CN111863181A - Medicine recommendation method and device, computer equipment and storage medium - Google Patents

Medicine recommendation method and device, computer equipment and storage medium Download PDF

Info

Publication number
CN111863181A
CN111863181A CN202010679984.6A CN202010679984A CN111863181A CN 111863181 A CN111863181 A CN 111863181A CN 202010679984 A CN202010679984 A CN 202010679984A CN 111863181 A CN111863181 A CN 111863181A
Authority
CN
China
Prior art keywords
recommendation
patient
data
drug
medication
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
Application number
CN202010679984.6A
Other languages
Chinese (zh)
Inventor
王傲迪
张鹏
姚鸣
王凯
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai Zhiben Medical Laboratory Co ltd
Origimed Technology Shanghai Co ltd
Original Assignee
Shanghai Zhiben Medical Laboratory Co ltd
Origimed Technology Shanghai Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shanghai Zhiben Medical Laboratory Co ltd, Origimed Technology Shanghai Co ltd filed Critical Shanghai Zhiben Medical Laboratory Co ltd
Priority to CN202010679984.6A priority Critical patent/CN111863181A/en
Publication of CN111863181A publication Critical patent/CN111863181A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H20/00—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
    • G16H20/10—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to drugs or medications, e.g. for ensuring correct administration to patients
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/24323—Tree-organised classifiers
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
    • G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H70/00—ICT specially adapted for the handling or processing of medical references
    • G16H70/40—ICT specially adapted for the handling or processing of medical references relating to drugs, e.g. their side effects or intended usage

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Public Health (AREA)
  • Medical Informatics (AREA)
  • Theoretical Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Primary Health Care (AREA)
  • General Health & Medical Sciences (AREA)
  • Epidemiology (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Chemical & Material Sciences (AREA)
  • Medicinal Chemistry (AREA)
  • Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Databases & Information Systems (AREA)
  • Pathology (AREA)
  • Pharmacology & Pharmacy (AREA)
  • Toxicology (AREA)
  • Medical Treatment And Welfare Office Work (AREA)

Abstract

The application relates to a medicine recommendation method, a medicine recommendation device, computer equipment and a storage medium. The method comprises the following steps: acquiring personal characteristic data of a patient, physical condition data of the patient, historical medication data of the patient and a plurality of drug recommendation schemes; inputting the personal characteristic data, the physical condition data, the historical medication data, and the plurality of medication recommendation programs into a medication recommendation model; and for each drug recommendation scheme, acquiring a recommendation result output by the drug recommendation model, wherein the recommendation result is used for representing whether the drug recommendation scheme is applicable to the patient or not, and the recommendation result comprises a recommendation state and a non-recommendation state. By adopting the method, the accuracy of medicine recommendation can be improved.

Description

Medicine recommendation method and device, computer equipment and storage medium
Technical Field
The present application relates to the field of computer technologies, and in particular, to a method and an apparatus for recommending a medicine, a computer device, and a storage medium.
Background
Along with the improvement of the quality of life of people, people pay more attention to the health condition of the people, wherein when the people need to take medicine, people hope to obtain more accurate medicine recommendation.
In the conventional art, medicines are recommended to people by matching symptoms of a human body with known characteristics of the medicines. For example, a drug may be recommended during development and clinical trials when an individual shows corresponding or similar symptoms and needs to take the drug.
However, in this way, since the medicine is recommended to people simply according to the matching relationship between the symptoms and the medicine, the medicine recommendation result finally obtained by people is not accurate.
Disclosure of Invention
In view of the above, it is necessary to provide a medicine recommendation method, apparatus, computer device and storage medium capable of improving accuracy.
In a first aspect, a recommendation method is provided, the method including:
acquiring personal characteristic data of a patient, physical condition data of the patient, historical medication data of the patient and a plurality of drug recommendation schemes; the personal characteristic data is used for characterizing individual differences of the patients; the physical condition data is used to characterize the health condition of the patient; the historical medication data is used for reflecting the characteristics of the medicine historically used by the patient; each of the drug recommendation schemes includes a plurality of drugs;
Inputting the personal characteristic data, the physical condition data, the historical medication data and the plurality of medicine recommendation schemes into a medicine recommendation model;
and for each drug recommendation scheme, acquiring a recommendation result output by the drug recommendation model, wherein the recommendation result is used for representing whether the drug recommendation scheme is applicable to the patient or not, and the recommendation result comprises a recommendation state and a non-recommendation state.
In one embodiment, the drug recommendation model comprises a medication effect prediction model and a patient duration prediction model, wherein the medication effect prediction model is used for predicting the treatment effect of each drug recommendation scheme on the patient, and the patient duration prediction model is used for predicting the duration of life of the patient under each drug recommendation scheme; for each drug recommendation scheme, obtaining a recommendation result output by the drug recommendation model includes:
for each medicine recommendation scheme, respectively acquiring a first recommendation value output by the medication effect prediction model and a second recommendation value output by the patient survival time prediction model; the first recommended value is used for representing the treatment effect of the medicine recommended scheme on the patient, and the second recommended value is used for representing the survival time of the patient under the medicine recommended scheme;
And obtaining the recommendation result according to the first recommendation value and the second recommendation value.
In one embodiment, the obtaining the recommendation result according to the first recommendation value and the second recommendation value includes:
for each medicine recommendation scheme, if the first recommendation value is greater than a first preset threshold value and the second recommendation value is greater than a second preset threshold value, outputting the recommendation result as a recommendation state;
and if the first recommended value is smaller than the first preset threshold value and/or the second recommended value is smaller than the second recommended threshold value, outputting the recommended result as a non-recommended state.
In one embodiment, the medication effect prediction model is constructed from a random forest model; the patient survival time prediction model is constructed by a multi-factor survival equation.
In one embodiment, the personal characteristic data includes at least one of an age of the patient, a sex of the patient, and a cancer stage data of the patient, the physical condition data includes at least one of a vital sign data of the patient and a genomic data of the patient, the historical medication data includes at least one of a type of a drug historically used by the patient, a dose of a drug historically used by the patient, and a treatment historically used by the patient, each of the drug recommendations includes at least one of a plurality of different types of drugs, a dose of each type of drug, and a plurality of treatment modalities including surgical resection, radiation therapy, chemotherapy, immunotherapy, and targeted therapy.
In one embodiment, before the inputting the personal characteristic data, the physical condition data, the historical medication data, and the plurality of drug recommendations into a drug recommendation model, the method further comprises:
acquiring a training data set, wherein the training data set comprises personal characteristic data of at least two patients, physical condition data of at least two patients and historical medication data of at least two patients;
and inputting the training data set into an initial medicine recommendation model to obtain the medicine recommendation model, wherein the initial medicine recommendation model comprises an initial medicine effect prediction model and an initial patient survival time prediction model.
In one embodiment, the inputting the training data set into the initial drug recommendation model comprises:
preprocessing the data in the training data set to obtain a preprocessed training data set, wherein the preprocessing comprises normalization processing and data cleaning;
inputting the preprocessed training data set into the initial drug recommendation model.
In a second aspect, there is provided a medication recommendation device, the device comprising:
the system comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring personal characteristic data of a patient, physical condition data of the patient, historical medication data of the patient and a plurality of medicine recommendation schemes; the personal characteristic data is used for characterizing individual differences of the patients; the physical condition data is used to characterize the health condition of the patient; the historical medication data is used for reflecting the characteristics of the medicine historically used by the patient; each of the drug recommendation schemes includes a plurality of drugs;
The input module is used for inputting the personal characteristic data, the physical condition data, the historical medication data and the plurality of medicine recommendation schemes into a medicine recommendation model;
and the second obtaining module is used for obtaining a recommendation result output by the medicine recommendation model for each medicine recommendation scheme, wherein the recommendation result is used for representing whether the medicine recommendation scheme is applicable to the patient or not, and the recommendation result comprises a recommendation state and a non-recommendation state.
In a third aspect, a computer device is provided, comprising a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program:
acquiring personal characteristic data of a patient, physical condition data of the patient, historical medication data of the patient and a plurality of drug recommendation schemes; the personal characteristic data is used for characterizing individual differences of the patients; the physical condition data is used to characterize the health condition of the patient; the historical medication data is used for reflecting the characteristics of the medicine historically used by the patient; each of the drug recommendation schemes includes a plurality of drugs;
inputting the personal characteristic data, the physical condition data, the historical medication data and the plurality of medicine recommendation schemes into a medicine recommendation model;
And for each drug recommendation scheme, acquiring a recommendation result output by the drug recommendation model, wherein the recommendation result is used for representing whether the drug recommendation scheme is applicable to the patient or not, and the recommendation result comprises a recommendation state and a non-recommendation state.
In a fourth aspect, there is provided a computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of:
acquiring personal characteristic data of a patient, physical condition data of the patient, historical medication data of the patient and a plurality of drug recommendation schemes; the personal characteristic data is used for characterizing individual differences of the patients; the physical condition data is used to characterize the health condition of the patient; the historical medication data is used for reflecting the characteristics of the medicine historically used by the patient; each of the drug recommendation schemes includes a plurality of drugs;
inputting the personal characteristic data, the physical condition data, the historical medication data and the plurality of medicine recommendation schemes into a medicine recommendation model;
and for each drug recommendation scheme, acquiring a recommendation result output by the drug recommendation model, wherein the recommendation result is used for representing whether the drug recommendation scheme is applicable to the patient or not, and the recommendation result comprises a recommendation state and a non-recommendation state.
According to the medicine recommendation method, the medicine recommendation device, the computer equipment and the storage medium, the personal characteristic data of the patient, the physical condition data of the patient, the historical medication data of the patient and a plurality of medicine recommendation schemes are obtained; the personal characteristic data is used for characterizing individual differences of the patients; the physical condition data is used to characterize the health condition of the patient; the historical medication data is used for reflecting the characteristics of the medicine historically used by the patient; each of the drug recommendation schemes includes a plurality of drugs; then inputting the personal characteristic data, the physical condition data, the historical medication data and the plurality of medicine recommendation schemes into a medicine recommendation model; and finally, for each medicine recommendation scheme, acquiring a recommendation result output by the medicine recommendation model, wherein the recommendation result is used for representing whether the medicine recommendation scheme is suitable for the patient or not, and the recommendation result comprises a recommendation state and an unremitting state. According to the medicine recommendation method, in the process of recommending the medicine for the patient, the information of multiple dimensions of the patient is fully considered, and then the information of the multiple dimensions of the patient is input into the medicine recommendation model, so that the medicine recommendation scheme finally output by the medicine recommendation model can be more suitable for the patient, and the final medicine recommendation result is more accurate.
Drawings
FIG. 1 is a flow chart illustrating a method for drug recommendation in one embodiment;
FIG. 2 is a flowchart illustrating a method for obtaining recommendation results output by the drug recommendation model for each of the drug recommendation scenarios in the drug recommendation method according to an embodiment;
FIG. 3 is a flowchart illustrating a method for obtaining the recommendation result according to the first recommendation value and the second recommendation value in the drug recommendation method according to an embodiment;
FIG. 4 is a flowchart illustrating a method for training a drug recommendation model in a drug recommendation method according to an embodiment;
FIG. 5 is a schematic flow chart diagram illustrating a method for inputting the training data set into an initial drug recommendation model in the drug recommendation method in one embodiment;
FIG. 6 is a block diagram of the structure of a drug recommendation device in one embodiment;
FIG. 7 is a block diagram of the structure of a drug recommendation device in one embodiment;
FIG. 8 is a diagram illustrating an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
In the embodiment of the present application, as shown in fig. 1, a medicine recommendation method is provided, and this embodiment is illustrated by applying the method to a terminal, it can be understood that the method may also be applied to a server, and may also be applied to a system including the terminal and the server, and is implemented through interaction between the terminal and the server. The terminal can be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers and portable wearable devices, and the server can be implemented by an independent server or a server cluster formed by a plurality of servers. In this embodiment, the method includes the steps of:
step 101, a terminal acquires personal characteristic data of a patient, physical condition data of the patient, historical medication data of the patient and a plurality of medicine recommendation schemes; the personal characteristic data is used for characterizing individual differences of the patients; the physical condition data is used to characterize the health condition of the patient; the historical medication data is used for reflecting the characteristics of the medicine historically used by the patient; each of the drug recommendations includes a plurality of drugs.
In this step, before recommending the medicine to the patient, personalized data of multiple dimensions of the patient needs to be acquired, which is equivalent to establishing a personalized file for each patient, so that it is beneficial to recommend the adapted medicine to the patients of different individuals. In general, personal characteristic data, physical condition data, historical medication data, and a plurality of drug recommendations provided for a patient are used as the basis for creating a personalized profile of the patient. Wherein, the personal characteristic data of the patient can be used as a fixed influence factor in the medication process; the physical condition data of the patient can be used as a precondition reference influence factor for medication; the patient's historical medication data may be used as a reference factor for medication.
And 102, inputting the personal characteristic data, the physical condition data, the historical medication data and the plurality of medicine recommendation schemes into a medicine recommendation model by the terminal.
After acquiring personal characteristic data, physical condition data, historical medication data and a plurality of drug recommendation schemes of a patient, the data needs to be input into a drug recommendation model. Wherein different weights may be set for different data.
When the data is input into the medicine recommendation model, different input modes are available, for example, personal characteristic data, physical condition data, historical medication data and a plurality of medicine recommendation schemes of a certain patient can be input into the medicine recommendation model as a group of data by taking the patient as a whole. Or, for a certain patient, according to different medicine recommendation schemes, personal characteristic data, physical condition data, historical medication data and a medicine recommendation scheme of the patient are input into the medicine recommendation model as a group of data.
103, the terminal obtains a recommendation result output by the drug recommendation model for each drug recommendation scheme, wherein the recommendation result is used for representing whether the drug recommendation scheme is applicable to the patient, and the recommendation result comprises a recommendation state and a non-recommendation state.
After the personal characteristic data, the physical condition data, the historical medication data and the plurality of drug recommendation schemes of the patient are input into the drug recommendation model in the step 102, the drug recommendation model outputs a recommendation result of each drug recommendation scheme, and the patient or the doctor can determine which drug recommendation scheme to select according to the recommendation result. Of course, the final recommendation result may be represented by a specific numerical value in addition to the recommendation state and the non-recommendation state, and the numerical value of the recommendation result represents the recommendation degree of each medicine recommendation scheme, and generally, the larger the numerical value, the higher the recommendation degree.
According to the medicine recommendation method, the medicine recommendation device, the computer equipment and the storage medium, the personal characteristic data of the patient, the physical condition data of the patient, the historical medication data of the patient and a plurality of medicine recommendation schemes are obtained; the personal characteristic data is used for characterizing individual differences of the patients; the physical condition data is used to characterize the health condition of the patient; the historical medication data is used for reflecting the characteristics of the medicine historically used by the patient; each of the drug recommendation schemes includes a plurality of drugs; then inputting the personal characteristic data, the physical condition data, the historical medication data and the plurality of medicine recommendation schemes into a medicine recommendation model; and finally, for each medicine recommendation scheme, acquiring a recommendation result output by the medicine recommendation model, wherein the recommendation result is used for representing whether the medicine recommendation scheme is suitable for the patient or not, and the recommendation result comprises a recommendation state and an unremitting state. According to the medicine recommendation method, in the process of recommending the medicine for the patient, the information of multiple dimensions of the patient is fully considered, and then the information of the multiple dimensions of the patient is input into the medicine recommendation model, so that the medicine recommendation scheme finally output by the medicine recommendation model can be more suitable for the patient, and the final medicine recommendation result is more accurate.
In this embodiment of the present application, the drug recommendation model includes a medication effect prediction model and a patient lifetime prediction model, the medication effect prediction model is used to predict the treatment effect of each drug recommendation scheme on the patient, the patient lifetime prediction model is used to predict the lifetime of the patient under each drug recommendation scheme, please refer to fig. 2, a method for obtaining the recommendation result output by the drug recommendation model for each drug recommendation scheme in the drug recommendation method is provided, and the method includes:
step 201, the terminal respectively obtains a first recommended value output by the medication effect prediction model and a second recommended value output by the patient survival time prediction model for each medicine recommendation scheme; the first recommended value is used for representing the treatment effect of the medicine recommended scheme on the patient, and the second recommended value is used for representing the survival time of the patient under the medicine recommended scheme.
In this step, the drug recommendation model includes two parts, one is a medication effect prediction model and the other is a patient lifetime prediction model. The medication effect prediction model is mainly used for predicting whether the patient is stressed or relieved within a certain period of time; two important reference factors for measuring whether a certain medicine recommendation scheme is suitable for a patient can be obtained by combining the two models, wherein the treatment effect of each medicine recommendation scheme predicted by the medication effect prediction model on the patient can be obtained according to the following indexes, wherein different indexes represent different treatment effects:
(1) Complete remission (English: complete response, CR for short): all target lesions disappeared, no new ones appeared, and tumor markers were normal for at least 4 weeks.
(2) Partial response (English: partial response, abbreviation: PR): the sum of the maximum diameters of the target lesions is reduced by more than or equal to 30 percent and is maintained for at least 4 weeks.
(3) Stable disease (English: stable disease, abbreviated as SD): the sum of the maximum diameters of the target lesions reduces the underreached PR or increases the underreached PD.
(4) Disease progression (English: progressive disease, abbreviated as PD): the sum of the maximum diameters of the target lesions is at least increased by more than or equal to 20 percent, or new lesions appear.
(5) CR + PR ═ OR-objective remission: a reduction in underreached PR (greater than or equal to 30% reduction in sum of length and diameter of baseline lesions) or an increase in underreached PD (greater than or equal to 20% increase in sum of length and diameter of baseline lesions, appearance of new lesions, or/and progression of non-target lesions), one or more non-target lesions, and/or marker abnormalities.
The survival time of the patient predicted by the patient survival time prediction model under each medicine recommendation scheme can be obtained according to the following indexes, wherein different indexes represent different time meanings, and the different time meanings can be used for representing the treatment effect of different medicine recommendation schemes on the patient:
(1) Total survival time (English: overall survival, abbreviated as OS): from the start of randomization to the time of death due to any cause.
(2) Total remission (English: Duration of over stress, abbreviation: DOR): from the first appearance of CR or PR to the time of first diagnosis of PD or relapse).
(3) The stable period of the disease (DSD): refers to the period of time from the start of treatment to when disease progression is assessed.
(4) Disease-free survival (English: Disease-free survival, abbreviated as DFS): or disease-free survival time, is the time from random group entry to first relapse or death.
(5) Progression-free survival (abbreviated as PFS): time from start of enrollment to tumor progression or death.
(6) Time To Progression (TTP): refers to the time from the onset of randomization to the onset of disease progression or death.
(7) Treatment failure time (English: time to failure, abbreviated as TTF): time from initiation of randomization to cessation/termination of treatment, including any cause of cessation/termination.
In this step, different values can be established for different indicators of the therapeutic effect, and generally, the larger the value, the better the therapeutic effect. In this way, the treatment effect of each drug recommendation output to the patient can be quantified, i.e. a specific numerical value of the first recommendation can be obtained. For the patient survival time prediction model, each corresponding index is represented by a corresponding numerical value, that is, a specific numerical value of the second recommended value can be directly obtained, so that a corresponding numerical value does not need to be set for each index.
Step 202, the terminal obtains the recommendation result according to the first recommendation value and the second recommendation value.
After a first recommended value output by a medication effect prediction model and a second recommended value output by a patient survival time prediction model in the medicine recommendation model are obtained, the terminal combines the first recommended value and the second recommended value to comprehensively judge whether each medicine recommendation scheme is suitable for a patient, so that a final recommendation result is obtained.
In the embodiment of the application, the first recommended value output by the medication effect prediction model and the second recommended value output by the patient survival time are combined, and the first recommended value and the second recommended value are combined, so that the final obtained recommended result is more accurate.
In an embodiment of the present application, please refer to fig. 3, which provides a method for obtaining the recommendation result according to the first recommendation value and the second recommendation value in a drug recommendation method, the method includes:
step 301, for each of the medicine recommendation schemes, if the first recommendation value is greater than a first preset threshold value and the second recommendation value is greater than a second preset threshold value, the terminal outputs the recommendation result as a recommendation state.
In this step, after obtaining the first recommended value and the second recommended value, the terminal needs to determine the magnitude of the first recommended value and the magnitude of the second recommended value, and generally, the larger the numerical values of the first recommended value and the second recommended value are, the more suitable the drug recommendation scheme is for the patient. Therefore, in order to obtain a final recommendation result, the first recommendation value and the second recommendation value need to be determined first, in this step, the determination is performed by setting a first preset threshold and a second preset threshold, and when the final first recommendation value of a certain drug recommendation scheme is greater than the first preset threshold and the second recommendation value is greater than the second preset threshold, the drug recommendation scheme is marked as a recommendation state and is recommended to a patient finally.
In some possible cases, if no corresponding numerical value is set for each index corresponding to the medication effect prediction model, a specific numerical value of the first recommended value cannot be finally obtained, in this case, the form of the first recommended value is not a specific numerical value form, but a category form, and different categories represent different indexes, in this case, the process of obtaining the final recommended result according to the first recommended value and the second recommended value specifically includes: each index corresponds to a category, and which categories are recommended categories and which are not recommended categories can be specified in advance. For a certain drug recommendation scheme, when the category of the index corresponding to the prediction result output by the medication effect prediction model belongs to the recommendation category, and the second recommendation value output by the patient survival time prediction model is greater than a second preset threshold, the terminal marks the recommendation result of the drug recommendation scheme as a recommendation state. Otherwise, the recommendation result of the medicine recommendation scheme is marked as a non-recommendation state.
Step 302, if the first recommended value is smaller than the first preset threshold value and/or the second recommended value is smaller than the second recommended threshold value, the terminal outputs the recommended result as a non-recommended state.
In this step, the recommendation process except for step 301 is given, that is, as long as two conditions, that is, the final first recommended value of a certain drug recommendation scheme is smaller than the first preset threshold, and the second recommended value is smaller than the second preset threshold, satisfy any one of the two conditions, the drug recommendation scheme is finally marked as a non-recommended state.
In the embodiment of the application, by setting the first preset threshold and the second preset threshold, whether each medicine recommendation scheme is marked as a recommendation state or not can be judged through accurate numerical comparison, and the medicine recommendation efficiency is improved.
In the embodiment of the application, the medication effect prediction model is constructed by a random forest model; the patient survival time prediction model is constructed by a multi-factor survival equation.
In the embodiment of the application, the medication effect prediction model is constructed by a random forest model. The random forest model is characterized in that a forest is established in a random mode, a lot of decision trees are arranged in the forest, and each decision tree of the random forest model is not related. After a forest is obtained, when a new input sample enters, each decision tree in the forest is judged once, the class to which the sample belongs is analyzed, and then the class is analyzed and selected most, so that the class of the sample is predicted. The random forest model may handle both quantities with discrete values for the attribute and quantities with continuous values for the attribute. In addition, the random forest model can be used for unsupervised learning clustering and abnormal point detection. The decision tree is a tree structure (which may be a binary tree or a non-binary tree). Each non-leaf node represents a test on a feature attribute, each branch represents the output of the feature attribute over a range of values, and each leaf node stores a category. The process of using the decision tree to make a decision is to start from the root node, test the corresponding characteristic attributes in the items to be classified, select an output branch according to the value of the characteristic attributes until the leaf node is reached, and take the category stored by the leaf node as a decision result.
Many researches show that the combined classifier has better classification effect than a single classifier, and the random forest model is a method for distinguishing and classifying data by using a plurality of classification trees, can give importance scores of all variables while classifying the data, and evaluates the role of all variables in classification.
For the patient survival time prediction model, a multi-factor survival equation is constructed and used for survival analysis, wherein the survival analysis refers to a series of statistical methods for exploring the occurrence time of the interested event. The survival time analysis of cancer patients and the failure time analysis of engineering progress and the like are common.
Given an example of the fact that i is,we denote by a triple (X)i,i,Ti) Wherein X isiThe feature vector, T, representing this exampleiIndicating the time of occurrence of the event for this example.
If the instance has an event of interest to us, then TiIndicating the time between the time point of occurrence of the event and the reference time point, andi＝1，
if the instance does not have an event of interest to us, then TiIndicating the time from the event occurrence time point to the observation end time point, and i＝0。
The objective of survival analysis is to find a new example XjTo estimate the time at which it occurred for the event of interest.
In the survival analysis study, for some instances, it may occur that during our study no event of interest has occurred, which we call deletion. Possible reasons for the deletion are: examples in the study phase were no events of interest; during the study period, the instance was lost; the instance has experienced other events that result in the inability to continue tracking.
Returning to our multifactor survival equation, X begins withjThe method is characterized in that the method is not a specific group, but a matrix data, has a lot of data, can not be directly classified and constructed, and the second difficulty is that the finally obtained value is not a specific classification but a continuous data. T isiAnd classification dataiMachine learning models cannot be used directly to classify or regress the predicted results. To solve the above problem, we construct a functional relation here:
h(t,Xi)＝h0(t)*exp(Xiβ)
wherein XiIs the eigenvector of example i, β is the parameter matrix, β is obtained by maximizing the classification variable partial likelihood. Wherein h is0(t) is an initial vector and is not particularly required.
We can estimate T that instance i has corresponded to what happenediTake place ofThe probability of time is:
Figure BDA0002585479180000111
this equation above primarily determines the instance i at time TiThe above occurring situation is again based on the time TiWill time TiAll previous events were statistically summed.
Substituting the formula to obtain:
Figure BDA0002585479180000121
the latter form is a generalized logistic regression method. The method can solve the relevant parameters by only calculating the maximum value of L (beta) to obtain the optimal segmentation site.
In the embodiment of the application, the medication effect prediction model and the patient survival time prediction model are constructed by using the random forest model and the multi-factor survival equation, and the random forest model has the advantages that the random forest model is well represented on a data set, and overfitting is not easy to occur on the random forest model due to the introduction of two randomness properties; compared with other algorithms, the random forest model has great advantages on the basis of a great number of data sets at present, and due to the introduction of two randomness properties, the random forest model has good anti-noise capability; and the random forest model can process data with very high dimensionality, and has strong adaptability to a data set: the method can process discrete data and continuous data, and a data set does not need to be normalized; in the training process, the random forest model can detect the mutual influence among the characteristics. And by adopting a multi-factor survival equation, the survival time of the sample can be used as an index, and the patients are grouped through the input multiple items of data, so that the drug recommendation scheme is more suitable for the patients.
In an embodiment of the application, the personal characteristic data comprises at least one of an age of the patient, a sex of the patient, and a cancer staging data of the patient, the physical condition data comprises at least one of a vital sign data of the patient and a genomic data of the patient, the historical medication data comprises at least one of a type of a drug historically used by the patient, a dose of a drug historically used by the patient, and a treatment historically used by the patient, each of the drug recommendations comprises at least one of a plurality of different types of drugs, a dose of each type of drug, and a plurality of treatment modalities including surgical resection, radiation therapy, chemotherapy, immunotherapy, and targeted therapy.
In the embodiment of the present application, the age data of the patient directly adopts the real age of the patient. The sex of the patient is according to male and female as two input matrices. When the patient is a cancer patient, the cancer staging data for the patient is determined by the patient's cancer belonging to stage I, II, III, IV and unknown, constituting a five-dimensional matrix as input. The vital sign data of the patient comprises common indexes of the heart rate, the body temperature, the blood pressure, the blood sugar and the like of the patient. The genomic data of a patient generally refers to genomic DNA data of the patient, including the following 4:
(1) Amino acid variation data: counting according to the amino acid variation condition of each gene; counting common variation, wherein more than 0.5 percent of mutations in ten thousand people are sorted into an input matrix, wherein the mutation is represented by 1 and not represented by 0; common variations include single nucleotide polymorphisms, synonymous mutations, missense mutations, nonsense mutations, stop codon mutations. Wherein: single Nucleotide Polymorphism (SNP): mainly refers to DNA sequence polymorphisms at the genomic level caused by variations of a single nucleotide. It is the most common one of the heritable variations in humans. Accounting for more than 90% of all known polymorphisms. The SNP exists widely in human genome, 1 base pair exists in each 500-1000 base pairs on average, and the total number of the SNP is estimated to be 300 ten thousand or more; synonymous Mutation (English: Samesense Mutation): after the base substitution, each codon is changed to another codon, but the amino acids encoded by the codons before and after the change are not changed due to the degeneracy of the codon, and thus, a mutation effect does not occur in practice. For example, when the third position G of GCG in the template strand of a DNA molecule is replaced with A and converted to GCA, the corresponding codon CGC in mRNA is converted to CGU, and since CGC and CGU are both codons encoding arginine, the gene products (proteins) before and after mutation are identical. Synonymous mutations account for about 25% of the total number of base substitution mutations; missense Mutation (English: Missense Mutation): a mutation in which a base pair substitution changes one codon of the mRNA to a codon encoding another amino acid is called a missense mutation. Missense mutation may cause structural and functional abnormality of certain protein or enzyme in body, resulting in disease. If the sixth position of the beta chain of normal hemoglobin in human is glutamic acid, the codon is GAA or GAG, if the second base A is replaced by U, the codon is GUA or GUG, and the glutamic acid is replaced by valine to form abnormal hemoglobin HbS, so that the individual has sickle cell anemia and has mutation effect; nonsense Mutation (English: Nonsense Mutation): mutations in a codon encoding an amino acid are stop codons, which prematurely terminate polypeptide chain synthesis, resulting in a biologically inactive polypeptide fragment, referred to as a nonsense mutation. For example, when G in ATG in a DNA molecule is replaced by T, the codon on the corresponding mRNA chain is changed from UAC to UAA, thereby stopping translation and causing the peptide chain to be shortened. Such mutations will in most cases affect the function of the protein or enzyme; stop codon Mutation (English: terminator codon Mutation): a mutation of a stop codon in a gene to a codon encoding an amino acid is referred to as a stop codon mutation. Since peptide chain synthesis does not stop until the next stop code appears, too long polypeptide chains are synthesized, and are therefore also referred to as elongation mutations. For example, human hemoglobin alpha chain mutant Hb Constant Spring has 31 more amino acids than normal human alpha globin chain;
(2) Gene variation data: counting according to the number of mutations generated on each gene, wherein the number of mutations generated on a certain gene is counted to form an input matrix;
(3) copy number variation: each gene forms a two-bit matrix according to whether an insertion fragment occurs and whether a deletion fragment occurs; if no insertion or deletion occurs, the column is 0. Copy Number Variation (CNV) is caused by genome rearrangement, generally refers to the increase or decrease of Copy number of large genome fragments with length of 1kb or more, and mainly shows deletion and duplication at sub-microscopic level. CNV is an important component of Structural genomic Variation (SV). The mutation rate of the CNV locus is far higher than that of SNP, and the CNV locus is one of important pathogenic factors of human diseases;
(4) and (3) fusion structural variation, wherein more than 1% of gene fusion conditions in ten thousand people are counted as an input matrix, and whether the sample has the mutation type is used as input. Gene fusion: it is a chimeric gene formed by connecting the coding regions of two or more genes end to end and placing them under the control of same set of regulatory sequence (including promoter, enhancer, ribosome binding sequence and terminator), and the expression product of the fusion gene is fusion protein.
The terminal normalized the above gene variation data by dividing by the same number. In the data with the number of 1 thousand persons, the maximum value of the number of the mutated genes can ensure the stability of each dimension of the input data of the drug recommendation model through standardization processing. The gene data can be subjected to data cleaning, and the data cleaning mainly comprises the following steps: missing data supplement and repeated data deletion; checking abnormal value data, and evaluating through the range of the test data; data distribution inspection, such as whether the age data distribution meets a positive distribution, a poisson distribution and the like; and converting the characteristics and continuously dispersing data.
Surgical resection, radiation therapy, chemotherapy, immunotherapy and targeted therapy are now described as follows:
(1) surgical resection refers to the removal of diseased tissue from a patient by way of an intervention inside the patient's body.
(2) Radiation therapy is a local treatment for tumors using radiation. The radiation includes alpha, beta and gamma rays generated by radioactive isotopes, and x-rays, electron beams, proton beams and other particle beams generated by various x-ray therapeutic machines or accelerators. About 70% of cancer patients require radiation therapy in the course of cancer treatment, and about 40% of cancers can be cured by radiation therapy.
(3) Chemotherapy is a short term for chemotherapy, and achieves the purpose of treatment by killing cancer cells with chemotherapeutic drugs. Chemotherapy is one of the most effective means for treating cancer at present, and is also called three major treatment means of cancer together with surgery and radiotherapy. Surgery and radiotherapy belong to local treatment, are only effective on tumors at the treatment part, and are difficult to effectively treat potential metastatic lesions (cancer cells actually have metastasized but cannot be clinically detected due to the limitation of the current technical means) and cancers with clinical metastasis. Chemotherapy is a systemic treatment means, and no matter what route is adopted (oral administration, intravenous administration, body cavity administration and the like), chemotherapy drugs are distributed throughout most organs and tissues of the whole body along with blood circulation. Therefore, chemotherapy is the main treatment for some tumors prone to systemic dissemination and for tumors in the middle and late stages that have metastasized.
(4) Immunotherapy (english: Immunotherapy) refers to a therapeutic method for artificially enhancing or suppressing the immune function of a body to treat a disease in an immune state with a low or high immune level. There are many methods of immunotherapy and are applicable to the treatment of a variety of diseases. Immunotherapy of tumors aims to activate the human immune system, relying on autoimmune functions to kill cancer cells and tumor tissues. Unlike previous surgery, chemotherapy, radiation therapy and targeted therapies, immunotherapy targets not tumor cells and tissues, but the human body's own immune system.
(5) Targeted therapy is a therapeutic approach at the cellular molecular level to a well-defined oncogenic site (which may be a protein molecule or a gene fragment inside a tumor cell). Corresponding therapeutic drugs can be designed, and the drugs enter the body and specifically select carcinogenic sites to combine to take effect, so that tumor cells are specifically killed without affecting normal tissue cells around the tumor, and the molecular targeted therapy is also called as 'biological missile'.
In the embodiment of the application, personal characteristic data, physical condition data, historical medication data and a plurality of drug recommendation schemes of patients are further refined, so that the difference among the patients can be more accurately represented by data input into a drug recommendation model, and the finally obtained recommendation result can be more accurate.
In an embodiment of the present application, please refer to fig. 4, which provides a method for training a drug recommendation model in a drug recommendation method, where the method includes:
in step 401, the terminal obtains a training data set, wherein the training data set comprises personal characteristic data of at least two patients, physical condition data of at least two patients and historical medication data of at least two patients.
Before using the drug recommendation model, it is necessary to train the drug recommendation model, and when acquiring the training data set, training data needs to be acquired for a large number of patients, for example, more than 200 patients. Generally speaking, the richer the training data, the better the training effect on the drug recommendation model. The specific content and meaning of each item of data in the training data set mentioned in this step have been described in detail in the above embodiments, and are not described herein again.
Step 402, the terminal inputs the training data set into an initial drug recommendation model to obtain the drug recommendation model, wherein the initial drug recommendation model comprises an initial drug effect prediction model and an initial patient survival time prediction model.
Correspondingly, because the medicine recommendation model comprises the medicine effect prediction model and the patient survival time prediction model, when the initial medicine recommendation model is trained, the initial medicine effect prediction model and the initial patient survival time prediction model are also trained equivalently. After training, a final drug recommendation model can be obtained.
In the embodiment of the application, the initial medicine recommendation model is trained by utilizing a rich training data set, so that the finally obtained medicine recommendation model has good performance.
In an embodiment of the present application, please refer to fig. 5, which provides a method for inputting the training data set into an initial drug recommendation model in a drug recommendation method, the method includes:
step 501, the terminal preprocesses the data in the training data set to obtain a preprocessed training data set, where the preprocessing includes normalization processing and data cleaning.
Step 502, the terminal inputs the preprocessed training data set into the initial drug recommendation model.
In the embodiment of the present application, after the training data set is obtained, normalization and data cleaning are performed on the data in the training data set. The normalization method has two forms, one is to change a number to a decimal between 0 and 1, and the other is to change a dimensional expression to a dimensionless expression. Data cleansing is used primarily to discover and correct recognizable errors in training data sets, and generally includes checking data consistency, processing invalid and missing values, and the like. The data after normalization and data cleaning can be input into the initial drug recommendation model.
In the embodiment of the application, the normalization processing and the data cleaning operation are set, so that the data finally input into the initial drug recommendation model is more accurate, the result of training failure caused by inaccurate data is avoided, and the efficiency of training the initial drug recommendation model is improved.
It should be understood that, although the steps in the flowcharts of fig. 1 to 5 are shown in sequence as indicated by the arrows, the steps are not necessarily performed in sequence as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 1 to 5 may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of performing the steps or stages is not necessarily sequential, but may be performed alternately or alternately with other steps or at least some of the other steps or stages.
In an embodiment of the present application, as shown in fig. 6, there is provided a medicine recommendation device 600 including: a first obtaining module 601, an input module 602, and a second obtaining module 603, wherein:
a first obtaining module 601, configured to obtain personal characteristic data of a patient, physical condition data of the patient, historical medication data of the patient, and a plurality of drug recommendation schemes; the personal characteristic data is used for characterizing individual differences of the patients; the physical condition data is used to characterize the health condition of the patient; the historical medication data is used for reflecting the characteristics of the medicine historically used by the patient; each of the drug recommendation schemes includes a plurality of drugs;
An input module 602, configured to input the personal characteristic data, the physical condition data, the historical medication data, and the plurality of drug recommendation schemes into a drug recommendation model;
a second obtaining module 603, configured to obtain, for each of the drug recommendation schemes, a recommendation result output by the drug recommendation model, where the recommendation result is used to characterize whether the drug recommendation scheme is applicable to the patient, and the recommendation result includes a recommendation status and a non-recommendation status.
In the embodiment of the application, the medicine recommendation model comprises a medication effect prediction model and a patient life-time prediction model, wherein the medication effect prediction model is used for predicting the treatment effect of each medicine recommendation scheme on the patient, and the patient life-time prediction model is used for predicting the life time of the patient under each medicine recommendation scheme; the second obtaining module 603 is specifically configured to: for each medicine recommendation scheme, respectively acquiring a first recommendation value output by the medication effect prediction model and a second recommendation value output by the patient survival time prediction model; the first recommended value is used for representing the treatment effect of the medicine recommended scheme on the patient, and the second recommended value is used for representing the survival time of the patient under the medicine recommended scheme; and obtaining the recommendation result according to the first recommendation value and the second recommendation value.
In this embodiment of the application, the second obtaining module 603 is specifically configured to: for each medicine recommendation scheme, if the first recommendation value is greater than a first preset threshold value and the second recommendation value is greater than a second preset threshold value, outputting the recommendation result as a recommendation state; and if the first recommended value is smaller than the first preset threshold value and/or the second recommended value is smaller than the second recommended threshold value, outputting the recommended result as a non-recommended state.
In the embodiment of the application, the medication effect prediction model is constructed by a random forest model; the patient survival time prediction model is constructed by a multi-factor survival equation.
In an embodiment of the application, the personal characteristic data comprises at least one of an age of the patient, a sex of the patient, and a cancer staging data of the patient, the physical condition data comprises at least one of a vital sign data of the patient and a genomic data of the patient, the historical medication data comprises at least one of a type of a drug historically used by the patient, a dose of a drug historically used by the patient, and a treatment historically used by the patient, each of the drug recommendations comprises at least one of a plurality of different types of drugs, a dose of each type of drug, and a plurality of treatment modalities including surgical resection, radiation therapy, chemotherapy, immunotherapy, and targeted therapy.
In the embodiment of the present application, please refer to fig. 7, another medicine recommending apparatus 700 is provided, where the medicine recommending apparatus 700 includes, in addition to the modules included in the medicine recommending apparatus 600, optionally, the medicine recommending apparatus 700 further includes: a training module 604.
In this embodiment, the training module 604 is configured to: acquiring a training data set, wherein the training data set comprises personal characteristic data of at least two patients, physical condition data of at least two patients and historical medication data of at least two patients; and inputting the training data set into an initial medicine recommendation model to obtain the medicine recommendation model, wherein the initial medicine recommendation model comprises an initial medicine effect prediction model and an initial patient survival time prediction model.
In this embodiment of the application, the training module 604 is specifically configured to: preprocessing the data in the training data set to obtain a preprocessed training data set, wherein the preprocessing comprises normalization processing and data cleaning; inputting the preprocessed training data set into the initial drug recommendation model
For specific limitations of the drug recommendation device, reference may be made to the above limitations of the drug recommendation method, which are not described herein again. The modules in the medicine recommending device can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In the embodiment of the present application, a computer device is provided, where the computer device may be a terminal, and its internal structure diagram may be as shown in fig. 8. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless communication can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program is executed by a processor to implement a drug recommendation method. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
Those skilled in the art will appreciate that the architecture shown in fig. 8 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In an embodiment of the present application, there is provided a computer device including a memory and a processor, the memory storing a computer program, and the processor implementing the following steps when executing the computer program:
acquiring personal characteristic data of a patient, physical condition data of the patient, historical medication data of the patient and a plurality of drug recommendation schemes; the personal characteristic data is used for characterizing individual differences of the patients; the physical condition data is used to characterize the health condition of the patient; the historical medication data is used for reflecting the characteristics of the medicine historically used by the patient; each of the drug recommendation schemes includes a plurality of drugs;
inputting the personal characteristic data, the physical condition data, the historical medication data and the plurality of medicine recommendation schemes into a medicine recommendation model;
And for each drug recommendation scheme, acquiring a recommendation result output by the drug recommendation model, wherein the recommendation result is used for representing whether the drug recommendation scheme is applicable to the patient or not, and the recommendation result comprises a recommendation state and a non-recommendation state.
In the embodiment of the application, the medicine recommendation model comprises a medication effect prediction model and a patient life-time prediction model, wherein the medication effect prediction model is used for predicting the treatment effect of each medicine recommendation scheme on the patient, and the patient life-time prediction model is used for predicting the life time of the patient under each medicine recommendation scheme; the processor, when executing the computer program, further performs the steps of:
for each medicine recommendation scheme, respectively acquiring a first recommendation value output by the medication effect prediction model and a second recommendation value output by the patient survival time prediction model; the first recommended value is used for representing the treatment effect of the medicine recommended scheme on the patient, and the second recommended value is used for representing the survival time of the patient under the medicine recommended scheme; and obtaining the recommendation result according to the first recommendation value and the second recommendation value.
In the embodiment of the present application, the processor, when executing the computer program, further implements the following steps:
For each medicine recommendation scheme, if the first recommendation value is greater than a first preset threshold value and the second recommendation value is greater than a second preset threshold value, outputting the recommendation result as a recommendation state; and if the first recommended value is smaller than the first preset threshold value and/or the second recommended value is smaller than the second recommended threshold value, outputting the recommended result as a non-recommended state.
In the embodiment of the application, the medication effect prediction model is constructed by a random forest model; the patient survival time prediction model is constructed by a multi-factor survival equation.
In an embodiment of the application, the personal characteristic data comprises at least one of an age of the patient, a sex of the patient, and a cancer staging data of the patient, the physical condition data comprises at least one of a vital sign data of the patient and a genomic data of the patient, the historical medication data comprises at least one of a type of a drug historically used by the patient, a dose of a drug historically used by the patient, and a treatment historically used by the patient, each of the drug recommendations comprises at least one of a plurality of different types of drugs, a dose of each type of drug, and a plurality of treatment modalities including surgical resection, radiation therapy, chemotherapy, immunotherapy, and targeted therapy.
In the embodiment of the present application, the processor, when executing the computer program, further implements the following steps:
acquiring a training data set, wherein the training data set comprises personal characteristic data of at least two patients, physical condition data of at least two patients and historical medication data of at least two patients; and inputting the training data set into an initial medicine recommendation model to obtain the medicine recommendation model, wherein the initial medicine recommendation model comprises an initial medicine effect prediction model and an initial patient survival time prediction model.
In the embodiment of the present application, the processor, when executing the computer program, further implements the following steps:
preprocessing the data in the training data set to obtain a preprocessed training data set, wherein the preprocessing comprises normalization processing and data cleaning; inputting the preprocessed training data set into the initial drug recommendation model.
In an embodiment of the application, a computer-readable storage medium is provided, on which a computer program is stored, which computer program, when being executed by a processor, realizes the steps of:
acquiring personal characteristic data of a patient, physical condition data of the patient, historical medication data of the patient and a plurality of drug recommendation schemes; the personal characteristic data is used for characterizing individual differences of the patients; the physical condition data is used to characterize the health condition of the patient; the historical medication data is used for reflecting the characteristics of the medicine historically used by the patient; each of the drug recommendation schemes includes a plurality of drugs;
Inputting the personal characteristic data, the physical condition data, the historical medication data and the plurality of medicine recommendation schemes into a medicine recommendation model;
and for each drug recommendation scheme, acquiring a recommendation result output by the drug recommendation model, wherein the recommendation result is used for representing whether the drug recommendation scheme is applicable to the patient or not, and the recommendation result comprises a recommendation state and a non-recommendation state.
In the embodiment of the application, the medicine recommendation model comprises a medication effect prediction model and a patient life-time prediction model, wherein the medication effect prediction model is used for predicting the treatment effect of each medicine recommendation scheme on the patient, and the patient life-time prediction model is used for predicting the life time of the patient under each medicine recommendation scheme; the computer program when executed by the processor further realizes the steps of:
for each medicine recommendation scheme, respectively acquiring a first recommendation value output by the medication effect prediction model and a second recommendation value output by the patient survival time prediction model; the first recommended value is used for representing the treatment effect of the medicine recommended scheme on the patient, and the second recommended value is used for representing the survival time of the patient under the medicine recommended scheme; and obtaining the recommendation result according to the first recommendation value and the second recommendation value.
In an embodiment of the application, the computer program when executed by the processor further performs the steps of:
for each medicine recommendation scheme, if the first recommendation value is greater than a first preset threshold value and the second recommendation value is greater than a second preset threshold value, outputting the recommendation result as a recommendation state; and if the first recommended value is smaller than the first preset threshold value and/or the second recommended value is smaller than the second recommended threshold value, outputting the recommended result as a non-recommended state.
In the embodiment of the application, the medication effect prediction model is constructed by a random forest model; the patient survival time prediction model is constructed by a multi-factor survival equation.
In an embodiment of the application, the personal characteristic data comprises at least one of an age of the patient, a sex of the patient, and a cancer staging data of the patient, the physical condition data comprises at least one of a vital sign data of the patient and a genomic data of the patient, the historical medication data comprises at least one of a type of a drug historically used by the patient, a dose of a drug historically used by the patient, and a treatment historically used by the patient, each of the drug recommendations comprises at least one of a plurality of different types of drugs, a dose of each type of drug, and a plurality of treatment modalities including surgical resection, radiation therapy, chemotherapy, immunotherapy, and targeted therapy.
In an embodiment of the application, the computer program when executed by the processor further performs the steps of:
acquiring a training data set, wherein the training data set comprises personal characteristic data of at least two patients, physical condition data of at least two patients and historical medication data of at least two patients; and inputting the training data set into an initial medicine recommendation model to obtain the medicine recommendation model, wherein the initial medicine recommendation model comprises an initial medicine effect prediction model and an initial patient survival time prediction model.
In an embodiment of the application, the computer program when executed by the processor further performs the steps of:
preprocessing the data in the training data set to obtain a preprocessed training data set, wherein the preprocessing comprises normalization processing and data cleaning; inputting the preprocessed training data set into the initial drug recommendation model.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims (10)

1. A method for drug recommendation, the method comprising:
acquiring personal characteristic data of a patient, physical condition data of the patient, historical medication data of the patient and a plurality of drug recommendation schemes; the personal characteristic data is used to characterize individual differences of the patients; the physical condition data is used to characterize the health condition of the patient; the historical medication data is used for reflecting the characteristics of the medicines used by the patient historically; each of the drug recommendations comprises a plurality of drugs;
Inputting the personal characteristic data, the physical condition data, the historical medication data, and the plurality of medication recommendation programs into a medication recommendation model;
and for each drug recommendation scheme, acquiring a recommendation result output by the drug recommendation model, wherein the recommendation result is used for representing whether the drug recommendation scheme is applicable to the patient or not, and the recommendation result comprises a recommendation state and a non-recommendation state.
2. The method of claim 1, wherein the drug recommendation models comprise a medication effect prediction model for predicting the treatment effect of each of the drug recommendations on the patient and a patient length-of-life prediction model for predicting the length of life of the patient for each of the drug recommendations; for each drug recommendation scheme, obtaining the recommendation result output by the drug recommendation model includes:
for each drug recommendation scheme, respectively acquiring a first recommendation value output by the medication effect prediction model and a second recommendation value output by the patient survival time prediction model; the first recommended value is used for representing the treatment effect of the drug recommendation scheme on the patient, and the second recommended value is used for representing the survival time of the patient under the drug recommendation scheme;
And obtaining the recommendation result according to the first recommendation value and the second recommendation value.
3. The method of claim 2, wherein obtaining the recommendation based on the first recommendation value and the second recommendation value comprises:
for each drug recommendation scheme, if the first recommendation value is greater than a first preset threshold value and the second recommendation value is greater than a second preset threshold value, outputting the recommendation result as a recommendation state;
and if the first recommended value is smaller than the first preset threshold value and/or the second recommended value is smaller than the second recommended threshold value, outputting the recommended result as a non-recommended state.
4. The method of claim 2, wherein the medication effect prediction model is constructed from a random forest model; the patient survival time prediction model is constructed by a multi-factor survival equation.
5. The method of claim 1, wherein the personal characteristic data comprises at least one of an age of the patient, a gender of the patient, and a cancer staging data of the patient, the physical condition data comprises at least one of vital sign data of the patient and genomic data of the patient, the historical medication data comprises at least one of a type of medication historically used by the patient, a dose of medication historically used by the patient, and a treatment historically used by the patient, each of the medication recommendations comprises at least one of a plurality of different types of medication, a dose of each type of medication, and a plurality of treatment modalities including surgical resection, radiation therapy, chemotherapy, immunotherapy, and targeted therapy.
6. The method of claim 1, wherein prior to entering the personal characteristic data, the physical condition data, the historical medication data, and the plurality of medication recommendations into a medication recommendation model, the method further comprises:
acquiring a training data set, wherein the training data set comprises personal characteristic data of at least two patients, physical condition data of at least two patients and historical medication data of at least two patients;
and inputting the training data set into an initial medicine recommendation model to obtain the medicine recommendation model, wherein the initial medicine recommendation model comprises an initial medicine effect prediction model and an initial patient survival time prediction model.
7. The method of claim 6, wherein the inputting the training data set into an initial drug recommendation model comprises:
preprocessing the data in the training data set to obtain a preprocessed training data set, wherein the preprocessing comprises normalization processing and data cleaning;
inputting the preprocessed training data set into the initial drug recommendation model.
8. A medication recommendation device, the device comprising:
The system comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring personal characteristic data of a patient, physical condition data of the patient, historical medication data of the patient and a plurality of medicine recommendation schemes; the personal characteristic data is used to characterize individual differences of the patients; the physical condition data is used to characterize the health condition of the patient; the historical medication data is used for reflecting the characteristics of the medicines used by the patient historically; each of the drug recommendations comprises a plurality of drugs;
an input module for inputting the personal characteristic data, the physical condition data, the historical medication data and the plurality of medication recommendation programs into a medication recommendation model;
and the second obtaining module is used for obtaining a recommendation result output by the medicine recommendation model for each medicine recommendation scheme, wherein the recommendation result is used for representing whether the medicine recommendation scheme is applicable to the patient or not, and the recommendation result comprises a recommendation state and a non-recommendation state.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method of any of claims 1 to 7.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
CN202010679984.6A 2020-07-15 2020-07-15 Medicine recommendation method and device, computer equipment and storage medium Pending CN111863181A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010679984.6A CN111863181A (en) 2020-07-15 2020-07-15 Medicine recommendation method and device, computer equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010679984.6A CN111863181A (en) 2020-07-15 2020-07-15 Medicine recommendation method and device, computer equipment and storage medium

Publications (1)

Publication Number Publication Date
CN111863181A true CN111863181A (en) 2020-10-30

Family

ID=72984670

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010679984.6A Pending CN111863181A (en) 2020-07-15 2020-07-15 Medicine recommendation method and device, computer equipment and storage medium

Country Status (1)

Country Link
CN (1) CN111863181A (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112447270A (en) * 2020-11-30 2021-03-05 泰康保险集团股份有限公司 Medication recommendation method, device, equipment and storage medium
CN113555079A (en) * 2021-09-16 2021-10-26 台州市中心医院（台州学院附属医院） Medication prediction system based on neural network
CN113707262A (en) * 2021-08-31 2021-11-26 平安医疗健康管理股份有限公司 Medicine use recommendation method and device, computer equipment and storage medium
CN113782146A (en) * 2021-09-16 2021-12-10 平安科技（深圳）有限公司 General medicine recommendation method, device, equipment and medium based on artificial intelligence
CN116612852A (en) * 2023-07-20 2023-08-18 青岛美迪康数字工程有限公司 Method, device and computer equipment for realizing drug recommendation

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108986879A (en) * 2018-05-31 2018-12-11 平安医疗科技有限公司 Drug recommended method, device, computer equipment and storage medium
CN109087691A (en) * 2018-08-02 2018-12-25 科大智能机器人技术有限公司 A kind of OTC drugs recommender system and recommended method based on deep learning
US20190326002A1 (en) * 2018-04-23 2019-10-24 Diane R. MOULD Systems and methods for modifying adaptive dosing regimens
CN110689939A (en) * 2019-08-27 2020-01-14 天津开心生活科技有限公司 Recommendation method and device for medication sequence, readable medium and electronic equipment
CN110880361A (en) * 2019-10-16 2020-03-13 平安科技（深圳）有限公司 Personalized accurate medication recommendation method and device

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190326002A1 (en) * 2018-04-23 2019-10-24 Diane R. MOULD Systems and methods for modifying adaptive dosing regimens
CN108986879A (en) * 2018-05-31 2018-12-11 平安医疗科技有限公司 Drug recommended method, device, computer equipment and storage medium
CN109087691A (en) * 2018-08-02 2018-12-25 科大智能机器人技术有限公司 A kind of OTC drugs recommender system and recommended method based on deep learning
CN110689939A (en) * 2019-08-27 2020-01-14 天津开心生活科技有限公司 Recommendation method and device for medication sequence, readable medium and electronic equipment
CN110880361A (en) * 2019-10-16 2020-03-13 平安科技（深圳）有限公司 Personalized accurate medication recommendation method and device

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112447270A (en) * 2020-11-30 2021-03-05 泰康保险集团股份有限公司 Medication recommendation method, device, equipment and storage medium
CN113707262A (en) * 2021-08-31 2021-11-26 平安医疗健康管理股份有限公司 Medicine use recommendation method and device, computer equipment and storage medium
CN113555079A (en) * 2021-09-16 2021-10-26 台州市中心医院（台州学院附属医院） Medication prediction system based on neural network
CN113782146A (en) * 2021-09-16 2021-12-10 平安科技（深圳）有限公司 General medicine recommendation method, device, equipment and medium based on artificial intelligence
CN113555079B (en) * 2021-09-16 2022-02-01 台州市中心医院（台州学院附属医院） Medication prediction system based on neural network
CN113782146B (en) * 2021-09-16 2023-08-22 平安科技（深圳）有限公司 Artificial intelligence-based general medicine recommendation method, device, equipment and medium
CN116612852A (en) * 2023-07-20 2023-08-18 青岛美迪康数字工程有限公司 Method, device and computer equipment for realizing drug recommendation
CN116612852B (en) * 2023-07-20 2023-10-31 青岛美迪康数字工程有限公司 Method, device and computer equipment for realizing drug recommendation

Similar Documents

Publication Publication Date Title
CN111863181A (en) Medicine recommendation method and device, computer equipment and storage medium
Pousset et al. A 22-year follow-up study of long-term cardiac outcome and predictors of survival in Friedreich ataxia
Zhavoronkov et al. Deep biomarkers of aging and longevity: from research to applications
Tomatis et al. Late rectal bleeding after 3D-CRT for prostate cancer: development of a neural-network-based predictive model
US11975218B2 (en) Radiotherapy plan parameters with privacy guarantees
Beltran et al. Inexpensive, non-invasive biomarkers predict Alzheimer transition using machine learning analysis of the Alzheimer’s Disease Neuroimaging (ADNI) database
EP4260340A1 (en) Predicting fractional flow reserve from electrocardiograms and patient records
Megna et al. A Comparison among Different Machine Learning Pretest Approaches to Predict Stress‐Induced Ischemia at PET/CT Myocardial Perfusion Imaging
US20220015698A1 (en) Method of identifying tumor drug resistance during treatment
Chen et al. Dual-Stream Subspace Clustering Network for revealing gene targets in Alzheimer's disease
WO2023232762A1 (en) Machine learning predictive models of treatment response
Wang et al. Comparative transcriptome analysis reveals substantial tissue specificity in human aortic valve
US20230242992A1 (en) Methods of predicting cancer progression
Shao et al. A new machine learning algorithm with high interpretability for improving the safety and efficiency of thrombolysis for stroke patients: A hospital-based pilot study
Roberti et al. Bias correction for estimates from linear excess relative risk models in small case‐control studies
CN106503489A (en) The acquisition methods and device in the mutational site of the corresponding gene of cardiovascular system
Simarmata et al. Grouping diabetes diagnosis based on age range with k-means algorithm
Ersoy et al. Nursing Strategies for Diabetic Patient Management: Predicting Parameter Values Post-Exenatide Treatment with Machine Learning Algorithm
Yan et al. Survival prediction for patients with glioblastoma multiforme using a Cox proportional hazards denoising autoencoder network
Varma et al. Identifying predominant clinical and genomic features for glioblastoma multiforme using sequential backward selection
AU2018102039A4 (en) A method of risk prediction and pharmacodynamic prediction of furosemide based on EXOME analysis and panel design
Heilbroner et al. AJS-1 Identifying subgroups at high-risk of ICI associated cardiac adverse events: An approach using topological data analysis
Pinto et al. Machine Learning Techniques to Determine Mutation Impact in Proteins Associated to Neurofibromatosis
Mishra NeuroXNet: Creating A Novel Deep Learning Architecture that Diagnoses Neurological Disorders, Finds New Blood Biomarkers, and Assesses Surgical, Drugs, and Radiation Treatment Plans Using Medical Imaging and Genomic Data
Ye et al. Predicting radiation pneumonitis in lung cancer: a EUD-based machine learning approach for volumetric modulated arc therapy patients

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