An artificial intelligence-based clinical test data management method and system

Abstract

The application belongs to the technical field of data management, and discloses a clinical test data management method and system based on artificial intelligence. The method comprises the following steps: a cloud data center constructs a clinical test data management engine and a clinical test knowledge graph, and deploys a blockchain storage network; a data server uses an OPC UA protocol to upload real-time clinical test data to the cloud data center; the cloud data center uses the clinical test data management engine to perform standardization processing; the cloud data center uses the clinical test data management engine to perform semantic expansion according to the clinical test knowledge graph; the cloud data center uses the clinical test data management engine to perform data analysis; the cloud data center performs encryption; and the cloud data center uses the blockchain storage network to perform data storage. The application solves the problems of great management difficulty, low data security and poor analysis accuracy in the prior art.

Description

Technical Field

[0001] This invention belongs to the field of data management technology, specifically relating to a clinical laboratory data management method and system based on artificial intelligence. Background Technology

[0002] Clinical laboratory testing is an important branch of medicine, primarily involving the laboratory analysis of patients' blood, body fluids, and tissues to aid in disease diagnosis, disease monitoring, and treatment efficacy evaluation. Clinical laboratory data refers to information obtained from samples (such as blood, urine, and tissues) collected from patients through laboratory testing methods, used for disease diagnosis, disease monitoring, and treatment effectiveness assessment.

[0003] Clinical laboratory data is a crucial basis for medical diagnosis, treatment, and disease prevention. With the continuous development of medical technology and the rapid pace of informatization, the volume of clinical laboratory data is growing exponentially. Furthermore, clinical laboratory data involves patients' private information; therefore, its management is becoming increasingly important. However, existing clinical laboratory data management technologies have revealed several shortcomings in practical applications, including:

[0004] 1) High management difficulty: In the existing technology, the formats, units, naming conventions and other aspects of clinical laboratory data from different data sources are often inconsistent, which makes it difficult to integrate and compare data. The lack of a unified data standardization process creates obstacles when exchanging and sharing data, which increases the difficulty of managing clinical laboratory data.

[0005] 2) Low data security: Traditional databases use a centralized storage method, which makes them easy targets for hackers. Once compromised, they may lead to a large amount of data leakage, posing a security risk to centralized storage.

[0006] 3) Poor analytical accuracy: Traditional data analysis methods often rely on human experience or simple statistical tools, which cannot deeply mine and predict data, and lack the ability of deep learning and big data analysis, resulting in inaccurate and incomplete analysis results that are difficult to meet complex clinical needs. Summary of the Invention

[0007] To address the problems of high management difficulty, low data security, and poor analytical accuracy in existing technologies, the present invention aims to provide a clinical laboratory data management method and system based on artificial intelligence.

[0008] The technical solution adopted in this invention is as follows:

[0009] An artificial intelligence-based method for managing clinical laboratory data includes the following steps:

[0010] The cloud data center uses artificial intelligence algorithms to build a clinical laboratory data management engine and a clinical laboratory knowledge graph, and uses blockchain technology to deploy a blockchain storage network;

[0011] The data server collects real-time clinical laboratory data and uploads it to the cloud data center using the OPC UA protocol.

[0012] The cloud data center uses a clinical laboratory data management engine to standardize real-time clinical laboratory data from different data sources to obtain standardized real-time clinical laboratory data.

[0013] In the cloud data center, based on the clinical laboratory knowledge graph, the clinical laboratory data management engine performs semantic expansion on standard real-time clinical laboratory data to obtain expanded real-time clinical laboratory data.

[0014] The cloud data center uses a clinical laboratory data management engine to perform data analysis on extended real-time clinical laboratory data, and obtain corresponding real-time search tags, real-time access permissions, and real-time data analysis results.

[0015] The cloud data center, based on real-time access permissions, uses the corresponding encryption key to encrypt extended real-time clinical test data, obtaining encrypted real-time clinical test data, and manages the encryption key in a controlled manner;

[0016] The cloud data center uses a blockchain storage network to store encrypted real-time clinical test data, corresponding real-time search tags, real-time access permissions, and real-time data analysis results.

[0017] Furthermore, the cloud data center uses artificial intelligence algorithms to build a clinical laboratory data management engine and a clinical laboratory knowledge graph, and uses blockchain technology to deploy a blockchain storage network, including the following steps:

[0018] The cloud data center collects some clinical laboratory knowledge and some historical clinical laboratory data, and performs preprocessing to obtain some preprocessed clinical laboratory knowledge and some preprocessed historical clinical laboratory data.

[0019] Based on some preprocessed clinical laboratory knowledge, a named entity and entity relationship extraction model is constructed using natural language processing algorithms, and a clinical laboratory knowledge graph is obtained.

[0020] Based on several preprocessed historical clinical laboratory data, a standardized processing model was constructed using a fusion algorithm of deep learning and reinforcement learning, and several standard historical clinical laboratory data were generated.

[0021] Based on the clinical laboratory knowledge graph, semantic expansion was performed on several standard historical clinical laboratory data to obtain several extended historical clinical laboratory data.

[0022] Based on several extended historical clinical laboratory data, a data analysis model was constructed using a multi-output deep learning algorithm;

[0023] By integrating named entity and entity relationship extraction models, standardization processing models, and data analysis models, a clinical laboratory data management engine is obtained.

[0024] By distributing and connecting several data nodes in a cloud data center, a blockchain consensus network is constructed. An IPFS system and smart contracts are then set up for the blockchain consensus network to obtain a blockchain storage network.

[0025] Furthermore, based on some preprocessed clinical laboratory knowledge, a named entity and entity relationship extraction model is constructed using natural language processing algorithms, resulting in a clinical laboratory knowledge graph, including the following steps:

[0026] We use natural language processing algorithms to build an initial named entity and entity relationship extraction model;

[0027] Based on some preprocessed clinical laboratory knowledge, the initial named entity and entity relationship extraction model is optimized and trained to obtain the final named entity and entity relationship extraction model, and generate several knowledge named entities and several knowledge entity relationships.

[0028] Based on several named entities and relationships between these entities, a knowledge graph is constructed to obtain a clinical laboratory knowledge graph.

[0029] Furthermore, the named entity and entity relationship extraction model is constructed based on the BERT-CRF-SVM algorithm;

[0030] The standardized processing model is built based on the GCN-MOPPO-SPA algorithm;

[0031] The data analysis model is built based on the GCN-Attention-DBN algorithm.

[0032] Furthermore, the data server collects real-time clinical laboratory data and uploads it to the cloud data center using the OPC UA protocol, including the following steps:

[0033] The data server constructs an OPC UA information model based on historical clinical test data, extracts the model metadata of the OPC UA information model, and sends the model metadata to the cloud data center.

[0034] In the cloud data center, deploy an OPC UA server and build an OPC UA instance in the address space of the OPC UA server based on the model metadata;

[0035] The data server collects raw real-time clinical test data according to the OPC UA information model and uploads the raw real-time clinical test data to the cloud data center using the OPC UA protocol.

[0036] The cloud data center writes the raw real-time clinical test data into the OPC UA instance, and obtains the final real-time clinical test data based on the OPC UA instance.

[0037] Furthermore, the cloud data center uses a clinical laboratory data management engine to standardize real-time clinical laboratory data from different data sources to obtain standardized real-time clinical laboratory data, including the following steps:

[0038] The cloud data center extracts real-time clinical laboratory data from OPC UA instances from different data sources and inputs the real-time clinical laboratory data into the standardized processing model of the clinical laboratory data management engine;

[0039] Using a standardized processing model, the first real-time graph structure features of real-time clinical laboratory data are extracted.

[0040] Based on the structural features of the first real-time graph, a standardization processing strategy is generated to obtain the real-time standardization processing strategy.

[0041] Based on the real-time standardization processing strategy, extract several target SPA algorithms;

[0042] Based on several target SPA algorithms, real-time clinical laboratory data are standardized to obtain corresponding standard real-time clinical laboratory data.

[0043] Furthermore, in the cloud data center, based on the clinical laboratory knowledge graph, the clinical laboratory data management engine performs semantic expansion on the standard real-time clinical laboratory data to obtain expanded real-time clinical laboratory data, including the following steps:

[0044] The cloud data center inputs standard real-time clinical laboratory data into the named entity and entity relationship extraction model of the clinical laboratory data management engine;

[0045] Using a named entity and entity relationship extraction model, we extract real-time semantic features from standard real-time clinical laboratory data.

[0046] Based on real-time semantic features, several corresponding named data entities are obtained;

[0047] Obtain the similarity between each data named entity and several knowledge named entities in the clinical laboratory knowledge graph, and take the knowledge named entity with the highest similarity as the first target knowledge named entity of the data named entity;

[0048] Several knowledge entity relations of the first target knowledge named entity are taken as several target knowledge entity relations of the data named entity, and the knowledge named entity on the other side of the target knowledge entity relation is taken as the second target knowledge named entity of the data named entity.

[0049] Based on the first target knowledge named entity, several target knowledge entity relationships, and several second target knowledge named entities for each data named entity, the standard real-time clinical laboratory data is semantically expanded to obtain expanded real-time clinical laboratory data.

[0050] Furthermore, the cloud data center uses a clinical laboratory data management engine to perform data analysis on the extended real-time clinical laboratory data, obtaining corresponding real-time search tags, real-time access permissions, and real-time data analysis results, including the following steps:

[0051] The cloud data center will expand the data analysis model of the clinical laboratory data management engine by inputting real-time clinical laboratory data;

[0052] Extract the second real-time graph structural features from extended real-time clinical laboratory data;

[0053] Based on the preset first attention weight value, the second real-time graph structure features are weighted and fused to obtain the first real-time weighted fused features;

[0054] Based on the preset second attention weight value, the second real-time graph structural features are weighted and fused to obtain the second real-time weighted fused features;

[0055] Based on the preset third attention weight value, the second real-time graph structural features are weighted and fused to obtain the third real-time weighted fused features;

[0056] Based on the first real-time weighted fusion feature, search tags are generated to obtain the corresponding real-time search tags;

[0057] Based on the second real-time weighted fusion feature, open permissions are generated to obtain the corresponding real-time open permissions;

[0058] Based on the third real-time weighted fusion feature, data analysis is performed to obtain the corresponding real-time data analysis results.

[0059] Furthermore, the cloud data center uses a blockchain storage network to store encrypted real-time clinical laboratory data, corresponding real-time search tags, real-time access permissions, and real-time data analysis results, including the following steps:

[0060] The cloud data center stores encrypted real-time clinical test data in the IPFS blockchain storage network, obtains real-time data hash values, and uses smart contracts to generate real-time storage requests.

[0061] The real-time storage request is sent to the blockchain consensus network of the blockchain storage network, and the data node that receives the real-time storage request is designated as the master node.

[0062] Based on the master node, a consensus algorithm is used to broadcast pre-preparation messages to other data nodes and to verify the legitimacy of real-time storage requests.

[0063] If the validity verification passes, the master node broadcasts a preparation message containing the master node's voting information to other data nodes and writes the preparation message to the message log.

[0064] Based on all data nodes, confirmation messages are exchanged. If the master node receives more than the required number of confirmation messages, the consensus is successful and the real-time consensus timestamp is extracted; otherwise, the consensus fails.

[0065] Once consensus is reached, a smart contract is used to generate real-time transaction data based on the real-time storage request, the real-time consensus timestamp, and the real-time data hash value.

[0066] Using a master node, real-time transaction data is converted into data blocks, and the data blocks are linked onto the blockchain using a blockchain consensus network.

[0067] An artificial intelligence-based clinical laboratory data management system is provided to implement a clinical laboratory data management method. The system includes a cloud data center and several data servers, all of which are communicatively connected to the cloud data center. The cloud data center is equipped with a clinical laboratory data management engine and a clinical laboratory knowledge graph. The cloud data center includes an initialization unit, a standardization processing unit, a semantic expansion unit, a data analysis unit, a data encryption unit, and a data storage unit, which are connected in sequence.

[0068] The beneficial effects of this invention are as follows:

[0069] This invention discloses an artificial intelligence-based clinical laboratory data management method and system. By constructing a unified OPC UA information model, it achieves standardized data transmission between cloud data centers and different data sources, simplifying subsequent standardization processes and improving transmission efficiency. Through a clinical laboratory data management engine, it standardizes data processing, achieving a unified format and standardization for data from different data sources, eliminating data heterogeneity, improving data processing and analysis efficiency, reducing data management difficulty, and offering a high degree of automation. It can automatically complete steps such as data collection, uploading, standardization processing, semantic expansion, data analysis, and storage, reducing manual intervention and improving work efficiency. A blockchain storage network built using blockchain technology is used for data storage, achieving decentralized storage and distributed management, effectively preventing data leakage and unauthorized access, and improving data security. Semantic expansion of clinical laboratory data is performed, combining it with clinical knowledge graphs to enrich data semantic information, improving the depth and breadth of data analysis. Furthermore, the clinical laboratory data management engine performs in-depth mining and prediction of clinical laboratory data, improving the accuracy and comprehensiveness of analysis results, and providing a more reliable basis for subsequent clinical diagnosis, treatment, and prevention.

[0070] Other beneficial effects of the present invention will be further explained in the specific embodiments. Attached Figure Description

[0071] Figure 1 This is a flowchart of the artificial intelligence-based clinical laboratory data management method in this invention.

[0072] Figure 2 This is a structural block diagram of the AI-based clinical laboratory data management system of this invention. Detailed Implementation

[0073] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments.

[0074] Example 1:

[0075] like Figure 1 As shown in the figure, this embodiment provides a clinical laboratory data management method based on artificial intelligence, including the following steps:

[0076] S1: Cloud data center, using artificial intelligence algorithms to build a clinical laboratory data management engine and a clinical laboratory knowledge graph, and using blockchain technology to deploy a blockchain storage network, including the following steps:

[0077] S1-1: Cloud data center, which collects some clinical laboratory knowledge and some historical clinical laboratory data, and performs preprocessing to obtain some preprocessed clinical laboratory knowledge and some preprocessed historical clinical laboratory data.

[0078] In this embodiment, clinical laboratory data is collected via the Open Platform Communications Unified Architecture (OPC UA) protocol. Therefore, the clinical laboratory data itself is a graph structure data. The data includes: patient information (name, age, gender, medical record number, etc.); laboratory tests (e.g., complete blood count, urinalysis, biochemical indicators, etc.); test results (specific numerical or qualitative results, such as white blood cell count, blood glucose level, positive / negative, etc.); timestamps (the specific time of data collection); device information (device number, model, etc.); and quality information (the quality status of the data, such as valid, invalid, requiring retesting, etc.). The graph structure includes nodes and edges: patient nodes (representing the patient and containing basic patient information); and laboratory test node (representing the specific laboratory test). Examples include: a complete blood count; a test result node representing a specific test result, such as a white blood cell count; a device node representing the data acquisition device, containing device information; a time node representing the time point of data acquisition; a quality node representing the quality status of the data; alarm and event nodes representing abnormal alarms and events; and edge types such as: patient-test item edge indicating which tests the patient underwent; test item-test result edge indicating the specific result corresponding to a certain test item; device-test result edge indicating the test result acquired by a certain device; time-test result edge indicating the time of test result acquisition; and quality-test result edge indicating the quality status of the test result.

[0079] Preprocessing includes data cleaning and error filtering of raw data to improve data quality and provide data support for subsequent model building.

[0080] S1-2: Based on some preprocessed clinical laboratory knowledge, a named entity and entity relationship extraction model is constructed using natural language processing algorithms to obtain a clinical laboratory knowledge graph, including the following steps:

[0081] S1-2-1: Use natural language processing algorithms to construct an initial named entity and entity relationship extraction model;

[0082] The named entity and entity relationship extraction model is built on the Bidirectional Encoder Representations from Transformers (BERT)-Conditional Random Fields (CRF)-Support Vector Machine (SVM) algorithm. The named entity and entity relationship extraction model includes a semantic feature extraction module based on the BERT algorithm, a named entity extraction module based on the CRF algorithm, and an entity relationship extraction module based on the SVM algorithm, which are connected in sequence.

[0083] The BERT module can capture deep semantic information in knowledge text, which is very useful for identifying different types of named entities (such as clinical laboratory terms, names of clinical laboratory components, etc.). The CRF module can consider the dependencies between adjacent named entity labels, which can help the model learn the sequence dependencies of entity labels, thereby improving the accuracy of named entity labeling and realizing the extraction of named entities. The entity relationship extraction module is mainly used to classify the relationships between the extracted named entity pairs, determine whether there are specific relationships between them, and the type of relationship. It converts named entities and their contextual information into high-dimensional feature vectors. These vectors can effectively represent the relationships between entities. By combining the deep semantic information extracted by BERT and the sequence dependencies considered by CRF, it can handle complex entity relationships more effectively.

[0084] S1-2-2: Based on some preprocessed clinical laboratory knowledge, the initial named entity and entity relationship extraction model is optimized and trained to obtain the final named entity and entity relationship extraction model, and generate some knowledge named entities and some knowledge entity relationships.

[0085] S1-2-3: Based on several knowledge-named entities and several knowledge-entity relationships, a knowledge graph is constructed to obtain a clinical laboratory knowledge graph;

[0086] S1-3: Based on several preprocessed historical clinical laboratory data, a standardized processing model is constructed using a fusion algorithm of deep learning and reinforcement learning, and several standard historical clinical laboratory data are generated.

[0087] The standardized processing model is built on the Graph Convolutional Network (GCN) - Multi-Objective Proximal Policy Optimization (MOPPO) - Standardized Processing Algorithm (SPA) algorithm. The standardized processing model includes a first graph structure feature extraction module built on the GCN algorithm, a standardized processing strategy generation module built on the MOPPO algorithm, and a standardized processing algorithm storing several SPA algorithms, which are connected in sequence. The standardized processing strategy generation module includes a set of objective functions, an experience replay pool, an Actor network, a Critic network, and an agent. The agent is connected to the set of objective functions, the experience replay pool, the Actor network, and the Critic network, respectively.

[0088] The first graph structure feature extraction module extracts graph structure features from clinical laboratory data, transforming the relationships between nodes into feature representations usable for model learning. This module utilizes graph convolutional networks to capture complex dependencies between nodes, enhancing feature expressiveness, effectively leveraging graph structure information, and improving the model's understanding of complex relationships. This provides rich input information for the standardized processing strategy generation module. The Actor network in the standardized processing strategy generation module is responsible for outputting the probability distribution of actions to be taken in a given state. Its goal is to learn an optimal policy, maximizing long-term cumulative reward. In the continuous action space, the Actor network typically outputs a mean and an optional variance parameter to describe the probability distribution of actions. The Critic network is responsible for evaluating the value of a given state, i.e., predicting the period that can be obtained by starting from that state and following the current policy. The expected return is typically a scalar value representing the value of a state or a state-action value. The experience replay pool stores historical experience for reuse during training. The objective function set includes functions defining multiple standardization objectives, such as minimizing standardization time cost, minimizing standardization computational resource cost, and maximizing standardization efficiency. It can generate standardization strategies based on graph structure characteristics, guiding how to standardize data. These strategies include several SPA algorithm invocation decisions. The standardization algorithm itself stores several SPA algorithms to receive invocation decisions and standardize clinical laboratory data. SPA algorithms include normalization algorithms, format conversion algorithms, sequence conversion algorithms, regularization algorithms, and discretization algorithms, aiming to transform the raw data into a standard format for more efficient processing by subsequent models.

[0089] S1-4: Based on the clinical laboratory knowledge graph, semantic expansion is performed on several standard historical clinical laboratory data to obtain several extended historical clinical laboratory data.

[0090] S1-5: Integrate the named entity and entity relationship extraction model, the standardization processing model, and the data analysis model to obtain the clinical laboratory data management engine;

[0091] S1-6: Based on several extended historical clinical laboratory data, a data analysis model is constructed using a multi-output deep learning algorithm;

[0092] The data analysis model is built based on the GCN-Attention-Deep Belief Network (DBN) algorithm. The data analysis model includes a second graph structure feature extraction module built based on the GCN algorithm, an attention weight module built based on the attention mechanism, a retrieval label generation module built based on the DBN algorithm, an open permission generation module built based on the DBN algorithm, and a data analysis module built based on the DBN algorithm. The second graph structure feature extraction module is connected to the attention weight module. The attention weight module is equipped with a first weighted fusion channel, a second weighted fusion channel, and a third weighted fusion channel connected in parallel. The first weighted fusion channel is connected to the retrieval label generation module, the second weighted fusion channel is connected to the open permission generation module, and the third weighted fusion channel is connected to the data analysis module.

[0093] The second graph structure feature extraction module extracts graph structure features from extended clinical laboratory data, using GCN to capture complex relationships between nodes, effectively utilizing graph structure information to enhance feature expressiveness, and providing rich, structured feature representations for subsequent modules, thereby improving the model's ability to understand and analyze complex clinical data. The attention weighting module assigns attention weights to different features, highlighting important information. Through multiple parallel weighted fusion channels, it achieves multi-angle feature fusion, improving the model's focus on key information, enhancing analytical accuracy, and increasing the model's flexibility and adaptability. Multi-channel fusion enriches feature representations and improves model performance. The retrieval label generation module uses a deep belief network to generate retrieval labels. This module is used for quickly locating and retrieving relevant data, improving the efficiency and accuracy of data retrieval, providing fast and accurate information retrieval services, and automatically discovering potential patterns and relationships in the data. The open access generation module generates data open permissions to ensure data security and privacy protection, enabling fine-grained access control, ensuring data security, improving the flexibility and operability of data sharing, and reducing management costs and human error through automated permission generation. The data analysis module performs in-depth analysis on data after feature extraction and weighted fusion, generating valuable data analysis results such as trend prediction and anomaly detection, providing comprehensive and in-depth data analysis services, providing scientific and accurate basis for clinical decision-making, and uncovering potential within the data.

[0094] S1-7: Distribute and connect several data nodes in the cloud data center to build a blockchain consensus network, and set up the InterPlanetary File System (IPFS) and smart contracts for the blockchain consensus network to obtain the blockchain storage network;

[0095] S2: Data server, which collects real-time clinical laboratory data and uploads it to the cloud data center using the OPC UA protocol, including the following steps:

[0096] S2-1: Data server, which constructs an OPC UA information model based on historical clinical test data, extracts the model metadata of the OPC UA information model, and sends the model metadata to the cloud data center;

[0097] S2-2: Cloud data center, deploy OPC UA server, and build OPC UA instance in the address space of OPC UA server according to model metadata;

[0098] S2-3: Data server, which collects raw real-time clinical test data according to the OPC UA information model and uploads the raw real-time clinical test data to the cloud data center using the OPC UA protocol;

[0099] S2-4: Cloud data center, writes the raw real-time clinical test data to the OPC UA instance, and obtains the final real-time clinical test data based on the OPC UA instance;

[0100] S3: Cloud Data Center, using a clinical laboratory data management engine, standardizes real-time clinical laboratory data from different data sources to obtain standardized real-time clinical laboratory data, including the following steps:

[0101] S3-1: Cloud data center, extracts real-time clinical laboratory data from OPC UA instances from different data sources, and inputs the real-time clinical laboratory data into the standardized processing model of the clinical laboratory data management engine;

[0102] S3-2: The first graph structure feature extraction module using the standardized processing model extracts the first real-time graph structure features of real-time clinical laboratory data;

[0103] S3-3: The standardization processing strategy generation module using the standardization processing model generates a standardization processing strategy based on the structural features of the first real-time graph, thus obtaining a real-time standardization processing strategy.

[0104] S3-4: Based on the real-time standardization processing strategy, extract several target SPA algorithms from the standardization processing algorithm library of the standardization processing model.

[0105] S3-5: Based on several target SPA algorithms, standardize the real-time clinical laboratory data to obtain the corresponding standard real-time clinical laboratory data;

[0106] S4: Cloud Data Center. Based on the clinical laboratory knowledge graph, the clinical laboratory data management engine performs semantic expansion on standard real-time clinical laboratory data to obtain expanded real-time clinical laboratory data. This includes the following steps:

[0107] S4-1: Cloud Data Center, which inputs standard real-time clinical laboratory data into the named entity and entity relationship extraction model of the clinical laboratory data management engine;

[0108] S4-2: Semantic feature extraction module using named entity and entity relationship extraction model to extract real-time semantic features of standard real-time clinical test data;

[0109] S4-3: The named entity extraction module uses the named entity and entity relationship extraction model to obtain several corresponding data named entities based on real-time semantic features;

[0110] S4-4: Obtain the similarity between each data named entity and several knowledge named entities in the clinical laboratory knowledge graph, and take the knowledge named entity with the highest similarity as the first target knowledge named entity of the data named entity.

[0111] S4-5: Take several knowledge entity relations of the first target knowledge named entity as several target knowledge entity relations of the data named entity, and take the knowledge named entity on the other side of the target knowledge entity relation as the second target knowledge named entity of the data named entity.

[0112] S4-6: Based on the first target knowledge named entity, several target knowledge entity relationships, and several second target knowledge named entities of each data named entity, semantic expansion is performed on the standard real-time clinical laboratory data to obtain expanded real-time clinical laboratory data.

[0113] S5: Cloud Data Center, using a clinical laboratory data management engine, performs data analysis on extended real-time clinical laboratory data to obtain corresponding real-time search tags, real-time access permissions, and real-time data analysis results, including the following steps:

[0114] S5-1: Cloud Data Center, which will expand the data analysis model of the clinical laboratory data management engine by inputting real-time clinical laboratory data;

[0115] S5-2: The second graph structure feature extraction module using the data analysis model extracts the second real-time graph structure features of the extended real-time clinical laboratory data;

[0116] S5-3: Based on the preset first attention weight value, the first weighted fusion channel of the attention weight module of the data analysis model is used to perform weighted fusion on the second real-time graph structure features to obtain the first real-time weighted fusion features.

[0117] S5-4: Based on the preset second attention weight value, the second weighted fusion channel of the attention weight module of the data analysis model is used to perform weighted fusion on the second real-time graph structure features to obtain the second real-time weighted fusion features.

[0118] S5-5: Based on the preset third attention weight value, the third weighted fusion channel of the attention weight module of the data analysis model is used to perform weighted fusion on the second real-time graph structure features to obtain the third real-time weighted fusion features.

[0119] S5-6: The retrieval tag generation module using the data analysis model generates retrieval tags based on the first real-time weighted fusion features, thus obtaining the corresponding real-time retrieval tags;

[0120] S5-7: The open permission generation module of the data analysis model generates open permissions based on the second real-time weighted fusion feature to obtain the corresponding real-time open permissions;

[0121] S5-8: The data analysis module using the data analysis model performs data analysis based on the third real-time weighted fusion feature to obtain the corresponding real-time data analysis results;

[0122] Real-time data analysis results include real-time trend prediction results of test indicators (predicting the future trend of specific test indicators, such as blood glucose, cholesterol, etc.), real-time abnormal value prediction results of clinical test data indicators (such as abnormal white blood cell count indicators, etc.), and real-time clinical test data classification levels (such as disease risk level, severity, etc.).

[0123] S6: The cloud data center, based on real-time access permissions, uses the corresponding encryption key to encrypt extended real-time clinical laboratory data, obtaining encrypted real-time clinical laboratory data, and manages the encryption key in a controlled manner, including the following steps:

[0124] S6-1: Cloud Data Center, set encryption keys with different preset open permissions. The encryption keys with different preset open permissions have different levels of complexity. The higher the level of preset open permission, the more complex the encryption key.

[0125] S6-2: Based on the real-time access permissions, match among several preset encryption keys to obtain the corresponding preset access permission encryption key;

[0126] S6-3: Use the corresponding encryption key to encrypt the extended real-time clinical laboratory data to obtain encrypted real-time clinical laboratory data;

[0127] S6-4: Store encryption keys with different preset open permissions in a controlled database in the cloud data center for controlled management. Only users with greater permissions than the real-time open permissions can call the corresponding preset open permission encryption key.

[0128] S7: Cloud Data Center, using a blockchain storage network, stores encrypted real-time clinical laboratory data, corresponding real-time search tags, real-time access permissions, and real-time data analysis results, including the following steps:

[0129] S7-1: Cloud Data Center, which stores encrypted real-time clinical test data in the IPFS blockchain storage network, obtains real-time data hash values, and uses smart contracts to generate real-time storage requests;

[0130] S7-2: The blockchain consensus network that sends the real-time storage request to the blockchain storage network will designate the data node that receives the real-time storage request as the master node;

[0131] S7-3: Based on the master node, using a consensus algorithm, broadcasts pre-preparation messages to other data nodes and verifies the legitimacy of real-time storage requests;

[0132] S7-4: If the validity verification passes, the master node broadcasts a preparation message containing the master node's voting information to other data nodes and writes the preparation message to the message log.

[0133] S7-5: Based on all data nodes, exchange confirmation messages. If the master node receives more than the required number of confirmation messages, the consensus is successful and the real-time consensus timestamp is extracted; otherwise, the consensus fails.

[0134] S7-6: After successful consensus, use smart contracts to generate real-time transaction data based on real-time storage requests, real-time consensus timestamps, and real-time data hash values;

[0135] S7-7: Using a master node, real-time transaction data is converted into data blocks, and the data blocks are linked to the blockchain consensus network.

[0136] Example 2:

[0137] like Figure 2 As shown, this embodiment provides an artificial intelligence-based clinical laboratory data management system for implementing a clinical laboratory data management method. The system includes a cloud data center and several data servers, all of which are communicatively connected to the cloud data center. The cloud data center is equipped with a clinical laboratory data management engine and a clinical laboratory knowledge graph. The cloud data center includes an initialization unit, a standardization processing unit, a semantic extension unit, a data analysis unit, a data encryption unit, and a data storage unit connected in sequence.

[0138] The data server is used to collect real-time clinical test data and upload the real-time clinical test data to the cloud data center using the OPC UA protocol.

[0139] The initialization unit is used to build a clinical laboratory data management engine and a clinical laboratory knowledge graph using artificial intelligence algorithms, and to deploy a blockchain storage network using blockchain technology.

[0140] The standardization processing unit is used to standardize real-time clinical laboratory data from different data sources using the clinical laboratory data management engine to obtain standard real-time clinical laboratory data.

[0141] The semantic extension unit is used to semantically extend standard real-time clinical laboratory data based on the clinical laboratory knowledge graph and the clinical laboratory data management engine to obtain extended real-time clinical laboratory data.

[0142] The data analysis unit is used to perform data analysis on extended real-time clinical laboratory data using the clinical laboratory data management engine, and obtain corresponding real-time search tags, real-time access permissions, and real-time data analysis results.

[0143] The data encryption unit is used to encrypt extended real-time clinical test data using the corresponding encryption key according to the real-time access permissions, so as to obtain encrypted real-time clinical test data, and to manage the encryption key in a controlled manner.

[0144] The data storage unit is used to store encrypted real-time clinical test data, corresponding real-time search tags, real-time access permissions, and real-time data analysis results using a blockchain storage network.

[0145] This invention discloses an artificial intelligence-based clinical laboratory data management method and system. By constructing a unified OPC UA information model, it achieves standardized data transmission between cloud data centers and different data sources, simplifying subsequent standardization processes and improving transmission efficiency. Through a clinical laboratory data management engine, it standardizes data processing, achieving a unified format and standardization for data from different data sources, eliminating data heterogeneity, improving data processing and analysis efficiency, reducing data management difficulty, and offering a high degree of automation. It can automatically complete steps such as data collection, uploading, standardization processing, semantic expansion, data analysis, and storage, reducing manual intervention and improving work efficiency. A blockchain storage network built using blockchain technology is used for data storage, achieving decentralized storage and distributed management, effectively preventing data leakage and unauthorized access, and improving data security. Semantic expansion of clinical laboratory data is performed, combining it with clinical knowledge graphs to enrich data semantic information, improving the depth and breadth of data analysis. Furthermore, the clinical laboratory data management engine performs in-depth mining and prediction of clinical laboratory data, improving the accuracy and comprehensiveness of analysis results, and providing a more reliable basis for subsequent clinical diagnosis, treatment, and prevention.

[0146] This invention is not limited to the optional embodiments described above, and anyone can derive other various forms of products based on the teachings of this invention. The specific embodiments described above should not be construed as limiting the scope of protection of this invention; the scope of protection of this invention should be determined by the claims, and the specification can be used to interpret the claims.

Claims

1. A clinical laboratory data management method based on artificial intelligence, characterized in that: Includes the following steps: The cloud data center uses artificial intelligence algorithms to build a clinical laboratory data management engine and a clinical laboratory knowledge graph, and uses blockchain technology to deploy a blockchain storage network, including the following steps: The cloud data center collects some clinical laboratory knowledge and some historical clinical laboratory data, and performs preprocessing to obtain some preprocessed clinical laboratory knowledge and some preprocessed historical clinical laboratory data. Based on some preprocessed clinical laboratory knowledge, a named entity and entity relationship extraction model is constructed using natural language processing algorithms, and a clinical laboratory knowledge graph is obtained. The named entity and entity relationship extraction model described above is constructed based on the BERT-CRF-SVM algorithm; Based on several preprocessed historical clinical laboratory data, a standardized processing model was constructed using a fusion algorithm of deep learning and reinforcement learning, and several standard historical clinical laboratory data were generated. The standardized processing model described above is constructed based on the GCN-MOPPO-SPA algorithm; Based on the clinical laboratory knowledge graph, semantic expansion was performed on several standard historical clinical laboratory data to obtain several extended historical clinical laboratory data. Based on several extended historical clinical laboratory data, a data analysis model was constructed using a multi-output deep learning algorithm; The data analysis model described above is built based on the GCN-Attention-DBN algorithm; By integrating named entity and entity relationship extraction models, standardization processing models, and data analysis models, a clinical laboratory data management engine is obtained. By distributing and connecting several data nodes in a cloud data center, a blockchain consensus network is constructed, and an IPFS system and smart contracts are set up for the blockchain consensus network to obtain a blockchain storage network. The data server collects real-time clinical laboratory data and uploads it to the cloud data center using the OPC UA protocol. The cloud data center uses a clinical laboratory data management engine to standardize real-time clinical laboratory data from different data sources, obtaining standardized real-time clinical laboratory data. This process includes the following steps: The cloud data center extracts real-time clinical laboratory data from OPC UA instances from different data sources and inputs the real-time clinical laboratory data into the standardized processing model of the clinical laboratory data management engine; Using a standardized processing model, the first real-time graph structure features of real-time clinical laboratory data are extracted. Based on the structural features of the first real-time graph, a standardization processing strategy is generated to obtain the real-time standardization processing strategy. Based on the real-time standardization processing strategy, extract several target SPA algorithms; Based on several target SPA algorithms, real-time clinical laboratory data is standardized to obtain corresponding standard real-time clinical laboratory data. In the cloud data center, based on the clinical laboratory knowledge graph, a clinical laboratory data management engine is used to semantically expand standard real-time clinical laboratory data to obtain expanded real-time clinical laboratory data. This process includes the following steps: The cloud data center inputs standard real-time clinical laboratory data into the named entity and entity relationship extraction model of the clinical laboratory data management engine; Using a named entity and entity relationship extraction model, we extract real-time semantic features from standard real-time clinical laboratory data. Based on real-time semantic features, several corresponding named data entities are obtained; Obtain the similarity between each data named entity and several knowledge named entities in the clinical laboratory knowledge graph, and take the knowledge named entity with the highest similarity as the first target knowledge named entity of the data named entity; Several knowledge entity relations of the first target knowledge named entity are taken as several target knowledge entity relations of the data named entity, and the knowledge named entity on the other side of the target knowledge entity relation is taken as the second target knowledge named entity of the data named entity. Based on the first target knowledge named entity, several target knowledge entity relationships, and several second target knowledge named entities of each data named entity, the standard real-time clinical laboratory data is semantically expanded to obtain expanded real-time clinical laboratory data. The cloud data center uses a clinical laboratory data management engine to perform data analysis on extended real-time clinical laboratory data, and obtain corresponding real-time search tags, real-time access permissions, and real-time data analysis results. The cloud data center, based on real-time access permissions, uses the corresponding encryption key to encrypt extended real-time clinical test data, obtaining encrypted real-time clinical test data, and manages the encryption key in a controlled manner. The cloud data center uses a blockchain storage network to store encrypted real-time clinical test data, corresponding real-time search tags, real-time access permissions, and real-time data analysis results.

2. The method for managing clinical laboratory data based on artificial intelligence according to claim 1, characterized in that: Based on some preprocessed clinical laboratory knowledge, a named entity and entity relationship extraction model is constructed using natural language processing algorithms, resulting in a clinical laboratory knowledge graph. The process includes the following steps: We use natural language processing algorithms to build an initial named entity and entity relationship extraction model; Based on some preprocessed clinical laboratory knowledge, the initial named entity and entity relationship extraction model is optimized and trained to obtain the final named entity and entity relationship extraction model, and generate a number of knowledge named entities and a number of knowledge entity relationships. Based on several named entities and relationships between these entities, a knowledge graph is constructed to obtain a clinical laboratory knowledge graph.

3. The method for managing clinical laboratory data based on artificial intelligence according to claim 2, characterized in that: The data server collects real-time clinical laboratory data and uploads it to the cloud data center using the OPC UA protocol, including the following steps: The data server constructs an OPC UA information model based on historical clinical test data, extracts the model metadata of the OPC UA information model, and sends the model metadata to the cloud data center. In the cloud data center, deploy an OPC UA server and build an OPC UA instance in the address space of the OPC UA server based on the model metadata; The data server collects raw real-time clinical test data according to the OPC UA information model and uploads the raw real-time clinical test data to the cloud data center using the OPC UA protocol. The cloud data center writes the raw real-time clinical test data into the OPC UA instance, and obtains the final real-time clinical test data based on the OPC UA instance.

4. The method for managing clinical laboratory data based on artificial intelligence according to claim 3, characterized in that: The cloud data center uses a clinical laboratory data management engine to perform data analysis on extended real-time clinical laboratory data, obtaining corresponding real-time search tags, real-time access permissions, and real-time data analysis results, including the following steps: The cloud data center will expand the data analysis model of the clinical laboratory data management engine by inputting real-time clinical laboratory data; Using a data analysis model, we extract the second real-time graph structure features of extended real-time clinical laboratory data; Based on the preset first attention weight value, the second real-time graph structure features are weighted and fused to obtain the first real-time weighted fused features; Based on the preset second attention weight value, the second real-time graph structure features are weighted and fused to obtain the second real-time weighted fused features; Based on the preset third attention weight value, the second real-time graph structural features are weighted and fused to obtain the third real-time weighted fused features; Based on the first real-time weighted fusion feature, search tags are generated to obtain the corresponding real-time search tags; Based on the second real-time weighted fusion feature, open permissions are generated to obtain the corresponding real-time open permissions; Based on the third real-time weighted fusion feature, data analysis is performed to obtain the corresponding real-time data analysis results.

5. The clinical laboratory data management method based on artificial intelligence according to claim 4, characterized in that: The cloud data center uses a blockchain storage network to store encrypted real-time clinical laboratory data, corresponding real-time search tags, real-time access permissions, and real-time data analysis results, including the following steps: The cloud data center stores encrypted real-time clinical test data in the IPFS blockchain storage network, obtains real-time data hash values, and uses smart contracts to generate real-time storage requests. The real-time storage request is sent to the blockchain consensus network of the blockchain storage network, and the data node that receives the real-time storage request is designated as the master node. Based on the master node, a consensus algorithm is used to broadcast pre-preparation messages to other data nodes and to verify the legitimacy of real-time storage requests. If the validity verification passes, the master node broadcasts a preparation message containing the master node's voting information to other data nodes and writes the preparation message to the message log. Based on all data nodes, confirmation messages are exchanged. If the master node receives more than the required number of confirmation messages, the consensus is successful and the real-time consensus timestamp is extracted; otherwise, the consensus fails. Once consensus is reached, a smart contract is used to generate real-time transaction data based on the real-time storage request, the real-time consensus timestamp, and the real-time data hash value. Using a master node, real-time transaction data is converted into data blocks, and the data blocks are linked onto the blockchain using a blockchain consensus network.

6. An artificial intelligence-based clinical laboratory data management system, used to implement the clinical laboratory data management method as described in any one of claims 1-5, characterized in that: The system includes a cloud data center and several data servers, all of which are communicatively connected to the cloud data center. The cloud data center is equipped with a clinical laboratory data management engine and a clinical laboratory knowledge graph, and includes an initialization unit, a standardization processing unit, a semantic extension unit, a data analysis unit, a data encryption unit, and a data storage unit connected in sequence.

Images