Medical service terminal data processing method based on cloud computing

Abstract

The invention discloses a medical service terminal data processing method based on cloud computing, and relates to the technical field of data processing, and the method comprises the steps: collecting the electronic medical record data of a patient from each medical device; integrating the collected data according to a preset data standard; transmitting the processed medical record data set to a cloud computing platform by using a private network; performing data storage on the medical record data set of each patient based on a distributed file system technology; selecting a corresponding storage strategy based on the data type and the access frequency; performing deep mining on the medical record data set of each patient by using an association rule mining algorithm; constructing a decision support model by using a machine learning algorithm; and regularly monitoring and evaluating the data processing performance of the cloud computing platform. The medical record data set of each patient is deeply mined through an association rule mining algorithm, so that a scientific basis is provided for diagnosis and treatment decisions of clinicians, and the pertinence and effectiveness of diagnosis and treatment can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and more particularly to a method for processing medical service terminal data based on cloud computing. Background Art

[0002] Cloud computing technology provides a new data storage and processing solution for medical institutions with its powerful storage capacity, high reliability and fault tolerance, convenient data access and sharing methods, and strong data processing capabilities. The method for processing medical service terminal data based on cloud computing combines cloud computing technology with medical services to achieve efficient, secure, and reliable management of medical data, and further improve the quality and efficiency of medical services.

[0003] With the continuous progress of medical technology and the wide application of medical devices, medical institutions generate a large amount of electronic medical record data every day, including key information such as patients' physiological parameters, medical images, and diagnostic records. The traditional local storage and processing methods can no longer meet the growing data volume of medical institutions and the demand for efficient data utilization. Summary of the Invention

[0004] To solve the above technical problems, a method for processing medical service terminal data based on cloud computing is provided, and the present technical solution solves the problems raised in the above background art.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] A method for processing medical service terminal data based on cloud computing, comprising:

[0007] Collecting patients' electronic medical record data from various medical devices, where the electronic medical record data includes physiological parameters, medical images, and diagnostic record data;

[0008] Integrating the collected data according to a preset data standard to form a medical record data set, removing duplicate data in the medical record data set, and processing missing values in the medical record data set;

[0009] Transmitting the processed medical record data set to a cloud computing platform using a dedicated network;

[0010] In the cloud computing platform, storing the medical record data sets of each patient based on the distributed file system technology;

[0011] Selecting a corresponding storage strategy based on the type and access frequency of the data, where the storage strategy includes hot data storage and cold data storage;

[0012] Using the association rule mining algorithm to deeply mine the medical record data sets of each patient, and extracting disease prediction, risk assessment, and treatment effect assessment information;

[0013] Based on the results of in-depth mining and analysis, a decision support model is constructed using machine learning algorithms to provide personalized diagnosis and treatment suggestions and medication guidance for clinicians;

[0014] Regularly monitor and evaluate the data processing performance of the cloud computing platform, and optimize and upgrade the cloud computing platform based on the evaluation results.

[0015] Preferably, the integration of the collected data according to the preset data standard to form a medical record data set, removing duplicate data in the medical record data set, and processing missing values in the medical record data set specifically includes:

[0016] Based on the norms and standards of the medical industry, formulate data standards, and the data standards include data naming rules, storage formats, and coding methods;

[0017] Based on the preset data standard, integrate the collected data to form a medical record data set;

[0018] Perform a mapping operation on each data in the medical record data set to establish an association relationship between the data in the medical record data set;

[0019] Select the patient's visit number as the unique identifier, and use a data comparison tool to identify duplicate data in the medical record data set;

[0020] Remove duplicate data in the medical record data set, and only retain the data identified by the visit number of the same patient;

[0021] Traverse the medical record data set to identify fields and records with missing values;

[0022] Perform linear interpolation on the data with missing values by calculating the average value of adjacent data points of the missing values.

[0023] Preferably, in the cloud computing platform, the data storage of the medical record data set of each patient based on the distributed file system technology specifically includes:

[0024] Based on the data capacity of the medical service terminal, configure the number of servers, storage capacity, and network bandwidth of the server cluster of the distributed file system;

[0025] Set up the metadata server and data nodes of the distributed file system;

[0026] Chunk each data point in the medical record data set according to the data type;

[0027] Obtain the ASCII code comparison table, and based on the ASCII code comparison table, obtain the ASCII codes of the characters in each data block;

[0028] Based on the ASCII codes of the characters in each data block, use the hash value calculation formula to perform hash calculation on each data block to generate the hash value of each data block;

[0029] Use the data transfer protocol to upload the chunked data to the data nodes of the distributed file system;

[0030] Use the hash values of each data block to establish an index for the stored data blocks;

[0031] The hash value calculation formula is:

[0032]

[0033] In the formula, H is the hash value of each data block, a(i) is the ASCII code of the i-th character in the data block, and n is the length of the string in the data block.

[0034] Preferably, the selecting the corresponding storage strategy based on the type and access frequency of the data specifically includes:

[0035] Analyze and calculate the access frequency of each data type;

[0036] Judge whether the access frequency of this data type is higher than the preset frequency threshold. If so, select the hot data storage method to store the data of this data type. If not, select the cold data storage method to store the data of this data type;

[0037] Use solid-state drives as the hardware for hot data storage;

[0038] Configure the hot data storage parameters, and the hot data storage parameters include data block size, cache strategy, and replication times;

[0039] Use traditional hard disks as the hardware for cold data storage;

[0040] Configure the cold data storage parameters, and the cold data storage parameters include data compression, archiving strategy, and deduplication times.

[0041] Preferably, the using the association rule mining algorithm to deeply mine the medical record data sets of each patient and extract disease prediction, risk assessment, and treatment effect assessment information specifically includes:

[0042] S101: Set the support threshold and confidence threshold based on business requirements, and use the support threshold to screen the frequent item sets in the medical record data sets of each patient;

[0043] S102: Generate at least one association rule from each frequent item set as a candidate item set, and the association rules include the association rules between diseases and symptoms, diseases and medications, and diseases and treatment effects;

[0044] S103: Scan the database and calculate the support of each candidate itemset using the support calculation formula;

[0045] S104: Determine whether there is an itemset with a support greater than or equal to the support threshold. If so, retain the itemset with a support greater than or equal to the support threshold as the frequent itemset and return to step S102. If not, end the iteration;

[0046] S105: For each frequent itemset, generate all disease-related datasets as the premise of the rule, and symptom-related datasets, medication-related datasets, and treatment effect-related datasets as the results of the rule to form association rules;

[0047] S106: Calculate the confidence of each association rule using the confidence calculation method;

[0048] S107: Determine whether the confidence of each association rule is greater than the confidence threshold. If so, output it as a valuable association rule. If not, do not output it;

[0049] The support calculation formula is:

[0050] S(X→Y) = P(X∩Y),

[0051] where X is the premise of the association rule, Y is the result of the association rule, S(X→Y) is the support of this association rule, and P(X×Y) is the probability that the premise and result of this association rule occur simultaneously;

[0052] The confidence calculation formula is:

[0053] C(X→Y) = P(Y|X),

[0054] where C(X→Y) is the confidence of this association rule, and P(Y|X) is the probability that the result of this association rule occurs when the premise of this association rule has occurred.

[0055] Preferably, based on the in-depth mining analysis results, using machine learning algorithms to construct a decision support model to provide personalized diagnosis and treatment suggestions and medication guidance for clinicians specifically includes:

[0056] Select features that have an impact on decision-making from the medical record dataset, and the features include gender, age, physiological indicators, and medical history;

[0057] Select gender, age, physiological indicators, and medical history from each feature in turn as the optimal features and use them as the root nodes;

[0058] Divide the medical record dataset into at least two subsets according to the feature values of the root nodes;

[0059] Select the next optimal feature as the child node and continue to partition the medical record dataset;

[0060] Use the data in the medical record dataset and the extracted features to train a decision support model to obtain the relationship between the input and the output;

[0061] Apply the trained decision support model to the new patient data to obtain personalized diagnosis and treatment suggestions and treatment plans for the new patients.

[0062] Preferably, the method for regularly monitoring and evaluating the data processing performance of the cloud computing platform and optimizing and upgrading the cloud computing platform based on the evaluation results specifically includes:

[0063] Regularly monitor the data processing performance of the cloud computing platform to obtain the real-time operation data of the cloud computing platform;

[0064] Based on the real-time operation data of the cloud computing platform, calculate the average speed of the cloud computing platform to process requests and output it as the response speed;

[0065] Based on the real-time operation data of the cloud computing platform, calculate the amount of data processed by the cloud computing platform per unit time and output it as the throughput;

[0066] Based on the real-time operation data of the cloud computing platform, obtain the number of services that the cloud computing platform can provide during operation and output it as the availability;

[0067] Based on the real-time operation data of the cloud computing platform, calculate the error frequency that occurs when the cloud computing platform processes data and output it as the error rate;

[0068] Based on the real-time operation data of the cloud computing platform, calculate the occupancy ratios of the CPU, memory, and storage resources of the cloud computing platform and output it as the resource utilization rate;

[0069] Preset the weights of each evaluation criterion and use the performance formula to calculate the data processing performance of the cloud computing platform;

[0070] Judge whether the data processing performance of the cloud computing platform is lower than the preset performance threshold. If so, output that the cloud computing platform needs to be optimized. If not, do not output;

[0071] Based on the business requirements and workload, dynamically adjust the resource allocation of the cloud computing platform and optimize and upgrade the cloud computing platform by upgrading the data structure;

[0072] The performance formula is:

[0073] η=ω1v+ω2α+ω3β+ω4r+ω5λ,

[0074] Where η is the data processing performance of the cloud computing platform, v, α, β, r, and λ are respectively the response speed, throughput, availability, error rate, and resource utilization rate of the cloud computing platform, and ω1, ω2, ω3, ω4, and ω5 are respectively the evaluation weights of the response speed, throughput, availability, error rate, and resource utilization rate of the cloud computing platform.

[0075] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0076] By using the association rule mining algorithm to deeply mine the medical record data sets of each patient, it provides a scientific basis for the diagnosis and treatment decisions of clinicians, helps improve the pertinence and effectiveness of diagnosis and treatment. Using the machine learning algorithm to construct a decision support model can provide personalized diagnosis and treatment suggestions and medication guidance for clinicians, which not only improves the accuracy of diagnosis and treatment, but also enhances the personalization and humanization of medical services. Regularly monitoring and evaluating the data processing performance of the cloud computing platform ensures that the platform can continuously and stably provide efficient data processing services, meeting the growing demand of the medical service terminal for data processing capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 It is a flowchart of the data processing method for the medical service terminal based on cloud computing of the present invention;

[0078] Figure 2 It is a flowchart of the method for integrating the collected data according to the preset data standard of the present invention;

[0079] Figure 3 It is a flowchart of the data storage method for the medical record data sets of each patient based on the distributed file system technology of the present invention;

[0080] Figure 4 It is a flowchart of the method for selecting the corresponding storage strategy based on the type and access frequency of data of the present invention;

[0081] Figure 5 It is a flowchart of the method for deeply mining the medical record data sets of each patient by using the association rule mining algorithm of the present invention;

[0082] Figure 6 It is a flowchart of the method for constructing a decision support model by using the machine learning algorithm of the present invention;

[0083] Figure 7 It is a flowchart of the method for regularly monitoring and evaluating the data processing performance of the cloud computing platform of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0084] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments in the following description are only examples, and those skilled in the art can think of other obvious variations.

[0085] Referring to Figure 1 shown, a method for processing medical service terminal data based on cloud computing includes:

[0086] Collecting electronic medical record data of patients from various medical devices, where the electronic medical record data includes physiological parameters, medical images, and diagnostic record data;

[0087] Integrating the collected data according to a preset data standard to form a medical record data set, removing duplicate data in the medical record data set, and processing missing values in the medical record data set;

[0088] Transmitting the processed medical record data set to a cloud computing platform using a dedicated network;

[0089] In the cloud computing platform, storing the medical record data sets of each patient based on the distributed file system technology;

[0090] Selecting a corresponding storage strategy based on the type and access frequency of the data, where the storage strategy includes hot data storage and cold data storage;

[0091] Using the association rule mining algorithm to deeply mine the medical record data sets of each patient, and extracting disease prediction, risk assessment, and treatment effect assessment information;

[0092] Based on the deep mining analysis results, using machine learning algorithms to build a decision support model to provide personalized diagnosis and treatment suggestions and medication guidance for clinicians;

[0093] Regularly monitoring and evaluating the data processing performance of the cloud computing platform, and optimizing and upgrading the cloud computing platform based on the evaluation results.

[0094] Referring to Figure 2 shown, integrating the collected data according to a preset data standard to form a medical record data set, removing duplicate data in the medical record data set, and processing missing values in the medical record data set specifically includes:

[0095] Formulating a data standard based on the norms and standards of the medical industry, where the data standard includes data naming rules, storage formats, and coding methods;

[0096] Integrating the collected data based on the preset data standard to form a medical record data set;

[0097] Performing a mapping operation on each data in the medical record data set to establish an association relationship between the data in the medical record data set;

[0098] Select the patient's visit number as the unique identifier, and use a data comparison tool to identify duplicate data existing in the medical record dataset;

[0099] Remove the duplicate data in the medical record dataset, and only retain the data identified by the visit number of the same patient;

[0100] Traverse the medical record dataset to identify fields and records with missing values;

[0101] Perform linear interpolation on the data with missing values by calculating the average value of adjacent data points of the missing values.

[0102] Based on the norms and standards of the medical industry, formulate detailed data standards, which include data naming rules (such as uniformly using English abbreviations or full names to name variables), storage formats (such as CSV, JSON, or database formats), and encoding methods (such as UTF-8 encoding), to ensure that all medical devices, systems, or platforms participating in data integration can understand and follow these data standards, so as to facilitate the subsequent data integration and processing work.

[0103] Refer to Figure 3 As shown, in the cloud computing platform, the data storage of the medical record datasets of each patient based on the distributed file system technology specifically includes:

[0104] Configure the number of servers, storage capacity, and network bandwidth of the server cluster of the distributed file system based on the data capacity of the medical service terminal;

[0105] Set up the metadata server and data nodes of the distributed file system;

[0106] Chunk each data point in the medical record dataset according to the data type;

[0107] Obtain the ASCII code comparison table, and based on the ASCII code comparison table, obtain the ASCII codes of the characters in each data chunk;

[0108] Based on the ASCII codes of the characters in each data chunk, perform a hash calculation on each data chunk using the hash value calculation formula to generate the hash values of each data chunk;

[0109] Use the data transfer protocol to upload the chunked data to the data nodes of the distributed file system;

[0110] Use the hash values of each data chunk to establish an index for the stored data chunks;

[0111] The hash value calculation formula is:

[0112]

[0113] Where H is the hash value of each data block, a(i) is the ASCII code of the i-th character in the data block, and n is the length of the string in the data block.

[0114] ASCII (American Standard Code for Information Interchange) is a character encoding based on the Latin alphabet. It includes 128 characters and can be stored in one byte. The hash value, that is, the HASH value, is usually represented by a short string of random letters and digits. It is the "data fingerprint" obtained by a hash algorithm for a group of input information of any length, that is, a group of binary values obtained through encryption operations. Since the underlying machine code of a computer uses a binary mode, any binary value of any length obtained through the hash algorithm is mapped to a shorter fixed-length binary value, that is, the hash value.

[0115] Refer to Figure 4 As shown, based on the type and access frequency of the data, selecting the corresponding storage strategy specifically includes:

[0116] Analyze and calculate the access frequency of each data type;

[0117] Judge whether the access frequency of this data type is higher than the preset frequency threshold. If so, select the hot data storage method to store the data of this data type. If not, select the cold data storage method to store the data of this data type;

[0118] Use a solid-state drive as the hardware for hot data storage;

[0119] Configure the hot data storage parameters, and the hot data storage parameters include data block size, cache policy, and replication times;

[0120] Use a traditional hard disk as the hardware for cold data storage;

[0121] Configure the cold data storage parameters, and the cold data storage parameters include data compression, archiving policy, and deduplication times.

[0122] A solid-state drive (SSD) is a hard disk made of a solid-state electronic storage chip array. It consists of a control unit and a storage unit (FLASH chip, DRAM chip). The solid-state drive is exactly the same as a common hard disk in terms of interface specifications and definitions, functions, and usage methods, and is also exactly the same as a common hard disk in terms of product shape and size. Advantages: fast read and write speed, shock and drop resistance, low power consumption, no noise, large working temperature range, light weight. Disadvantages: small capacity, limited lifespan, high price; A traditional hard disk (HDD) is the basic computer memory.

[0123] Refer to Figure 5 As shown, use the association rule mining algorithm to deeply mine the medical record data sets of each patient, and extract disease prediction, risk assessment, and treatment effect assessment information, specifically including:

[0124] S101: Set the support threshold and confidence threshold based on business requirements, and use the support threshold to filter the frequent item sets in the medical record datasets of each patient;

[0125] S102: Generate at least one association rule from each frequent item set as a candidate item set, where the association rules include the association rules between diseases and symptoms, diseases and medications, and diseases and treatment effects;

[0126] S103: Scan the database and calculate the support of each candidate item set using the support calculation formula;

[0127] S104: Determine whether there is an item set with a support greater than or equal to the support threshold. If so, retain the item set with a support greater than or equal to the support threshold as a frequent item set and return to step S102. If not, end the iteration;

[0128] S105: For each frequent item set, generate all disease-related datasets as the premise of the rule, and symptom-related datasets, medication-related datasets, and treatment effect-related datasets as the results of the rule to form association rules;

[0129] S106: Calculate the confidence of each association rule using the confidence calculation method;

[0130] S107: Determine whether the confidence of each association rule is greater than the confidence threshold. If so, output it as a valuable association rule. If not, do not output it;

[0131] The support calculation formula is:

[0132] S(X→Y) = P(X∩Y),

[0133] where X is the premise of the association rule, Y is the result of the association rule, S(X→Y) is the support of this association rule, and P(X∩Y) is the probability that the premise and result of this association rule occur simultaneously;

[0134] The confidence calculation formula is:

[0135] C(X→Y) = P(Y|X),

[0136] where C(X→Y) is the confidence of this association rule, and P(Y|X) is the probability that the result of this association rule occurs when the premise of this association rule has occurred.

[0137] Support represents the frequency of occurrence of a certain item set, that is, the ratio of the number of transactions containing this item set to the total number of transactions. Confidence represents the frequency of occurrence of item B when item A appears. In other words, confidence refers to the ratio of the number of transactions containing both item A and item B to the number of transactions containing item A.

[0138] Refer to Figure 6 As shown, based on the results of in-depth mining and analysis, a decision support model is constructed using machine learning algorithms to provide personalized diagnosis and treatment suggestions and medication guidance for clinicians, specifically including:

[0139] Select features that have an impact on decision-making from the medical record dataset, and the features include gender, age, physiological indicators, and medical history;

[0140] Select gender, age, physiological indicators, and medical history from each feature in turn as the optimal features and use them as the root nodes;

[0141] Divide the medical record dataset into at least two subsets according to the feature values of the root nodes;

[0142] Select the next optimal feature as the child node and continue to divide the medical record dataset;

[0143] Use the data in the medical record dataset and the extracted features to train the decision support model to obtain the relationship between input and output;

[0144] Apply the trained decision support model to new patient data to obtain personalized diagnosis and treatment suggestions and treatment plans for the new patients.

[0145] To prevent overfitting of the decision tree, pruning of the generated decision tree is required. The pruning strategies include pre-pruning and post-pruning. Pre-pruning is to stop the growth of the tree in advance during the construction of the decision tree, and post-pruning is to simplify it after the decision tree is generated.

[0146] Refer to Figure 7 As shown, regularly monitor and evaluate the data processing performance of the cloud computing platform, and optimize and upgrade the cloud computing platform based on the evaluation results, specifically including:

[0147] Regularly monitor the data processing performance of the cloud computing platform to obtain the real-time operation data of the cloud computing platform;

[0148] Based on the real-time operation data of the cloud computing platform, calculate the average speed of the cloud computing platform to process requests and output it as the response speed;

[0149] Based on the real-time operation data of the cloud computing platform, calculate the amount of data processed by the cloud computing platform per unit time and output it as the throughput;

[0150] Based on the real-time operation data of the cloud computing platform, obtain the number of services that the cloud computing platform can provide during operation and output it as the availability;

[0151] Based on the real-time operation data of the cloud computing platform, calculate the error frequency that occurs when the cloud computing platform processes data and output it as the error rate;

[0152] Based on the real-time operation data of the cloud computing platform, calculate the occupancy ratios of the CPU, memory, and storage resources of the cloud computing platform, and output them as resource utilization rates;

[0153] Preset the weights of each evaluation criterion, and calculate the data processing performance of the cloud computing platform using the performance formula;

[0154] Judge whether the data processing performance of the cloud computing platform is lower than the preset performance threshold. If so, output that the cloud computing platform needs to be optimized; if not, do not output;

[0155] Based on business requirements and workloads, dynamically adjust the resource allocation of the cloud computing platform, and optimize and upgrade the cloud computing platform by upgrading the data structure;

[0156] The performance formula is:

[0157] η = ω1v + ω2α + ω3β + ω4r + ω5λ,

[0158] In the formula, η is the data processing performance of the cloud computing platform, v, α, β, r, and λ are the response speed, throughput, availability, error rate, and resource utilization rate of the cloud computing platform respectively, and ω1, ω2, ω3, ω4, and ω5 are the evaluation weights of the response speed, throughput, availability, error rate, and resource utilization rate of the cloud computing platform respectively.

[0159] Configure an alarm mechanism, set reasonable thresholds according to business requirements, and automatically send an alarm when key indicators exceed the set thresholds to notify relevant personnel to take timely actions to verify the optimized and upgraded cloud computing platform to ensure that its performance is improved.

[0160] Furthermore, this solution also proposes a computer-readable storage medium, on which a computer-readable program is stored. When the computer-readable program is called, it executes the above-mentioned computing power optimization method for cloud computing data center transmission.

[0161] It can be understood that the storage medium can be a magnetic medium, such as a floppy disk, hard disk, or magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid-state drive (SSD).

[0162] In summary, the advantages of the present invention are as follows: By using the association rule mining algorithm to deeply mine the medical record datasets of each patient, it provides a scientific basis for the diagnosis and treatment decisions of clinicians, helps improve the pertinence and effectiveness of diagnosis and treatment. By using machine learning algorithms to construct a decision support model, it can provide personalized diagnosis and treatment suggestions and medication guidance for clinicians, which not only improves the accuracy of diagnosis and treatment, but also enhances the personalization and humanization of medical services. Regularly monitoring and evaluating the data processing performance of the cloud computing platform ensures that the platform can continuously and stably provide efficient data processing services, meeting the growing demand of medical service terminals for data processing capabilities.

[0163] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification is only the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for processing medical service terminal data based on cloud computing, characterized in that, Including: Collecting the electronic medical record data of patients from various medical devices, where the electronic medical record data includes physiological parameters, medical images, and diagnostic record data; Integrating the collected data according to preset data standards to form a medical record data set, removing duplicate data in the medical record data set, and processing missing values in the medical record data set; Transmitting the processed medical record data set to a cloud computing platform using a dedicated network; In the cloud computing platform, storing the medical record data sets of each patient based on the distributed file system technology; Selecting corresponding storage strategies based on the type and access frequency of the data, where the storage strategies include hot data storage and cold data storage; Using the association rule mining algorithm to deeply mine the medical record data sets of each patient, and extracting information on disease prediction, risk assessment, and treatment effect assessment; Based on the results of in-depth mining analysis, using machine learning algorithms to build a decision support model to provide personalized diagnosis and treatment suggestions and medication guidance for clinicians; Regularly monitoring and evaluating the data processing performance of the cloud computing platform, and optimizing and upgrading the cloud computing platform based on the evaluation results.

2. The method for processing medical service terminal data based on cloud computing according to claim 1, wherein The specific steps of integrating the collected data according to preset data standards to form a medical record data set, removing duplicate data in the medical record data set, and processing missing values in the medical record data set include: Formulating data standards based on the norms and standards of the medical industry, where the data standards include data naming rules, storage formats, and coding methods; Integrating the collected data based on the preset data standards to form a medical record data set; Performing mapping operations on each data in the medical record data set to establish the association relationships between the data in the medical record data set; Selecting the patient's visit number as the unique identifier, and using a data comparison tool to identify duplicate data existing in the medical record data set; Removing the duplicate data in the medical record data set, and only retaining the data identified by the visit number of the same patient; Traversing the medical record data set to identify the fields and records with missing values; Performing linear interpolation processing on the data with missing values by calculating the average value of the adjacent data points of the missing values.

3. A data processing method for a medical service terminal based on cloud computing according to claim 2, characterized in that, The specific steps of storing the medical record data sets of each patient based on the distributed file system technology in the cloud computing platform include: Configuring the number of servers, storage capacity, and network bandwidth of the server cluster of the distributed file system based on the data capacity of the medical service terminal; Setting up the metadata server and data nodes of the distributed file system; Chunking each data point in the medical record data set according to the data type; Obtaining the ASCII code comparison table, and based on the ASCII code comparison table, obtaining the ASCII codes of the characters in each data chunk; Based on the ASCII codes of the characters in each data chunk, performing hash calculation on each data chunk using the hash value calculation formula to generate the hash values of each data chunk; Uploading the chunked data to the data nodes of the distributed file system using a data transfer protocol; Establishing an index for the stored data chunks using the hash values of each data chunk; The hash value calculation formula is: In the formula, H is the hash value of each data chunk, a(i) is the ASCII code of the i-th character in the data chunk, and n is the length of the string in the data chunk.

4. A method for processing medical service terminal data based on cloud computing according to claim 3, characterized in that, Selecting the corresponding storage strategy based on the type and access frequency of data specifically includes: Analyze and calculate the access frequency of each data type; Determine whether the access frequency of this data type is higher than the preset frequency threshold. If so, select the hot data storage method to store the data of this data type. If not, select the cold data storage method to store the data of this data type; Use a solid-state drive as the hardware for hot data storage; Configure hot data storage parameters, and the hot data storage parameters include data block size, cache policy, and replication times; Use a traditional hard disk as the hardware for cold data storage; Configure cold data storage parameters, and the cold data storage parameters include data compression, archiving policy, and deduplication times.

5. A method for processing medical service terminal data based on cloud computing according to claim 4, characterized in that, Deeply mining the medical record datasets of each patient using the association rule mining algorithm to extract disease prediction, risk assessment, and treatment effect assessment information specifically includes: S101: Set the support threshold and confidence threshold based on business requirements, and use the support threshold to screen the frequent item sets in the medical record datasets of each patient; S102: Generate at least one association rule from each frequent item set as a candidate item set, and the association rules include the association rules between diseases and symptoms, diseases and medications, and diseases and treatment effects; S103: Scan the database and calculate the support of each candidate item set using the support calculation formula; S104: Determine whether there is an item set with a support greater than or equal to the support threshold. If so, retain the item set with a support greater than or equal to the support threshold as a frequent item set, and return to step S102. If not, end the iteration; S105: For each frequent item set, generate all disease-related datasets as the premise of the rule, and symptom-related datasets, medication-related datasets, and treatment effect-related datasets as the results of the rule to form association rules; S106: Calculate the confidence of each association rule using the confidence calculation method; S107: Determine whether the confidence of each association rule is greater than the confidence threshold. If so, output it as a valuable association rule. If not, do not output it; The support calculation formula is: S(X→Y) = P(X∩Y), where X is the premise of the association rule, Y is the result of the association rule, S(X→Y) is the support of this association rule, and P(X∩Y) is the probability that the premise and result of this association rule occur simultaneously; The confidence calculation formula is: C(X→Y) = P(Y|X), where C(X→Y) is the confidence of this association rule, and P(Y|X) is the probability that the result of this association rule occurs when the premise of this association rule has occurred.

6. A data processing method for a medical service terminal based on cloud computing according to claim 5, characterized in that Based on the results of in-depth mining analysis, using machine learning algorithms to construct a decision support model to provide personalized diagnosis and treatment suggestions and medication guidance for clinicians specifically includes: Select the features that have an impact on the decision from the medical record datasets, and the features include gender, age, physiological indicators, and medical history; Select gender, age, physiological indicators, and medical history from each feature in turn as the optimal features and use them as the root nodes; Divide the medical record datasets into at least two subsets according to the feature values of the root nodes; Select the next optimal feature as the child node and continue to partition the medical record dataset; Use the data in the medical record dataset and the extracted features to train a decision support model to obtain the relationship between the input and the output; Apply the trained decision support model to the new patient data to obtain personalized diagnosis and treatment suggestions and treatment plans for the new patient.

7. A method for processing medical service terminal data based on cloud computing according to claim 6, characterized in that, The monitoring and evaluation of the data processing performance of the cloud computing platform on a regular basis and the optimization and upgrade of the cloud computing platform based on the evaluation results specifically include: Monitor the data processing performance of the cloud computing platform regularly to obtain the real-time operation data of the cloud computing platform; Based on the real-time operation data of the cloud computing platform, calculate the average speed of the cloud computing platform to process requests and output it as the response speed; Based on the real-time operation data of the cloud computing platform, calculate the amount of data processed by the cloud computing platform per unit time and output it as the throughput; Based on the real-time operation data of the cloud computing platform, obtain the number of services that the cloud computing platform can provide during operation and output it as the availability; Based on the real-time operation data of the cloud computing platform, calculate the error frequency that occurs when the cloud computing platform processes data and output it as the error rate; Based on the real-time operation data of the cloud computing platform, calculate the occupancy ratios of the CPU, memory, and storage resources of the cloud computing platform and output them as the resource utilization rate; Preset the weights of each evaluation criterion and use the performance formula to calculate the data processing performance of the cloud computing platform; Judge whether the data processing performance of the cloud computing platform is lower than the preset performance threshold. If so, output that the cloud computing platform needs to be optimized. If not, do not output; Dynamically adjust the resource allocation of the cloud computing platform based on business requirements and workloads, and optimize and upgrade the cloud computing platform by upgrading the data structure; The performance formula is: η = ω1v + ω2α + ω3β + ω4r + ω5λ, where η is the data processing performance of the cloud computing platform, v, α, β, r, and λ are the response speed, throughput, availability, error rate, and resource utilization rate of the cloud computing platform respectively, and ω1, ω2, ω3, ω4, and ω5 are the evaluation weights of the response speed, throughput, availability, error rate, and resource utilization rate of the cloud computing platform respectively.

Images