CN113345571A

CN113345571A - Big data processing method based on intelligent medical service and block chain medical system

Info

Publication number: CN113345571A
Application number: CN202110682220.7A
Authority: CN
Inventors: 刘钢
Original assignee: Individual
Current assignee: Individual
Priority date: 2021-06-20
Filing date: 2021-06-20
Publication date: 2021-09-03

Abstract

The embodiment of the disclosure provides a big data processing method based on intelligent medical service and a block chain medical system, wherein the characteristic information of a query intention object can be used for representing partial intention information of query behavior data, different query intention objects correspond to different partial intention information, the information of jump excavation intention generated based on the characteristic information of each query intention object in each query behavior data is richer and more comprehensive, and the accuracy of the jump excavation intention can be effectively improved. Furthermore, the feature information of the query intention object comprises not only query behavior data information but also jump information, the query intention component represents the query behavior data information, the jump attribute component represents the jump information, and the fusion intention component fuses the query behavior data information and the jump information.

Description

Big data processing method based on intelligent medical service and block chain medical system

Technical Field

The disclosure relates to the technical field of intelligent medical treatment, in particular to a big data processing method based on intelligent medical service and a block chain medical system.

Background

The intelligent medical treatment is the upgrading development of medical informatization, and provides medical big data service for all parties by taking a medical cloud data center as a carrier through deep fusion with big data and cloud computing technology, so that six synergies among doctors and patients, doctors and nurses, large hospitals and community hospitals, medical treatment and insurance, medical institutions and health management departments, and medical institutions and drug management are realized, and an intelligent medical service system is gradually constructed. Wherein, the construction of the healthy big data platform lays a foundation for intelligent medical treatment.

Aiming at the role of big data in intelligent medical application, the fusion of big data technology and means can greatly perfect and accelerate the pace of pushing, managing and processing the health medical information. For example, the intention information of the user is analyzed based on mining the big data, so as to facilitate information push optimization and service iterative update, however, the current big data analysis strategy does not consider the jump information of the query behavior (for example, the jump process of the behavior intention of the user is also an important mining factor), so that the accuracy of the mining intention still needs to be improved.

Disclosure of Invention

In order to overcome at least the above disadvantages in the prior art, the present disclosure provides a big data processing method and a blockchain medical system based on intelligent medical services.

In a first aspect, the present disclosure provides a big data processing method based on intelligent medical services, which is applied to a blockchain medical system, where the blockchain medical system is communicatively connected to a plurality of intelligent medical distributed devices, and the method includes:

acquiring target intelligent medical service big data of the intelligent medical distributed equipment, determining target query behavior data from the target intelligent medical service big data, extracting query intention objects from the target query behavior data and jump query behavior data corresponding to the target query behavior data respectively, generating query intention components corresponding to the query intention objects, and obtaining a target query intention component sequence corresponding to the target query behavior data and a jump query intention component sequence corresponding to the jump query behavior data; the target query intention component sequence comprises query intention components corresponding to all target query intention objects in target query behavior data respectively, the skip query intention component sequence comprises query intention components corresponding to all skip query intention objects in skip query behavior data respectively, and the target intelligent medical service big data is obtained by collecting based on a plurality of block chain nodes of a block chain network where the intelligent medical distributed equipment is located;

matching the target query intention object and the skip query intention object based on the component correlation degree between the target query intention object and the skip query intention object, and generating a skip attribute component corresponding to the target query intention object based on the successfully matched skip correlation degree between the target query intention object and the skip query intention object;

fusing the jump attribute component and the query intention component corresponding to the same target query intention object to obtain a corresponding fused intention component, and updating the target query intention component sequence based on the fused intention component to obtain an updated query intention component sequence;

and obtaining a jump mining intention corresponding to the target intelligent medical service big data based on the updated query intention component sequence corresponding to the target query behavior data.

In a second aspect, the disclosed embodiments further provide a big data processing system based on intelligent medical services, where the big data processing system based on intelligent medical services includes a blockchain medical system and a plurality of intelligent medical distributed devices communicatively connected to the blockchain medical system;

the blockchain medical system is used for:

According to any one of the aspects, in the embodiments provided by the present disclosure, the feature information of the query intention object may be used to represent partial intention information of the query behavior data, different query intention objects correspond to different partial intention information, information of the jump mining intention generated based on the feature information of each query intention object in each query behavior data is richer and more comprehensive, and the accuracy of the jump mining intention can be effectively improved. Furthermore, the feature information of the query intention object comprises not only query behavior data information but also jump information, the query intention component represents the query behavior data information, the jump attribute component represents the jump information, and the fusion intention component fuses the query behavior data information and the jump information.

Drawings

FIG. 1 is a schematic diagram of an application scenario of a big data processing system based on intelligent medical services according to an embodiment of the present disclosure;

FIG. 2 is a schematic flowchart illustrating a big data processing method based on intelligent medical services according to an embodiment of the present disclosure;

FIG. 3 is a functional block diagram of a big data processing device based on intelligent medical services according to an embodiment of the present disclosure;

fig. 4 is a schematic block diagram of a blockchain medical system for implementing the big data processing method based on intelligent medical services according to an embodiment of the present disclosure.

Detailed Description

The following describes in detail aspects of embodiments of the present disclosure with reference to the drawings attached hereto. In the following description, for purposes of explanation and not limitation, specific details are set forth such as particular system structures, interfaces, techniques, etc. in order to provide a thorough understanding of the particular embodiments of the disclosure.

Fig. 1 is a schematic view of a big data processing system 10 based on intelligent medical services according to an embodiment of the present disclosure. The big data processing system 10 based on intelligent medical services can comprise a blockchain medical system 100 and an intelligent medical distributed device 200 which is connected with the blockchain medical system 100 in a communication mode. While the smart medical services-based big data processing system 10 shown in FIG. 1 is merely one possible example, in other possible embodiments, the smart medical services-based big data processing system 10 may include only at least some of the components shown in FIG. 1 or may include other components.

In this embodiment, the blockchain medical system 100 and the intelligent medical distributed device 200 in the smart medical service-based big data processing system 10 can cooperate to execute the smart medical service-based big data processing method described in the following method embodiments, and the detailed description of the method embodiments can be referred to in the following steps of the blockchain medical system 100 and the intelligent medical distributed device 200.

To solve the technical problem in the background art, fig. 2 is a flowchart illustrating a big data processing method based on smart medical services according to an embodiment of the present disclosure, which can be executed by the blockchain medical system 100 shown in fig. 1, and the big data processing method based on smart medical services is described in detail below.

Step S110, acquiring target intelligent medical service big data, determining target query behavior data from the target intelligent medical service big data, extracting query intention objects from the target query behavior data and jump query behavior data corresponding to the target query behavior data respectively, generating query intention components corresponding to the query intention objects, and acquiring a target query intention component sequence corresponding to the target query behavior data and a jump query intention component sequence corresponding to the jump query behavior data; the target query intention component sequence comprises query intention components corresponding to all target query intention objects in the target query behavior data respectively, and the jump query intention component sequence comprises query intention components corresponding to all jump query intention objects in the jump query behavior data respectively.

For example, the blockchain medical system 100 may obtain the target smart medical service big data locally or from the smart medical distributed device 200 or other blockchain medical systems 100, and generate the jump mining intention corresponding to the target smart medical service big data. The blockchain medical system 100 may perform query behavior data analysis on the target smart medical service big data to obtain a query behavior data sequence corresponding to the target smart medical service big data. The blockchain medical system 100 may determine at least one target query behavior data from the sequence of query behavior data, and generate jump mining intents based on each target query behavior data. In determining the target query behavior data, the blockchain medical system 100 may select a unit number of query behavior data from the query behavior data sequence as the target query behavior data, for example, each behavior trigger node acquires one query behavior data as the target query behavior data. Of course, the blockchain medical system 100 may also use each query behavior data in the query behavior data sequence as the target query behavior data, and the blockchain medical system 100 may also randomly select any number of query behavior data from the target query behavior data as the target query behavior data.

The skip query behavior data corresponding to the target query behavior data refers to at least one of forward query behavior data and backward query behavior data corresponding to the target query behavior data. In the query behavior data sequence, for example, several pieces of selected forward query behavior data may be spaced in front of the target query behavior data to serve as skip query behavior data, and several pieces of selected forward query behavior data may be spaced behind the target query behavior data to serve as skip query behavior data. For example, the first third jump of the target query behavior data is selected as the forward query behavior data, and the last third jump of the target query behavior data is selected as the backward query behavior data.

In the query of behavior data, there are some behavior objects with query intentions, for example, drug consultation behavior intentions on a behavior object querying a disease object topic, and the patterns of these behavior objects have specific features, which can be distinguished from other regions, and these features still have invariance under some common data processing processes of querying behavior data. Once there may be many such query intents in the query behavior data, the objects of which are referred to as query intent objects. The query intention component corresponding to the query intention object refers to query behavior data characteristics of the behavior object corresponding to the query intention object, and can be presented in a coded form.

The target query intention object is a query intention object extracted from the target query behavior data, and the jump query intention object is a query intention object extracted from the jump query behavior data. The target query intention component sequence comprises query intention components corresponding to the target query intention objects in the target query behavior data. The jump query intention component sequence comprises query intention components corresponding to all jump query intention objects in the jump query behavior data.

For example, the blockchain medical system 100 may extract a query intention object from the target query behavior data, obtain at least one target query intention object, and generate a query intention component corresponding to each target query intention object, thereby obtaining a target query intention component sequence corresponding to the target query behavior data. The blockchain medical system 100 may extract the query intention object from the jump query behavior data to obtain at least one jump query intention object, and generate a query intention component corresponding to each jump query intention object, thereby obtaining a jump query intention component sequence corresponding to the jump query behavior data.

In one embodiment, the blockchain medical system 100 may extract query intent objects from the query behavior data through a query intent object extraction algorithm, such as a query intent classification model trained based on artificial intelligence.

And step S120, matching the target query intention object and the jump query intention object based on the component correlation degree between the target query intention object and the jump query intention object, and generating a jump attribute component corresponding to the target query intention object based on the jump correlation degree between the target query intention object and the jump query intention object which are successfully matched.

For example, the component correlation between the target query intent object and the jump query intent object refers to a quantitative value of a similar behavior between a query intent component corresponding to the target query intent object and a query intent component corresponding to the jump query intent object. The blockchain medical system 100 may perform matching of the target query intent object and the skip query intent object based on component relevance between the target query intent object and the skip query intent object to determine the target query intent object and the skip query intent object that match each other.

In one embodiment, the matching of the target query intention object and the jump query intention object may adopt a nearest neighbor algorithm based on query intention components, calculate component correlations between one target query intention object and each jump query intention object, and select a jump query intention object corresponding to the component correlation with the smallest quantization value of the similar behavior from the multiple component correlations as the jump query intention object matched with the target query intention object. In addition, in order to improve the matching accuracy, a threshold may be set, and when the component relevance degree with the smallest quantization value of the similar behavior is greater than the component relevance degree threshold, the jump query intention object corresponding to the component relevance degree with the smallest quantization value of the similar behavior is finally determined as the jump query intention object matched with the target query intention object.

The jump relevance of the target query intention object and the jump query intention object refers to a quantitative value of similar behaviors of the intentions of the target query intention object and the jump query intention object in the jump process. The jump attribute component is jump information for representing a target query intent object.

For example, when matching between the target query intention object and the jump query intention object is performed, some target query intention objects may find the jump query intention objects matched with each other, and some target query intention objects may not find the jump query intention objects matched with each other. And when the target query intention object and the skip query intention object are matched with each other, indicating that the target query intention object and the skip query intention object are successfully matched. The target query intention object and the jump query intention object which are successfully matched can be regarded as the same query intention object on different query behavior data, so that the jump attribute component corresponding to the target query intention object can be generated based on the jump relevance of the target query intention object and the jump query intention object which are successfully matched.

In one embodiment, the jump query behavior data includes at least one of forward query behavior data and backward query behavior data. The query intention object in the forward query behavior data can be used as a forward query intention object, and a forward jump attribute component corresponding to the target query intention object can be generated based on the jump relevance of the successfully matched target query intention object and the forward query intention object. Similarly, the query intention object in the backward query behavior data may be used as a backward query intention object, and a backward jump attribute component corresponding to the target query intention object may be generated based on the jump relevance between the target query intention object and the backward query intention object that are successfully matched. Because the target query behavior data usually includes a plurality of target query intention objects, a plurality of forward jump attribute components and a plurality of backward jump attribute components can be finally obtained, each forward jump attribute component can form a forward jump attribute component sequence, and each backward jump attribute component can form a backward jump attribute component sequence.

In one embodiment, the jump relevance of the successfully matched target query intention object and jump query intention object can be directly used as the jump attribute component of the corresponding target query intention object.

And S130, fusing the jump attribute component and the query intention component corresponding to the same target query intention object to obtain a corresponding fused intention component, and updating the target query intention component sequence based on the fused intention component to obtain an updated query intention component sequence.

For example, after determining the jump attribute component corresponding to the target query intention object, the blockchain medical system 100 may fuse the jump attribute component corresponding to the target query intention object with the query intention component to obtain a fused intention component corresponding to the target query intention object. The jump attribute component can represent jump information of the target query intention object, the characteristic can represent query behavior data information of a local query behavior data part corresponding to the target query intention object, and a fusion intention component obtained by fusing the jump attribute component and the query intention component can simultaneously represent the jump information and the query behavior data information of the target query intention object. Further, the blockchain medical system 100 may update the target query intent component sequence based on the fused intent component, resulting in an updated query intent component sequence.

In one embodiment, the sequence of updated query intent components includes updated query intent components corresponding to respective target query intent objects. The updated query intent component may be a fused intent component, i.e., the updated query intent component sequence includes fused intent components corresponding to respective target query intent objects. When updating the target query intent component sequence based on the fused intent components, the blockchain medical system 100 may replace the query intent components corresponding to the target query intent objects with the corresponding fused intent components and filter the target query intent objects without the corresponding fused intent components to relieve the storage pressure of the blockchain medical system 100. The update query intent component may be a fusion intent component or a query intent component, that is, the update query intent component sequence may also include a fusion intent component corresponding to a target query intent object having a jump attribute component and a query intent component corresponding to a target query intent object not having a jump attribute component. The blockchain medical system 100 may replace the query intent component corresponding to the target query intent object with the corresponding fused intent component and retain the query intent component corresponding to the target query intent object without the corresponding fused intent component when updating the target query intent component sequence based on the fused intent component.

Step S140, obtaining the jump mining intention corresponding to the target intelligent medical service big data based on the updated query intention component sequence corresponding to the target query behavior data.

The skip mining intention is intention mining information used for representing the content of the intelligent medical service big data, and the skip mining intention can be applied to application scenes such as recommendation of the intelligent medical service big data and search of the intelligent medical service big data.

For example, the blockchain medical system 100 obtains the jump mining intention corresponding to the target smart medical service big data based on the updated query intention component sequence corresponding to the target query behavior data. It can be understood that, when there are a plurality of target query behavior data, the blockchain medical system 100 may obtain an updated query intention component sequence corresponding to each target query behavior data, and each updated query intention component sequence constitutes a jump mining intention corresponding to the target smart medical service big data. The updated query intent component sequence may also be used as a query behavior data intent corresponding to the target query behavior data.

The jump mining intent can be applied to intelligent medical service big data recommendations. The blockchain medical system 100 can obtain first smart medical service big data browsed by a user in history and second smart medical service big data to be recommended, generate a first skip mining intention corresponding to the first smart medical service big data and a second skip mining intention corresponding to the second smart medical service big data, and determine a smart medical service big data matching result of the first smart medical service big data and the second smart medical service big data based on a matching degree of the first skip mining intention and the second skip mining intention. When the matching result of the intelligent medical service big data is that the intelligent medical service big data are similar, the second intelligent medical service big data are not recommended to the user, and therefore repeated intelligent medical service big data are prevented from being recommended to the user. Of course, the blockchain medical system 100 may also obtain an initial recommendation result sequence actively recommended to the user by the current recommendation system. When the fact that similar intelligent medical service big data exist in the initial recommendation result sequence is determined based on the jumping mining intention of each intelligent medical service big data in the initial recommendation result sequence, the initial recommendation result sequence is updated, only one of the similar intelligent medical service big data is reserved, a target recommendation result sequence is obtained, and the target recommendation result sequence is displayed to a user.

The jump mining intent can be applied to intelligent medical service big data search. The blockchain medical system 100 may obtain an initial sequence of search results that the current search system recommends to the user. The initial search result sequence is generated based on search information of the user. When the similar intelligent medical service big data exist in the initial search result sequence is determined based on the jumping excavation intention of each intelligent medical service big data in the initial search result sequence, the initial search result sequence is updated, only one of the similar intelligent medical service big data is reserved to obtain a target search result sequence, and the target search result sequence is displayed to a user. Therefore, the user can obtain more various search results, and the search effectiveness is improved.

When the intelligent medical service big data is matched, the update component correlation degree between query intention objects in the query behavior data of the first intelligent medical service big data and the query behavior data of the second intelligent medical service big data can be calculated, so that the number of similar query intention objects between the query behavior data is determined, the matching query intention object proportion is calculated based on the number of similar query intention objects between the query behavior data, and whether the query behavior data is similar query behavior data or not is determined based on the matching query intention object proportion. And further, determining a similar query behavior data ratio based on the quantity of the similar query behavior data, and determining a smart medical service big data matching result of the first smart medical service big data and the second smart medical service big data based on the similar query behavior data ratio. The specific process of matching the smart medical service big data may refer to a big data processing method based on the smart medical service described in each related embodiment of the smart medical service big data matching method described later, and details thereof are not repeated herein.

In the method for generating the jump excavation intention, the characteristic information of the query intention object can be used for representing partial intention information of the query behavior data, different query intention objects correspond to different partial intention information, the information of the jump excavation intention generated based on the characteristic information of each query intention object in each query behavior data is richer and more comprehensive, and the accuracy of the jump excavation intention can be effectively improved. Furthermore, the characteristic information of the query intention object comprises query behavior data information and skip information, correspondingly, the skip excavation intention comprises the query behavior data information and the skip information, and therefore when the query behavior data are matched based on the skip excavation intention, the accuracy of intelligent medical service big data matching can be effectively improved.

In one embodiment, matching the target query intent object and the skip query intent object based on component relevance between the target query intent object and the skip query intent object comprises: calculating component relevance between the query intention components corresponding to the current target query intention object and the query intention components corresponding to the skip query intention objects respectively to obtain a plurality of component relevance corresponding to the current target query intention object; and when the maximum component correlation degree is greater than the preset correlation degree, taking the jump query intention object corresponding to the maximum component correlation degree as a jump query intention object matched with the current target query intention object.

For example, when performing query intent object matching, the blockchain medical system 100 may randomly select one target query intent object from the multiple target query intent objects as a current target query intent object, and calculate component relevance between query intent components corresponding to the current target query intent object and query intent components corresponding to each skip query intent object, so as to obtain multiple component relevance corresponding to the current target query intent object. And then selecting the maximum component correlation degree from the multiple component correlation degrees, and when the maximum component correlation degree is greater than the preset correlation degree, taking the jump query intention object corresponding to the maximum component correlation degree as the jump query intention object matched with the current target query intention object. And if the maximum component correlation degree is less than or equal to the preset correlation degree, indicating that the current target query intention object does not have a corresponding jump query intention object. When the target query behavior data comprises a plurality of target query intention objects, each target query intention object searches the corresponding matched jump query intention object according to the same method. The preset correlation degree can be set according to actual needs.

In this embodiment, the query intent component of the query intent object may represent local query behavior data information, and the query intent objects that are matched with each other should be located in the same or very similar local query behavior data portions, so that when the maximum component correlation degree is greater than the preset correlation degree, it may be quickly determined that the jump query intent object corresponding to the maximum component correlation degree and the current target query intent object are located in the same local query behavior data portion, and the two are successfully matched.

In one embodiment, the step of using the jump query intention object corresponding to the maximum component relevance as the jump query intention object matched with the current target query intention object comprises the following steps: and when the jump relevance between the jump query intention object corresponding to the maximum component relevance and the current target query intention object is greater than the target relevance, taking the jump query intention object corresponding to the maximum component relevance as the jump query intention object matched with the current target query intention object.

For example, when performing query intent object matching, the blockchain medical system 100 may calculate component relevance between query intent components corresponding to the current target query intent object and query intent components corresponding to the respective jump query intent objects, obtain a plurality of component relevance corresponding to the current target query intent object, select a maximum component relevance from the plurality of component relevance, and when the maximum component relevance is greater than a preset relevance and the jump relevance between the jump query intent object corresponding to the maximum component relevance and the current target query intent object is less than the target relevance, take the jump query intent object corresponding to the maximum component relevance as the jump query intent object matched with the current target query intent object. If the jump relevance between the jump query intention object corresponding to the maximum component relevance and the current target query intention object is smaller than or equal to the target relevance, the difference between the jump query intention object and the current target query intention object is large, the intention change between query behavior data is too obvious, the query behavior data may be abnormal, therefore, the jump query intention object and the current target query intention object are not considered to be matched with each other, and the target relevance can be set according to actual needs.

In this embodiment, when the maximum component correlation degree is greater than the preset correlation degree and the jump correlation degree between the jump query intention object corresponding to the maximum component correlation degree and the current target query intention object is greater than the target correlation degree, the jump query intention object corresponding to the maximum component correlation degree is used as the jump query intention object matched with the current target query intention object. In this way, the jump query intention object matched with the current target query intention object can be accurately found.

In one embodiment, the jump query behavior data corresponding to the target query behavior data includes at least one of forward query behavior data and backward query behavior data corresponding to the target query behavior data, the jump query intention object corresponding to the forward query behavior data is a forward query intention object, the jump query intention object corresponding to the backward query behavior data is a backward query intention object, and the jump attribute component includes at least one of a forward jump attribute component and a backward jump attribute component. Generating a jump attribute component corresponding to the target query intention object based on the jump relevance of the successfully matched target query intention object and the jump query intention object, wherein the jump attribute component comprises the following steps: when the successfully matched jump query intention object is the forward query intention object of the target query intention object, generating a forward jump attribute component corresponding to the target query intention object based on the jump relevance of the successfully matched target query intention object and the jump query intention object; and when the successfully matched jump query intention object is the backward query intention object of the target query intention object, generating a backward jump attribute component corresponding to the target query intention object based on the jump relevance of the successfully matched target query intention object and the jump query intention object.

For example, the jump query behavior data corresponding to the target query behavior data includes at least one of forward query behavior data and backward query behavior data corresponding to the target query behavior data, and the jump attribute component includes at least one of a forward jump attribute component and a backward jump attribute component. And when the skip query behavior data corresponding to the target query behavior data is the forward query behavior data corresponding to the target query behavior data, the skip query intention object in the skip query behavior data is the forward query intention object. And when the successfully matched jump query intention object is the forward query intention object of the target query intention object, generating a forward jump attribute component corresponding to the target query intention object based on the jump relevance of the successfully matched target query intention object and the jump query intention object. And when the jump query behavior data corresponding to the target query behavior data is backward query behavior data corresponding to the target query behavior data, the jump query intention object in the jump query behavior data is a backward query intention object. And when the successfully matched jump query intention object is the backward query intention object of the target query intention object, generating a backward jump attribute component corresponding to the target query intention object based on the jump relevance of the successfully matched target query intention object and the jump query intention object.

In one embodiment, the generating of the current jump attribute component corresponding to the target query intention object based on the jump relevance between the target query intention object and the jump query intention object that are successfully matched includes:

step S121, generating initial jump attribute components corresponding to all target query intention objects based on the jump relevance of the target query intention objects and the jump query intention objects which are successfully matched; the initial hop attribute component carries the hop dimension.

And S122, clustering the initial jump attribute components corresponding to the same jump dimension to obtain clustering objects corresponding to all jump dimensions respectively.

And S123, counting the number of the initial jump attribute components in the same clustering object to obtain the reference number corresponding to each clustering object.

And step S124, taking the jump dimensionality corresponding to the clustering objects with the maximum reference number as a target jump dimensionality.

And step S125, adaptively updating each initial jump attribute component based on the target jump dimension to obtain the current jump attribute component corresponding to each target query intention object.

In order to counteract the influence caused by the service update of the intelligent medical service big data, the skip dimension of the skip attribute component needs to be normalized, and the skip dimension normalization needs to be respectively carried out on the forward skip attribute component and the backward skip attribute component. The blockchain medical system 100 may generate initial jump attribute components corresponding to the target query intent objects based on the jump relevance of the successfully matched target query intent objects and the jump query intent objects, where the initial jump attribute components carry jump dimensions. And then, clustering the initial jump attribute components corresponding to the same jump dimension to obtain a clustering object corresponding to each jump dimension. One cluster object comprises at least one initial jump attribute component corresponding to the same jump dimension. The blockchain medical system 100 may count the number of initial jump attribute components in the same cluster object to obtain a reference number corresponding to each cluster object, and use a jump dimension corresponding to the cluster object with the largest reference number as a target jump dimension, that is, use a jump dimension corresponding to most of the initial jump attribute components as a main motion jump dimension of the target query behavior data. Finally, the blockchain medical system 100 may adaptively update each initial jump attribute component based on the target jump dimension to obtain a current jump attribute component corresponding to each target query intent object.

And when the successfully matched jump query intention object is a forward query intention object of the target query intention object, the initial jump attribute component is a forward initial jump attribute component, and the forward initial jump attribute component carries forward jump dimensionality. The blockchain medical system 100 may cluster the forward initial skip attribute components corresponding to the same forward skip dimension to obtain cluster objects corresponding to each forward skip dimension, count the number of the forward initial skip attribute components in the same cluster object to obtain a reference number corresponding to each cluster object, and use the forward skip dimension corresponding to the cluster object with the largest reference number as a target forward skip dimension, that is, use skip dimensions corresponding to most of the forward initial skip attribute components as main forward motion skip dimensions of target query behavior data, and adjust each forward initial skip attribute component based on the target forward skip dimension to obtain a forward target skip attribute component corresponding to each target query intention object.

And when the successfully matched jump query intention object is a backward query intention object of the target query intention object, the initial jump attribute component is a backward initial jump attribute component, and the backward initial jump attribute component carries a backward jump dimension. The blockchain medical system 100 may cluster the backward initial jump attribute components corresponding to the same backward jump dimension to obtain cluster objects corresponding to each backward jump dimension, count the number of the backward initial jump attribute components in the same cluster object to obtain the reference number corresponding to each cluster object, and use the backward jump dimension corresponding to the cluster object with the largest reference number as the target backward jump dimension, that is, use the jump dimensions corresponding to most of the backward initial jump attribute components as the main backward movement jump dimension of the target query behavior data, and adjust each backward initial jump attribute component based on the target backward jump dimension to obtain the backward target jump attribute component corresponding to each target query intention object.

In this embodiment, the influence caused by the service update of the smart medical service big data can be offset by normalizing the skip dimensions of the skip attribute components, so that the accuracy of the query behavior data intention of the target query behavior data is improved, and the accuracy of the skip excavation intention is improved.

In one embodiment, adaptively updating each initial jump attribute component based on a target jump dimension to obtain a current jump attribute component corresponding to each target query intention object, includes:

and S1251, adaptively updating each initial jump attribute component based on the target jump dimension to obtain an intermediate jump attribute component corresponding to each target query intention object.

Step S1252, perform regularization processing on each intermediate jump attribute component to obtain a current jump attribute component corresponding to each target query intention object.

For example, after the skip dimension normalization is performed, the normalization is performed, and the limitation information of all skip attribute components is normalized to the unit vector skip cost, so that an error caused by the limitation information of the skip attribute components can be avoided, and the features are focused on the service update dimension components. The blockchain medical system 100 may, for example, adaptively update each initial jump attribute component based on the target jump dimension to obtain an intermediate jump attribute component corresponding to each target query intention object, and perform regularization processing on each intermediate jump attribute component to obtain a current jump attribute component corresponding to each target query intention object. It can be understood that all the forward initial jump attribute components are adjusted to the dimension component of a main forward motion jump dimension based on the main forward motion jump dimension, and all the backward initial jump attribute components are adjusted to the dimension component of a main backward motion jump dimension based on the main backward motion jump dimension, so that the intermediate jump attribute component with the jump dimension normalized is obtained. And then, carrying out regularization processing on each intermediate jump attribute component to obtain the current jump attribute component corresponding to each target query intention object. The position of the query intention object is changed due to noise, which in turn causes a change in the constraint information (i.e., the hop cost) of the hop attribute component. When the constraint information is unreliable, if the quantized value of the vector similarity behavior is calculated based on the constraint information, the quantized value of the similarity behavior has a problem, so that the regularization processing is required to avoid the problem.

In the embodiment, the normalization of the jump dimension is performed and then the regularization is performed, so that the accuracy of the query behavior data intention of the target query behavior data can be improved, and the accuracy of the jump mining intention is improved.

In one embodiment, performing regularization processing on each intermediate jump attribute component to obtain a current jump attribute component corresponding to each target query intention object includes:

(1) and acquiring the intermediate jump attribute component with the jump cost smaller than the target cost from each intermediate jump attribute component as a key jump attribute component, and randomly assigning the key jump attribute component as a random jump attribute component.

(2) And carrying out regularization processing on the random jump attribute component and the intermediate jump attribute component with the jump cost larger than or equal to the target cost to obtain the current jump attribute component corresponding to each target query intention object.

For example, in order to improve the accuracy of matching the smart medical service big data based on the jump mining intention, random disturbance processing can be performed on the jump attribute component of the key target query intention object, so that mismatching of the smart medical service big data caused by mismatching of the key target query intention object is avoided. The blockchain medical system 100 may obtain the intermediate jump attribute component with the jump cost less than the target cost from the intermediate jump attribute components as a key jump attribute component, where the key jump attribute component represents a jump attribute component corresponding to a key target query intention object. Next, the blockchain medical system 100 may randomly assign the key jump attribute component to a random jump attribute component, that is, randomly initialize the value of the key jump attribute component to generate a random jump attribute component, where the vector jump cost of the random jump attribute component is not limited. And finally, carrying out regularization processing on the random jump attribute component and the intermediate jump attribute component with the jump cost larger than or equal to the target cost, normalizing the limiting information of all the jump attribute components to the unit vector jump cost, and obtaining the current jump attribute component corresponding to each target query intention object. Wherein, the target cost can be set according to actual needs.

It can be understood that each key jump attribute component may also be randomly assigned as a random jump attribute component of the unit vector jump cost, that is, the jump cost of the random jump attribute component is the unit vector jump cost, but each vector value in different random jump attribute components may be different. Subsequently, the random jump attribute component can directly carry out regularization processing on the intermediate jump attribute component with the jump cost larger than or equal to the target cost without carrying out regularization processing.

In one embodiment, the method for generating the fusion intention component includes any one of the following modes: embedding the jump attribute components corresponding to the same target query intention object into the preset positions in the corresponding query intention components to obtain corresponding fusion intention components; and carrying out vector splicing on the jump attribute component and the query intention component corresponding to the same target query intention object to obtain a corresponding fusion intention component.

For example, fusing the jump attribute component and the query intent component corresponding to the target query intent object to generate the fusion intent component may be embedding the jump attribute component corresponding to the target query intent object into a preset position in the query intent component. Of course, the jump attribute component and the query intention component corresponding to the target query intention object are fused to generate the fusion intention component, or the jump attribute component and the query intention component are vector-spliced to obtain the fusion intention component.

In this embodiment, the skip attribute component is embedded into the query intention component to obtain the fusion intention component, and it can be ensured that the number of bits of the fusion intention component is the same as that of the query intention component, thereby ensuring compatibility of the blockchain medical system 100 in processing smart medical service big data, and developers only need to design a software program with the number of bits of the query intention component. The skip attribute component and the query intention component are spliced to obtain a fusion intention component, the integrity of the query intention component can be guaranteed, more information can be considered in the following intelligent medical service big data matching based on skip mining intention, and therefore the accuracy of the intelligent medical service big data matching is improved.

In one embodiment, the embodiment of the present disclosure further provides a big data processing method based on intelligent medical services, including the following steps:

step S210, obtaining the first smart medical service big data and the second smart medical service big data, and generating a first skip mining intention corresponding to the first smart medical service big data and a second skip mining intention corresponding to the second smart medical service big data.

For example, the first smart medical service big data and the second smart medical service big data may be smart medical service big data issued by the same platform, or smart medical service big data issued by different platforms. The blockchain medical system 100 may obtain first smart medical service big data and second smart medical service big data locally or from the smart medical distributed device 200 and other blockchain medical systems 100, generate a first skip mining intention corresponding to the first smart medical service big data and a second skip mining intention corresponding to the second smart medical service big data, perform smart medical service big data matching of the first smart medical service big data and the second smart medical service big data based on the first skip mining intention and the second skip mining intention, and determine whether the first smart medical service big data and the second smart medical service big data are similar smart medical service big data.

The method for generating the skip excavation intention comprises the following steps: determining target query behavior data from current intelligent medical service big data, extracting query intention objects from the target query behavior data and jump query behavior data corresponding to the target query behavior data respectively, generating query intention components corresponding to the query intention objects respectively, obtaining a target query intention component sequence corresponding to the target query behavior data and a jump query intention component sequence corresponding to the jump query behavior data, wherein the target query intention component sequence comprises the query intention components corresponding to the target query intention objects in the target query behavior data respectively, the jump query intention component sequence comprises the query intention components corresponding to the jump query intention objects in the jump query behavior data respectively, and matching the target query intention objects and the jump query intention objects based on the component correlation degree between the target query intention objects and the jump query intention objects, generating a jump attribute component corresponding to the target query intention object based on the jump relevance of the successfully matched target query intention object and the jump relevance of the jump query intention object, fusing the jump attribute component and the query intention component corresponding to the same target query intention object to obtain a fused intention component, updating a target query intention component sequence based on the fused intention component to obtain an updated query intention component sequence, and obtaining a jump mining intention corresponding to the current intelligent medical service big data based on the updated query intention component sequence corresponding to the target query behavior data.

The current intelligent medical service big data is first intelligent medical service big data or second intelligent medical service big data. If the current intelligent medical service big data is the first intelligent medical service big data, the target query behavior data is the first query behavior data, and the target query intention object is the first query intention object. If the current intelligent medical service big data is second intelligent medical service big data, the target query behavior data is second query behavior data, and the target query intention object is a second query intention object.

It is understood that, the specific process of generating the jump mining intention may refer to the big data processing method based on the intelligent medical service described in the foregoing related embodiments of the jump mining intention generation method, and will not be described herein again.

Step S220, determining an intelligent medical service big data matching result of the first intelligent medical service big data and the second intelligent medical service big data based on the matching degree of the first jump mining intention and the second jump mining intention.

For example, the blockchain medical system 100 may match the first jump mining intention and the second jump mining intention, and determine an intelligent medical service big data matching result of the first intelligent medical service big data and the second intelligent medical service big data based on a matching degree of the first jump mining intention and the second jump mining intention. The matching result of the intelligent medical service big data comprises similarity of the intelligent medical service big data and dissimilarity of the intelligent medical service big data. In the scene of intelligent medical service big data recommendation and intelligent medical service big data search, when the matching result of the intelligent medical service big data is that the intelligent medical service big data is similar, one of the first intelligent medical service big data and the second intelligent medical service big data is not repeatedly recommended to a user, and when the matching result of the intelligent medical service big data is that the intelligent medical service big data is not similar, the first intelligent medical service big data and the second intelligent medical service big data can be recommended to the user.

Based on the steps, the characteristic information of the query intention object can be used for representing partial intention information of the query behavior data, different query intention objects correspond to different partial intention information, the information of the jump mining intention generated based on the characteristic information of each query intention object in each query behavior data is richer and more comprehensive, and the accuracy of the jump mining intention can be effectively improved. Furthermore, the feature information of the query intention object comprises not only query behavior data information but also jump information, the query intention component represents the query behavior data information, the jump attribute component represents the jump information, and the fusion intention component fuses the query behavior data information and the jump information. Therefore, when query behavior data are matched based on the jump mining intention, the accuracy of intelligent medical service big data matching can be effectively improved.

In one embodiment, the first jump mining intention comprises a first updated query intention component sequence corresponding to each first query behavior data, and the second jump mining intention comprises a second updated query intention component sequence corresponding to each second query behavior data. Step S220 includes:

step S221, each first updating query intention component sequence is matched with each second updating query intention component sequence, and a matching result of each first query behavior data and the query behavior data of each second query behavior data is determined according to the matching result.

Step S222, calculating a similar query behavior data ratio for the successfully matched query behavior data logarithm based on the query behavior data matching result, and determining a smart medical service big data matching result of the first smart medical service big data and the second smart medical service big data based on the similar query behavior data ratio.

The first jump mining intention comprises first updating query intention component sequences corresponding to the first query behavior data respectively, one first updating query intention component sequence can represent the query behavior data intention of the first query behavior data, and the query behavior data intention of the first query behavior data forms the first jump mining intention. A first sequence of update query intent components includes update query intent components corresponding to respective first query intent objects in a first query behavior data. The second jump mining intention comprises second updating query intention component sequences respectively corresponding to the second query behavior data, one second updating query intention component sequence can represent the query behavior data intention of one second query behavior data, and the query behavior data intention of each second query behavior data forms the second jump mining intention. And a second updating query intention component sequence comprises updating query intention components corresponding to the second query intention objects in the second query behavior data respectively.

For example, when performing jump mining intent matching, the blockchain medical system 100 may match each first updated query intent component sequence with each second updated query intent component sequence, and determine a result of matching each first query behavior data with the query behavior data of each second query behavior data according to the matching result. When matching a first updated query intent component sequence and a second updated query intent component sequence, the blockchain medical system 100 may calculate an updated component correlation degree between each first query intent object and each second query intent object, determine similar first query intent objects and second query intent objects in the first query behavior data and the second query behavior data, and determine whether the first query behavior data and the second query behavior data are successfully matched according to the matching query intent object ratio. The query behavior data matching result comprises matching success and matching failure. And when the matching result of the query behavior data of the first query behavior data and the second query behavior data is successful, indicating that the first query behavior data and the second query behavior data are similar query behavior data to form a pair of query behavior data. When the matching result of the query behavior data of the first query behavior data and the second query behavior data is failure, the first query behavior data and the second query behavior data are not similar query behavior data. After determining the query behavior data matching result, the blockchain medical system 100 may calculate a similar query behavior data ratio for the successfully matched query behavior data logarithm based on the query behavior data matching result, and determine a smart medical service big data matching result of the first smart medical service big data and the second smart medical service big data based on the similar query behavior data ratio. The similar query behavior data proportion is used for representing the proportion of the similar query behavior data in the intelligent medical service big data, when the similar query behavior data proportion is larger than the query behavior data proportion threshold value, the intelligent medical service big data matching result can be determined to be similar to the intelligent medical service big data, and when the similar query behavior data proportion is smaller than or equal to the query behavior data proportion threshold value, the intelligent medical service big data matching result can be determined to be dissimilar to the intelligent medical service big data. It can be understood that the matching degree of the first jump mining intention and the second jump mining intention comprises the matching result of each query behavior data and the proportion of similar query behavior data.

In this embodiment, the similar query behavior data between the first smart medical service big data and the second smart medical service big data is determined, the similar query behavior data ratio is calculated based on the log of the similar query behavior data, and the smart medical service big data matching result is determined based on the similar query behavior data ratio, so that an accurate smart medical service big data matching result can be obtained.

In one embodiment, the first sequence of updated query intent components includes updated query intent components corresponding to a plurality of first query intent objects and the second sequence of updated query intent components includes updated query intent components corresponding to a plurality of second query intent objects. Step S221 includes:

(1) and determining a first target updating query intention component sequence from the first updating query intention component sequences, and determining a second target updating query intention component sequence from the second updating query intention component sequences.

(2) In the first target update query intent component sequence and the second target update query intent component sequence, the first query intent object and the second query intent object are matched based on an update component relevance between the first query intent object and the second query intent object.

(3) And calculating a matching query intention object ratio based on the successfully matched query intention object pairs, and determining a query behavior data matching result of first query behavior data corresponding to the first target updating query intention component sequence and second query behavior data corresponding to the second target updating query intention component sequence based on the matching query intention object ratio.

For example, the blockchain medical system 100 may randomly select one first update query intent component sequence from the first update query intent component sequences as a first target update query intent component sequence and one second update query intent component sequence from the second update query intent component sequences as a second target update query intent component sequence. In the first target update query intent component sequence and the second target update query intent component sequence, the first query intent object and the second query intent object are matched based on an update component relevance between the first query intent object and the second query intent object. When the update component relevance between a first query intent object and a second query intent object is greater than the target relevance, it may be determined that the first query intent object and the second query intent object are a pair of similar query intent objects, the first query intent object and the second query intent object matching successfully. And calculating a matching query intention object ratio based on the successfully matched query intention object pairs, and determining a query behavior data matching result of first query behavior data corresponding to the first target updating query intention component sequence and second query behavior data corresponding to the second target updating query intention component sequence based on the matching query intention object ratio. The matching query intention object proportion is used for representing the proportion of similar query intention objects in the query behavior data, when the matching query intention object proportion is larger than a query intention object proportion threshold value, the matching result of the query behavior data can be determined to be successful in matching, the two pieces of query behavior data are similar query behavior data, when the matching query intention object proportion is smaller than or equal to the query intention object proportion threshold value, the matching result of the query behavior data can be determined to be failed in matching, and the two pieces of query behavior data are not similar query behavior data. That is, the blockchain medical system 100 may calculate the update component correlation between the query intention objects in any two query behavior data, thereby finding the number of similar query intention objects between the two query behavior data, and further determining whether the two query behavior data are similar query behavior data.

In the embodiment, the similar query intention object between the two pieces of query behavior data is determined, the matching query intention object ratio is calculated based on the similar query intention object logarithm, and the query behavior data matching result is determined based on the similar key ratio, so that the accurate query behavior data matching result can be obtained.

In one embodiment, matching a first query intent object and a second query intent object based on an update component relevance between the first query intent object and the second query intent object in a first target update query intent component sequence and a second target update query intent component sequence comprises: and when the update component relevance is greater than the target relevance, determining that the corresponding first query intention object and the second query intention object are successfully matched.

For example, in performing query intent object matching, when an update component relevance between a first query intent object and a second query intent object is greater than a target relevance, it may be determined that the first query intent object and the second query intent object match successfully. The target correlation degree can be set according to actual needs.

In this embodiment, when matching query intent objects between different pieces of intelligent medical service big data is performed, only the update component correlation degree between the query intent objects is considered, and the skip correlation degree between the query intent objects is not considered. In this way, query intent objects that match each other can also be found in the translated query behavior data.

In one embodiment, calculating a matching query intention object ratio based on a successfully matched query intention object logarithm, and determining a query behavior data matching result of first query behavior data corresponding to a first target update query intention component sequence and second query behavior data corresponding to a second target update query intention component sequence based on the matching query intention object ratio, includes:

(1) the number of the first query intention objects is determined based on the number of the updated query intention components in the first target updated query intention component sequence, and the number of the second query intention objects is determined based on the number of the updated query intention components in the second target updated query intention component sequence.

(2) And taking the query intention object number with the smaller number in the first query intention object number and the second query intention object number as the jump query intention object number.

(3) And determining the number of similar query intention objects based on the successfully matched query intention object logarithm, and obtaining the proportion of the matched query intention objects based on the number of the similar query intention objects and the number of the jump query intention objects.

(4) And when the matching query intention object proportion is larger than the target proportion, determining the matching result of the query behavior data of the corresponding first query behavior data and the second query behavior data as successful matching.

For example, in order to calculate the matching query intention object ratio, the number of similar query intention objects and the number of jump query intention objects need to be determined, so that the matching query intention object ratio is calculated based on the ratio of the number of similar query intention objects and the number of jump query intention objects. Because one update query intent component in one sequence of update query intent components corresponds to one target query intent object, the blockchain medical system 100 may determine a first number of query intent objects based on the number of update query intent components in a first sequence of target update query intent components and a second number of query intent objects based on the number of update query intent components in a second sequence of target update query intent components. The blockchain medical system 100 may determine the number of similar query intent objects based on the number of pairs of successfully matched query intent objects, e.g., the number of similar query intent objects is 6 when the number of pairs of successfully matched query intent objects is 6. Then, the blockchain medical system 100 may use the smaller number of the first number of query intent objects and the second number of query intent objects as the number of jump query intent objects, and use the ratio of the number of similar query intent objects to the number of jump query intent objects as the matching query intent object ratio. Further, when the matching query intention object ratio is greater than the target ratio, the matching result of the query behavior data of the corresponding first query behavior data and the second query behavior data can be determined to be successful. When the matching query intention object proportion is smaller than or equal to the target proportion, the matching result of the query behavior data of the corresponding first query behavior data and the second query behavior data can be determined as the matching failure. Wherein, the target proportion can be set according to actual needs.

For example, assuming that the target proportion is 80%, the first query behavior data includes 8 query intention objects, the second query behavior data includes 10 query intention objects, and the number of pairs of similar query intention objects between the first query behavior data and the second query behavior data is 7. Then, the number of similar query intention objects between the first query behavior data and the second query behavior data is 7, the number of skip query intention objects is 8, and the matching query intention object ratio is 87.5% >80%, so that it can be determined that the first query behavior data and the second query behavior data are similar query behavior data, and the matching result of the query behavior data is matching success.

In this embodiment, the number of query intent objects with a smaller number of the first number of query intent objects and the second number of query intent objects is used as the number of jump query intent objects, and an accurate matching query intent object ratio can be obtained based on the number of similar query intent objects and the number of jump query intent objects. Thus, even if one query behavior data is a simple extension of the other query behavior data, it can be determined that the two query behavior data include the same content and are similar query behavior data.

In one embodiment, step S222 includes:

(1) and acquiring a first inquiry behavior data quantity corresponding to the first intelligent medical service big data and acquiring a second inquiry behavior data quantity corresponding to the second intelligent medical service big data.

(2) And taking the query behavior data quantity with the smaller quantity in the first query behavior data quantity and the second query behavior data quantity as the jump query behavior data quantity.

(3) And determining the number of similar query behavior data for the query behavior data logarithm successfully matched based on the query behavior data matching result, and obtaining the proportion of the similar query behavior data based on the number of the similar query behavior data and the number of the jump query behavior data.

(4) And when the proportion of the similar query behavior data is larger than the proportion of the target data, determining that the matching result of the intelligent medical service big data is similar to the intelligent medical service big data.

(5) And when the proportion of the similar query behavior data is smaller than or equal to the proportion of the target data, determining that the matching result of the intelligent medical service big data is that the intelligent medical service big data is not similar.

For example, in order to calculate the similar query behavior data ratio, the number of the similar query behavior data and the number of the jump query behavior data need to be determined, so that the similar query behavior data ratio is calculated based on the ratio of the number of the similar query behavior data and the number of the jump query behavior data. The blockchain medical system 100 may determine the number of similar query behavior data for the successfully matched log of query behavior data based on the query behavior data matching result, for example, when the successfully matched log of query behavior data is 10 pairs, the number of similar query behavior data is 10. The blockchain medical system 100 may obtain a first query behavior data amount corresponding to the first smart medical service big data, obtain a second query behavior data amount corresponding to the second smart medical service big data, and use a query behavior data amount with a smaller amount of the first query behavior data amount and the second query behavior data amount as the skip query behavior data amount. Thus, the blockchain medical system 100 may use the ratio of the number of similar query behavior data to the number of jump query behavior data as the similar query behavior data ratio. When the ratio of the similar query behavior data is larger than the target data ratio, it can be determined that the intelligent medical service big data matching result of the first intelligent medical service big data and the second intelligent medical service big data is the intelligent medical service big data similarity. When the ratio of the similar query behavior data is smaller than or equal to the target data ratio, it can be determined that the smart medical service big data matching result of the first smart medical service big data and the second smart medical service big data is that the smart medical service big data is not similar. Wherein, the target data proportion can be set according to actual needs.

For example, assuming that the target data ratio is 80%, the first smart medical service big data includes 50 pieces of first query behavior data, the second smart medical service big data includes 100 pieces of second query behavior data, and the logarithm of the similar query behavior data between the first smart medical service big data and the second smart medical service big data is 42. Then, the number of the similar query behavior data between the first smart medical service big data and the second smart medical service big data is 42, the number of the skip query intention objects is 50, and the matching query intention object ratio is 84% >80%, so that it can be determined that the first smart medical service big data and the second smart medical service big data are similar smart medical service big data, and the matching result of the smart medical service big data is that the smart medical service big data are similar.

In this embodiment, the query behavior data number with the smaller number of the first query behavior data number and the second query behavior data number is used as the jump query behavior data number, and the similar query behavior data ratio is obtained based on the similar query behavior data number and the jump query behavior data number. Thus, even if one smart medical service big data is simply extended to another smart medical service big data, it can be determined that the two smart medical service big data include the same content and are similar smart medical service big data.

In one embodiment, the method further comprises: and when the matching result of the intelligent medical service big data is that the intelligent medical service big data is similar, forbidding to recommend the second intelligent medical service big data to the browsing user intelligent medical distributed equipment 200 corresponding to the first intelligent medical service big data, and forbidding to recommend the first intelligent medical service big data to the browsing user intelligent medical distributed equipment 200 corresponding to the second intelligent medical service big data.

For example, when the matching result of the smart medical service big data is that the smart medical service big data is similar, it indicates that the first smart medical service big data and the second smart medical service big data are similar smart medical service big data. Then, the blockchain medical system 100 may prohibit the smart medical distributed device 200 of the browsing user corresponding to the first smart medical service big data from recommending the second smart medical service big data, and prohibit the smart medical distributed device 200 of the browsing user corresponding to the second smart medical service big data from recommending the first smart medical service big data, thereby avoiding repeated recommendation of the smart medical service big data to the same user.

In an embodiment, on the basis of the foregoing embodiment, an embodiment of the present disclosure further provides an information push configuration method based on smart medical service big data, where the method may include the following steps.

Step A110, determining effective matching degree of corresponding first hotspots based on multiple hotspot feedback behaviors of each medical intention hotspot in a first medical intention hotspot sequence generated when each intelligent medical recommendation entry is applied by the jump mining intention corresponding to the target intelligent medical service big data, and selecting a preset number of multiple medical intention hotspots from descending results of the effective matching degree of the first hotspots to form a second medical intention hotspot sequence.

Step A120, performing feature extraction on the multiple hot feedback behaviors of each medical intention hot spot in the second medical intention hot spot sequence to obtain multiple hot feedback knowledge point features corresponding to each medical intention hot spot.

Step a130, determining a corresponding second hotspot effective matching degree based on the features of the multiple hotspot feedback knowledge points of each medical intention hotspot in the second medical intention hotspot sequence.

Step a140, based on the descending result of the effective matching degree of the second hotspot of each medical intention hotspot in the second medical intention hotspot sequence, respectively executing policy rule configuration of the information push policy of each medical intention hotspot.

The medical intention hotspot can refer to the matched push hotspot information which is possibly generated when each intelligent medical recommendation entry is applied and the skip excavation intention is generated, and the hotspot feedback behavior can be used for representing the behavior characteristic condition when hotspot feedback is generated. The hotspot effective matching degree can be used for representing the probability that the behavior can be converted to have effective push value in the process of generating hotspot feedback. The hotspot feedback knowledge point feature can be used for representing a feature vector of an exponential feature condition when a hotspot feedback behavior is generated, wherein the feature vector takes a preset graph network as a vector filling unit.

In this embodiment, in step a140, policy rule configuration of the information push policy of each medical intention hotspot may be respectively executed based on a descending result of the effective matching degree of the second hotspot of each medical intention hotspot in the second medical intention hotspot sequence, so that the levels of the information push policy of each medical intention hotspot may be sequentially configured according to a sequence from high to low of the effective matching degree of the second hotspot, and therefore, in a subsequent threat protection process, push may be preferentially enabled by the medical intention hotspot with a higher effective matching degree of the hotspot, so that pertinence and accuracy of the information push policy may be ensured.

Based on the above steps, the embodiment uses the hotspot feedback behavior for determining the effective matching degree of the first hotspot, and reuses the hotspot feedback knowledge point characteristics corresponding to the same characteristics when determining the effective matching degree of the second hotspot, thereby reducing the workload caused by multiple data processing required in the determination process of the effective matching degree of the hotspots at different stages, and by determining the effective matching degree of the hotspots in stages, compared with the method of determining the effective matching degree of the hotspots at one time for a large amount of information, the efficiency is higher, and further determining the effective matching degree of the second hotspot based on the hotspot feedback knowledge point characteristics can further improve the precision of determining the effective matching degree of the hotspots and reduce the complexity of an information push strategy, thereby comprehensively sorting the medical intention hotspots based on the effective matching degree of the first hotspot and the effective matching degree of the second hotspot, and the subsequent hierarchy determination of the information pushing strategy is carried out according to the hierarchy determination, so that the pertinence and the precision of the information pushing strategy are ensured.

In one embodiment, for step a110, the first sequence of medical intent hotspots is obtained by:

and a substep A111, acquiring application recommendation content data of each application service node of the intelligent medical recommendation data source in the application service range, which is analyzed when the intelligent medical recommendation entries are applied by the jump mining intention corresponding to the target intelligent medical service big data.

And a substep A112, determining each application data source target corresponding to the intelligent medical recommended data source according to the application recommended content data of each application service node, and respectively determining target application data source targets which have linkage application with the current application data source target from the application recommended content data of the rest application service nodes for each application data source target.

And a substep A113, performing hotspot updating characteristic search on the current application data source target, and performing hotspot updating characteristic search on the target application data source target to respectively obtain first hotspot updating characteristic search information of the current application data source target and second hotspot updating characteristic search information of the target application data source target.

And a substep a114, generating a hotspot update association characteristic of each current application data source target and the corresponding target application data source target according to the first hotspot update characteristic search information and the second hotspot update characteristic search information.

And a substep A115, analyzing each current application data source target and the corresponding target application data source target respectively according to the hot spot updating correlation characteristics, combining the analyzed source hot spot label information according to a time sequence arrangement mode to obtain a plurality of combined source hot spot label information sequences, and generating the source hot spot classification attribute of the intelligent medical recommendation data source for each combined source hot spot label information sequence based on a random forest tree model.

And a substep A116, acquiring a corresponding medical intention hotspot according to hotspot matching information corresponding to the source hotspot classification attribute, and using the acquired hotspot matching information as a first medical intention hotspot sequence.

In this embodiment, the intelligent medical recommendation data source may be a data source specified by a push rule pre-configured for the jump mining intention 200 corresponding to the target intelligent medical service big data, and the application service range may be flexibly set according to different service requirements, which is not particularly limited. Each application service node may be configured to perform application matching for different application parameters in an application process, and may be flexibly designed according to an actual application environment, which is not specifically limited herein.

In this embodiment, the application data source target may be understood as a data source object that may perform subsequent push information configuration.

In this embodiment, the first hotspot update characteristic search information and the second hotspot update characteristic search information respectively include update log information of respective corresponding hotspot update sources, and the hotspot update sources may respectively serve a plurality of preset update source objects associated with respective corresponding hotspot application services.

Based on the design, each application data source target of each application service node and the target application data source target which has linkage application with the current application data source target are determined, so that the judgment accuracy of the medical intention hotspot behavior of the intelligent medical recommendation data source on the hotspot update source in the application service range can be improved, and the accuracy of the medical intention hotspot identification can be improved.

In one embodiment, still with respect to step a110, the hotspot feedback behavior of each medical intent hotspot can be obtained by the following detailed implementation:

in sub-step a117, for each medical intention hotspot in the first medical intention hotspot sequence, a hotspot feedback behavior corresponding to a hotspot characteristic attribute of the medical intention hotspot is queried from a hotspot feedback behavior data set of the intention hotspot knowledge graph.

Wherein the intended hotspot knowledge graph can be used to determine a first hotspot effective match based on hotspot feedback behavior. For example, a large number of samples of hotspot feedback behaviors and corresponding hotspot effective matching degrees can be collected in advance for model training, so that an intention hotspot knowledge graph can be obtained, and the obtained intention hotspot knowledge graph can have the prediction capability of the hotspot effective matching degrees.

In the substep a118, when the hotspot characteristic attribute of the medical intention hotspot is the hotspot characteristic attribute of the corresponding intention hotspot knowledge graph and the hotspot characteristic attribute is not queried from the hotspot feedback behavior data set of the intention hotspot knowledge graph, converting hotspot operation behavior data of the hotspot characteristic attribute into operation behavior characteristic information, and encoding the operation behavior characteristic information to obtain operation behavior characteristic encoded information.

And a substep A119, encoding the hot spot label information of the hot spot characteristic attribute to obtain hot spot label encoded information, and fusing the hot spot label encoded information and the operation behavior characteristic encoded information to obtain a hot spot feedback behavior of the medical intention hot spot.

As a possible example, for step a115, in the process of generating the source hot point classification attribute of the intelligent medical recommendation data source for each combined source hot point tag information sequence based on the random forest tree model, the following exemplary sub-steps may be implemented, and are described in detail as follows.

And a substep A1151, extracting the feature information of each combined source hotspot tag information sequence based on a random forest tree model, inputting the feature information of the combined source hotspot tag information sequence into a classification layer for classification, and outputting the confidence coefficient of the feature information of the combined source hotspot tag information sequence in the matching information of each hotspot.

And a substep A1152, obtaining the source heat point classification attribute of the intelligent medical recommendation data source according to the confidence degree of the feature information of the combined source heat point label information sequence in each hot point matching information.

For example, if the confidence of the feature information of the combined source hotspot tag information sequence in a certain hotspot matching information a is greater than the set confidence, it indicates that the source hotspot classification attribute is the hotspot matching information a, so as to facilitate the selection of the subsequent information push strategy. For example, an information push policy corresponding to different hotspot matching information may be selected, and a specific configuration process may be configured according to a specific situation of an actual service, which is not limited specifically herein.

Optionally, the random forest tree model may be obtained by training a pre-configured training sample sequence and training hotspot matching information corresponding to each training sample in the training sample sequence based on a deep learning network, where the training sample is a source hotspot tag information sequence, and the specific training mode may refer to a conventional training mode in the prior art, and the training process is not a key point of the embodiment of the disclosure and is not described herein again.

In an embodiment, for step a120, when the hotspot feedback behavior is a single feedback behavior, weight calculation may be performed on the influence weights respectively corresponding to the multiple hotspot feedback behaviors and the corresponding hotspot feedback behavior to obtain multiple hotspot feedback knowledge point features corresponding to each medical intention hotspot.

In this embodiment, the weight calculation mode may be that the fusion processing is performed after multiplying the weight parameters by the weight parameters corresponding to the respective weights. It should be noted that, in the following process of calculating the weight, reference may be made to the above example, and details are not described in the following description.

For another example, in another possible case, when the hotspot feedback behavior is a composite feedback behavior, weight calculation is performed on the influence weights respectively corresponding to the hotspot feedback behaviors and the behavior components of the corresponding hotspot feedback behaviors, and weighting processing is performed on the weight calculation result, so as to obtain a plurality of hotspot feedback knowledge point characteristics corresponding to each medical intention hotspot.

In another embodiment, for step a120, an expected feedback behavior of the medical intention hotspot may be obtained for each medical intention hotspot in the second sequence of medical intention hotspots, and the expected feedback behavior is associated with each generated candidate feedback behavior.

For example, when the hotspot feedback behavior and the expected feedback behavior are single feedback behaviors, the influence weights respectively corresponding to the multiple expected feedback behaviors are subjected to weight calculation with the corresponding expected feedback behaviors, so as to obtain multiple hotspot feedback knowledge point characteristics corresponding to each medical intention hotspot. On the basis, the influence weights respectively corresponding to the plurality of hotspot feedback behaviors and the corresponding hotspot feedback behaviors can be subjected to weight calculation to obtain a plurality of hotspot feedback knowledge point characteristics corresponding to each medical intention hotspot.

For another example, in another possible case, when the hotspot feedback behavior and the expected feedback behavior are composite feedback behaviors, weight calculation is performed on the influence weights respectively corresponding to the multiple expected feedback behaviors and the multiple behavior components corresponding to the expected feedback behaviors, weighting processing is performed on the obtained weight calculation results of the multiple behavior components corresponding to the expected feedback behaviors, so as to obtain multiple hotspot feedback knowledge point features corresponding to each medical intention hotspot, weight calculation is performed on the influence weights respectively corresponding to the multiple hotspot feedback behaviors and the multiple behavior components corresponding to the hotspot feedback behaviors, and weighting processing is performed on the obtained weight calculation results of the multiple behavior components corresponding to the hotspot feedback behaviors, so as to obtain multiple hotspot feedback knowledge point features corresponding to each medical intention hotspot.

In one embodiment, the types of expected feedback behavior for the medical intent hotspot may include single-dimensional vectors and multi-dimensional vectors. Based on this, in the process of acquiring the expected feedback behavior of the medical intention hotspot, the feedback feature vectors of the plurality of feedback units of the generated candidate feedback behavior of the medical intention hotspot can be acquired, and the feedback feature vector of each feedback unit is taken as a single-dimensional vector.

Thus, the multidimensional vector can be obtained by at least one of the following combinations:

for example, the feedback feature vector of the at least one feedback unit of the generated candidate feedback behavior is combined with the feedback feature vector associated with the at least one feedback unit of the medical intent hotspot.

As another example, the feedback feature vector of the at least one feedback unit of the generated candidate feedback behavior is combined with the feedback feature vector of the at least one feedback unit of the environment-dependent.

In one embodiment, step a130 may be implemented by the following exemplary sub-steps, which are described in detail below.

In sub-step a131, based on the multiple hotspot feedback knowledge point features of each medical intention hotspot in the second medical intention hotspot sequence and the incidence relations among the multiple hotspot feedback knowledge point features, a corresponding third hotspot effective matching degree is determined.

And a substep A132, fusing the characteristics of the plurality of hot feedback knowledge points of the medical intention hot points, and fusing the fusion result with the classification attribute of the random forest tree classification network to obtain the classification attribute characteristics of the corresponding medical intention hot points.

And a substep A133, mapping the classification attribute features from the classification attribute feature space to a hot spot effective matching degree space, so as to obtain a fourth hot spot effective matching degree corresponding to the medical intention hot spot.

In the substep a134, the third hotspot effective matching degree and the fourth hotspot effective matching degree of each medical intention hotspot in the second medical intention hotspot sequence are weighted to obtain the corresponding second hotspot effective matching degree.

Exemplarily, in the sub-step a131, it can be realized by the following detailed embodiments, which are described as follows.

(1) Performing the following for each medical intent hotspot in the second sequence of medical intent hotspots:

(2) combining the characteristics of a plurality of hotspot feedback knowledge points of the medical intention hotspots according to at least one of the following modes to obtain corresponding fusion characteristic vectors:

1) and performing weight calculation on the hotspot feedback knowledge point characteristics corresponding to at least two hotspot feedback behaviors, and taking the obtained multiplication result as a corresponding fusion feature vector.

2) And performing weight calculation on the hotspot feedback knowledge point characteristics corresponding to at least one hotspot feedback behavior and the hotspot feedback knowledge point characteristics corresponding to at least one expected feedback behavior, and taking the obtained multiplication result as a corresponding fusion feature vector.

3) And performing weight calculation on the hotspot feedback knowledge point characteristics respectively corresponding to at least two expected feedback behaviors, and taking the obtained weight calculation result as a corresponding fusion feature vector, wherein the hotspot feedback knowledge point characteristics used in each combination are partially or completely different, so as to form a plurality of fusion feature vectors of the medical intention hotspots.

4) Weighting the multiple fusion feature vectors of the medical intention hot spot to obtain first weighting information, and weighting the multiple hot spot feedback behaviors and the multiple expected feedback behaviors by taking hot spot effective matching degree environment influence parameters respectively corresponding to the multiple hot spot feedback behaviors and the multiple expected feedback behaviors as weighting parameters to obtain second weighting information.

5) And taking the first weighted information as the third hotspot effective matching degree of the medical intention hotspot, or carrying out weighted processing on the sum of the first weighted information and the second weighted information, and taking the weighted processing result as the third hotspot effective matching degree of the medical intention hotspot.

Therefore, a series of weighting processing is carried out by combining the expected feedback behaviors to consider the conditions of different dimensions, so that the judgment accuracy of the effective matching degree of the third hotspot can be improved.

Fig. 3 is a schematic functional block diagram of a big data processing apparatus 300 based on smart medical services according to an embodiment of the present disclosure, and the functions of the functional blocks of the big data processing apparatus 300 based on smart medical services are described in detail below.

The first generating module 310 is configured to obtain target smart medical service big data of the smart medical distributed device, determine target query behavior data from the target smart medical service big data, extract query intent objects from the target query behavior data and jump query behavior data corresponding to the target query behavior data, generate query intent components corresponding to the query intent objects, and obtain a target query intent component sequence corresponding to the target query behavior data and a jump query intent component sequence corresponding to the jump query behavior data. The target query intention component sequence comprises query intention components corresponding to all target query intention objects in the target query behavior data respectively, and the jump query intention component sequence comprises query intention components corresponding to all jump query intention objects in the jump query behavior data respectively.

And the second generating module 320 is configured to match the target query intention object with the skip query intention object based on the component correlation between the target query intention object and the skip query intention object, and generate a skip attribute component corresponding to the target query intention object based on the skip correlation between the target query intention object and the skip query intention object that are successfully matched.

The updating module 330 is configured to fuse the jump attribute component and the query intent component corresponding to the same target query intent object to obtain a corresponding fused intent component, and update the target query intent component sequence based on the fused intent component to obtain an updated query intent component sequence.

And a third generating module 340, configured to obtain a skip mining intention corresponding to the target smart medical service big data based on the updated query intention component sequence corresponding to the target query behavior data.

Fig. 4 is a schematic diagram illustrating a hardware structure of a blockchain medical system 100 for implementing the big data processing method based on intelligent medical services according to an embodiment of the present disclosure, and as shown in fig. 4, the blockchain medical system 100 may include a processor 110, a machine-readable storage medium 120, a bus 130, and a communication unit 140.

In a specific implementation process, at least one processor 110 executes computer-executable instructions stored in the machine-readable storage medium 120, so that the processor 110 may execute the big data processing method based on smart medical services according to the above method embodiment, the processor 110, the machine-readable storage medium 120, and the communication unit 140 are connected through the bus 130, and the processor 110 may be configured to control transceiving actions of the communication unit 140, so as to perform data transceiving with the smart medical distributed device 200.

The specific implementation process of the processor 110 can be referred to in the embodiments of the methods executed by the blockchain medical system 100, which have similar implementation principles and technical effects, and the detailed description of the embodiments is omitted here.

In addition, the embodiment of the disclosure also provides a readable storage medium, in which a computer executing instruction is preset, and when a processor executes the computer executing instruction, the big data processing method based on the intelligent medical service is implemented.

Finally, it should be understood that the examples in this specification are only intended to illustrate the principles of the examples in this specification. Other variations are also possible within the scope of this description. Accordingly, by way of example, and not limitation, alternative configurations of the embodiments of the specification can be seen as matching the teachings of the specification. Accordingly, the embodiments of the present description are not limited to only those embodiments explicitly described and depicted herein.

Claims

1. A big data processing method based on intelligent medical service is applied to a block chain medical system, the block chain medical system is in communication connection with a plurality of intelligent medical distributed devices, and the method comprises the following steps:

2. The big data processing method based on intelligent medical service of claim 1, wherein the matching of the target query intention object and the skip query intention object based on the component correlation degree between the target query intention object and the skip query intention object comprises:

calculating component relevance between the query intention components corresponding to the current target query intention object and the query intention components corresponding to the skip query intention objects respectively to obtain a plurality of component relevance corresponding to the current target query intention object;

and when the maximum component correlation degree is smaller than the preset correlation degree, taking the jump query intention object corresponding to the maximum component correlation degree as the jump query intention object matched with the current target query intention object.

3. The big data processing method based on intelligent medical service according to claim 2, wherein the using the skip query intention object corresponding to the maximum component correlation as the skip query intention object matched with the current target query intention object comprises:

and when the jump relevance between the jump query intention object corresponding to the maximum component relevance and the current target query intention object is greater than the target relevance, taking the jump query intention object corresponding to the maximum component relevance as the jump query intention object matched with the current target query intention object.

4. The big data processing method based on intelligent medical services according to claim 1, wherein the jump query behavior data corresponding to the target query behavior data includes at least one of forward query behavior data and backward query behavior data corresponding to the target query behavior data, the jump query intent object corresponding to the forward query behavior data is a forward query intent object, the jump query intent object corresponding to the backward query behavior data is a backward query intent object, and the jump attribute component includes at least one of a forward jump attribute component and a backward jump attribute component;

the step of generating the jump attribute component corresponding to the target query intention object based on the jump relevance of the successfully matched target query intention object and the jump query intention object comprises the following steps:

when the successfully matched jump query intention object is the forward query intention object of the target query intention object, generating a forward jump attribute component corresponding to the target query intention object based on the jump relevance of the successfully matched target query intention object and the jump query intention object;

and when the successfully matched jump query intention object is the backward query intention object of the target query intention object, generating a backward jump attribute component corresponding to the target query intention object based on the jump relevance of the successfully matched target query intention object and the jump query intention object.

5. The big data processing method based on intelligent medical service of claim 4, wherein the current jump attribute component is a forward jump attribute component or a backward jump attribute component, and the current jump attribute component corresponding to the target query intention object is generated based on the jump correlation between the target query intention object and the jump query intention object which are successfully matched, including:

generating initial jump attribute components corresponding to the target query intention objects based on the jump relevance of the successfully matched target query intention objects and the jump query intention objects; the initial jump attribute component carries a jump dimension;

clustering initial jump attribute components corresponding to the same jump dimension to obtain clustering objects corresponding to all the jump dimensions respectively;

counting the number of initial jump attribute components in the same clustering object to obtain the reference number corresponding to each clustering object;

taking the jump dimensionality corresponding to the clustering objects with the maximum reference number as a target jump dimensionality;

and updating each initial jump attribute component based on the target jump dimension adaptability to obtain the current jump attribute component corresponding to each target query intention object.

6. The big data processing method based on intelligent medical service of claim 5, wherein the adaptively updating each initial jump attribute component based on target jump dimension to obtain a current jump attribute component corresponding to each target query intention object comprises:

adaptively updating each initial jump attribute component based on the target jump dimension to obtain a middle jump attribute component corresponding to each target query intention object;

carrying out regularization processing on each intermediate jump attribute component to obtain a current jump attribute component corresponding to each target query intention object;

the regularizing each intermediate jump attribute component to obtain a current jump attribute component corresponding to each target query intention object includes:

acquiring intermediate jump attribute components with jump costs smaller than target costs from the intermediate jump attribute components as key jump attribute components, and randomly assigning the key jump attribute components as random jump attribute components;

and carrying out regularization processing on the random jump attribute component and the intermediate jump attribute component with the jump cost larger than or equal to the target cost to obtain the current jump attribute component corresponding to each target query intention object.

7. The big data processing method based on intelligent medical services, according to any one of claims 1 to 6, characterized in that the method further comprises:

determining a corresponding first hotspot effective matching degree based on a plurality of hotspot feedback behaviors of each medical intention hotspot in a first medical intention hotspot sequence generated when each intelligent medical recommendation entry is applied by the jump mining intention corresponding to the target intelligent medical service big data, and selecting a plurality of medical intention hotspots in a preset number from descending results of the first hotspot effective matching degree to form a second medical intention hotspot sequence;

performing feature extraction on a plurality of hotspot feedback behaviors of each medical intention hotspot in the second medical intention hotspot sequence to obtain a plurality of hotspot feedback knowledge point features corresponding to each medical intention hotspot;

determining a corresponding second hot spot effective matching degree based on a plurality of hot spot feedback knowledge point characteristics of each medical intention hot spot in the second medical intention hot spot sequence;

and respectively executing strategy rule configuration of an information pushing strategy of each medical intention hotspot based on a descending result of the effective matching degree of the second hotspot of each medical intention hotspot in the second medical intention hotspot sequence.

8. The big data processing method based on intelligent medical service as claimed in claim 7, wherein the hot spot feedback behavior of each medical intention hot spot is obtained by:

for each medical intention hotspot in the first medical intention hotspot sequence, inquiring hotspot feedback behaviors corresponding to hotspot characteristic attributes of the medical intention hotspot from a hotspot feedback behavior data set of a corresponding intention hotspot knowledge graph; wherein the intended hotspot knowledge graph is used to determine the first hotspot effective match based on the hotspot feedback behavior;

when the hotspot characteristic attribute of the medical intention hotspot is the hotspot characteristic attribute corresponding to the intention hotspot knowledge graph and the hotspot characteristic attribute is not inquired from a hotspot feedback behavior data set of the intention hotspot knowledge graph, converting hotspot operation behavior data of the hotspot characteristic attribute into operation behavior characteristic information, and coding the operation behavior characteristic information to obtain operation behavior characteristic coding information;

and coding the hotspot tag information of the hotspot characteristic attribute to obtain hotspot tag coding information, and fusing the hotspot tag coding information and the operation behavior characteristic coding information to obtain hotspot feedback behavior of the medical intention hotspot.

9. The big data processing method based on intelligent medical service of claim 1, wherein the first sequence of medical intention hotspots is obtained by:

acquiring application recommendation content data of each application service node of an intelligent medical recommendation data source in an application service range, wherein the application recommendation content data is analyzed when the intelligent medical recommendation entries are applied to the jump mining intents corresponding to the target intelligent medical service big data;

determining each application data source target corresponding to the intelligent medical recommended data source according to the application recommended content data of each application service node, and respectively determining a target application data source target which has linkage application with the current application data source target from the application recommended content data of the rest application service nodes for each application data source target;

performing hotspot updating feature searching on the current application data source target, and performing hotspot updating feature searching on the target application data source target to respectively obtain first hotspot updating feature searching information of the current application data source target and second hotspot updating feature searching information of the target application data source target, wherein the first hotspot updating feature searching information and the second hotspot updating feature searching information respectively comprise updating log information of respective corresponding hotspot updating sources, and the hotspot updating sources respectively serve a plurality of preset updating source objects associated with respective corresponding hotspot application services;

generating hotspot updating correlation characteristics of each current application data source target and the corresponding target application data source target according to the first hotspot updating characteristic searching information and the second hotspot updating characteristic searching information;

analyzing each current application data source target and the corresponding target application data source target respectively according to the hot spot updating correlation characteristics, combining the analyzed source hot spot label information according to a time sequence arrangement mode to obtain a plurality of combined source hot spot label information sequences, and generating a source hot spot classification attribute of the intelligent medical recommended data source for each combined source hot spot label information sequence based on a random forest tree model;

and acquiring a corresponding medical intention hotspot according to hotspot matching information corresponding to the source hotspot classification attribute to serve as the first medical intention hotspot sequence.

10. A blockchain medical system, comprising a processor, a machine-readable storage medium, and a communication unit, wherein the machine-readable storage medium, the communication unit, and the processor are associated with each other through a bus system, the communication unit is configured to be communicatively connected to at least one smart medical distribution device, the machine-readable storage medium is configured to store computer instructions, and the processor is configured to execute the computer instructions in the machine-readable storage medium to perform the smart medical service-based big data processing method according to any one of claims 1 to 9.