Medical supplies refined management method and system based on big data

Abstract

The present invention discloses a method and system for fine management of medical supplies based on big data, and the method includes data collection, data preprocessing, medical supplies inventory prediction, effective monitoring of medical supplies, hyperparameter optimization and fine management of medical supplies. The present invention belongs to the field of medical information management technology, specifically a method and system for fine management of medical supplies based on big data. The present invention proposes a strategy for dual management of medical supplies purchase and shipment, and realizes fine control and management of the whole process of medical supplies from purchase to shipment; introduces momentum acceleration optimization technology and adaptive dynamic learning rate, smoothes the update of model weights, improves the stability of the prediction model, and improves the accuracy of medical supplies inventory prediction results; this scheme adopts dynamic cosine chaos initialization method, adaptive dynamic weight mechanism and dynamic selection reverse learning strategy to improve the algorithm for obtaining the optimal parameters of the monitoring model, and obtains the optimal parameter combination.

Description

Technical Field

[0001] The present invention belongs to the technical field of medical information management, and specifically relates to a method and system for fine-grained management of medical supplies based on big data. Background Art

[0002] The refined management method and system of medical supplies based on big data use data analysis to predict demand and optimize inventory to ensure the balance of supply and demand of materials. The system can timely identify insufficient inventory, improve the scientific nature of procurement decisions, and ensure the efficiency and safety of medical services. However, in the existing refined management methods for medical supplies, there are problems such as the inability to accurately control the purchase volume of medical supplies in purchase management and the lax monitoring of the validity period of materials in shipment management, which leads to waste of medical resources and safety hazards; in the existing models suitable for medical supplies inventory prediction, there are technical problems such as slow model convergence speed and difficulty in converging to the global minimum, which leads to inaccurate medical supplies inventory prediction results; in the existing models suitable for effective monitoring of medical supplies, there are problems such as improper settings of built-in parameters, which leads to inaccurate output results of the monitoring model. Summary of the invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a method and system for the refined management of medical supplies based on big data. In view of the problems in the existing refined management methods for medical supplies, such as the inability to accurately control the purchase quantity of medical supplies in purchase management and the lax monitoring of the validity period of materials in shipment management, which cause waste of medical resources and potential safety hazards, this solution creatively proposes a strategy for dual management of purchase and shipment of medical supplies, realizes refined control and management of the entire process of medical supplies from purchase to shipment, improves the efficiency of the use of medical supplies, reduces the waste of medical resources, ensures the stability and continuity of the supply of medical supplies, and provides a strong guarantee for the high-quality development of medical services; in view of the existing methods applicable to the refined management of medical supplies, the present invention provides a method and system for the refined management of medical supplies based on big data, and solves the problem that the purchase quantity of medical supplies in purchase management and the validity period of materials in shipment management cannot be accurately controlled, which causes waste of medical resources and potential safety hazards. In the inventory prediction model, there are technical problems such as slow model convergence and difficulty in converging to the global minimum, which leads to inaccurate medical supplies inventory prediction results. This scheme introduces momentum acceleration optimization technology and adaptive dynamic learning rate to smooth the update of model weights, accelerate the convergence speed of the model, and improve the stability of the prediction model, thereby improving the accuracy of medical supplies inventory prediction results; in view of the existing models suitable for effective monitoring of medical supplies, there are improper built-in parameter settings, which leads to inaccurate output results of the monitoring model. This scheme adopts dynamic cosine chaos initialization method, adaptive dynamic weight mechanism and dynamic selection reverse learning strategy to improve the algorithm for obtaining the optimal parameters of the monitoring model, obtain the optimal parameter combination, and improve the accuracy of the model output results.

[0004] The technical solution adopted by the present invention is as follows: The present invention provides a method for fine management of medical supplies based on big data, the method comprising the following steps:

[0005] Step S1: data collection;

[0006] Step S2: data preprocessing;

[0007] Step S3: Medical supplies inventory forecast;

[0008] Step S4: Effective monitoring of medical supplies;

[0009] Step S5: hyperparameter optimization;

[0010] Step S6: Refined management of medical supplies.

[0011] Furthermore, in step S1, the data collection is specifically to obtain the original data of medical supplies management from the internal information system of the hospital; the original data of medical supplies management includes historical medical supplies inventory data, historical medical supplies consumption data, real-time medical supplies inventory data, real-time basic information of medical supplies and external environment data; the historical medical supplies inventory data contains the historical inventory quantity, usage frequency, inventory fluctuation and loss record of each medical supply; the historical medical supplies consumption data includes consumption quantity, time, purpose and consumption type; the real-time medical supplies inventory data includes inventory quantity, status information and inventory warning information; the real-time medical basic information includes material name, model, specification, manufacturer, production date, expiration date, batch number, storage conditions and supplier information; the external environment data includes supplier's delivery time, delivery frequency, supply stability and seasonal epidemic diseases.

[0012] Furthermore, in step S2, the data preprocessing is used to perform data cleaning, standardization and feature selection on the original data of medical supplies management to obtain preliminary data of medical supplies management; the data cleaning is to process missing values, abnormal values ​​and duplicate values; the standardization is to standardize the data based on the maximum and minimum normalization method; the feature selection is to use the correlation analysis method to screen out the features related to the refined management of medical supplies from the original data of medical supplies management.

[0013] Further, in step S3, the medical supplies inventory prediction is specifically to establish a medical supplies inventory prediction model through momentum acceleration optimization technology and adaptive dynamic learning rate to obtain medical supplies inventory prediction results, including the following steps:

[0014] Step S31: Establishing a medical supplies inventory forecasting model, specifically including the following steps:

[0015] Step S311: Initialize model parameters, specifically, set the number of input layer neurons d, the number of hidden layer neurons q, the number of output layer neurons L, the loss function threshold and the maximum number of iterations;

[0016] Step S312: Model forward calculation, specifically including the following steps:

[0017] Step S3121: Calculate the hidden layer input using the following formula:

[0018] ;

[0019] In the formula, represents the hidden layer input value, represents the weight from the input layer to the hidden layer, Represents the input value of the i-th input neuron;

[0020] Step S3122: Calculate the output of the hidden layer using the following formula:

[0021] ;

[0022] In the formula, represents the output value of the hidden layer neuron, Bias term of hidden layer;

[0023] Step S3123: Calculate the output layer input using the following formula:

[0024] ;

[0025] In the formula, represents the output layer input value, Represents the weight from the hidden layer to the output layer;

[0026] Step S3124: Calculate the output layer output, the formula used is as follows:

[0027] ;

[0028] In the formula, represents the output value of the output layer neuron, The bias term of the output layer;

[0029] Step S313: Design a loss function, the formula used is as follows:

[0030] ;

[0031] In the formula, represents the loss function of the model, represents the actual target value, w represents the set of all weights, represents the regularization parameter;

[0032] Step S314: reverse adjustment of the model, specifically including the following steps:

[0033] Step S3141: Calculate the gradient using the following formula:

[0034] ;

[0035] In the formula, represents the gradient of the loss function with respect to the weight, Indicates the sensitivity of the residual to weight changes, represents the residual value of a single sample, r represents the residual vector of all samples, represents the Jacobian matrix The transposed matrix of

[0036] Step S3142: Update the weight change. Specifically, the weight change is updated by using momentum acceleration optimization technology. The formula used is as follows:

[0037] ;

[0038] ;

[0039] In the formula, Indicates the weight change from the input layer to the hidden layer at the nth iteration, represents the weight change from the hidden layer to the output layer in the nth iteration, represents the momentum factor, which is used to accelerate convergence and reduce oscillation during training. represents the learning rate, represents the gradient of the loss function from the input layer to the hidden layer weights, represents the gradient of the loss function with respect to the weights from the hidden layer to the output layer, Indicates the weight change from the input layer to the hidden layer at the n-1th iteration, Indicates the weight change from the hidden layer to the output layer in the n-1th iteration;

[0040] Step S3143: Obtain the adaptive dynamic learning rate, the formula used is as follows:

[0041] ;

[0042] ;

[0043] In the formula, represents the learning rate at the n+1th iteration, represents the learning rate at the nth iteration, represents the loss function value of the bias term at the n+1th iteration, Represents the loss function value of the bias term at the nth iteration, and represents the upper and lower limits of the learning rate, represents the set of all bias terms;

[0044] Step S3144: Calculate the error using the following formula:

[0045] ;

[0046] ;

[0047] In the formula, represents the error value of the output layer, Represents the error value of the hidden layer;

[0048] Step S315: Update weights and biases using the following formula:

[0049] ;

[0050] In the formula, Represents the weight update value from the input layer to the hidden layer, Represents the weight update value from the hidden layer to the output layer, represents the updated value of the bias term of the hidden layer, Represents the updated value of the bias item of the output layer;

[0051] Step S316: Iteration determination, specifically monitoring the change of the loss function and the current number of iterations. If the loss function is lower than the set threshold, the iterative training is stopped, otherwise the iterative training is continued. If the maximum number of iterations is reached, the iterative training is stopped, otherwise the iterative training is continued;

[0052] Step S32: training a medical supplies inventory prediction model, specifically using the historical medical supplies inventory data and external environment data in the preliminary medical supplies management data as input data to train the medical supplies inventory prediction model to obtain a trained medical supplies inventory prediction model;

[0053] Step S33: Medical supplies inventory forecasting, specifically, using the real-time medical supplies inventory data and external environmental factors in the preliminary medical supplies management data as input data of the trained medical supplies inventory forecasting model to obtain medical supplies inventory forecasting results, wherein the medical supplies inventory forecasting results specifically refer to the material demand for various types of medical supplies in the future.

[0054] Further, in step S4, the monitoring of the validity period of medical supplies is specifically carried out by establishing a monitoring model for the validity period of medical supplies and training the monitoring model, and taking the real-time basic information of medical supplies in the preliminary data of medical supplies management and the medical supplies inventory prediction results as input data of the trained monitoring model to obtain the monitoring results of the validity period of medical supplies; the steps include:

[0055] Step S41: Establishing a medical supplies validity monitoring model, specifically including the following steps:

[0056] Step S411: Initialize model parameters, set the number of model hidden units, input vector dimension and output vector dimension;

[0057] Step S412: Forget gate calculation process, the formula used is as follows:

[0058] ;

[0059] In the formula, f t represents the output of the forget gate; X t Indicates input data; H t-1 Indicates the hidden state at the last moment; W xf and W hf They represent the weight matrices corresponding to the forget gate processing input data and processing hidden states, respectively. represents the bias parameter of the forget gate, represents the sigmoid function, represents the Hadamard product;

[0060] Step S413: Input gate calculation process, the formula used is as follows:

[0061] ;

[0062] ;

[0063] ;

[0064] In the formula, i t represents the output of the input gate, W xi and W hi They represent the weight matrices corresponding to the input gate processing input data and processing hidden states, respectively, and b i represents the bias parameter of the input gate, c t It represents the cell state at the current moment. represents the candidate value of the new cell state, c t-1 Indicates the cell state at the last moment, W xc and W hcThey represent the weight matrices for processing input data and hidden states of corresponding cell states, and b c The bias parameter representing the cell state, represents element-wise multiplication, represents the hyperbolic tangent function;

[0065] Step S414: Output gate calculation process, the formula used is as follows:

[0066] ;

[0067] In the formula, o t represents the output of the output gate, W xo and W ho They represent the weight matrices corresponding to the output gate processing input data and processing hidden states respectively; b o Represents the bias parameter of the output gate;

[0068] Step S415: hidden state calculation process, the formula used is as follows:

[0069] ;

[0070] In the formula, Indicates the hidden state at the current moment;

[0071] Step S416: Get the final output value, the formula used is as follows:

[0072] ;

[0073] In the formula, Represents the output value of the monitoring model, represents the weight matrix corresponding to the output, represents the activation function, Represents the bias parameter of the output layer;

[0074] Step S42: training a medical supplies validity period monitoring model, specifically using the historical medical supplies consumption data in the preliminary medical supplies management data as input data to train the medical supplies validity period monitoring model, thereby obtaining a trained medical supplies validity period monitoring model;

[0075] Step S43: Monitoring the validity period of medical supplies, specifically using the real-time basic information of medical supplies in the preliminary data of medical supplies management and the medical supplies inventory forecast results as input data of the trained medical supplies validity period monitoring model to obtain medical supplies validity period monitoring results, and the medical supplies validity period monitoring results specifically are the expiration risk of current medical supplies in the future.

[0076] Furthermore, in step S5, the hyperparameter optimization specifically includes the following steps:

[0077] Step S51: Initialize parameters, specifically by constructing initial algorithm parameters; the initial algorithm parameters include population size N, search tank ratio P N , observe the tank ratio S N , maximum number of iterations T, warning value R 2 and safety threshold ST;

[0078] Step S52: Initializing the population, specifically using a dynamic cosine chaos initialization method to initialize the tank group, including the following steps:

[0079] Step S521: Generate N chaotic sequence values ​​and initialize the random value z 0 , the formula used is as follows:

[0080] ;

[0081] In the formula, represents the i-th chaotic sequence value of the d-th dimension, represents the i+1th chaotic sequence value of the dth dimension, and round represents a random number between 0 and 1;

[0082] Step S522: The chaotic sequence value is converted to search individuals. The formula is as follows:

[0083] ;

[0084] In the formula, represents the initialization position of the dth dimension of the i-th individual, lb d andub d Respectively represent the search lower and upper limits of dimension d;

[0085] Step S53: Population division: Calculate the individual fitness value f in the population i , the performance of the medical supplies validity monitoring model established based on the individual position is used as the individual fitness value, the individuals are sorted from best to worst according to the fitness value, and the search tanks and auxiliary tanks are divided; the individual with the best fitness f is found b and its position X b , and the individual f with the worst fitness w and its position X w ;

[0086] Step S54: Obtain the weight factor, specifically, obtain the weight factor through an adaptive dynamic weight mechanism, and the formula used is as follows:

[0087] ;

[0088] In the formula, ω(t) represents the adaptive weight factor, t represents the number of iterations, Represents a random number within the range [0, 1];

[0089] Step S55: Update the search tank position, using the following formula:

[0090] ;

[0091] In the formula, represents the position of the i-th individual in the d-th dimension of the tank group in the (t + 1)-th generation, represents the position of the i-th individual in the d-th dimension of the tank group in the t-th generation, represents the corresponding position of the best fitness in the current iteration, and represent random numbers within the range [0, 1], R 2 <ST indicates that there are no enemies nearby and it is in a safe position, R 2 ≥ST indicates that there are enemies nearby and it is necessary to change to another area, Q represents a random number subject to a normal distribution, represents the position of a random individual in the population, represents the smoothing term, which is a constant;

[0092] Step S56: Update the auxiliary tank position, specifically by updating the position through the ground-hugging flight strategy; the formula used is as follows:

[0093] ;

[0094] In the formula, i represents the index of the search individual, n represents the number of individuals, represents the corresponding position of the best fitness value in the (t + 1)-th iteration of the aircraft group, represents the corresponding position of the worst fitness in the current iteration, A + represents the dimension pseudo-inverse matrix, and the elements are randomly assigned 1 or -1; represents a random number uniformly distributed within the range [0, 1], represents the position scaling factor, which is a fixed parameter;

[0095] Step S57: Update the observation tank position, randomly select 8% of the individuals in the aircraft group as observation tanks, and perform position update; the formula used is as follows:

[0096] ;

[0097] In the formula, represents a standard normal distribution random number with a mean of 0 and a variance of 1, and K represents a random number uniformly distributed in the interval [-1, 1];

[0098] Step S58: Individual mutation, recalculate the fitness value fi of each individual, and perform individual mutation by dynamically selecting the reverse learning strategy; the formula used is as follows:

[0099] ;

[0100] ;

[0101] ;

[0102] In the formula, represents the position of the search individual after mutation in the t+1th iteration, represents the position of the optimal individual after mutation in the tth iteration, represents a random variable that satisfies the Cauchy distribution, and represent the maximum and minimum values ​​of the solution space respectively, , and Represents a random number in the range (0,1], represents the probability of selection;

[0103] Step S59: search determination, specifically, by constructing a search termination condition, searching and determining the optimal individual position, and obtaining the optimal individual position data setting;

[0104] The search termination conditions include threshold termination and iteration termination;

[0105] The threshold termination is specifically to set the fitness threshold, when the individual fitness value f i When it is above the fitness threshold, the hyperparameter search is completed;

[0106] The iteration termination specifically refers to terminating the iteration and obtaining the optimal individual position when the maximum number of iterations is reached;

[0107] The optimal individual position is specifically a parameter of a medical supplies validity period monitoring model.

[0108] Furthermore, in step S6, the medical supplies are finely managed, specifically determining a reasonable purchase quantity of medical supplies based on the output results of the medical supplies inventory prediction model, and ensuring that the medical supplies are shipped within the validity period based on the output results of the materials validity period monitoring model, and giving priority to medical supplies with a high risk of expiration. By combining the two output results, the entire process of medical supplies from purchase to shipment can be finely controlled and managed, thereby improving the efficiency of the use of medical supplies, reducing waste, and maximizing the value of each piece of medical supplies, thereby ensuring the stability and continuity of the supply of medical supplies.

[0109] The technical solution adopted by the present invention is as follows: the medical supplies refined management system based on big data provided by the present invention includes a data acquisition module, a data preprocessing module, a medical supplies inventory prediction module, a medical supplies effective monitoring module, a hyperparameter optimization module and a medical supplies refined management module;

[0110] The data acquisition module obtains raw medical material management data by collecting historical medical material inventory data, historical medical material consumption data, real-time medical material inventory data, real-time basic medical material information and external environment data from the hospital's internal information system, and sends the raw medical material management data to the data preprocessing module;

[0111] The data preprocessing module receives the data sent by the data acquisition module, and performs data cleaning, standardization and feature selection on the raw data of medical material management to obtain preliminary data of medical material management, and sends the preliminary data of medical material management to the medical material inventory prediction module and the medical material effective monitoring module;

[0112] The medical supplies inventory prediction module receives the preliminary medical supplies management data sent by the data preprocessing module, establishes a medical supplies inventory prediction model to predict the future medical supplies inventory, obtains the medical supplies inventory prediction result, and sends the medical supplies inventory prediction result to the medical supplies effective monitoring module and the medical supplies refined management module;

[0113] The medical supplies validity monitoring module receives the preliminary medical supplies management data sent by the data preprocessing module and the medical supplies inventory prediction results sent by the medical supplies inventory prediction module, and establishes a medical supplies validity period monitoring model to monitor the validity period of medical supplies, obtains medical supplies validity period monitoring results, and sends the medical supplies validity period monitoring results to the medical supplies refined management module;

[0114] The hyperparameter optimization module obtains the optimal solution combination of the medical supplies validity monitoring model parameters by establishing a search algorithm, and sends the optimal solution combination of the medical supplies validity monitoring model parameters to the medical supplies validity monitoring module;

[0115] The medical supplies refined management module receives the output results sent by the medical supplies inventory prediction module and the medical supplies effective monitoring module, and through the combination of the two output results, it can achieve refined control and management of the entire process of medical supplies from purchase to shipment.

[0116] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0117] (1) In view of the problems in the existing refined management methods for medical supplies, such as the inability to accurately control the purchase quantity of medical supplies in purchase management and the lax monitoring of the validity period of materials in shipment management, which leads to waste of medical resources and safety hazards, this plan creatively proposes a strategy of dual management of purchase and shipment of medical supplies, realizing refined control and management of the entire process of medical supplies from purchase to shipment, improving the efficiency of the use of medical supplies, reducing waste, ensuring the stability and continuity of the supply of medical supplies, and providing a strong guarantee for the high-quality development of medical services.

[0118] (2) In view of the technical problems of slow model convergence and difficulty in converging to the global minimum in the existing models suitable for medical supplies inventory forecasting, which leads to inaccurate medical supplies inventory forecasting results, this scheme introduces momentum acceleration optimization technology and adaptive dynamic learning rate to smooth the update of model weights, accelerate the convergence speed of the model, and improve the stability of the forecasting model, thereby improving the accuracy of medical supplies inventory forecasting results.

[0119] (3) In view of the fact that the existing models suitable for effective monitoring of medical supplies have improper built-in parameter settings, which leads to inaccurate output results of the monitoring model, this scheme adopts dynamic cosine chaos initialization method, adaptive dynamic weight mechanism and dynamic selection reverse learning strategy to improve the algorithm for obtaining the optimal parameters of the monitoring model, obtain the optimal parameter combination, and improve the accuracy of the model output results. BRIEF DESCRIPTION OF THE DRAWINGS

[0120] Figure 1 A schematic diagram of the process of the medical supplies refined management method based on big data provided by the present invention;

[0121] Figure 2 A schematic diagram of the modules of the medical supplies refined management system based on big data provided by the present invention;

[0122] Figure 3 This is a schematic diagram of the process of predicting the inventory of medical supplies in step S3;

[0123] Figure 4 This is a schematic diagram of the process of effective monitoring of medical supplies in step S4;

[0124] Figure 5 A schematic diagram of the process of hyperparameter optimization in step S5;

[0125] Figure 6 A schematic diagram of the process of establishing a medical supplies inventory forecasting model in step S31;

[0126] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0127] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0128] In the description of the present invention, it is necessary to understand that terms such as “upper”, “lower”, “front”, “back”, “left”, “right”, “top”, “bottom”, “inside” and “outside” indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the referred system or element must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as a limitation on the present invention.

[0129] Example 1, see Figure 1 The present invention provides a method for fine management of medical supplies based on big data, the method comprising the following steps:

[0130] Step S1: data collection, specifically, obtaining original medical material management data by collecting historical medical material inventory data, historical medical material consumption data, real-time medical material inventory data, real-time basic information of medical materials and external environment data;

[0131] Step S2: data preprocessing, specifically, performing data cleaning, standardization processing and feature selection on the raw data of medical material management to obtain preliminary data of medical material management;

[0132] Step S3: medical supplies inventory forecasting, specifically forecasting the future medical supplies inventory by establishing a medical supplies inventory forecasting model to obtain a medical supplies inventory forecasting result;

[0133] Step S4: monitoring the validity of medical supplies, specifically monitoring the validity period of medical supplies by establishing a medical supplies validity period monitoring model to obtain the medical supplies validity period monitoring results;

[0134] Step S5: Hyperparameter optimization, specifically, obtaining the optimal solution combination of parameters of the medical supplies validity period monitoring model by establishing a search algorithm;

[0135] Step S6: Fine-grained management of medical supplies, specifically, through the output results of the medical supplies inventory prediction model and the medical supplies effective monitoring model, fine-grained control and management of the entire process of medical supplies from purchase to shipment can be achieved.

[0136] Example 2, see Figure 1 and Figure 2 This embodiment is based on the above embodiment. In step S1, the data collection is specifically to obtain the original data of medical material management from the information system within the hospital by collection; the original data of medical material management includes historical medical material inventory data, historical medical material consumption data, real-time medical material inventory data, real-time basic information of medical materials and external environment data; the historical medical material inventory data contains the historical inventory quantity, usage frequency, inventory fluctuation and loss record of each medical material; the historical medical material consumption data includes consumption quantity, time, purpose and consumption type; the real-time medical material inventory data includes inventory quantity, status information and inventory warning information; the real-time medical basic information includes material name, model, specification, manufacturer, production date, expiration date, batch number, storage conditions and supplier information; the external environment data includes supplier's delivery time, delivery frequency, supply stability and seasonal epidemic diseases.

[0137] Example 3, see Figure 1 and Figure 2 This embodiment is based on the above embodiment. In step S2, the data preprocessing is used to perform data cleaning, standardization and feature selection on the original data of medical material management to obtain preliminary data of medical material management; the data cleaning is to process missing values, abnormal values ​​and duplicate values; the standardization is to standardize the data based on the maximum and minimum normalization method; the feature selection is to use the correlation analysis method to screen out the features related to the refined management of medical materials from the original data of medical material management.

[0138] Example 4, see Figure 1 , Figure 2 , Figure 3 and Figure 6 This embodiment is based on the above embodiment. In step S3, the medical supplies inventory forecasting is to establish a medical supplies inventory forecasting model through momentum acceleration optimization technology and adaptive dynamic learning rate to obtain the medical supplies inventory forecasting result, including the following steps:

[0139] Step S31: Establishing a medical supplies inventory forecasting model, specifically including the following steps:

[0140] Step S311: Initialize model parameters, specifically, set the number of input layer neurons d, the number of hidden layer neurons q, the number of output layer neurons L, the loss function threshold and the maximum number of iterations;

[0141] Step S312: Model forward calculation, specifically including the following steps:

[0142] Step S3121: Calculate the hidden layer input using the following formula:

[0143] ;

[0144] In the formula, represents the hidden layer input value, represents the weight from the input layer to the hidden layer, Represents the input value of the i-th input neuron;

[0145] Step S3122: Calculate the output of the hidden layer using the following formula:

[0146] ;

[0147] In the formula, represents the output value of the hidden layer neuron, Bias term of hidden layer;

[0148] Step S3123: Calculate the output layer input using the following formula:

[0149] ;

[0150] In the formula, represents the output layer input value, Represents the weight from the hidden layer to the output layer;

[0151] Step S3124: Calculate the output layer output; the formula used is as follows:

[0152] ;

[0153] In the formula, represents the output value of the output layer neuron, The bias term of the output layer;

[0154] Step S313: Design a loss function, the formula used is as follows:

[0155] ;

[0156] In the formula, represents the loss function of the model, represents the actual target value, w represents the set of all weights, represents the regularization parameter;

[0157] Step S314: reverse adjustment of the model, specifically including the following steps:

[0158] Step S3141: Calculate the gradient using the following formula:

[0159] ;

[0160] In the formula, represents the gradient of the loss function with respect to the weight, Indicates the sensitivity of the residual to weight changes, represents the residual value of a single sample, r represents the residual vector of all samples, represents the Jacobian matrix The transposed matrix of

[0161] Step S3142: Update the weight change. Specifically, the weight change is updated by using momentum acceleration optimization technology. The formula used is as follows:

[0162] ;

[0163] ;

[0164] In the formula, Indicates the weight change from the input layer to the hidden layer at the nth iteration, represents the weight change from the hidden layer to the output layer in the nth iteration, represents the momentum factor, which is used to accelerate convergence and reduce oscillation during training. represents the learning rate, represents the gradient of the loss function from the input layer to the hidden layer weights, represents the gradient of the loss function with respect to the weights from the hidden layer to the output layer, Indicates the weight change from the input layer to the hidden layer at the n-1th iteration, Indicates the weight change from the hidden layer to the output layer in the n-1th iteration;

[0165] Step S3143: Obtain the adaptive dynamic learning rate, the formula used is as follows:

[0166] ;

[0167] ;

[0168] In the formula, represents the learning rate at the n+1th iteration, represents the learning rate at the nth iteration, represents the loss function value of the bias term at the n+1th iteration, Represents the loss function value of the bias term at the nth iteration, and represents the upper and lower limits of the learning rate, represents the set of all bias terms;

[0169] Step S3144: Calculate the error using the following formula:

[0170] ;

[0171] ;

[0172] In the formula, represents the error value of the output layer, Represents the error value of the hidden layer;

[0173] Step S315: Update weights and biases using the following formula:

[0174] ;

[0175] In the formula, Represents the weight update value from the input layer to the hidden layer, Represents the weight update value from the hidden layer to the output layer, represents the updated value of the bias term of the hidden layer, Represents the updated value of the bias item of the output layer;

[0176] Step S316: Iteration determination, specifically monitoring the change of the loss function and the current number of iterations. If the loss function is lower than the set threshold, the iterative training is stopped, otherwise the iterative training is continued. If the maximum number of iterations is reached, the iterative training is stopped, otherwise the iterative training is continued;

[0177] Step S32: training a medical supplies inventory prediction model, specifically using the historical medical supplies inventory data and external environment data in the preliminary medical supplies management data as input data to train the medical supplies inventory prediction model to obtain a trained medical supplies inventory prediction model;

[0178] Step S33: Medical supplies inventory forecasting, specifically, using the real-time medical supplies inventory data and external environmental factors in the preliminary medical supplies management data as input data of the trained medical supplies inventory forecasting model to obtain medical supplies inventory forecasting results, wherein the medical supplies inventory forecasting results specifically refer to the material demand for various types of medical supplies in the future.

[0179] By performing the above operations, the existing models suitable for medical supplies inventory forecasting have technical problems such as slow model convergence and difficulty in converging to the global minimum, which leads to inaccurate medical supplies inventory forecasting results. This solution introduces momentum acceleration optimization technology and adaptive dynamic learning rate to smooth the update of model weights, accelerate the convergence speed of the model, and improve the stability of the forecasting model, thereby improving the accuracy of medical supplies inventory forecasting results.

[0180] Example 5, see Figure 1 , Figure 2 and Figure 4This embodiment is based on the above embodiment. In step S4, the monitoring of the validity period of medical supplies is specifically carried out by establishing a monitoring model for the validity period of medical supplies and training the monitoring model. The real-time basic information of medical supplies in the preliminary data of medical supplies management and the forecast results of the inventory of medical supplies are used as input data of the monitoring model after training to obtain the monitoring results of the validity period of medical supplies. The monitoring includes the following steps:

[0181] Step S41: Establishing a medical supplies validity monitoring model, specifically including the following steps:

[0182] Step S411: Initialize model parameters, set the number of model hidden units, input vector dimension and output vector dimension;

[0183] Step S412: Forget gate calculation process, the formula used is as follows:

[0184] ;

[0185] In the formula, f t represents the output of the forget gate; X t Indicates input data; H t-1 Indicates the hidden state at the last moment; W xf and W hf They represent the weight matrices corresponding to the forget gate processing input data and processing hidden states, respectively. represents the bias parameter of the forget gate, represents the sigmoid function, represents the Hadamard product;

[0186] Step S413: Input gate calculation process, the formula used is as follows:

[0187] ;

[0188] ;

[0189] ;

[0190] In the formula, i t represents the output of the input gate, W xi and W hi They represent the weight matrices corresponding to the input gate processing input data and processing hidden states, respectively, and b i represents the bias parameter of the input gate, c t It represents the cell state at the current moment. represents the candidate value of the new cell state, c t-1 Indicates the cell state at the last moment, W xc and W hcRespectively represent the weight matrices corresponding to the cell state processing input data and processing hidden state, b c The bias parameter representing the cell state, represents element-wise multiplication, represents the hyperbolic tangent function;

[0191] Step S414: Output gate calculation process, the formula used is as follows:

[0192] ;

[0193] In the formula, o t represents the output of the output gate, W xo and W ho They represent the weight matrices corresponding to the output gate processing input data and processing hidden states respectively; b o Represents the bias parameter of the output gate;

[0194] Step S415: hidden state calculation process, the formula used is as follows:

[0195] ;

[0196] In the formula, Indicates the hidden state at the current moment;

[0197] Step S416: Get the final output value, the formula used is as follows:

[0198] ;

[0199] In the formula, Represents the output value of the monitoring model, represents the weight matrix corresponding to the output, represents the activation function, Represents the bias parameter of the output layer;

[0200] Step S42: training a medical supplies validity period monitoring model, specifically using the historical medical supplies consumption data in the preliminary medical supplies management data as input data to train the medical supplies validity period monitoring model, thereby obtaining a trained medical supplies validity period monitoring model;

[0201] Step S43: Monitoring the validity period of medical supplies, specifically using the real-time basic information of medical supplies in the preliminary data of medical supplies management and the medical supplies inventory forecast results as input data of the trained medical supplies validity period monitoring model to obtain medical supplies validity period monitoring results, and the medical supplies validity period monitoring results specifically are the expiration risk of current medical supplies in the future.

[0202] Example 6, see Figure 1 , Figure 2 and Figure 5 This embodiment is based on the above embodiment. In step S5, the hyperparameter optimization specifically includes the following steps:

[0203] Step S51: Initialize parameters, specifically by constructing initial algorithm parameters; the initial algorithm parameters include population size N, search tank ratio P N , observe the tank ratio S N , maximum number of iterations T, warning value R 2 and safety threshold ST;

[0204] Step S52: Initializing the population, specifically using a dynamic cosine chaos initialization method to initialize the tank group, including the following steps:

[0205] Step S521: Generate N chaotic sequence values ​​and initialize the random value z 0 , the formula used is as follows:

[0206] ;

[0207] In the formula, represents the i-th chaotic sequence value of the d-th dimension, represents the i+1th chaotic sequence value of the dth dimension, and round represents a random number between 0 and 1;

[0208] Step S522: The chaotic sequence value is converted to search individuals. The formula is as follows:

[0209] ;

[0210] In the formula, represents the initialization position of the dth dimension of the i-th individual, lb d andub d Respectively represent the search lower and upper limits of dimension d;

[0211] Step S53: Population division: Calculate the individual fitness value f in the population i , the performance of the medical supplies validity monitoring model established based on the individual position is used as the individual fitness value, the individuals are sorted from best to worst according to the fitness value, and the search tanks and auxiliary tanks are divided; the individual with the best fitness f is found b and its position X b , and the individual f with the worst fitness w and its position X w ;

[0212] Step S54: Obtain the weight factor, specifically, obtain the weight factor through an adaptive dynamic weight mechanism, and the formula used is as follows:

[0213] ;

[0214] In the formula, ω(t) represents the adaptive weight factor, and t represents the number of iterations. represents a random number within the range of [0, 1];

[0215] Step S55: Update the position of the search tank. The formula used is as follows:

[0216] ;

[0217] In the formula, represents the position of the i-th individual in the d-th dimension of the tank group in the (t + 1)-th generation, represents the position of the i-th individual in the d-th dimension of the tank group in the t-th generation, represents the corresponding position of the best fitness in the current iteration, and represents a random number within the range of [0, 1], R 2 <ST indicates that there is no enemy nearby and it is in a safe position; R 2 ≥ST indicates that there is an enemy nearby and it needs to change to another area. Q represents a random number subject to a normal distribution. represents the position of a random individual in the population, represents the smoothing term, which is a constant;

[0218] Step S56: Update the position of the auxiliary tank. Specifically, update the position through the ground-hugging flight strategy. The formula used is as follows:

[0219] ;

[0220] In the formula, i represents the index of the search individual, and n represents the number of individuals. represents the corresponding position of the best fitness value of the aircraft group in the (t + 1)-th iteration, represents the corresponding position of the worst fitness in the current iteration, A + represents the dimension pseudo-inverse matrix, and the elements are randomly assigned 1 or -1; represents a random number uniformly distributed within the range of [0, 1], represents the position scaling factor, which is a fixed parameter;

[0221] Step S57: Update the position of the observation tank. Randomly select 8% of the individuals in the aircraft group as the observation tanks and perform position updates. The formula used is as follows:

[0222] ;

[0223] In the formula, represents a standard normal distribution random number with mean 0 and variance 1, and K represents a uniformly distributed random number in the interval [−1,1];

[0224] Step S58: Individual mutation, recalculate the fitness value fi of each individual, and perform individual mutation by dynamically selecting the reverse learning strategy; the formula used is as follows:

[0225] ;

[0226] ;

[0227] ;

[0228] In the formula, represents the position of the search individual after mutation in the t+1th iteration, represents the position of the optimal individual after mutation in the tth iteration, represents a random variable that satisfies the Cauchy distribution, and represent the maximum and minimum values ​​of the solution space respectively, , and Represents a random number in the range (0,1], represents the probability of selection;

[0229] Step S59: search determination, specifically, by constructing a search termination condition, searching and determining the optimal individual position, and obtaining the optimal individual position data setting;

[0230] The search termination conditions include threshold termination and iteration termination;

[0231] The threshold termination is specifically to set the fitness threshold, when the individual fitness value f i When it is above the fitness threshold, the hyperparameter search is completed;

[0232] The iteration termination specifically refers to terminating the iteration and obtaining the optimal individual position when the maximum number of iterations is reached;

[0233] The optimal individual position is specifically a parameter of a medical supplies validity period monitoring model.

[0234] By performing the above operations, it is found that the existing models suitable for effective monitoring of medical supplies have improper built-in parameter settings, which leads to inaccurate output results of the monitoring model. This scheme adopts dynamic cosine chaos initialization method, adaptive dynamic weight mechanism and dynamic selection reverse learning strategy to improve the algorithm for obtaining the optimal parameters of the monitoring model, obtain the optimal parameter combination, and improve the accuracy of the model output results.

[0235] Embodiment 7, see Figure 1 and Figure 2 This embodiment is based on the above embodiment. In step S6, the medical supplies are finely managed, specifically determining a reasonable purchase quantity of medical supplies according to the output results of the medical supplies inventory prediction model, and at the same time ensuring that the medical supplies are shipped within the validity period according to the output results of the materials validity period monitoring model, and giving priority to medical supplies with a high risk of expiration. By combining the two output results, the whole process of medical supplies from purchase to shipment can be finely controlled and managed, the use efficiency of medical supplies can be improved, waste can be reduced, and each piece of medical supplies can play the maximum value, ensuring the stability and continuity of the supply of medical supplies.

[0236] By executing the above operations, in view of the problems in the existing refined management methods for medical supplies, such as the inability to accurately control the purchase quantity of medical supplies in purchase management and the lax monitoring of the validity period of materials in shipment management, which causes waste of medical resources and safety hazards, this solution creatively proposes a strategy for dual management of purchase and shipment of medical supplies, realizes refined control and management of the entire process of medical supplies from purchase to shipment, improves the efficiency of the use of medical supplies, reduces waste, ensures the stability and continuity of the supply of medical supplies, and provides a strong guarantee for the high-quality development of medical services.

[0237] Embodiment 8, see Figure 1 and Figure 2 , this embodiment is based on the above embodiment, and the medical supplies refined management system based on big data provided by the present invention includes a data acquisition module, a data preprocessing module, a medical supplies inventory prediction module, a medical supplies effective monitoring module, a hyperparameter optimization module and a medical supplies refined management module;

[0238] The data acquisition module obtains raw medical material management data by collecting historical medical material inventory data, historical medical material consumption data, real-time medical material inventory data, real-time basic medical material information and external environment data from the hospital's internal information system, and sends the raw medical material management data to the data preprocessing module;

[0239] The data preprocessing module receives the data sent by the data acquisition module, and performs data cleaning, standardization and feature selection on the raw data of medical material management to obtain preliminary data of medical material management, and sends the preliminary data of medical material management to the medical material inventory prediction module and the medical material effective monitoring module;

[0240] The medical supplies inventory prediction module receives the preliminary medical supplies management data sent by the data preprocessing module, establishes a medical supplies inventory prediction model to predict the future medical supplies inventory, obtains the medical supplies inventory prediction result, and sends the medical supplies inventory prediction result to the medical supplies effective monitoring module and the medical supplies refined management module;

[0241] The medical supplies validity monitoring module receives the preliminary medical supplies management data sent by the data preprocessing module and the medical supplies inventory prediction results sent by the medical supplies inventory prediction module, and establishes a medical supplies validity period monitoring model to monitor the validity period of medical supplies, obtains medical supplies validity period monitoring results, and sends the medical supplies validity period monitoring results to the medical supplies refined management module;

[0242] The hyperparameter optimization module obtains the optimal solution combination of the medical supplies validity monitoring model parameters by establishing a search algorithm, and sends the optimal solution combination of the medical supplies validity monitoring model parameters to the medical supplies validity monitoring module;

[0243] The medical supplies refined management module receives the output results sent by the medical supplies inventory prediction module and the medical supplies effective monitoring module, and through the combination of the two output results, it can achieve refined control and management of the entire process of medical supplies from purchase to shipment.

[0244] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0245] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

[0246] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. A refined management method for medical supplies based on big data, characterized by: The method comprises the following steps: Step S1: data collection, specifically, obtaining original medical material management data by collecting historical medical material inventory data, historical medical material consumption data, real-time medical material inventory data, real-time basic information of medical materials and external environment data; Step S2: data preprocessing, specifically, performing data cleaning, standardization processing and feature selection on the raw data of medical material management to obtain preliminary data of medical material management; Step S3: medical supplies inventory prediction, specifically, establishing a medical supplies inventory prediction model to predict future medical supplies inventory through momentum acceleration optimization technology and adaptive dynamic learning rate, and obtaining medical supplies inventory prediction results; Step S4: monitoring the validity of medical supplies, specifically monitoring the validity period of medical supplies by establishing a medical supplies validity period monitoring model to obtain the medical supplies validity period monitoring results; Step S5: Hyperparameter optimization, specifically using a dynamic cosine chaos initialization method, an adaptive dynamic weight mechanism, and a dynamic selection reverse learning strategy to improve the algorithm for obtaining the optimal parameters of the monitoring model, and obtain the optimal solution combination of the parameters of the medical supplies validity period monitoring model; Step S6: Refined management of medical supplies, specifically, realizing refined control and management of the entire process of medical supplies from purchase to shipment through the output results of the medical supplies inventory prediction model and the medical supplies effective monitoring model; The medical supplies inventory forecast specifically includes the following steps: Step S31: Establishing a medical supplies inventory forecasting model, specifically including the following steps: Step S311: Initialize model parameters, specifically, set the number of input layer neurons d, the number of hidden layer neurons q, the number of output layer neurons L, the loss function threshold and the maximum number of iterations; Step S312: Model forward calculation, specifically including the following steps: Step S3121: Calculate the hidden layer input using the following formula: ； In the formula, represents the hidden layer input value, represents the weight from the input layer to the hidden layer, Represents the input value of the i-th input neuron; Step S3122: Calculate the output of the hidden layer using the following formula: ； In the formula, represents the output value of the hidden layer neuron, Bias term of hidden layer; Step S3123: Calculate the output layer input using the following formula: ； In the formula, represents the output layer input value, Represents the weight from the hidden layer to the output layer; Step S3124: Calculate the output layer output, the formula used is as follows: ； In the formula, represents the output value of the output layer neuron, The bias term of the output layer; Step S313: Design a loss function, the formula used is as follows: ； In the formula, represents the loss function of the model, represents the actual target value, w represents the set of all weights, represents the regularization parameter; Step S314: reverse adjustment of the model, specifically including the following steps: Step S3141: Calculate the gradient using the following formula: ； In the formula, represents the gradient of the loss function with respect to the weight, Indicates the sensitivity of the residual to weight changes, represents the residual value of a single sample, r represents the residual vector of all samples, represents the Jacobian matrix The transposed matrix of Step S3142: Update the weight change. Specifically, the weight change is updated by using momentum acceleration optimization technology. The formula used is as follows: ； ； In the formula, Indicates the weight change from the input layer to the hidden layer at the nth iteration, represents the weight change from the hidden layer to the output layer in the nth iteration, represents the momentum factor, which is used to accelerate convergence and reduce oscillation during training. represents the learning rate, represents the gradient of the loss function from the input layer to the hidden layer weights, represents the gradient of the loss function with respect to the weights from the hidden layer to the output layer, Indicates the weight change from the input layer to the hidden layer at the n-1th iteration, Indicates the weight change from the hidden layer to the output layer in the n-1th iteration; Step S3143: Obtain the adaptive dynamic learning rate, the formula used is as follows: ； ； In the formula, represents the learning rate at the n+1th iteration, represents the learning rate at the nth iteration, represents the loss function value of the bias term at the n+1th iteration, represents the loss function value of the bias term at the nth iteration, and represents the upper and lower limits of the learning rate, represents the set of all bias terms; Step S3144: Calculate the error using the following formula: ； ； In the formula, represents the error value of the output layer, Represents the error value of the hidden layer; Step S315: Update weights and biases using the following formula: ； In the formula, Represents the updated value of the weight from the input layer to the hidden layer, Represents the weight update value from the hidden layer to the output layer, represents the updated value of the bias term of the hidden layer, Represents the updated value of the bias item of the output layer; Step S316: Iteration determination, specifically monitoring the change of the loss function and the current number of iterations. If the loss function is lower than the set threshold, the iterative training is stopped, otherwise the iterative training is continued. If the maximum number of iterations is reached, the iterative training is stopped, otherwise the iterative training is continued; Step S32: training a medical supplies inventory prediction model, specifically using the historical medical supplies inventory data and external environment data in the preliminary medical supplies management data as input data to train the medical supplies inventory prediction model to obtain a trained medical supplies inventory prediction model; Step S33: medical material inventory forecasting, specifically, using the real-time medical material inventory data and external environmental factors in the preliminary medical material management data as input data of the trained medical material inventory forecasting model to obtain a medical material inventory forecasting result, wherein the medical material inventory forecasting result specifically refers to the material demand of various medical materials in the future; The effective monitoring of medical supplies specifically includes the following steps: Step S41: Establishing a medical supplies validity monitoring model, specifically including the following steps: Step S411: Initialize model parameters, set the number of model hidden units, input vector dimension and output vector dimension; Step S412: Forget gate calculation process, the formula used is as follows: ； In the formula, f t Represents the output of the forget gate; Indicates input data; H t-1 Indicates the hidden state at the last moment; W xf and W hf They represent the weight matrices corresponding to the forget gate processing input data and processing hidden states, respectively. represents the bias parameter of the forget gate, represents the sigmoid function, represents the Hadamard product; Step S413: Input gate calculation process, the formula used is as follows: ； ； ； In the formula, i t represents the output of the input gate, W xi and W hi They represent the weight matrices corresponding to the input gate processing input data and processing hidden states, respectively, and b i represents the bias parameter of the input gate, c t Indicates the current cell state. represents the candidate value of the new cell state, c t-1 Indicates the cell state at the last moment, W xc and W hc They represent the weight matrices for processing input data and hidden states of corresponding cell states, and b c The bias parameter representing the cell state, represents element-wise multiplication, represents the hyperbolic tangent function; Step S414: Output gate calculation process, the formula used is as follows: ； In the formula, o t represents the output of the output gate, W xo and W ho They represent the weight matrices corresponding to the output gate processing input data and processing hidden states respectively; b o Represents the bias parameter of the output gate; Step S415: hidden state calculation process, the formula used is as follows: ； In the formula, Indicates the hidden state at the current moment; Step S416: Get the final output value, the formula used is as follows: ； In the formula, Represents the output value of the monitoring model, represents the weight matrix corresponding to the output, represents the activation function, Represents the bias parameter of the output layer; Step S42: training a medical supplies validity period monitoring model, specifically using the historical medical supplies consumption data in the preliminary medical supplies management data as input data to train the medical supplies validity period monitoring model, thereby obtaining a trained medical supplies validity period monitoring model; Step S43: monitoring the validity period of medical supplies, specifically, using the real-time basic information of medical supplies in the preliminary data of medical supplies management and the medical supplies inventory prediction results as input data of the trained medical supplies validity period monitoring model to obtain the medical supplies validity period monitoring results, and the medical supplies validity period monitoring results are specifically the expiration risk of current medical supplies in the future; The hyperparameter optimization specifically includes the following steps: Step S51: Initialize parameters, specifically by constructing initial algorithm parameters; the initial algorithm parameters include population size N, search tank ratio P N , observe the tank proportion S N , maximum number of iterations T, warning value R2 and safety threshold ST; Step S52: Initializing the population, specifically using a dynamic cosine chaos initialization method to initialize the tank group, including the following steps: Step S521: Generate N chaotic sequence values ​​and initialize the random value z0. The formula used is as follows: ； In the formula, represents the i-th chaotic sequence value of the d-th dimension, represents the i+1th chaotic sequence value of the dth dimension, and round represents a random number between 0 and 1; Step S522: The chaotic sequence value is converted to search individuals. The formula is as follows: ； In the formula, represents the initialization position of the dth dimension of the i-th individual, lb d andub d Respectively represent the search lower and upper limits of dimension d; Step S53: Population division: Calculate the individual fitness value f in the population i , the performance of the medical supplies validity monitoring model established based on the individual position is used as the individual fitness value, the individuals are sorted from best to worst according to the fitness value, and the search tanks and auxiliary tanks are divided; the individual with the best fitness f is found b and its position X b , and the individual f with the worst fitness w and its position X w ; Step S54: Obtain the weight factor, specifically, obtain the weight factor through an adaptive dynamic weight mechanism, and the formula used is as follows: ； In the formula, ω(t) represents the adaptive weight factor, t represents the number of iterations, Represents a random number in the range [0,1]; Step S55: Update the search tank position, the formula used is as follows: ； In the formula, represents the d-th dimensional position of the i-th individual in the tank group at the (t + 1)-th generation, represents the d-th dimensional position of the i-th individual in the tank group at the t-th generation, represents the corresponding position of the best fitness in the current iteration, and represent random numbers within the range of [0, 1]. R2 < ST indicates that there are no enemies nearby and the position is safe. R2 ≥ ST indicates that there are enemies nearby and a different area needs to be changed. Q represents a random number subject to a normal distribution, represents the position of a random individual in the population, represents the smoothing term, which is a constant; Step S56: Update the position of the auxiliary tank, specifically by updating the position through a ground-flying strategy; the formula used is as follows: ； In the formula, i represents the index of the search individual, n represents the number of individuals, Indicates the corresponding position of the best fitness value of the cluster in the t+1th iteration, Indicates the corresponding position of the worst fitness in the current iteration, A + express Dimensional pseudo-inverse matrix, elements are randomly assigned 1 or -1; Represents a uniformly distributed random number in the range [0,1]. Represents the position scaling factor, which is a fixed parameter; Step S57: Update the position of the observation tank, randomly select 8% of the individuals in the fleet as observation tanks, and update the position; the formula used is as follows: ； In the formula, represents a standard normal distribution random number with mean 0 and variance 1, and K represents a uniformly distributed random number in the interval [−1,1]; Step S58: Individual mutation, recalculate the fitness value fi of each individual, and perform individual mutation by dynamically selecting the reverse learning strategy; the formula used is as follows: ； ； ； In the formula, represents the position of the search individual after mutation in the t+1th iteration, represents the position of the optimal individual after mutation in the tth iteration, represents a random variable that satisfies the Cauchy distribution, and represent the maximum and minimum values ​​of the solution space respectively, , and Represents a random number in the range (0,1], represents the probability of selection; Step S59: search determination, specifically, by constructing a search termination condition, searching and determining the optimal individual position, and obtaining the optimal individual position data setting; The search termination conditions include threshold termination and iteration termination; The threshold termination is specifically to set the fitness threshold, when the individual fitness value f i When it is above the fitness threshold, the hyperparameter search is completed; The iteration termination specifically refers to terminating the iteration and obtaining the optimal individual position when the maximum number of iterations is reached; The optimal individual position is specifically a parameter of a medical supplies validity period monitoring model.

2. The method for refined management of medical supplies based on big data according to claim 1, characterized in that: In step S6, the medical supplies are finely managed, specifically determining a reasonable purchase quantity of medical supplies based on the output results of the medical supplies inventory prediction model, and ensuring that medical supplies are shipped within the validity period based on the output results of the materials validity period monitoring model, and giving priority to medical supplies with a high risk of expiration. By combining the two output results, the entire process of medical supplies from purchase to shipment can be finely controlled and managed, thereby improving the efficiency of the use of medical supplies, reducing waste, and maximizing the value of each medical supply, thereby ensuring the stability and continuity of the supply of medical supplies.

3. The method for refined management of medical supplies based on big data according to claim 1, characterized in that: In step S1, the data collection is specifically to obtain the original data of medical material management from the information system within the hospital by collection; the original data of medical material management includes historical medical material inventory data, historical medical material consumption data, real-time medical material inventory data, real-time basic information of medical materials and external environment data; the historical medical material inventory data includes the historical inventory quantity, usage frequency, inventory fluctuation and loss record of each medical material; the historical medical material consumption data includes the consumption quantity, time, purpose and consumption type; the real-time medical material inventory data includes the inventory quantity, status information and inventory warning information; The real-time basic medical information includes material name, model, specification, manufacturer, production date, expiration date, batch number, storage conditions and supplier information; the external environment data includes the supplier's delivery time, delivery frequency, supply stability and seasonal epidemic diseases.

4. The method for refined management of medical supplies based on big data according to claim 1, characterized in that: In step S2, the data preprocessing is used to perform data cleaning, standardization and feature selection on the original data of medical supplies management to obtain preliminary data of medical supplies management; the data cleaning is to process missing values, abnormal values ​​and duplicate values; the standardization is to standardize the data based on the maximum and minimum normalization method; the feature selection is to use the correlation analysis method to screen out the features related to the refined management of medical supplies from the original data of medical supplies management.

5. A medical supplies refined management system based on big data, used to implement the medical supplies refined management method based on big data as described in any one of claims 1 to 4, characterized in that: It includes data acquisition module, data preprocessing module, medical supplies inventory forecasting module, medical supplies effective monitoring module, hyperparameter optimization module and medical supplies refined management module.

6. The medical supplies refined management system based on big data according to claim 5 is characterized by: The data acquisition module obtains raw data of medical material management by collecting historical medical material inventory data, historical medical material consumption data, real-time medical material inventory data, real-time basic information of medical materials and external environment data from the hospital's internal information system, and sends the raw data of medical material management to the data preprocessing module; The data preprocessing module receives the data sent by the data acquisition module, and performs data cleaning, standardization and feature selection on the raw data of medical material management to obtain preliminary data of medical material management, and sends the preliminary data of medical material management to the medical material inventory prediction module and the medical material effective monitoring module; The medical supplies inventory prediction module receives the preliminary medical supplies management data sent by the data preprocessing module, establishes a medical supplies inventory prediction model to predict the future medical supplies inventory, obtains the medical supplies inventory prediction result, and sends the medical supplies inventory prediction result to the medical supplies effective monitoring module and the medical supplies refined management module; The medical supplies validity monitoring module receives the preliminary medical supplies management data sent by the data preprocessing module and the medical supplies inventory prediction results sent by the medical supplies inventory prediction module, and establishes a medical supplies validity period monitoring model to monitor the validity period of medical supplies, obtains medical supplies validity period monitoring results, and sends the medical supplies validity period monitoring results to the medical supplies refined management module; The hyperparameter optimization module obtains the optimal solution combination of the medical supplies validity monitoring model parameters by establishing a search algorithm, and sends the optimal solution combination of the medical supplies validity monitoring model parameters to the medical supplies validity monitoring module; The medical supplies refined management module receives the output results sent by the medical supplies inventory prediction module and the medical supplies effective monitoring module, and through the combination of the two output results, it can achieve refined control and management of the entire process of medical supplies from purchase to shipment.

Images