Medical flow direction data transmission safety early warning method and system based on big data

Abstract

The invention relates to the technical field of data processing, and discloses a medicine flow data transmission safety early warning method and system based on big data. Firstly, an improved neighbor sorting algorithm and an improved clustering algorithm are used for carrying out similar repeated data cleaning and missing data filling on an original medicine flow direction data set, and a processed medicine flow direction data set is obtained; secondly, dimension reduction processing is carried out on the processed medicine flow direction data set, and a final medicine flow direction data set is obtained; a kernel parameter and a penalty factor in the support vector machine are optimized by using a parrot optimization algorithm, and Cauchy variation improvement is introduced to obtain a support vector machine early warning model; and finally, establishing a medicine flow direction data transmission safety early warning model, checking the final medicine flow direction data set to obtain an early warning result, and completing safety early warning. By processing and analyzing the medicine flow direction data, the purpose of medicine flow direction data transmission safety early warning is achieved, and the method is accurate and objective.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method and system for early warning of medical flow data transmission security based on big data. Background Art

[0002] Chinese patent CN116506205B discloses a data processing method and system for a smart medical platform. The method specifically includes: based on cloud computing technology, the collection end collects medical data from multiple data sources to obtain medical data, processes the medical data and inputs it into the database; during the data transmission process, the processing end obtains the warning coefficient in real time, sets the warning threshold, compares the warning coefficient with the warning threshold, and when the warning coefficient is less than the warning threshold, the data transmission stops, the data is further cached in the memory, and a warning signal is issued; after receiving the warning signal, the maintenance personnel maintain the equipment, and when the equipment is restored, the processed medical data continues to be input into the database; the smart medical platform performs data interaction and data sharing based on the Internet. However, the invention does not further clean the medical data, and the results are erroneous.

[0003] Traditional methods for early warning of medical data transmission security usually need to be improved in terms of data redundancy processing due to the complexity of big data and data security processing. At the same time, traditional methods for early warning of medical data transmission security do not use high-tech technologies such as artificial intelligence, and do not establish neural network models to warn of data security, and cannot effectively achieve data security transmission. Summary of the invention

[0004] In response to the problems in the related technology, the present invention provides a method and system for security early warning of medicine flow data transmission based on big data to overcome the above-mentioned technical problems existing in the existing related technology.

[0005] In order to solve the above technical problems, the present invention is achieved through the following technical solutions: The present invention is a method for early warning of medical flow data transmission security based on big data, comprising the following steps: S1. Obtaining a medicine flow indicator to obtain an original medicine flow data set, and performing similar duplicate data cleaning on the original medicine flow data set based on an improved nearest neighbor sorting algorithm to obtain a processed medicine flow data set; S2, using the improved clustering algorithm to find the optimized cluster center for the processed medicine flow data set, and then filling in the missing data to obtain a processed medicine flow data set; S3, using principal component analysis to perform dimensionality reduction processing on the processed medicine flow data set to obtain a final medicine flow data set; S4. Use the Parrot optimization algorithm to optimize the kernel parameters and penalty factors in the support vector machine, improve the Parrot optimization algorithm by introducing Cauchy mutation, obtain the optimized kernel parameters and penalty factors, and obtain the support vector machine early warning model; S5. Add a classification function to the support vector machine warning model, establish a medicine flow data transmission security warning model, combine the final medicine flow data set, obtain a warning result, and complete the safety warning.

[0006] The invention first obtains the original medicine flow data set, uses the improved nearest neighbor sorting algorithm to clean the original medicine flow data set for similar duplicate data, and finds similar duplicate data by calculating the weight of the original medicine flow data; compared with the traditional nearest neighbor sorting algorithm, the algorithm effectively improves the search accuracy and recall rate, and solves the problem of mixed detection of text and digital data; secondly, uses the improved clustering algorithm to find the optimized clustering center, and by optimizing the clustering center based on distance and density, the missing medicine flow data is classified and then filled; this method overcomes the problem of traditional clustering algorithms being too dependent on the initial k value, and improves the stability and reliability of data processing The data after cleaning similar repeated data and filling missing data are processed by dimensionality reduction to obtain the final medical flow data set. The principal components after dimensionality reduction can effectively describe the data characteristics, reduce the redundant information of the data, and reduce the calculation amount of the support vector machine; the support vector machine is trained, and the kernel parameters and penalty factors in the support vector machine are optimized using the parrot optimization algorithm. This algorithm imitates the behavior of parrots when they are domesticated, and can explore the search space well to avoid falling into the local optimal solution. At the same time, the Cauchy mutation is introduced for improvement, and the global search capability is enhanced by increasing the step size; finally, a medical flow data transmission security early warning model is established, the early warning results are output, and the security early warning is completed.

[0007] Preferably, the S1 comprises the following steps: S11, obtaining drug flow indicators, wherein the drug flow indicators include drug production date, drug number, drug sales quantity, etc., collecting drug flow indicator related data, and unifying the data format of the drug flow indicator related data to obtain original drug flow data, forming an original drug flow data set, wherein the original drug flow data set includes m Medical flow indicators, each of which includes n Original medical flow data; S12, using the modified nearest neighbor sorting algorithm to clean similar duplicate data from the original medicine flow data set to obtain a processed medicine flow data set, the specific steps are as follows: S121, converting the original medicine flow data set into a unified code to obtain a medicine flow data set to be processed, setting the medicine flow data set to be processed The first indicator of the medical flow The pending medical flow data is , then the medicine flow data to be processed The attribute value and The entropy calculation formula for the medical flow data to be processed is as follows: , ; in, Indicates the medicine flow data to be processed The attribute value of Indicates The entropy value of the medical flow data to be processed, ; According to The entropy value calculation of the medical flow data to be processed The weight of the pending medical flow data is calculated as follows: ; in, Indicates The weight of the medical flow data to be processed, ; Score each of the to-be-processed medicine flow direction data in the to-be-processed medicine flow direction data set to obtain an evaluation score set, and calculate the first The weight of the evaluation score, combined with the The weight of the evaluation score and the The weight calculation of the pending medical flow data The comprehensive weight of the pending medical flow data; n The comprehensive weights of the to-be-processed medical flow data are sorted in ascending order to obtain a comprehensive weight set; S122, grouping according to the comprehensive weight set, setting an initial group, selecting any comprehensive weight in the comprehensive weight set to add to the initial group, and then calculating the similarity between the initial group and the comprehensive weight in the comprehensive weight set, and setting the similarity threshold to , when the similarity of the comprehensive weight is greater than the similarity threshold, the corresponding comprehensive weight is added to the initial group; otherwise, a new group is established to obtain several groups; The Euclidean distance of the comprehensive weights in several groups is calculated, and a distance threshold is set. The medical flow data to be processed corresponding to the Euclidean distance of the comprehensive weights less than the distance threshold is regarded as similar duplicate data and deleted to obtain a processed medical flow data set.

[0008] This invention uses an improved nearest neighbor sorting algorithm to clean similar duplicate data from the original medical flow data set, and calculates the weight of the original medical flow data. Compared with the traditional nearest neighbor sorting algorithm, the search accuracy and recall rate are effectively improved, similar duplicate data are effectively found, and the problem of mixed detection of text and digital data is solved.

[0009] Preferably, S2 comprises the following steps: S21, according to whether the processed medicine flow data set contains missing values, dividing the processed medicine flow data set into a complete medicine flow data set and a missing medicine flow data set; setting the number of cluster centers to 1, calculating the sum of squares of errors of the cluster centers and the complete medicine flow data in the complete medicine flow data set, recorded as the first sum of squares of errors, increasing the number of cluster centers, and calculating the second sum of squares of errors; sequentially increasing the number of cluster centers until the sum of squares of errors converges, and obtaining the optimized cluster centers; The optimized cluster center is used as the initial cluster center of the complete drug flow data set, and the complete drug flow data set is clustered using an improved clustering algorithm to obtain k cluster centers. The specific steps are as follows: S211, select any complete medicine flow data from the complete medicine flow data set, record it as the first complete data and the second complete data, calculate the distance between the first complete data and the second complete data, and the average distance calculation formula is as follows: ; in, represents the average distance between the first complete data and the second complete data, Represents the first complete data, Represents the second complete data, Represents the distance between the first complete data and the second complete data, Indicates the number of pairs; According to the average distance between the first complete data and the second complete data, the number of complete medicine flow data closest to the first complete data is counted, and recorded as the closest distance coefficient; S212. Calculate the density of the first complete data by combining the closest distance coefficient and the distance between the first complete data and the second complete data. The calculation formula is as follows: ; in, represents the density of the first complete data, c represents the closest distance coefficient, e represents a constant, ; sequentially calculating the densities of other complete medicine flow data in the complete medicine flow data set to obtain a complete medicine flow data density set; S213, selecting the maximum complete medicine flow data density from the complete medicine flow data density set, adding it to the new density set, and then selecting the complete medicine flow data density farthest from the new density set from the complete medicine flow data density set as the cluster center, and updating the initial cluster center until k cluster centers are obtained; S22, calculating the distance between the missing medicine flow data and k cluster centers in the missing medicine flow data set, obtaining the cluster center corresponding to the minimum distance between the missing medicine flow data and the k cluster centers, recording it as the minimum cluster center, and classifying the corresponding missing medicine flow data into the minimum cluster center to obtain k clusters; The Lagrange interpolation method is used to fill in the missing medicine flow data in k classes at the same time, and k filled classes are output. Then, the complete medicine flow data set is combined to obtain the processed medicine flow data set.

[0010] This invention uses an improved clustering algorithm to find the optimized cluster center, optimizes the cluster center based on distance and density, classifies the missing medical flow data, and then fills it in. It overcomes the problem of traditional clustering algorithms being too dependent on the initial k value and improves the stability and reliability of data processing.

[0011] Preferably, S3 comprises the following steps: S31, the processed medicine flow data set m The first indicator of the medical flow n The original drug flow data is recorded as , generate the processed medicine flow data matrix B as follows: ; Set the transpose of the processed medicine flow data matrix to , then the covariance matrix , calculate the eigenvalues ​​of the covariance matrix, and calculate the eigenvectors corresponding to the eigenvalues ​​to obtain an eigenvector matrix; S32, calculating the product of the processed medicine flow data matrix and the eigenvector matrix to obtain a principal component matrix, selecting the principal component in the principal component matrix, and calculating the principal component contribution rate according to the eigenvalue corresponding to the principal component, and the calculation formula is as follows: ; in, Indicates c The principal component contribution rate of the principal components, Indicates cThe eigenvalues ​​corresponding to the principal components are ; The principal component contribution values ​​of all principal components are accumulated to obtain the previous c The cumulative contribution rate of the principal components is calculated, a contribution rate threshold is set, and the principal component whose cumulative contribution rate is greater than the contribution rate threshold is recorded as the key principal component to complete the dimensionality reduction process. The key principal components constitute the final medicine flow data set.

[0012] The invention obtains the final medicine flow data set by performing dimensionality reduction processing on the processed medicine flow data. The principal components after dimensionality reduction can effectively describe the data characteristics, reduce the redundant information of the data, and reduce the calculation amount of the support vector machine.

[0013] Preferably, S4 comprises the following steps: S41, setting the kernel function of the support vector machine to a Gaussian kernel function, the Gaussian kernel function including a kernel parameter, establishing an objective function according to the input sample, the objective function including a penalty factor; reacquiring a new medicine flow data set, the new medicine flow data set undergoing data preprocessing and principal component analysis, and normalization processing to obtain a medicine flow data sample set, dividing the medicine flow data sample set into a sample training set and a sample test set, inputting the sample training set into the support vector machine, and obtaining an output result error; S42, taking the output result error as the fitness function, using the Parrot optimization algorithm to optimize the kernel parameters and penalty factors in the support vector machine, and introducing Cauchy mutation to improve the Parrot optimization algorithm to obtain the optimized kernel parameters and penalty factors. The specific steps are as follows: S421. Assume that there is a parrot population in the search space. Each parrot individual in the parrot population represents a solution of the fitness function. The coordinates of the parrot individual represent the kernel parameter and the penalty factor. Assume that the number of parrot populations is D. The lower and upper limits of the search space are and , represents a random number and , then the calculation formula for the initial position of the e-th parrot in the parrot population is as follows: ; in, represents the initial position of the e-th parrot in the parrot population; In the process of foraging, the current iteration number is set to h, the maximum iteration number is set to H, and the h-th iteration position of the e-th parrot in the parrot population is , represents a random number and , the h-th iteration average position of the parrot population is , the best position of the h-th iteration of the parrot population is , Lévy flight is introduced to describe the flight of parrots, and the Lévy flight coefficient is , where dim represents the individual dimension of each parrot in the parrot population, and the individual position of the parrot is updated. The calculation formula is as follows: ; in, The h+1th iteration position of the eth parrot in the parrot population; Calculate the fitness function value at the hth iteration, find the current best fitness function value, the parrot individual position corresponding to the current best fitness function value, and the best kernel parameter and penalty factor at the hth iteration; when the current best fitness function value is greater than the fitness function value at the h+1th iteration, use the fitness function value at the h+1th iteration to replace the current best fitness function value; S422, during the communication process, parrots randomly stay at a certain part of the owner, and the vector with all 1s of dimension dim is recorded as , represents a random number and , at this time, the individual position of the parrot is updated again, and the calculation formula is as follows: ; Set the flight probability. When the flight probability is greater than 0.5, the e-th parrot in the parrot population will move away from the parrot population. When the flight probability is less than or equal to 0.5, the e-th parrot in the parrot population will stay in the parrot population. During the fear process, the e-th parrot in the parrot population chooses to stay away from strangers and get closer to the owner, and the position of the parrot is updated synchronously; during the communication process and the fear process, the Cauchy mutation is introduced, setting Represents a random number that follows the Cauchy distribution, and updates the individual position of the parrot. The calculation formula is as follows: ; When the current number of iterations reaches the maximum number of iterations, the iteration is stopped, and the final position of the e-th parrot in the parrot population is obtained. The two-dimensional coordinates of the final position correspond to the optimized kernel parameters and penalty factors respectively; S43, the support vector machine uses the optimized kernel parameters and penalty factors, and inputs the sample test set into the support vector machine until the support vector machine converges to obtain a support vector machine early warning model.

[0014] The invention optimizes the kernel parameters and penalty factors in the support vector machine by using the parrot optimization algorithm. By imitating the behavior of parrots during training, it can explore the search space well and avoid falling into the local optimal solution. At the same time, the Cauchy mutation is introduced for improvement, and the global search capability is enhanced by increasing the step size.

[0015] Preferably, the S5 comprises the following steps: S51, after normalizing the final medicine flow data set, input it into the support vector machine early warning model, output the classification result, and add a classification function to the support vector machine early warning model to obtain a medicine flow data transmission security early warning model. The classification function formula is as follows: ; Where f represents the classification function, and represents the warning threshold, and p represents the classification result; S52. In the medical flow data transmission security warning model, a medical flow data transmission security warning is performed according to the classification result to obtain a warning result; when the classification function is 0, it indicates that the medical flow data transmission is in a safe state; when the classification function is 0.5, it indicates that the medical flow data transmission security is within the warning range; when the classification function is 1, it indicates that the medical flow data transmission is in a dangerous state, and the safety warning is completed.

[0016] The present invention also discloses a system for a safety early warning method for medicine flow data transmission based on big data, which specifically includes: a data cleaning and filling module, a data dimension reduction processing module, a support vector machine parameter optimization module and a data safety early warning module; The data cleaning and filling module is used to clean similar repeated data and fill missing data in the original medicine flow data set; The data dimension reduction processing module is used to perform dimension reduction processing on the processed medicine flow data set using principal component analysis; The support vector machine parameter optimization module is used to optimize the kernel parameters and penalty factors in the support vector machine using the parrot optimization algorithm; The data security warning module is used to establish a medicine flow data transmission security warning model and output warning results.

[0017] The present invention has the following beneficial effects: 1. This invention uses an improved nearest neighbor sorting algorithm to clean similar duplicate data from the original medical flow data set. Compared with the traditional nearest neighbor sorting algorithm, the search accuracy and recall rate are effectively improved, similar duplicate data can be effectively found, and the problem of mixed detection of text and digital data is solved.

[0018] 2. The invention uses an improved clustering algorithm to find the optimized cluster center, optimizes the cluster center based on distance and density, classifies the missing medicine flow data, and then fills it in. It overcomes the problem of traditional clustering algorithms being too dependent on the initial k value and improves the stability and reliability of data processing.

[0019] 3. The invention obtains the final medicine flow data set by performing dimensionality reduction processing on the processed medicine flow data. The principal components after dimensionality reduction can effectively describe the data characteristics, reduce the redundant information of the data, and reduce the calculation amount of the support vector machine.

[0020] 4. The invention optimizes the kernel parameters and penalty factors in the support vector machine by using the Parrot optimization algorithm. The algorithm can explore the search space well and avoid falling into the local optimal solution. At the same time, the Cauchy mutation is introduced for improvement, and the global search capability is enhanced by increasing the step size. Then a medical flow data transmission security early warning model is established, the early warning results are output, and the safety early warning is completed.

[0021] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the drawings required for describing the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, they can also obtain drawings based on these drawings without paying creative work.

[0023] Figure 1 A schematic diagram of the process of providing a security warning for medicine flow data transmission by a medicine flow data transmission security warning system based on big data provided by the present invention. DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] In the description of the present invention, it is necessary to understand that the terms "opening", "upper", "lower", "top", "middle", "inside" and the like indicating orientation or positional relationship are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the invention.

[0026] Example 1 Please refer to Figure 1 The present invention provides a method for early warning of medical flow data transmission security based on big data, comprising the following steps: S1. Obtaining a medicine flow indicator to obtain an original medicine flow data set, and performing similar duplicate data cleaning on the original medicine flow data set based on an improved nearest neighbor sorting algorithm to obtain a processed medicine flow data set; The S1 comprises the following steps: S11, obtaining drug flow indicators, wherein the drug flow indicators include drug production date, drug number, drug sales quantity, etc., collecting drug flow indicator related data, and unifying the data format of the drug flow indicator related data to obtain original drug flow data, forming an original drug flow data set, wherein the original drug flow data set includes m Medical flow indicators, each of which includes n Original medical flow data; S12, using the modified nearest neighbor sorting algorithm to clean similar duplicate data from the original medicine flow data set to obtain a processed medicine flow data set, the specific steps are as follows: S121, converting the original medicine flow data set into a unified code to obtain a medicine flow data set to be processed, setting the medicine flow data set to be processed The first indicator of the medical flow The pending medical flow data is , then the medicine flow data to be processed The attribute value and The entropy calculation formula for the medical flow data to be processed is as follows: , ; in, Indicates the medicine flow data to be processed The attribute value of Indicates The entropy value of the medical flow data to be processed, ; According to The entropy value calculation of the medical flow data to be processed The weight of the pending medical flow data is calculated as follows: ; in, Indicates The weight of the medical flow data to be processed, ; Score each of the to-be-processed medicine flow direction data in the to-be-processed medicine flow direction data set to obtain an evaluation score set, and calculate the first The weight of the evaluation score, combined with the The weight of the evaluation score and the The weight calculation of the pending medical flow data The comprehensive weight of the pending medical flow data; n The comprehensive weights of the to-be-processed medical flow data are sorted in ascending order to obtain a comprehensive weight set; S122, grouping according to the comprehensive weight set, setting an initial group, selecting any comprehensive weight in the comprehensive weight set to add to the initial group, and then calculating the similarity between the initial group and the comprehensive weight in the comprehensive weight set, and setting the similarity threshold to , when the similarity of the comprehensive weight is greater than the similarity threshold, the corresponding comprehensive weight is added to the initial group; otherwise, a new group is established to obtain several groups; Calculate the Euclidean distance of the comprehensive weights in several groups, set a distance threshold, regard the to-be-processed medicine flow data corresponding to the Euclidean distance of the comprehensive weights less than the distance threshold as similar duplicate data, and delete them to obtain a processed medicine flow data set; S2, using the improved clustering algorithm to find the optimized cluster center for the processed medicine flow data set, and then filling in the missing data to obtain a processed medicine flow data set; The S2 comprises the following steps: S21, according to whether the processed medicine flow data set contains missing values, dividing the processed medicine flow data set into a complete medicine flow data set and a missing medicine flow data set; setting the number of cluster centers to 1, calculating the sum of squares of errors of the cluster centers and the complete medicine flow data in the complete medicine flow data set, recorded as the first sum of squares of errors, increasing the number of cluster centers, and calculating the second sum of squares of errors; sequentially increasing the number of cluster centers until the sum of squares of errors converges, and obtaining the optimized cluster centers; The optimized cluster center is used as the initial cluster center of the complete drug flow data set, and the complete drug flow data set is clustered using an improved clustering algorithm to obtain k cluster centers. The specific steps are as follows: S211, select any complete medicine flow data from the complete medicine flow data set, record it as the first complete data and the second complete data, calculate the distance between the first complete data and the second complete data, and the average distance calculation formula is as follows: ; in, represents the average distance between the first complete data and the second complete data, Represents the first complete data, Represents the second complete data, Represents the distance between the first complete data and the second complete data, Indicates the number of pairs; According to the average distance between the first complete data and the second complete data, the number of complete medicine flow data closest to the first complete data is counted, and recorded as the closest distance coefficient; S212. Calculate the density of the first complete data by combining the closest distance coefficient and the distance between the first complete data and the second complete data. The calculation formula is as follows: ; in, represents the density of the first complete data, c represents the closest distance coefficient, e represents a constant, ; sequentially calculating the densities of other complete medicine flow data in the complete medicine flow data set to obtain a complete medicine flow data density set; S213, selecting the maximum complete medicine flow data density from the complete medicine flow data density set, adding it to the new density set, and then selecting the complete medicine flow data density farthest from the new density set from the complete medicine flow data density set as the cluster center, and updating the initial cluster center until k cluster centers are obtained; S22, calculating the distance between the missing medicine flow data and k cluster centers in the missing medicine flow data set, obtaining the cluster center corresponding to the minimum distance between the missing medicine flow data and the k cluster centers, recording it as the minimum cluster center, and classifying the corresponding missing medicine flow data into the minimum cluster center to obtain k clusters; Lagrange interpolation method is used to fill in the missing drug flow data in k classes at the same time, and k filled classes are output, which are then combined with the complete drug flow data set to obtain the processed drug flow data set; S3, using principal component analysis to perform dimensionality reduction processing on the processed medicine flow data set to obtain a final medicine flow data set; The S3 comprises the following steps: S31, the processed medicine flow data set m The first indicator of the medical flow n The original drug flow data is recorded as , generate the processed medicine flow data matrix B as follows: ; Set the transpose of the processed medicine flow data matrix to , then the covariance matrix , calculate the eigenvalues ​​of the covariance matrix, and calculate the eigenvectors corresponding to the eigenvalues ​​to obtain an eigenvector matrix; S32, calculating the product of the processed medicine flow data matrix and the eigenvector matrix to obtain a principal component matrix, selecting the principal component in the principal component matrix, and calculating the principal component contribution rate according to the eigenvalue corresponding to the principal component, and the calculation formula is as follows: ; in, Indicates c The principal component contribution rate of the principal components, Indicates c The eigenvalues ​​corresponding to the principal components are ; The principal component contribution values ​​of all principal components are accumulated to obtain the previous c The cumulative contribution rate of the principal components is calculated, and a contribution rate threshold is set. The principal component whose cumulative contribution rate is greater than the contribution rate threshold is recorded as the key principal component, and the dimension reduction process is completed. The key principal components constitute the final drug flow data set; S4. Use the Parrot optimization algorithm to optimize the kernel parameters and penalty factors in the support vector machine, improve the Parrot optimization algorithm by introducing Cauchy mutation, obtain the optimized kernel parameters and penalty factors, and obtain the support vector machine early warning model; The S4 comprises the following steps: S41, setting the kernel function of the support vector machine to a Gaussian kernel function, the Gaussian kernel function including a kernel parameter, establishing an objective function according to the input sample, the objective function including a penalty factor; reacquiring a new medicine flow data set, the new medicine flow data set undergoing data preprocessing and principal component analysis, and normalization processing to obtain a medicine flow data sample set, dividing the medicine flow data sample set into a sample training set and a sample test set, inputting the sample training set into the support vector machine, and obtaining an output result error; S42, taking the output result error as the fitness function, using the Parrot optimization algorithm to optimize the kernel parameters and penalty factors in the support vector machine, and introducing Cauchy mutation to improve the Parrot optimization algorithm to obtain the optimized kernel parameters and penalty factors. The specific steps are as follows: S421. Assume that there is a parrot population in the search space. Each parrot individual in the parrot population represents a solution of the fitness function. The coordinates of the parrot individual represent the kernel parameter and the penalty factor. Assume that the number of parrot populations is D. The lower and upper limits of the search space are and , represents a random number and , then the calculation formula for the initial position of the e-th parrot in the parrot population is as follows: ; in, represents the initial position of the e-th parrot in the parrot population; In the process of foraging, the current iteration number is set to h, the maximum iteration number is set to H, and the h-th iteration position of the e-th parrot in the parrot population is , represents a random number and , the h-th iteration average position of the parrot population is , the best position of the h-th iteration of the parrot population is , Lévy flight is introduced to describe the flight of parrots, and the Lévy flight coefficient is , where dim represents the individual dimension of each parrot in the parrot population, and the individual position of the parrot is updated. The calculation formula is as follows: ; in, The h+1th iteration position of the eth parrot in the parrot population; Calculate the fitness function value at the hth iteration, find the current best fitness function value, the parrot individual position corresponding to the current best fitness function value, and the best kernel parameter and penalty factor at the hth iteration; when the current best fitness function value is greater than the fitness function value at the h+1th iteration, use the fitness function value at the h+1th iteration to replace the current best fitness function value; S422, during the communication process, parrots randomly stay at a certain part of the owner, and the vector with all 1s of dimension dim is recorded as , represents a random number and , at this time, the individual position of the parrot is updated again, and the calculation formula is as follows: ; Set the flight probability. When the flight probability is greater than 0.5, the e-th parrot in the parrot population will move away from the parrot population. When the flight probability is less than or equal to 0.5, the e-th parrot in the parrot population will stay in the parrot population. During the fear process, the e-th parrot in the parrot population chooses to stay away from strangers and get closer to the owner, and the position of the parrot is updated synchronously; during the communication process and the fear process, the Cauchy mutation is introduced, setting Represents a random number that follows the Cauchy distribution, and updates the individual position of the parrot. The calculation formula is as follows: ; When the current number of iterations reaches the maximum number of iterations, the iteration is stopped, and the final position of the e-th parrot in the parrot population is obtained. The two-dimensional coordinates of the final position correspond to the optimized kernel parameters and penalty factors respectively; S43, the support vector machine uses the optimized kernel parameters and penalty factors, inputs the sample test set into the support vector machine, until the support vector machine converges, and obtains a support vector machine early warning model; S5, adding a classification function to the support vector machine warning model, establishing a medicine flow data transmission security warning model, combining the final medicine flow data set, obtaining a warning result, and completing a safety warning; The S5 comprises the following steps: S51, after normalizing the final medicine flow data set, input it into the support vector machine early warning model, output the classification result, and add a classification function to the support vector machine early warning model to obtain a medicine flow data transmission security early warning model. The classification function formula is as follows: ; Where f represents the classification function, and represents the warning threshold, and p represents the classification result; S52. In the medical flow data transmission security warning model, a medical flow data transmission security warning is performed according to the classification result to obtain a warning result; when the classification function is 0, it indicates that the medical flow data transmission is in a safe state; when the classification function is 0.5, it indicates that the medical flow data transmission security is within the warning range; when the classification function is 1, it indicates that the medical flow data transmission is in a dangerous state, and the safety warning is completed.

[0027] Example 2 The present invention also discloses a system for a safety early warning method for medicine flow data transmission based on big data, which specifically includes: a data cleaning and filling module, a data dimension reduction processing module, a support vector machine parameter optimization module and a data safety early warning module; The data cleaning and filling module is used to clean similar repeated data and fill missing data in the original medicine flow data set; The data dimension reduction processing module is used to perform dimension reduction processing on the processed medicine flow data set using principal component analysis; The support vector machine parameter optimization module is used to optimize the kernel parameters and penalty factors in the support vector machine using the parrot optimization algorithm; The data security warning module is used to establish a medicine flow data transmission security warning model and output warning results.

[0028] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0029] The preferred embodiments of the invention disclosed above are only used to help explain the invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the invention, so that those skilled in the art can understand and use the invention well.

Claims

1. A method for early warning of medical flow data transmission security based on big data, characterized in that: The steps include: S1. Obtaining a medicine flow indicator to obtain an original medicine flow data set, and performing similar duplicate data cleaning on the original medicine flow data set based on an improved nearest neighbor sorting algorithm to obtain a processed medicine flow data set; S2, using the improved clustering algorithm to find the optimized cluster center for the processed medicine flow data set, and then filling in the missing data to obtain a processed medicine flow data set; S3, using principal component analysis to perform dimensionality reduction processing on the processed medicine flow data set to obtain a final medicine flow data set; S4. Use the optimization algorithm to optimize the kernel parameters and penalty factors in the support vector machine, improve the optimization algorithm by introducing Cauchy mutation, obtain the optimized kernel parameters and penalty factors, and obtain the support vector machine early warning model; S5. Add a classification function to the support vector machine warning model, establish a medicine flow data transmission security warning model, combine the final medicine flow data set, obtain a warning result, and complete the safety warning.

2. According to the big data-based medicine flow data transmission security early warning method of claim 1, it is characterized by: The S1 comprises the following steps: S11, obtaining a medicine flow indicator to obtain an original medicine flow data set; S12. Use the modified nearest neighbor sorting algorithm to clean similar duplicate data from the original medicine flow data set to obtain a processed medicine flow data set.

3. According to the big data-based medicine flow data transmission security early warning method of claim 2, it is characterized in that: The S12 comprises the following steps: S121, converting the original medicine flow data set through unified code encoding to obtain a medicine flow data set to be processed, calculating the entropy value and weight of the medicine flow data to be processed in the medicine flow data set to be processed, and obtaining a comprehensive weight set; S122. Grouping is performed according to the comprehensive weight set to obtain several groups; the Euclidean distance of the comprehensive weights in the several groups is calculated, a distance threshold is set, and the to-be-processed medical flow data corresponding to the Euclidean distance less than the distance threshold is regarded as similar duplicate data, and deleted to obtain a processed medical flow data set.

4. According to the big data-based medicine flow data transmission security early warning method of claim 3, it is characterized in that: The S2 comprises the following steps: S21, dividing the processed medicine flow data set into a complete medicine flow data set and a missing medicine flow data set; finding an optimized clustering center in the complete medicine flow data set by iteration, using the optimized clustering center as the initial clustering center of the complete medicine flow data set, and using an improved clustering algorithm to perform clustering processing on the complete medicine flow data set to obtain k clustering centers; S22. Calculate the distance between the missing medicine flow data and k cluster centers in the missing medicine flow data set, classify the missing medicine flow data, and obtain k classes; use the Lagrange interpolation method to fill the missing medicine flow data in the k classes at the same time, output k filled classes, and then combine them with the complete medicine flow data set to obtain a processed medicine flow data set.

5. According to the big data-based medicine flow data transmission security early warning method of claim 4, it is characterized in that: The S3 comprises the following steps: S31, generating a processed medicine flow data matrix according to the processed medicine flow data set, calculating a covariance matrix according to the processed medicine flow data matrix, and calculating eigenvalues ​​and eigenvectors of the covariance matrix to obtain an eigenvector matrix; S32. Calculate the product of the processed pharmaceutical flow data matrix and the eigenvector matrix to obtain the principal component matrix, select the principal components in the principal component matrix, and calculate the principal component contribution rate according to the eigenvalues ​​corresponding to the principal components; add up the principal component contribution values ​​to obtain the cumulative contribution rate, set the contribution rate threshold, and record the principal component corresponding to the cumulative contribution rate greater than the contribution rate threshold as the key principal component to complete the dimensionality reduction process. The key principal components constitute the final pharmaceutical flow data set.

6. According to claim 5, a method for early warning of medical flow data transmission security based on big data is characterized in that: The S4 comprises the following steps: S41, setting the kernel function of the support vector machine to a Gaussian kernel function, the Gaussian kernel function including a kernel parameter, establishing an objective function according to the input sample, the objective function including a penalty factor; reacquiring a new medicine flow data set, the new medicine flow data set undergoing data preprocessing and principal component analysis, and normalization processing to obtain a medicine flow data sample set, dividing the medicine flow data sample set into a sample training set and a sample test set, inputting the sample training set into the support vector machine, and obtaining an output result error; S42, taking the output result error as the fitness function, using the Parrot optimization algorithm to optimize the kernel parameters and penalty factors in the support vector machine, and introducing the Cauchy mutation to improve the Parrot optimization algorithm to obtain the optimized kernel parameters and penalty factors; S43, the support vector machine uses the optimized kernel parameters and penalty factors, and inputs the sample test set into the support vector machine until the support vector machine converges to obtain a support vector machine early warning model.

7. A method for early warning of medical flow data transmission security based on big data according to claim 6, characterized in that: The S42 comprises the following steps: S421. Assume that there is a parrot population in the search space. Each parrot individual in the parrot population represents a solution of the fitness function. The coordinates of the parrot individual represent the kernel parameter and the penalty factor. Assume that the number of parrot populations is D. The lower and upper limits of the search space are and , represents a random number and , then the calculation formula for the initial position of the e-th parrot in the parrot population is as follows: ； in, represents the initial position of the e-th parrot in the parrot population; In the process of foraging, the current iteration number is set to h, the maximum iteration number is set to H, and the h-th iteration position of the e-th parrot in the parrot population is , represents a random number and , the h-th iteration average position of the parrot population is , the best position of the h-th iteration of the parrot population is , Lévy flight is introduced to describe the flight of parrots, and the Lévy flight coefficient is , where dim represents the individual dimension of each parrot in the parrot population, and the individual position of the parrot is updated. The calculation formula is as follows: ； in, The h+1th iteration position of the eth parrot in the parrot population; Calculate the fitness function value at the hth iteration, find the current best fitness function value, the parrot individual position corresponding to the current best fitness function value, and the best kernel parameter and penalty factor at the hth iteration; when the current best fitness function value is greater than the fitness function value at the h+1th iteration, use the fitness function value at the h+1th iteration to replace the current best fitness function value; S422, during the communication process, parrots randomly stay at a certain part of the owner, and the vector with all 1s of dimension dim is recorded as , represents a random number and , at this time, the individual position of the parrot is updated again, and the calculation formula is as follows: ； Set the flight probability. When the flight probability is greater than 0.5, the e-th parrot in the parrot population will move away from the parrot population. When the flight probability is less than or equal to 0.5, the e-th parrot in the parrot population will stay in the parrot population. During the fear process, the e-th parrot in the parrot population chooses to stay away from strangers and get closer to the owner, and the position of the parrot is updated synchronously; during the communication process and the fear process, the Cauchy mutation is introduced, setting Represents a random number that follows the Cauchy distribution, and updates the individual position of the parrot. The calculation formula is as follows: ； When the current number of iterations reaches the maximum number of iterations, the iteration is stopped and the final position of the e-th parrot in the parrot population is obtained. The coordinates of the final position correspond to the optimized kernel parameters and penalty factors respectively.

8. The method for early warning of medical flow data transmission security based on big data according to claim 7 is characterized in that: The S5 comprises the following steps: S51, after normalizing the final medicine flow data set, input it into a support vector machine early warning model, output the classification result, and add a classification function to the support vector machine early warning model to obtain a medicine flow data transmission security early warning model; S52. In the medicine flow data transmission security warning model, a medicine flow data transmission security warning is performed according to the classification result, a warning result is obtained, and a security warning is completed.

9. A system for implementing the big data-based medicine flow data transmission security early warning method as described in any one of claims 1 to 8, characterized in that: Specifically include: Data cleaning and filling module, data dimension reduction processing module, support vector machine parameter optimization module and data security early warning module; The data cleaning and filling module is used to clean similar repeated data and fill missing data in the original medicine flow data set; The data dimension reduction processing module is used to perform dimension reduction processing on the processed medicine flow data set using principal component analysis; The support vector machine parameter optimization module is used to optimize the kernel parameters and penalty factors in the support vector machine using the parrot optimization algorithm; The data security warning module is used to establish a medicine flow data transmission security warning model and output warning results.

Images