CN114678114A - Big data mining evaluation method and big data mining system applied to intelligent medical treatment - Google Patents

Big data mining evaluation method and big data mining system applied to intelligent medical treatment Download PDF

Info

Publication number
CN114678114A
CN114678114A CN202210322591.9A CN202210322591A CN114678114A CN 114678114 A CN114678114 A CN 114678114A CN 202210322591 A CN202210322591 A CN 202210322591A CN 114678114 A CN114678114 A CN 114678114A
Authority
CN
China
Prior art keywords
medical
requirement
data
variable
sample data
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.)
Withdrawn
Application number
CN202210322591.9A
Other languages
Chinese (zh)
Inventor
丁菊
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.)
Individual
Original Assignee
Individual
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 Individual filed Critical Individual
Priority to CN202210322591.9A priority Critical patent/CN114678114A/en
Publication of CN114678114A publication Critical patent/CN114678114A/en
Withdrawn legal-status Critical Current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
    • G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
    • G16H40/20—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/047—Probabilistic or stochastic networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/70—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

Abstract

The embodiment of the application provides a big data mining evaluation method and a big data mining system applied to intelligent medical treatment, which can not only improve the subsequent medical requirement correlation analysis efficiency, but also enable the requirement variable of each piece of medical requirement positioning data to be fully fused with the depth characteristic information of a target medical behavior sample data and a contrast medical behavior sample data by respectively conducting the requirement variable aggregation of different pieces of medical requirement positioning data in the same piece of medical behavior sample data and the requirement variable aggregation of different pieces of medical requirement positioning data in different pieces of medical behavior sample data, thereby being capable of extracting a long-span variable in the medical behavior sample data so as to be convenient for determining the requirement correlation value of any piece of medical requirement positioning data in the target medical behavior sample data and any piece of medical requirement positioning data in the contrast medical behavior sample data, the accuracy of matching medical requirements in medical behavior sample data is improved.

Description

Big data mining evaluation method and big data mining system applied to intelligent medical treatment
Technical Field
The application relates to the technical field of intelligent medical treatment, in particular to a big data mining evaluation method and a big data mining system applied to intelligent medical treatment.
Background
The intelligent medical service uses medical users as the center and medical requirements as the starting point, and applies technologies such as big data, artificial intelligence, cloud computing and the Internet of things to medical scenes, so that the efficiency and the quality of the Internet medical service are improved in all directions. Based on the medical resource network, when a user needs to make a medical resource reservation, the medical resource allocation service can be provided for the user based on the medical requirement of the user.
In the related art, medical requirements of a single user are generally allocated one by one in the process of allocating medical resources, which results in low medical resource allocation efficiency, that is, in the related art, medical requirements are not matched for different users, and uniform allocation cannot be performed for a certain user group with associated medical requirements.
Disclosure of Invention
In order to overcome at least the above disadvantages in the prior art, the present application is directed to a big data mining evaluation method and a big data mining system applied to smart medical treatment.
In a first aspect, the present application provides a big data mining evaluation method applied to smart medical treatment, applied to a big data mining system, the method including:
respectively analyzing medical requirements for target medical behavior sample data of a target user and comparison medical behavior sample data of a comparison user to obtain a plurality of medical requirement positioning data in the target medical behavior sample data and a plurality of medical requirement positioning data in the comparison medical behavior sample data, wherein each piece of medical requirement positioning data comprises a medical requirement to be mined;
for the target medical behavior sample data and the comparison medical behavior sample data, respectively performing demand variable aggregation of different medical demand positioning data in the same medical behavior sample data and demand variable aggregation of different medical demand positioning data in different medical behavior sample data to obtain a demand variable of each medical demand positioning data in the target medical behavior sample data and a demand variable of each medical demand positioning data in the comparison medical behavior sample data;
determining a requirement correlation value of any one medical requirement positioning data in the target medical behavior sample data and any one medical requirement positioning data in the comparison medical behavior sample data according to the requirement variable of each medical requirement positioning data in the target medical behavior sample data and the requirement variable of each medical requirement positioning data in the comparison medical behavior sample data;
generating medical requirement associated information based on the requirement associated value, and generating user requirement grouping evaluation information of the target user and the comparison user based on the medical requirement associated information, wherein the medical requirement associated information represents whether any one of medical requirement positioning data in the target medical behavior sample data and any one of medical requirement positioning data in the comparison medical behavior sample data include associated medical requirements.
In one possible implementation of the first aspect, the target users and the corresponding comparison users with the associated medical needs are added to the same medical needs user group according to the user needs grouping evaluation information of each target user and the corresponding comparison user;
aiming at each medical demand user group, distributing corresponding medical resources to each user in the medical demand user group according to a medical resource distribution strategy corresponding to medical demands, wherein the medical resources comprise medical science popularization push contents;
acquiring resource feedback information of the medical demand user group aiming at the distributed medical resources, and updating and optimizing the medical resource distribution strategy based on the resource feedback information;
in a possible implementation manner of the first aspect, the step of performing update optimization on the medical resource allocation policy based on the resource feedback information includes:
acquiring corresponding interest point trigger activity data according to the resource feedback information;
loading the interest point trigger activity data to a first attention intention decision model, and outputting a first attention intention variable of the interest point trigger activity data;
according to variable matching information between the first attention intention variable and second attention intention variables of the authentication interest point triggering activity data of each medical resource distribution tendency in the medical resource distribution tendency cluster, outputting a deviation probability value corresponding to each medical resource distribution tendency in the interest point triggering activity data and the medical resource distribution tendency in the medical resource distribution tendency cluster;
according to the deviation probability value, obtaining strategy optimization information corresponding to each medical resource distribution tendency included in the interest point trigger activity data from the medical resource distribution tendency cluster;
and updating and optimizing the medical resource allocation strategy according to the strategy optimization information corresponding to each medical resource allocation tendency.
For instance, in one possible implementation of the first aspect, the first attention intention variable comprises a plurality, and the second attention intention variable comprises a plurality;
the outputting of biased probability values corresponding to the interest point triggering activity data and the medical resource allocation tendencies in the medical resource allocation tendency clusters according to variable matching information between the first attention intention variable and second attention intention variables of the authentication interest point triggering activity data of the medical resource allocation tendencies in the medical resource allocation tendency clusters includes:
determining variable matching information between the first target attention intention variable and the second target attention intention variable of each medical resource allocation tendency, and outputting a first deviation probability value; the first target attention intention variable is one of a plurality of the first attention intention variables; the second target attention intention variable is one of a plurality of the second attention intention variables;
determining the first deviation probability value as the deviation probability value corresponding to the point of interest trigger activity data and each medical resource allocation tendency in the medical resource allocation tendency cluster; or
Determining variable matching information between each first attention intention variable and each second attention intention variable of each medical resource allocation tendency to respectively obtain a plurality of second deviation probability values;
outputting a third deviation probability value of each first attention intention variable and each second attention intention variable of each medical resource allocation tendency according to a plurality of second deviation probability values which are positioned in a set number interval from large to small in sequence in the plurality of second deviation probability values;
and determining the biased probability values corresponding to the interest point trigger activity data and the medical resource allocation trends in the medical resource allocation trend cluster according to the weighted biased probability values of the third biased probability values of the first attention intention variables.
For example, in a possible implementation manner of the first aspect, the first attention intention decision model includes a first attention intention variable parsing branch and a first attention intention decision branch, and the method further includes a step of performing model development optimization on the first attention intention decision model in advance, specifically including:
determining the second attention intention variable of the authentication interest point triggering activity data of each medical resource allocation tendency as first model development sample data, and determining the non-attention intention variable of the authentication interest point triggering activity data of each medical resource allocation tendency as second model development sample data;
updating the weight of the first attention intention decision branch according to the first model development sample data and the second model development sample data, and outputting the first attention intention decision branch after weight updating;
the loading the interest point trigger activity data into a first attention intention decision model and outputting a first attention intention variable of the interest point trigger activity data comprises:
loading the interest point triggering activity data to the first attention intention variable analysis branch, and outputting a first interest point triggering tendency variable of the interest point triggering activity data;
and loading the first interest point trigger tendency variable to a first attention intention decision branch after the weight is updated, and outputting the first attention intention variable of the interest point trigger activity data.
For example, in a possible implementation manner of the first aspect, before the determining the second attention intention variable of the authentication point of interest triggering activity data of each medical resource allocation tendency as first model development sample data and determining the non-attention intention variable of the authentication point of interest triggering activity data of each medical resource allocation tendency as second model development sample data, the method further includes:
loading the authentication interest point triggering activity data of each medical resource allocation tendency in the medical resource allocation tendency cluster to a second attention tendency decision model, and outputting a second interest point triggering tendency variable of each authentication interest point triggering activity data of the medical resource allocation tendency and a tendency variable member of each interest point triggering tendency variable in the second interest point triggering tendency variables of each medical resource allocation tendency;
determining model development evaluation indexes of the interest point triggering tendency variables in second interest point triggering tendency variables of the medical resource distribution tendency authentication interest point triggering activity data according to tendency variable members of the interest point triggering tendency variables;
and developing an evaluation index according to a model of each interest point triggering tendency variable in the second interest point triggering tendency variables, and obtaining the second attention intention variable and the non-attention intention variable from the second interest point triggering tendency variables of the authentication interest point triggering activity data of each medical resource allocation tendency.
For instance, in one possible implementation of the first aspect, the second attention intent decision model includes a second attention intent variable resolution branch and a second attention intent decision branch; the loading, to a second attention intention decision model, the authentication interest point trigger activity data of each medical resource allocation inclination in the medical resource allocation inclination cluster, and outputting a second interest point trigger inclination variable of each authentication interest point trigger activity data of the medical resource allocation inclination, and an inclination variable member of each interest point trigger inclination variable in the second interest point trigger inclination variables of each medical resource allocation inclination authentication interest point trigger activity data, includes:
loading the authentication interest point triggering activity data of each medical resource allocation tendency to the second attention intention variable analysis branch, and outputting the second interest point triggering tendency variable of the authentication interest point triggering activity data of each medical resource allocation tendency;
and updating the preset number of times of weights of the second attention intention decision branch by adopting the second interest point triggering tendency variable of the authentication interest point triggering activity data of each medical resource distribution tendency in the medical resource distribution tendency cluster, and outputting tendency variable members of each interest point triggering tendency variable in the second interest point triggering tendency variables of the authentication interest point triggering activity data of each medical resource distribution tendency.
For example, in one possible implementation of the first aspect, each of the second interest point trigger propensity variables corresponds to an initial loop training variable;
the second interest point triggering tendency variable of the authentication interest point triggering activity data adopting each medical resource distribution tendency in the medical resource distribution tendency cluster performs preset number of times of weight updating on the second attention intention decision branch, and outputs a tendency variable member of each interest point triggering tendency variable in the second interest point triggering tendency variables of each medical resource distribution tendency authentication interest point triggering activity data, including:
loading each interest point triggering tendency variable in second interest point triggering tendency variables of the medical resource allocation tendency data and an initial cycle training variable corresponding to each interest point triggering tendency variable in the medical resource allocation tendency cluster to a second attention intention decision branch, performing preset number of times of weight updating on the second attention intention decision branch, and outputting a target cycle training variable corresponding to each interest point triggering tendency variable in the second interest point triggering tendency variables of the medical resource allocation tendency authentication interest point triggering activity data and training variable convolution information corresponding to the target cycle training variable;
and taking the target cyclic training variable corresponding to each interest point triggering tendency variable in the second interest point triggering tendency variables of the second medical resource allocation tendency data and the training variable convolution information corresponding to the target cyclic training variable as a tendency variable member of each interest point triggering tendency variable in the second interest point triggering tendency variables of the second medical resource allocation tendency data.
For example, in one possible implementation manner of the first aspect, the target cyclic training variable corresponding to each of the second interest point triggering tendency variables, and the training variable convolution information corresponding to the target cyclic training variable respectively include: in a preset number of weight updating processes, obtaining a target cyclic training variable after each round of weight updating corresponding to each interest point triggering tendency variable in the second interest point triggering tendency variables is finished, and training variable convolution information corresponding to the target cyclic training variable;
the determining of the model development evaluation index of each interest point triggering tendency variable in the second interest point triggering tendency variables according to the tendency variable member of each interest point triggering tendency variable in the second interest point triggering tendency variables of the authentication interest point triggering activity data of each medical resource allocation tendency includes:
determining fusion information between the target cyclic training variable obtained after each round of weight updating corresponding to each interest point triggering tendency variable in second interest point triggering tendency variables of the authentication interest point triggering activity data of each medical resource allocation tendency is finished and the training variable convolution information corresponding to the target cyclic training variable, and outputting i partial model development evaluation indexes; wherein i is the total weight updating times, and the partial model development evaluation indexes are the model development evaluation indexes corresponding to each weight updating process;
and determining a comprehensive model development evaluation index of the i partial model development evaluation indexes as a model development evaluation index corresponding to each interest point triggering tendency variable in the second interest point triggering tendency variables.
For example, in a possible implementation manner of the first aspect, the loading the first point of interest trigger tendency variable to the first attention intention decision branch after the weight update, and outputting the first attention intention variable of the point of interest trigger activity data includes:
loading the first interest point trigger tendency variable to a first attention intention decision branch after the weight is updated, and outputting attention intention thermal decision data corresponding to each interest point trigger tendency variable in the first interest point trigger tendency variable;
and obtaining, from the first interest point trigger tendency variables, interest point trigger tendency variables of which the interest point thermal decision data matches preset interest point thermal decision data according to the interest point thermal decision data corresponding to each of the first interest point trigger tendency variables, as the first interest point variables of the interest point trigger activity data.
For instance, in one possible implementation of the first aspect, the first attention intent variable encompasses a corresponding attention intent thermodynamic decision data;
the outputting of the third biased probability values of each first attention intention variable and each second attention intention variable of each medical resource allocation bias according to a plurality of second biased probability values located in a set number interval in descending order from the first biased probability values includes:
determining weighted deviation probability values of a plurality of deviation probability values which are positioned in a set number interval from large to small in the plurality of second deviation probability values, and outputting target deviation probability values;
and taking fused information between the target deviation probability value and attention intention thermal decision data covered by first attention intention variables corresponding to the second deviation probability values as the third deviation probability value of each first attention intention variable and each second attention intention variable of each medical resource allocation tendency.
In a second aspect, the present invention further provides a big data mining system, which includes a processor and a machine-readable storage medium, where a computer program is stored in the machine-readable storage medium, and the computer program is loaded and executed by the processor to implement the big data mining evaluation method applied to smart medical treatment in the above first aspect.
Based on the above aspects, by respectively performing demand variable aggregation of different medical demand positioning data in the same medical behavior sample data and demand variable aggregation of different medical demand positioning data in different medical behavior sample data for the target medical behavior sample data and the comparison medical behavior sample data, not only can the subsequent medical demand correlation analysis efficiency be improved, but also the demand variable of each medical demand positioning data can be fully fused with the depth characteristic information of the target medical behavior sample data and the comparison medical behavior sample data, so that a long-span variable in the medical behavior sample data can be extracted, thereby facilitating the subsequent demand variable of each medical demand positioning data in the target medical behavior sample data and the demand variable of each medical demand positioning data in the comparison medical behavior sample data, when a requirement correlation value of any one medical requirement positioning data in the calibrated medical behavior sample data and any one medical requirement positioning data in the contrast medical behavior sample data is determined, the accuracy of matching of medical requirements in the medical behavior sample data is improved.
Drawings
Fig. 1 is a schematic diagram illustrating an architecture of a big data mining evaluation system applied to smart medical treatment according to an embodiment of the present application;
fig. 2 is a schematic flowchart of a big data mining evaluation method applied to smart medical treatment according to an embodiment of the present disclosure.
Detailed Description
Fig. 1 is a schematic diagram of an architecture of a big data mining evaluation system 10 applied to smart medicine according to an embodiment of the present application. The big data mining evaluation system 10 applied to the smart medicine may include a big data mining system 100 and a smart medicine online platform 200 communicatively connected to the big data mining system 100. The smart medical big data mining evaluation system 10 shown in fig. 1 is only one possible example, and in other possible embodiments, the smart medical big data mining evaluation system 10 may include only at least some of the components shown in fig. 1 or may also include other components.
In an embodiment according to the independent concept, the big data mining system 100 and the intelligent medical online platform 200 of the big data mining evaluation system 10 for intelligent medical treatment can be used to perform the big data mining evaluation method for intelligent medical treatment according to the following method embodiments, and the detailed description of the method embodiments can be referred to in the following steps of the big data mining system 100 and the intelligent medical online platform 200.
In step S101, medical requirement analysis is performed on target medical behavior sample data of a target user and comparison medical behavior sample data of a comparison user, respectively, to obtain a plurality of medical requirement positioning data in the target medical behavior sample data and a plurality of medical requirement positioning data in the comparison medical behavior sample data.
In this embodiment, each piece of medical requirement positioning data includes at least one medical requirement to be mined for reflecting the medical requirements of the target user and the comparison user, for example, for the medical requirements of rehabilitation training under a certain disease category, the target user and the comparison user correspond to two different medical users, and by performing medical requirement matching analysis on the target user and the comparison user, subsequent medical resource allocation can be facilitated, and the medical resource allocation efficiency is improved.
In some possible embodiments, the medical need decision model may implement the step S101 described above by: respectively positioning medical requirement related data of the target medical behavior sample data and the comparison medical behavior sample data to obtain a plurality of reference medical requirement positioning data in the target medical behavior sample data and a plurality of reference medical requirement positioning data in the comparison medical behavior sample data; respectively carrying out medical requirement decision (namely determining confidence degree of medical requirement category included by the medical requirement positioning data) and regression processing (namely determining positioning trigger node of the medical requirement positioning data) on a plurality of reference medical requirement positioning data in the target medical behavior sample data and a plurality of reference medical requirement positioning data in the contrast medical behavior sample data, obtaining a plurality of medical requirement positioning data of medical requirement in any medical requirement category included by the target medical behavior sample data and a plurality of medical requirement positioning data of medical requirement in any medical requirement category included by the contrast medical behavior sample data (namely aiming at each reference medical requirement positioning data, determining maximum confidence degree in confidence degrees of different categories corresponding to the medical requirements included by the reference medical requirement positioning data, and filtering out the reference medical requirement positioning data with the maximum confidence degree smaller than a confidence degree threshold value (namely filtering out the reference medical requirement positioning data not including the medical requirement to be mined Reference medical need location data for the need), a plurality of medical need location data obtained after the remaining reference medical need location data is used as target medical behavior sample data and medical need analysis is performed on the comparison medical behavior sample data).
It should be noted that, for the medical requirement analysis of the reference medical behavior sample data, the medical requirement analysis process of the reference medical behavior sample data may be implemented, and the embodiments of the present application are not described herein again.
In step S102, for the target medical behavior sample data and the comparison medical behavior sample data, respectively performing demand variable aggregation of different medical demand positioning data in the same medical behavior sample data and demand variable aggregation of different medical demand positioning data in different medical behavior sample data, to obtain a demand variable of each medical demand positioning data in the target medical behavior sample data and a demand variable of each medical demand positioning data in the comparison medical behavior sample data.
In some possible embodiments, step S102 may be implemented by step S1021 and step S1022.
In step S1021, a requirement variable of each medical requirement positioning data in the target medical behavior sample data is extracted, and a requirement variable of each medical requirement positioning data in the comparison medical behavior sample data is extracted.
In some possible embodiments, the requirement variable extraction model may extract the requirement variable of each medical requirement positioning data in the target medical behavior sample data by: extracting a first behavior tendency variable of the target medical behavior sample data; extracting second behavior tendency variables of each medical requirement positioning data in the target medical behavior sample data; and carrying out variable relation according to the first behavior tendency variable of the target medical behavior sample data and the second behavior tendency variable of each medical requirement positioning data in the target medical behavior sample data, and taking the initial requirement variable of each medical requirement positioning data in the obtained target medical behavior sample data as a requirement variable of each medical requirement positioning data in the target medical behavior sample data called by the first round of cyclic processing.
For example, assuming that after performing medical requirement analysis on the target medical behavior sample data in step S101, 3 pieces of medical requirement positioning data are obtained, which are medical requirement positioning data 1, medical requirement positioning data 2, and medical requirement positioning data 3, respectively, second behavior tendency variables of the medical requirement positioning data 1, the medical requirement positioning data 2, and the medical requirement positioning data 3 can be extracted through the CNN (for example, assuming that the second behavior tendency variables of the medical requirement positioning data 1 to the medical requirement positioning data 3 are M1, M2, and M3, respectively), then the first behavior tendency variable of the target medical behavior sample data (assuming that the first behavior tendency variable of the target medical behavior sample data is M) is extracted through the CNN (for example, a residual error network), and then the second behavior tendency variable of each medical requirement positioning data can be related to the first behavior tendency variable of the target medical behavior sample data, the initial demand variable of each medical demand positioning data in the target medical behavior sample data is obtained, for example, the initial demand variable of the medical demand positioning data 1 is M1+ M, the initial demand variable of the medical demand positioning data 2 is M2+ M, and the initial demand variable of the medical demand positioning data 3 is M3+ M. That is, the initial requirement variable of each medical requirement positioning data may be obtained by connecting only the first behavior tendency variable of the target medical behavior sample data and the second behavior tendency variable of the medical requirement positioning data.
In other possible embodiments, the variable association between the second behavior tendency variable of the positioning data according to the medical requirement in the target medical behavior sample data and the first behavior tendency variable of the target medical behavior sample data may be further implemented by: for each medical requirement positioning data in the target medical behavior sample data, executing the following steps: extracting a behavior trigger node vector of the medical requirement positioning data in the target medical behavior sample data; extracting a medical requirement label vector of the medical requirement to be mined, which is included in the medical requirement positioning data; and carrying out variable association on the second behavior tendency variable of the medical requirement positioning data, the first behavior tendency variable of the target medical behavior sample data, the behavior trigger node vector of the medical requirement positioning data in the target medical behavior sample data and the medical requirement label vector of the medical requirement to be mined, which are included in the medical requirement positioning data, to obtain the initial requirement variable of the medical requirement positioning data.
For example, it is assumed that after the medical requirement analysis is performed on the target medical behavior sample data in step S101, 3 pieces of medical requirement positioning data are obtained, which are respectively medical requirement positioning data 1, medical requirement positioning data 2 and medical requirement positioning data 3, where a behavior trigger node vector of each piece of medical requirement positioning data can be obtained through the regression processing in step S101, for example, it is assumed that a behavior trigger node vector of the medical requirement positioning data 1 in the target medical behavior sample data is l1, a behavior trigger node vector of the medical requirement positioning data 2 in the target medical behavior sample data is l2, a behavior trigger node vector of the medical requirement positioning data 3 in the target medical behavior sample data is l3, and meanwhile, a medical requirement label vector of the medical requirement included in each piece of medical requirement positioning data can be obtained through the medical requirement decision in step S101 (for example, a category corresponding to the maximum confidence degree can be determined as the medical requirement decision making The category of the medical requirement included in the bit data, and the medical requirement tag vector corresponding to the category is determined in a unique hot coding manner), for example, assuming that the medical requirement tag vector of the medical requirement included in the medical requirement positioning data 1 is o1, the medical requirement tag vector of the medical requirement included in the medical requirement positioning data 2 is o2, and the medical requirement tag vector of the medical requirement included in the medical requirement positioning data 3 is o3, the second behavior tendency variable of each medical requirement positioning data, the first behavior tendency variable of the target medical behavior sample data, the behavior trigger node vector of each medical requirement positioning data in the target medical behavior sample data, and the medical requirement tag vector of the medical requirement included in each medical requirement positioning data can be subjected to variable association to obtain the initial requirement variable of each medical requirement positioning data, for example, the initial demand variable of the medical demand positioning data 1 is M1+ M + l1+ o1, the initial demand variable of the medical demand positioning data 2 is M2+ M + l2+ o2, and the initial demand variable of the medical demand positioning data 3 is M3+ M + l3+ o 3. That is, the initial demand variable of each medical requirement positioning data can also be the first behavior tendency variable of the target medical behavior sample data, the second behavior tendency variable of the medical requirement positioning data, the behavior trigger node vector of the medical requirement positioning data in the target medical behavior sample data, and the medical requirement label vector of the medical requirement included in the medical requirement positioning data, and thus, the positioning trigger node information of the medical requirement positioning data and the category information of the medical requirement included in the medical requirement positioning data are comprehensively considered, and the precision in the subsequent medical requirement correlation analysis can be further improved.
In an actual implementation process, after the second behavior tendency variable of the medical requirement positioning data is connected to the first behavior tendency variable of the target medical behavior sample data, the second behavior tendency variable of the medical requirement positioning data may be further connected to at least one of the behavior trigger node vector of the medical requirement positioning data in the target medical behavior sample data and the medical requirement label vector of the medical requirement included in the medical requirement positioning data, so as to obtain an initial requirement variable of the medical requirement positioning data (for example, the medical requirement positioning data 1 may be the initial requirement variable of M1+ M + l1, or M1+ M + o 1), and for the comparison medical behavior positioning data, the initial requirement variable of each medical requirement in the comparison medical behavior sample data may be obtained by referring to the processing mode of the target medical behavior sample data, the embodiments of the present application are not described herein again.
In step S1022, the following steps are cyclically executed: performing feature aggregation on the demand variable of each medical requirement positioning data in the target medical behavior sample data and the demand variables of other medical requirement positioning data in the target medical behavior sample data to obtain an aggregated demand variable of each medical requirement positioning data in the target medical behavior sample data, and performing feature aggregation on the demand variable of each medical requirement positioning data in the comparison medical behavior sample data and the demand variable of other medical requirement positioning data in the comparison medical behavior sample data to obtain an aggregated demand variable of each medical requirement positioning data in the comparison medical behavior sample data; and performing feature aggregation on the aggregated demand variable of each medical demand positioning data in the target medical behavior sample data and the aggregated demand variable of each medical demand positioning data in the comparison medical behavior sample data to obtain a secondary aggregated demand variable of each medical demand positioning data in the target medical behavior sample data, performing feature aggregation on the aggregated demand variable of each medical demand positioning data in the comparison medical behavior sample data and the aggregated demand variable of each medical demand positioning data in the target medical behavior sample data, and obtaining a secondary aggregated demand variable of each medical demand positioning data in the comparison medical behavior sample data.
The number of times of the loop processing is at least one, the demand variable of each medical requirement positioning data in the target medical behavior sample data called by the first loop processing is an initial demand variable extracted from each medical requirement positioning data in the target medical behavior sample data, and the demand variable of each medical requirement positioning data in the comparison medical behavior sample data called by the first loop processing is an initial demand variable extracted from each medical requirement positioning data in the comparison medical behavior sample data; the demand variables of the medical requirement positioning data in the target medical behavior sample data called by the first round of loop processing and subsequent loop processing are aggregated demand variables after the previous round of loop processing (for example, the demand variables of the medical requirement positioning data in the target medical behavior sample data called by the second round of loop processing are aggregated demand variables after the first round of loop processing, the demand variables of the medical requirement positioning data in the target medical behavior sample data called by the third round of loop processing are aggregated demand variables again after the second round of loop processing, and so on), the demand variables of the medical requirement positioning data in the comparison medical behavior sample data called by the first round of loop processing and subsequent loop processing are aggregated demand variables after the previous round of loop processing (for example, the demand variables of the medical requirement positioning data in the comparison medical behavior sample data called by the second round of loop processing are aggregated demand variables after the first round of loop processing, the demand variables of the medical demand positioning data in the contrast medical behavior sample data called by the third loop processing are aggregated demand variables again after the second loop processing, and so on).
In some possible embodiments, the feature aggregation model may implement the feature aggregation of the requirement variable of each medical requirement positioning data in the target medical behavior sample data and the requirement variable of other medical requirement positioning data in the target medical behavior sample data by: for each medical requirement positioning data in the target medical behavior sample data, executing the following steps: knowledge vector mining is carried out on demand variables of the medical demand positioning data to obtain a corresponding basic knowledge vector and an extended knowledge vector; according to the basic knowledge vector and the extended knowledge vector corresponding to the medical requirement positioning data and the basic knowledge vector and the extended knowledge vector corresponding to other medical requirement positioning data in the target medical behavior sample data, the demand variables of the medical requirement positioning data and the demand variables of other medical requirement positioning data in the target medical behavior sample data are weighted and aggregated, and the weighted aggregated variables and the historical related demand variables are aggregated to obtain the aggregated demand variables of the medical requirement positioning data.
For example, the basic knowledge vector and the extended knowledge vector corresponding to the medical requirement positioning data, and the basic knowledge vector and the extended knowledge vector corresponding to other medical requirement positioning data in the target medical behavior sample data can be weighted and aggregated according to the following ways: calling frequent output according to a frequent item attention template (the principle of the frequent item attention template is to calculate attention frequent items between the expanded knowledge vector and each basic knowledge vector to obtain frequent item parameters) according to the basic knowledge vector and the expanded knowledge vector corresponding to the medical requirement positioning data and the basic knowledge vector and the expanded knowledge vector corresponding to other medical requirement positioning data in the target medical behavior sample data, and obtaining the medical requirement positioning data and the frequent item parameters corresponding to other medical requirement positioning data in the target medical behavior sample data; and weighting and aggregating the demand variables of the medical demand positioning data and the demand variables of other medical demand positioning data in the target medical behavior sample data according to the frequent parameters respectively corresponding to the medical demand positioning data and other medical demand positioning data in the target medical behavior sample data.
For example, assume that after performing medical requirement analysis on target medical behavior sample data in step S101, 3 pieces of medical requirement positioning data are obtained, which are medical requirement positioning data 1, medical requirement positioning data 2, and medical requirement positioning data 3, respectively, and after requirement variable extraction and variable association in step S1021, the requirement variables from which the 3 pieces of medical requirement positioning data can be obtained are: x1, x2 and x3, for example, x1= M1+ M, x2= M2+ M, x3= M3+ M, that is, each requirement variable of the medical requirement positioning data is obtained by connecting the second behavior tendency variable of the medical requirement positioning data with the first behavior tendency variable of the target medical behavior sample data, then for each medical requirement positioning data, knowledge vector mining is performed on the requirement variable of the medical requirement positioning data to obtain a corresponding basic knowledge vector and an extended knowledge vector, for example, for the medical requirement positioning data 1, after knowledge vector mining is performed on the requirement variable x1 of the medical requirement positioning data 1, a corresponding basic knowledge vector k1 and an extended knowledge vector q1 can be obtained, correspondingly, after knowledge vector mining is performed on the requirement variable x2 of the medical requirement positioning data 2, a corresponding basic knowledge vector k2 and an extended knowledge vector q2 of the medical requirement positioning data 2 can be obtained, similarly, a basic knowledge vector k3 and an expanded knowledge vector q3 corresponding to the medical requirement positioning data 3 can be obtained. After obtaining the basic knowledge vector and the extended knowledge vector that 3 medical requirement positioning data correspond respectively, the self-attention convolutional layer can be called to carry out the frequent output according to the frequent item attention template on the basic knowledge vector and the extended knowledge vector that the 3 medical requirement positioning data correspond respectively, the frequent item parameters that the 3 medical requirement positioning data correspond respectively are obtained, for example, a1, a2 and a3, wherein, a1 is the frequent item parameter that medical requirement positioning data 1 corresponds, a2 is the frequent item parameter that medical requirement positioning data 2 corresponds, a3 is the frequent item parameter that medical requirement positioning data 3 corresponds, for example, take medical requirement positioning data 1 as an example, can be according to medical requirement positioning data 1, and the frequent item parameter that medical requirement positioning data 2 and medical requirement positioning data 3 correspond respectively, demand variable to medical requirement positioning data 1 and demand positioning data 2 and medical requirement positioning data 3 need positioning data Obtaining variables to carry out weighted aggregation, and aggregating the weighted aggregated variables and historical related demand variables b1 to obtain aggregated demand variables x11 of the medical demand positioning data 1, wherein the specific formula is as follows:
x11=a1* x1+a2* x2+a3* x3+b1
wherein, can adopt and locate data 1 with medical requirement the same processing mode obtain the aggregate demand variable of other medical requirement locating data (for example medical requirement locating data 2 and medical requirement locating data 3) in the subject medical action sample data, this application embodiment is no longer repeated here. In addition, it should be noted that, in practical application, in addition to calling the basic knowledge vector and the extended knowledge vector corresponding to the medical requirement positioning data by the above frequent item attention template and performing frequent output according to the frequent item attention template by the basic knowledge vector and the extended knowledge vector corresponding to other medical requirement positioning data in the target medical behavior sample data, LSTM or GRU may be used to perform frequent output on the basic knowledge vector and the extended knowledge vector corresponding to the medical requirement positioning data and the basic knowledge vector and the extended knowledge vector corresponding to other medical requirement positioning data in the target medical behavior sample data to obtain the medical requirement positioning data and frequent item parameters corresponding to other medical requirement positioning data in the target medical behavior sample data.
In step S103, a requirement associated value of any one of the medical requirement positioning data in the targeted medical behavior sample data and any one of the medical requirement positioning data in the comparison medical behavior sample data is determined according to the requirement variable of each of the medical requirement positioning data in the targeted medical behavior sample data and the requirement variable of each of the medical requirement positioning data in the comparison medical behavior sample data.
In some possible embodiments, step S103 may be implemented by steps S1031 to S1032.
In step S1031, a requirement association value array between different medical requirement positioning data in different medical behavior sample data is generated according to the requirement variable of each medical requirement positioning data in the target medical behavior sample data and the requirement variable of each medical requirement positioning data in the comparison medical behavior sample data.
In some possible embodiments, after obtaining the requirement variable of each medical requirement positioning data in the target medical behavior sample data and the requirement variable of each medical requirement positioning data in the comparison medical behavior sample data according to step S102, a requirement associated value array between different medical requirement positioning data in different medical behavior sample data may be generated by the requirement associated value calculation layer according to the requirement variable of each medical requirement positioning data in the target medical behavior sample data and the requirement variable of each medical requirement positioning data in the comparison medical behavior sample data. For example, the variable inner product between the demand variable of each medical demand positioning data in the target medical behavior sample data and the demand variable of each medical demand positioning data in the contrast medical behavior sample data can be calculated, the calculation result is determined as the demand correlation value between the two medical demand positioning data, and the demand correlation value array between different medical demand positioning data in different medical behavior sample data is generated according to the plurality of demand correlation values.
In other possible embodiments, the array of demand associated values may also be generated by: and generating a requirement association value array between different medical requirement positioning data in different medical behavior sample data according to the initial requirement variable of each medical requirement positioning data in the target medical behavior sample data and the initial requirement variable of each medical requirement positioning data in the contrast medical behavior sample data. For example, after the target medical behavior sample data and the comparison medical behavior sample data are respectively analyzed for medical requirements according to step S101 to obtain a plurality of medical requirement positioning data in the target medical behavior sample data and a plurality of medical requirement positioning data in the comparison medical behavior sample data, the second behavior tendency variable of each medical requirement positioning data in the target medical behavior sample data is in variable connection with the first behavior tendency variable of the target medical behavior sample data to obtain the initial requirement variable of each medical requirement positioning data in the target medical behavior sample data, and the second behavior tendency variable of each medical requirement positioning data in the comparison medical behavior sample data is in variable connection with the first behavior tendency variable of the comparison medical behavior sample data to obtain the initial requirement variable of each medical requirement positioning data in the comparison medical behavior sample data, and then directly determining a requirement correlation value of any one medical requirement positioning data in the targeted medical behavior sample data and any one medical requirement positioning data in the contrast medical behavior sample data according to the initial requirement variable of each medical requirement positioning data in the targeted medical behavior sample data and the initial requirement variable of each medical requirement positioning data in the contrast medical behavior sample data. For example, a variable inner product between an initial requirement variable of each medical requirement positioning data in the target medical behavior sample data and an initial requirement variable of each medical requirement positioning data in the comparison medical behavior sample data is calculated, and a calculation result is determined as a requirement correlation value between the two medical requirement positioning data.
In step S1032, different array members are traversed from the array of demand associated values.
Here, different array members characterize a demand association value between two medical demand positioning data corresponding to different medical behaviour sample data.
In some possible embodiments, it is assumed that, after the medical requirement analysis is performed on the target medical behavior sample data in step S101, 3 pieces of medical requirement positioning data are obtained, which are respectively medical requirement positioning data 1, medical requirement positioning data 2 and medical requirement positioning data 3, and after the medical requirement analysis is performed on the target medical behavior sample data, 3 pieces of medical requirement positioning data are also obtained, which are respectively medical requirement positioning data 4, medical requirement positioning data 5 and medical requirement positioning data 6, and after the requirement variables are aggregated in step S102, the requirement variables of the 3 pieces of medical requirement positioning data in the target medical behavior sample data are respectively t1, t2 and t3, correspondingly, the requirement variables of the 3 pieces of medical requirement positioning data in the target medical behavior sample data are respectively t4, t5 and t6, and then the requirement variables of each piece of medical requirement positioning data in the target medical behavior and the target medical behavior sample data can be calculated And determining a calculation result as a requirement correlation value between two medical requirement positioning data, and generating the following requirement correlation value array according to the requirement correlation value:
wherein the data indicating a need association value between the medical need positioning data 1 in the subject medical behavior sample data and the medical need positioning data 4 in the comparison medical behavior sample data, the data indicating a need association value between the medical need positioning data 1 in the subject medical behavior sample data and the medical need positioning data 5 in the comparison medical behavior sample data, the data indicating a need association value between the medical need positioning data 1 in the subject medical behavior sample data and the medical need positioning data 6 in the comparison medical behavior sample data, the data indicating a need association value between the medical need positioning data 2 in the subject medical behavior sample data and the medical need positioning data 4 in the comparison medical behavior sample data, the data indicating a need association value between the medical need positioning data 2 in the subject medical behavior sample data and the medical need positioning data 5 in the comparison medical behavior sample data, the medical action sample data comprises a requirement correlation value representing between medical requirement positioning data 2 in the target medical action sample data and medical requirement positioning data 6 in the comparison medical action sample data, a requirement correlation value representing between medical requirement positioning data 3 in the target medical action sample data and medical requirement positioning data 4 in the comparison medical action sample data, a requirement correlation value representing between medical requirement positioning data 3 in the target medical action sample data and medical requirement positioning data 5 in the comparison medical action sample data, and a requirement correlation value representing between medical requirement positioning data 3 in the target medical action sample data and medical requirement positioning data 6 in the comparison medical action sample data.
In other possible embodiments, after generating the array of requirement association values between different medical requirement positioning data in different medical behavior sample data, and when the target medical behavior sample data comprises a different amount of data segments of the medical requirement positioning data than the control medical behavior sample data, the following steps may be further performed: generating an array member as a blank cell in the requirement correlation value array; loading a demand variable of medical demand positioning data which only appears in the target medical behavior sample data or the contrast medical behavior sample data into a blank unit to obtain an updated demand correlation value array; and performing regularized conversion on each array member cycle in the updated requirement correlation value array, and removing blank units to obtain a requirement correlation value array updated again.
For example, assuming that after the target medical behavior sample data is analyzed for medical needs in step S101, 3 pieces of medical need positioning data, which are respectively medical need positioning data 1, medical need positioning data 2, and medical need positioning data 3, are obtained, and after the comparative medical behavior sample data is analyzed for medical needs, 2 pieces of medical need positioning data, which are respectively medical need positioning data 4 and medical need positioning data 5, are obtained, that is, the data fragment amount of the medical need positioning data included in the target medical behavior sample data is different from the data fragment amount of the medical need positioning data included in the comparative medical behavior sample data, after the requirement association value array between the target medical behavior sample data and the comparative medical behavior sample data is generated, an array member can be generated in the requirement association value array as a blank cell, and filling the demand variables of the medical demand positioning data 3 included in the target medical behavior sample data into the blank cells to obtain an updated demand associated value array, then circularly performing regularized conversion on each array member in the updated demand associated value array by adopting an optimal transmission-Sinkhorn algorithm, removing the newly added blank cells, and obtaining a re-updated demand associated value array.
In some possible embodiments, taking the medical service distribution service as an example (that is, the target medical behavior sample data is medical behavior sample data acquired by collecting medical behavior sample data of the medical service distribution service, and the comparison medical behavior sample data is medical behavior sample data of the same medical service distribution service acquired from past historical data), after obtaining a requirement associated value of any one of the medical requirement positioning data in the target medical behavior sample data and any one of the medical requirement positioning data in the comparison medical behavior sample data according to step S103, the following steps may be further performed: determining the data fragment quantity of the medical requirement positioning data of which the requirement correlation value is greater than the preset requirement correlation value in the comparison medical behavior sample data in the calibrated medical behavior sample data; and when the data fragment size is larger than the preset fragment size, determining that the target medical behavior sample data and the comparison medical behavior sample data comprise the same medical service object, and updating the comparison medical behavior sample data in the past historical data into the target medical behavior sample data.
In step S104, medical requirement related information is generated based on the requirement related value.
Here, the medical requirement related information represents whether any one of the medical requirement positioning data in the subject medical behavior sample data and any one of the medical requirement positioning data in the comparison medical behavior sample data include the related medical requirement.
In some possible embodiments, when a requirement correlation value between any one of the medical requirement positioning data in the target medical behavior sample data and any one of the medical requirement positioning data in the comparison medical behavior sample data is greater than a preset requirement correlation value, it is determined that the medical requirements included in the two pieces of medical requirement positioning data are the same medical requirement; when a requirement correlation value between any one of the medical requirement positioning data in the target medical behavior sample data and any one of the medical requirement positioning data in the comparison medical behavior sample data is smaller than a preset requirement correlation value, it is determined that the medical requirements included in the two pieces of medical requirement positioning data are not the same medical requirement.
Based on the steps, by respectively carrying out the demand variable aggregation of different medical demand positioning data in the same medical behavior sample data and the demand variable aggregation of different medical demand positioning data in different medical behavior sample data aiming at the target medical behavior sample data and the comparison medical behavior sample data, not only the subsequent medical demand correlation analysis efficiency can be improved, but also the demand variable of each medical demand positioning data can be fully fused with the depth characteristic information of the target medical behavior sample data and the comparison medical behavior sample data, so that the long-span variable in the medical behavior sample data can be extracted, and the demand variable of each medical demand positioning data in the target medical behavior sample data and the demand variable of each medical demand positioning data in the comparison medical behavior sample data can be conveniently obtained, when a requirement correlation value of any one medical requirement positioning data in the calibrated medical behavior sample data and any one medical requirement positioning data in the contrast medical behavior sample data is determined, the accuracy of matching of medical requirements in the medical behavior sample data is improved.
In some possible implementations, on the basis of the above embodiments, the embodiments of the present application may further include the following steps.
And S105, adding the target users with the associated medical requirements and the corresponding comparison users into the same medical requirement user group according to the user requirement grouping evaluation information of each target user and the corresponding comparison users.
Step S106, aiming at each medical requirement user group, distributing corresponding medical resources to each user in the medical requirement user group according to a medical resource distribution strategy corresponding to medical requirements, wherein the medical resources comprise medical science popularization push contents.
And S107, acquiring resource feedback information of the medical demand user group aiming at the distributed medical resources, and updating and optimizing the medical resource distribution strategy according to the resource feedback information.
Among them, step S107 may be realized by the following exemplary steps.
And step S1, acquiring corresponding interest point trigger activity data according to the resource feedback information.
And step S2, loading the interest point triggering activity data to a first attention intention decision model, and outputting a first attention intention variable of the interest point triggering activity data.
In some possible embodiments, the first attention intention variable triggers a tendency variable for a point of interest that matches a preset plurality of medical resource allocation tendencies.
In some possible embodiments, the point-of-interest trigger activity data may be loaded into a first attention intention decision model, which may extract respective point-of-interest trigger tendency variables from the point-of-interest trigger activity data, and obtain, from the various types of point-of-interest trigger tendency variables of the point-of-interest trigger activity data, point-of-interest trigger tendency variables that match a plurality of medical resource allocation tendencies that are preset previously as the first attention intention variables of the point-of-interest trigger activity data.
In some possible embodiments, each round may be loaded with one point of interest trigger activity data collected in response to an event gathering task, or may be loaded with one point of interest trigger activity data collected in response to a plurality of event gathering tasks at a time for analysis, which is not limited herein.
Step S3, determining variable matching information between the first attention intention variable and a second attention intention variable of the authentication interest point triggering activity data of each medical resource distribution tendency in the medical resource distribution tendency cluster, and outputting a deviation probability value corresponding to each medical resource distribution tendency in the interest point triggering activity data and the medical resource distribution tendency cluster.
In some possible embodiments, the authentication interest point triggering activity data may be a key interest point triggering activity data preset in advance for use as a basis for identifying the attention intention variable. The second point of interest intent variable is a point of interest trigger propensity variable that matches the medical resource allocation propensity included in the authentication point of interest trigger activity data.
For example, the second intention-of-interest variable of the authenticated interest point triggering activity data of each medical resource allocation tendency in the plurality of medical resource allocation tendency clusters may be determined in advance, and the medical resource allocation tendency included in the interest point triggering activity data may be included in the medical resource allocation tendency cluster, and the second intention-of-interest variable of each medical resource allocation tendency authenticated interest point triggering activity data may be an interest point triggering tendency variable that matches the medical resource allocation tendency included in the authenticated interest point triggering activity data of the medical resource allocation tendency. In this way, variable matching information (e.g., degree of association) between the first intention of interest variable of the interest point triggering activity data and each second intention of interest variable of the authentication interest point triggering activity data of each medical resource allocation tendency in the medical resource allocation tendency cluster can be analyzed, and then deviation probability values corresponding to each medical resource allocation tendency in the medical resource allocation tendency cluster can be obtained accordingly.
And step S4, obtaining strategy optimization information corresponding to each medical resource distribution tendency included in the interest point trigger activity data from the medical resource distribution tendency cluster according to the deviation probability value.
And step S5, updating and optimizing the medical resource allocation strategy according to the strategy optimization information corresponding to each medical resource allocation tendency.
In some possible embodiments, the authentication point of interest triggering activity data for each medical resource allocation trend may further include an attention-free intent variable; the non-attention intention variable of the authentication interest point trigger activity data of each medical resource allocation tendency may be an interest point trigger tendency variable which is included in each authentication interest point trigger activity data of each medical resource allocation tendency cluster and is not related to the medical resource allocation tendency included in the authentication interest point trigger activity data.
In some possible embodiments, the first attention intention decision model in S2 includes: a first attention intent variable resolution branch and a first attention intent decision branch. The first attention intention decision branch can be obtained by performing model development optimization on a second attention intention variable and a non-attention intention variable of the activity data according to the authentication interest point of each medical resource allocation tendency in the medical resource allocation tendency cluster.
In some possible embodiments, the above method may further include a step of performing model development optimization on the first attention intention decision model in advance, specifically including the contents of steps S11-S12 described below.
Step S11, determining a second attention intention variable of the authentication interest point triggering activity data of each medical resource allocation tendency as first model development sample data, and determining a non-attention intention variable of the authentication interest point triggering activity data of each medical resource allocation tendency as second model development sample data.
And step S12, updating the weight of the first attention intention decision branch according to the first model development sample data and the second model development sample data, and outputting the first attention intention decision branch after weight updating.
The first model development sample data may refer to positive sample data, and the second model development sample data may refer to negative sample data.
Thus, the way of outputting the first attention intention variable of the point of interest trigger activity data may be implemented by, for example, the following embodiments: firstly, the interest point triggering activity data is loaded to a first attention intention variable analysis branch, a first interest point triggering tendency variable of the interest point triggering activity data is output, then the first interest point triggering tendency variable is loaded to a first attention intention decision branch after the weight is updated, and the first attention intention variable of the interest point triggering activity data is output.
In some possible embodiments, after the first attention intention decision branch is adjusted by using network parameters obtained by network convergence, the first interest point triggering tendency variable of the interest point triggering activity data is loaded into the adjusted first attention intention decision branch, so as to analyze an attention intention variable and a non-attention intention variable in the first interest point triggering tendency variable of the interest point triggering activity data, and further obtain an attention intention variable in the first interest point triggering tendency variable as the first attention intention variable.
In some possible embodiments, the first interest point triggering tendency variable may be loaded to the first attention intention decision branch after the weight update, and attention intention thermal decision data corresponding to each of the interest point triggering tendency variables in the first interest point triggering tendency variable is output; then, according to attention intention thermal decision data corresponding to each interest point trigger tendency variable in the first interest point trigger tendency variables, obtaining an interest point trigger tendency variable with attention intention thermal decision data matched with preset attention intention thermal decision data from the first interest point trigger tendency variables, and taking the obtained interest point trigger tendency variable as the first attention intention variable of the interest point trigger activity data.
In some possible embodiments, the first attention intention variable may include a plurality of variables, and the second attention intention variable may also include a plurality of variables. Thus, the above S3 may include the following step S310 and step S320.
Step S310, determining variable matching information between the first target attention intention variable and the second target attention intention variable of each medical resource allocation tendency, and outputting a first deviation probability value; the first target attention intention variable is one of a plurality of first attention intention variables; the second target attention intention variable is one of a plurality of second attention intention variables.
Step S320, determining the first deviation probability value as the deviation probability value corresponding to the interest point trigger activity data and each medical resource distribution tendency in the medical resource distribution tendency cluster.
Wherein the first target attention intention variable may be one of a plurality of first attention intention variables of the point of interest triggering activity data, and the second target attention intention variable may also be one of a plurality of second attention intention variables of each medical resource allocation tendency.
In some possible embodiments, for the authentication interest point triggering activity data of each medical resource allocation tendency in the medical resource allocation tendency cluster, a first target interest intention variable and a pearson correlation coefficient between the authentication interest point triggering activity data and a second target interest intention variable may be calculated as biased probability values corresponding to each medical resource allocation tendency in the interest point triggering activity data and the medical resource allocation tendency in the medical resource allocation tendency cluster, so that mining efficiency of the interest point triggering activity data may be accelerated.
In some possible embodiments, the step S3 may further include the following steps S311 and S331:
step S311, determining variable matching information between each first attention intention variable and each second attention intention variable of each medical resource allocation tendency, and obtaining a plurality of second deviation probability values, respectively.
Step S321, outputting a third biased probability value of each first attention intention variable and each second attention intention variable of each medical resource allocation bias according to a plurality of second biased probability values located in a set number interval according to a descending order in the plurality of second biased probability values.
In some possible embodiments, the target biased probability value can be output according to the weighted biased probability values of the plurality of biased probability values which are positioned in the set number interval from large to small; and taking fused information between the target deviation probability value and attention intention thermal decision data of the first attention intention variable corresponding to the deviation probability values as a third deviation probability value of each first attention intention variable and each second attention intention variable of each medical resource allocation tendency.
In some possible embodiments, the first attention intention variable may encompass thermodynamic decision data with a corresponding attention intention. In this way, weighted biased probability values of a plurality of biased probability values located in a set number interval according to a descending order in the plurality of biased probability values may be determined first, a target biased probability value is output, and then fusion information between the target biased probability value and attention intention thermal decision data covered by first attention intention variables corresponding to the plurality of biased probability values is used as the third biased probability value of each first attention intention variable and each second attention intention variable of each medical resource allocation tendency.
S331, determining the biased probability values corresponding to the interest point triggering activity data and the medical resource allocation trends in the medical resource allocation trend clusters according to the weighted biased probability values of the third biased probability values of all the first attention intention variables. Wherein the weighted biased probability value may be a sum of corresponding coefficients.
In some possible embodiments, the second attention intention variable and the non-attention intention variable may be obtained before the above step S12 and specifically include the following.
(1) And loading the authentication interest point triggering activity data of each medical resource allocation tendency in the medical resource allocation tendency cluster to a second attention tendency decision model, and outputting a second interest point triggering tendency variable of each medical resource allocation tendency authentication interest point triggering activity data and a tendency variable member of each interest point triggering tendency variable in the second interest point triggering tendency variables of each medical resource allocation tendency authentication interest point triggering activity data.
(2) And determining model development evaluation indexes of the interest point triggering tendency variables in the second interest point triggering tendency variables according to the tendency variable members of the interest point triggering tendency variables in the second interest point triggering tendency variables of the authentication interest point triggering activity data of each medical resource distribution tendency.
In some possible embodiments, the trend variable members may include a cyclic training variable and training variable convolution information. Accordingly, the model development evaluation index is determined as follows: and outputting model development evaluation indexes corresponding to the interest point triggering tendency variables in the second interest point triggering tendency variables according to fusion information between the cyclic training variables corresponding to the interest point triggering tendency variables and the training variable convolution information in the second interest point triggering tendency variables of the authentication interest point triggering activity data of each medical resource distribution tendency. For example, fusion information between the cyclic training variables and the training variable convolution information corresponding to each interest point triggering tendency variable may be calculated according to a preset weighting coefficient, and model development evaluation indexes corresponding to each interest point triggering tendency variable in the second interest point triggering tendency variables may be output.
(3) And developing an evaluation index according to the model of each interest point triggering tendency variable in the second interest point triggering tendency variables, and obtaining second attention intention variables and non-attention intention variables from the second interest point triggering tendency variables of the authentication interest point triggering activity data of each medical resource allocation tendency.
In some possible embodiments, after obtaining the model development evaluation index of each of the second interest point trigger tendency variables, the interest point trigger tendency variables whose model development evaluation indexes meet the preset requirements are obtained from the second interest point trigger tendency variables according to the model development evaluation indexes of each of the second interest point trigger tendency variables; and determining the interest point triggering tendency variable with the model development evaluation index meeting the preset requirement as a second attention intention variable, and determining other interest point triggering tendency variables in the second interest point triggering tendency variable as non-attention intention variables.
In some possible embodiments, the second attention intention decision model includes a second attention intention variable resolution branch and a second attention intention decision branch. Loading the authentication interest point trigger activity data of each medical resource allocation tendency in the medical resource allocation tendency cluster into a second attention tendency decision model, outputting a second interest point trigger tendency variable of each authentication interest point trigger activity data of the medical resource allocation tendency, and tendency variable members of each interest point trigger tendency variable in the second interest point trigger tendency variables of each authentication interest point trigger activity data of the medical resource allocation tendency, wherein the tendency variable members comprise the following contents (11) and (12).
(11) And loading the authentication interest point triggering activity data of each medical resource allocation tendency to the second attention intention variable analysis branch, and outputting the second interest point triggering tendency variable of the authentication interest point triggering activity data of each medical resource allocation tendency.
(12) And updating the preset number of times of weights of the second attention intention decision branch by adopting the second interest point triggering tendency variable of the authentication interest point triggering activity data of each medical resource distribution tendency in the medical resource distribution tendency cluster, and outputting tendency variable members of each interest point triggering tendency variable in the second interest point triggering tendency variables of the authentication interest point triggering activity data of each medical resource distribution tendency.
In detail, in step (12), each interest point trigger tendency variable in the second interest point trigger tendency variables of the medical resource allocation tendency certification interest point trigger activity data of each medical resource allocation tendency in the medical resource allocation tendency cluster and the initial loop training variable corresponding to each interest point trigger tendency variable may be loaded to the second interest point decision branch, a preset number of times of weight updating may be performed on the second interest point decision branch, and a target loop training variable corresponding to each interest point trigger tendency variable in the second interest point trigger tendency variables of each medical resource allocation tendency certification interest point trigger activity data and convolution information of training variables corresponding to the target loop training variables may be output; then, the target cyclic training variable corresponding to each interest point trigger tendency variable in the second interest point trigger tendency variables of the second interest point trigger tendency data of each medical resource allocation tendency and the training variable convolution information corresponding to the target cyclic training variable are used as tendency variable members of each interest point trigger tendency variable in the second interest point trigger tendency variables of the second interest point trigger tendency data of each medical resource allocation tendency.
In some possible embodiments, the target cyclic training variable corresponding to each of the second interest point triggering tendency variables, and the training variable convolution information corresponding to the target cyclic training variable respectively include: in a preset number of weight updating processes, obtaining a target cyclic training variable after each round of weight updating corresponding to each interest point triggering tendency variable in the second interest point triggering tendency variables is finished, and training variable convolution information corresponding to the target cyclic training variable.
Accordingly, the trend variable members of the interest point trigger trend variables in the second interest point trigger trend variables of the authentication interest point trigger activity data according to each medical resource allocation trend determine model development evaluation indexes of the interest point trigger trend variables in the second interest point trigger trend variables, which may be implemented by the following embodiments, for example: determining fusion information between the target cyclic training variable and the training variable convolution information corresponding to the target cyclic training variable obtained after each round of weight updating corresponding to each interest point triggering tendency variable in the second interest point triggering tendency variables of the authentication interest point triggering activity data of each medical resource allocation tendency is finished, and outputting i partial model development evaluation indexes; wherein i is the total weight updating times, and the partial model development evaluation indexes are the model development evaluation indexes corresponding to each weight updating process; then, a comprehensive model development evaluation index of the i partial model development evaluation indexes is determined as a model development evaluation index corresponding to each of the second interest point triggering tendency variables. The comprehensive model development and evaluation index may be weighted evaluation information corresponding to each partial model development and evaluation index.
In an embodiment, which may be according to the independent concept, the big data mining system 100 may include: a processor 101 and a machine-readable storage medium 102. Wherein the machine-readable storage medium 102 is used for storing a program for supporting the big data mining system 100 to execute the big data mining evaluation method applied to smart medicine provided in any one of the above-mentioned embodiments, and the processor 101 is configured to execute the program stored in the machine-readable storage medium 102.
The program comprises one or more computer instructions, wherein the one or more computer instructions, when executed by the processor 101, enable all or part of the steps of any of the foregoing embodiments.
The architecture of the big data mining system 100 may further include a communication unit 103, which is used for the big data mining system 100 to communicate with other devices or systems (e.g., the intelligent medical online platform 200).
In addition, the present application provides a computer storage medium for storing computer software instructions for the big data mining system 100, which includes a program for executing the big data mining evaluation method applied to smart medical in any one of the above method embodiments.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions in the embodiments of the present application.

Claims (10)

1. A big data mining evaluation method applied to intelligent medical treatment is characterized in that the big data mining evaluation method is applied to a big data mining system, and the method comprises the following steps:
respectively analyzing medical requirements for target medical behavior sample data of a target user and comparison medical behavior sample data of a comparison user to obtain a plurality of medical requirement positioning data in the target medical behavior sample data and a plurality of medical requirement positioning data in the comparison medical behavior sample data, wherein each piece of medical requirement positioning data comprises a medical requirement to be mined;
for the target medical behavior sample data and the comparison medical behavior sample data, respectively performing demand variable aggregation of different medical demand positioning data in the same medical behavior sample data and demand variable aggregation of different medical demand positioning data in different medical behavior sample data to obtain a demand variable of each medical demand positioning data in the target medical behavior sample data and a demand variable of each medical demand positioning data in the comparison medical behavior sample data;
determining a requirement correlation value of any one medical requirement positioning data in the target medical behavior sample data and any one medical requirement positioning data in the comparison medical behavior sample data according to the requirement variable of each medical requirement positioning data in the target medical behavior sample data and the requirement variable of each medical requirement positioning data in the comparison medical behavior sample data;
generating medical requirement associated information based on the requirement associated value, and generating user requirement grouping evaluation information of the target user and the comparison user based on the medical requirement associated information, wherein the medical requirement associated information represents whether any one of medical requirement positioning data in the target medical behavior sample data and any one of medical requirement positioning data in the comparison medical behavior sample data include associated medical requirements.
2. The big data mining and evaluation method for intelligent medical treatment according to claim 1, wherein said respectively performing demand variable aggregation of different medical requirement positioning data in the same medical behavior sample data and demand variable aggregation of different medical requirement positioning data in different medical behavior sample data to obtain demand variable of each medical requirement positioning data in the target medical behavior sample data and demand variable of each medical requirement positioning data in the comparison medical behavior sample data comprises:
the following steps are executed in a circulating way:
performing feature aggregation on the demand variable of each medical requirement positioning data in the target medical behavior sample data and the demand variables of other medical requirement positioning data in the target medical behavior sample data to obtain an aggregated demand variable of each medical requirement positioning data in the target medical behavior sample data, and performing feature aggregation on the demand variable of each medical requirement positioning data in the comparison medical behavior sample data and the demand variables of other medical requirement positioning data in the comparison medical behavior sample data to obtain an aggregated demand variable of each medical requirement positioning data in the comparison medical behavior sample data;
performing feature aggregation on the aggregated demand variable of each medical demand positioning data in the target medical behavior sample data and the aggregated demand variable of each medical demand positioning data in the comparison medical behavior sample data to obtain a re-aggregated demand variable of each medical demand positioning data in the target medical behavior sample data, and performing feature aggregation on the aggregated demand variable of each medical demand positioning data in the comparison medical behavior sample data and the aggregated demand variable of each medical demand positioning data in the target medical behavior sample data to obtain a re-aggregated demand variable of each medical demand positioning data in the comparison medical behavior sample data;
the demand variable of each medical requirement positioning data in the target medical behavior sample data called by the first round of cyclic processing is an initial demand variable extracted from each medical requirement positioning data in the target medical behavior sample data, and the demand variable of each medical requirement positioning data in the comparison medical behavior sample data called by the first round of cyclic processing is an initial demand variable extracted from each medical requirement positioning data in the comparison medical behavior sample data; and the demand variable of each medical requirement positioning data in the target medical behavior sample data called by the subsequent loop processing is an aggregated demand variable after the previous loop processing, and the demand variable of each medical requirement positioning data in the contrast medical behavior sample data called by the subsequent loop processing is an aggregated demand variable after the previous loop processing.
3. The big data mining evaluation method for intelligent medical treatment according to claim 2, wherein the method further comprises:
acquiring a requirement variable of each medical requirement positioning data in the target medical behavior sample data called by the first round of circular processing in the following mode:
extracting a first behavior tendency variable of the target medical behavior sample data, and extracting a second behavior tendency variable of each medical requirement positioning data in the target medical behavior sample data;
and performing variable relation according to the first behavior tendency variable of the target medical behavior sample data and the second behavior tendency variable of each medical requirement positioning data in the target medical behavior sample data, and taking the obtained initial requirement variable of each medical requirement positioning data in the target medical behavior sample data as a requirement variable of each medical requirement positioning data in the target medical behavior sample data called by the first round of cyclic processing.
4. The big data mining evaluation method for intelligent medical treatment according to claim 3, wherein the variable association according to the first behavior tendency variable of the target medical behavior sample data and the second behavior tendency variable of each medical requirement positioning data in the target medical behavior sample data comprises:
for each medical requirement positioning data in the target medical behavior sample data, extracting a behavior trigger node vector of the medical requirement positioning data in the target medical behavior sample data;
extracting a medical requirement label vector of the medical requirement to be mined, which is included in the medical requirement positioning data;
and performing variable association on a second behavior tendency variable of the medical requirement positioning data, a first behavior tendency variable of the target medical behavior sample data, a behavior trigger node vector of the medical requirement positioning data in the target medical behavior sample data, and a medical requirement label vector of the medical requirement to be mined, which is included in the medical requirement positioning data, to obtain an initial requirement variable of the medical requirement positioning data.
5. The big data mining evaluation method for intelligent medical treatment according to claim 2, wherein the characteristic aggregation of the demand variable of each medical demand positioning data in the target medical behavior sample data and the demand variables of other medical demand positioning data in the target medical behavior sample data comprises:
for each medical requirement positioning data in the target medical behavior sample data, performing knowledge vector mining on a requirement variable of the medical requirement positioning data to obtain a corresponding basic knowledge vector and an extended knowledge vector;
according to the basic knowledge vector and the extended knowledge vector corresponding to the medical requirement positioning data and the basic knowledge vector and the extended knowledge vector corresponding to other medical requirement positioning data in the target medical behavior sample data, conducting weighted aggregation on the requirement variable of the medical requirement positioning data and the requirement variable of other medical requirement positioning data in the target medical behavior sample data, and aggregating the weighted aggregation variable and historical relevant requirement variables to obtain the aggregated requirement variable of the medical requirement positioning data.
6. The big data mining and evaluation method for intelligent medical treatment according to claim 5, wherein the weighted aggregation of the demand variables of the medical requirement positioning data and the demand variables of the other medical requirement positioning data in the target medical behavior sample data according to the basic knowledge vector and the extended knowledge vector corresponding to the medical requirement positioning data and the basic knowledge vector and the extended knowledge vector corresponding to the other medical requirement positioning data in the target medical behavior sample data comprises:
outputting the frequency degree according to a frequent item attention template according to the basic knowledge vector and the extended knowledge vector corresponding to the medical requirement positioning data and the basic knowledge vector and the extended knowledge vector corresponding to other medical requirement positioning data in the target medical behavior sample data to obtain frequent item parameters respectively corresponding to the medical requirement positioning data and other medical requirement positioning data in the target medical behavior sample data;
and according to the medical requirement positioning data and the frequent parameters respectively corresponding to other medical requirement positioning data in the target medical behavior sample data, carrying out weighted aggregation on the requirement variables of the medical requirement positioning data and the requirement variables of other medical requirement positioning data in the target medical behavior sample data.
7. The big data mining evaluation method for intelligent medical treatment according to claim 1, wherein the determining the requirement related value of any one of the medical requirement positioning data in the target medical behavior sample data and any one of the medical requirement positioning data in the comparison medical behavior sample data according to the requirement variable of each of the medical requirement positioning data in the target medical behavior sample data and the requirement variable of each of the medical requirement positioning data in the comparison medical behavior sample data comprises:
generating a demand association value array between different medical demand positioning data in different medical behavior sample data according to the demand variable of each medical demand positioning data in the target medical behavior sample data and the demand variable of each medical demand positioning data in the comparison medical behavior sample data;
traversing and resolving different array members from the array of requirement correlation values, wherein the different array members represent requirement correlation values between two pieces of medical requirement positioning data corresponding to different pieces of medical behavior sample data;
after generating an array of requirement associated values between different medical requirement positioning data in different medical behavior sample data, and when the data fragment quantity of the medical requirement positioning data included in the target medical behavior sample data is different from the data fragment quantity of the medical requirement positioning data included in the control medical behavior sample data, generating an array member serving as a blank unit in the array of requirement associated values;
loading a requirement variable of medical requirement positioning data only appearing in the target medical behavior sample data or the contrast medical behavior sample data into the blank unit to obtain an updated requirement correlation value array;
and circularly performing regularized conversion on each array member in the updated requirement correlation value array, and removing the blank unit to obtain a requirement correlation value array updated again.
8. The big data mining evaluation method for intelligent medical treatment according to claim 1, wherein the performing of medical requirement analysis on the target medical behavior sample data and the control medical behavior sample data respectively to obtain a plurality of medical requirement positioning data in the target medical behavior sample data and a plurality of medical requirement positioning data in the control medical behavior sample data comprises:
respectively positioning medical requirement related data of the target medical behavior sample data and the comparison medical behavior sample data to obtain a plurality of reference medical requirement positioning data in the target medical behavior sample data and a plurality of reference medical requirement positioning data in the comparison medical behavior sample data;
and respectively carrying out medical requirement type decision on a plurality of reference medical requirement positioning data in the target medical behavior sample data and a plurality of reference medical requirement positioning data in the comparison medical behavior sample data, and obtaining a plurality of medical requirement positioning data of the medical requirement in any medical requirement type in the target medical behavior sample data and a plurality of medical requirement positioning data of the medical requirement in any medical requirement type in the comparison medical behavior sample data.
9. The big data mining evaluation method applied to smart medicine according to claim 1, wherein the method further comprises:
grouping and evaluating information according to the user requirements of each target user and the corresponding comparison user, and adding the target users with the associated medical requirements and the corresponding comparison users into the same medical requirement user group;
aiming at each medical demand user group, distributing corresponding medical resources to each user in the medical demand user group according to a medical resource distribution strategy corresponding to medical demands, wherein the medical resources comprise medical science popularization push contents;
acquiring resource feedback information of the medical requirement user group aiming at the distributed medical resource, and updating and optimizing the medical resource distribution strategy based on the resource feedback information;
wherein the step of performing update optimization on the medical resource allocation strategy based on the resource feedback information comprises:
acquiring corresponding interest point trigger activity data according to the resource feedback information;
loading the interest point trigger activity data to a first attention intention decision model, and outputting a first attention intention variable of the interest point trigger activity data;
according to variable matching information between the first attention intention variable and second attention intention variables of the authentication interest point triggering activity data of each medical resource distribution tendency in the medical resource distribution tendency cluster, outputting a deviation probability value corresponding to each medical resource distribution tendency in the interest point triggering activity data and the medical resource distribution tendency in the medical resource distribution tendency cluster;
according to the deviation probability value, obtaining strategy optimization information corresponding to each medical resource distribution tendency included in the interest point trigger activity data from the medical resource distribution tendency cluster;
and updating and optimizing the medical resource allocation strategy according to the strategy optimization information corresponding to each medical resource allocation tendency.
10. A big data mining system, comprising a processor and a machine-readable storage medium, wherein the machine-readable storage medium stores a computer program, and the computer program is loaded and executed according to the processor to implement the big data mining evaluation method for smart medicine according to any one of claims 1 to 9.
CN202210322591.9A 2022-03-30 2022-03-30 Big data mining evaluation method and big data mining system applied to intelligent medical treatment Withdrawn CN114678114A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210322591.9A CN114678114A (en) 2022-03-30 2022-03-30 Big data mining evaluation method and big data mining system applied to intelligent medical treatment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210322591.9A CN114678114A (en) 2022-03-30 2022-03-30 Big data mining evaluation method and big data mining system applied to intelligent medical treatment

Publications (1)

Publication Number Publication Date
CN114678114A true CN114678114A (en) 2022-06-28

Family

ID=82075864

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210322591.9A Withdrawn CN114678114A (en) 2022-03-30 2022-03-30 Big data mining evaluation method and big data mining system applied to intelligent medical treatment

Country Status (1)

Country Link
CN (1) CN114678114A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115951588A (en) * 2023-03-10 2023-04-11 广东亿佛手健康科技有限公司 Massage mechanical action control method and system based on AI self-adaptive adjustment

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115951588A (en) * 2023-03-10 2023-04-11 广东亿佛手健康科技有限公司 Massage mechanical action control method and system based on AI self-adaptive adjustment
CN115951588B (en) * 2023-03-10 2023-09-01 广东亿佛手健康科技有限公司 Massaging mechanical action control method and system based on AI self-adaptive adjustment

Similar Documents

Publication Publication Date Title
CN109587713B (en) Network index prediction method and device based on ARIMA model and storage medium
CN113064879B (en) Database parameter adjusting method and device and computer readable storage medium
US20220414426A1 (en) Neural Architecture Search Method and Apparatus, Device, and Medium
Ramírez-Gallego et al. A distributed evolutionary multivariate discretizer for big data processing on apache spark
CN114117240B (en) Internet content pushing method based on big data demand analysis and AI system
CN108846695A (en) The prediction technique and device of terminal replacement cycle
CN112416369B (en) Intelligent deployment method oriented to heterogeneous mixed environment
Crawford et al. Dynamic selection of enumeration strategies for solving constraint satisfaction problems
Cheng et al. Tuning configuration of apache spark on public clouds by combining multi-objective optimization and performance prediction model
CN114678114A (en) Big data mining evaluation method and big data mining system applied to intelligent medical treatment
CN114647790A (en) Big data mining method and cloud AI (Artificial Intelligence) service system applied to behavior intention analysis
CN113626241B (en) Abnormality processing method, device, equipment and storage medium for application program
CN117236656B (en) Informationized management method and system for engineering project
CN110119399A (en) Work Flow Optimizing method based on machine learning
CN114564648A (en) Personalized service content optimization method based on big data and artificial intelligence cloud system
CN106708875B (en) Feature screening method and system
CN114492601A (en) Resource classification model training method and device, electronic equipment and storage medium
CN113282630A (en) Data query method and device based on interface switching
CN112541556A (en) Model construction optimization method, device, medium, and computer program product
CN115422486B (en) Cloud service online page optimization method based on artificial intelligence and big data system
CN110119268A (en) Workflow optimization method based on artificial intelligence
CN112463964B (en) Text classification and model training method, device, equipment and storage medium
CN111949530B (en) Test result prediction method and device, computer equipment and storage medium
CN114528973A (en) Method for generating business processing model, business processing method and device
CN114625967A (en) User information mining method based on big data service optimization and artificial intelligence system

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WW01 Invention patent application withdrawn after publication

Application publication date: 20220628

WW01 Invention patent application withdrawn after publication