obtained, and the number of samples among the reference datasets is balanced. Standardized automatic scaling is used to scale the features of samples in any reference data so that the features of samples in each reference data are of the same magnitude. Based on several reference datasets that have undergone feature scaling, the processed drug information is obtained.

9. The drug classification device based on drugs prohibited in clinical trials as described in claim 8, characterized in that, The expression corresponding to scaling the features of samples in any reference data using standardized automatic scaling is: S j '=(S j -avg) / d, where S j 'S represents the scaled feature value of feature j, where S is the feature value after scaling. j is the original feature value of feature j before scaling, avg is the average feature value of any reference dataset, and d is the variance of the feature value of any reference dataset.

10. A drug classification device based on drugs banned in clinical trials, characterized in that, The drug classification device based on drugs prohibited in clinical trials includes: a memory, a processor, and a drug classification program based on drugs prohibited in clinical trials stored on the memory and executable on the processor, wherein the drug classification program based on drugs prohibited in clinical trials is configured to implement the steps of the drug classification method based on drugs prohibited in clinical trials as described in any one of claims 1 to 6.

Images