Security risk management platform for multi-point collaborative analysis

Abstract

The invention discloses a security risk management platform for multi-point collaborative analysis, and relates to the technical field of data management, and the platform comprises a storage system interacting with a target drug storage bin and extracting drug storage distribution data; generating a plurality of drug feature sets; acquiring a plurality of storage monitoring index sets; determining a first risk identification result; determining a second risk identification result; and taking the plurality of first risk drugs, the plurality of first risk coefficients, the plurality of risk areas and the plurality of second risk coefficients as drug storage risk identification results. The technical problems that in the prior art, due to the fact that the drug storage safety risk management monitoring dimension is single, risk identification lags behind, multi-source data collaborative analysis is difficult to achieve, potential risks in the drug storage process cannot be accurately predicted, and drug storage safety is insufficient are solved. The technical effects of drug storage risk real-time dynamic monitoring, multi-dimensional collaborative early warning and intelligent accurate management are achieved.

Description

Technical Field

[0001] This application relates to the field of data management technology, specifically to a security risk management platform for multi-point collaborative analysis. Background Technology

[0002] The safe storage and effective management of drugs are not only related to their efficacy and stability, but also directly affect patient medication safety. In actual drug storage management, due to the wide variety of drugs and different storage conditions (such as different environmental requirements for temperature, humidity, and light), coupled with the possibility that different drugs may have similar appearances and names, risks such as confusion, deterioration, or substandard storage environments can easily occur during drug storage. Traditional drug storage management mainly relies on manual inspections and simple information records, which makes it difficult to achieve real-time and accurate risk monitoring and early warning. It is easy for safety hazards to be caused by human negligence or environmental fluctuations. However, drug storage management systems are often limited to single-dimensional monitoring, such as focusing only on environmental parameters or relying solely on manual verification of drug information. They lack multi-dimensional and collaborative risk analysis capabilities and are unable to comprehensively identify potential storage risks.

[0003] Therefore, current technologies related to drug storage risk identification suffer from several problems, including a single dimension for monitoring drug storage safety risks, delayed risk identification, difficulty in achieving collaborative analysis of multi-source data, and consequently, an inability to accurately predict potential risks during drug storage and insufficient drug storage safety. Summary of the Invention

[0004] This application provides a multi-point collaborative analysis security risk management platform, which adopts integrated intelligent sensing arrays, big data analysis technology and artificial intelligence algorithms to solve the technical problems of existing technologies, such as single monitoring dimensions, lagging risk identification, difficulty in achieving multi-source data collaborative analysis, and thus inability to accurately predict potential risks in the drug storage process and insufficient drug storage security. It achieves the technical effects of real-time dynamic monitoring of drug storage risks, multi-dimensional collaborative early warning, and intelligent and precise management.

[0005] This application provides a security risk management platform for multi-point collaborative analysis. The platform includes: a drug storage distribution data extraction module, used to interact with the storage system of a target drug storage warehouse to extract drug storage distribution data, wherein the drug storage distribution data includes multiple drug sets corresponding to multiple drug storage areas and multiple drug storage location identifiers; a drug feature set generation module, used to extract features from the multiple drug sets according to preset risk characteristics to generate multiple drug feature sets, wherein the multiple drug feature sets include multiple drug name sets, multiple outer packaging image sets, and multiple storage environment requirement sets; and a storage monitoring indicator set acquisition module, used to interact with multiple intelligent sensing arrays in the multiple drug storage areas to acquire multiple storage monitoring indicator sets, wherein the multiple storage monitoring indicator sets include multiple storage... The system includes: temperature, multiple storage humidity levels, and multiple storage radiation levels; a first risk identification result determination module, used to identify risks in the multiple storage monitoring indicator sets based on the multiple storage environment requirement sets, and determine a first risk identification result, wherein the first risk identification result includes multiple first-risk drugs and multiple first risk coefficients; a second risk identification result determination module, used to identify regional confusion risks based on the multiple drug name sets, multiple outer packaging image sets, the multiple drug sets, and the multiple drug storage location identifiers, and determine a second risk identification result, wherein the second risk identification result includes multiple risk areas and multiple second risk coefficients; and a safety risk management module, used to perform safety risk management based on the multiple first-risk drugs, multiple first risk coefficients, multiple risk areas, and multiple second risk coefficients.

[0006] In a possible implementation, the drug feature set generation module also performs the following processing: the preset risk features include drug name, outer packaging image, and storage environment requirements.

[0007] In a possible implementation, the module for acquiring the storage monitoring index set further performs the following processing: data acquisition is performed on the multiple intelligent sensing arrays within a preset verification window to generate multiple verification temperature sequence sets, multiple verification humidity sequence sets, and multiple verification radiation quantity sequence sets; the multiple verification temperature sequence sets, multiple verification humidity sequence sets, and multiple verification radiation quantity sequence sets are normalized to obtain multiple normalized verification temperature sequence sets, multiple normalized verification humidity sequence sets, and multiple normalized verification radiation quantity sequence sets; the multiple normalized verification temperature sequence sets, multiple normalized verification humidity sequence sets, and multiple normalized verification radiation quantity sequence sets are window-aligned and synchronously distributed to multiple verification fitting branches to generate multiple verification fitting spaces, wherein the multiple verification fitting spaces include multiple verification fitting particle sets; linear fitting is performed on the multiple verification fitting particle sets within the multiple verification fitting spaces; the feasibility of the multiple intelligent sensing arrays is verified based on the linear fitting results; when verification is successful, the multiple intelligent sensing arrays are used to monitor the multiple drug storage areas to generate the multiple storage monitoring index sets, wherein the linear fitting results include multiple fitted lines.

[0008] In a possible implementation, the module for storing and acquiring monitoring index sets further performs the following processing: randomly generating multiple initial straight lines within the multiple verification fitting spaces, counting the number of verification fitting particles whose distance to the multiple initial straight lines is a preset bandwidth, and generating multiple initial fitting values; iteratively updating the multiple initial straight lines within the multiple verification fitting spaces according to the preset bandwidth and the multiple initial fitting values ​​to obtain multiple iterative straight lines; iterating the multiple iterative straight lines within the multiple verification fitting spaces according to a preset angle step size to obtain multiple angle iterative straight lines; determining whether the difference between the fitting values ​​of the multiple angle iterative straight lines and the multiple iterative straight lines satisfies a preset fitting value difference, and if so, using the multiple angle iterative straight lines as multiple fitting straight lines.

[0009] In a possible implementation, the module for storing and acquiring monitoring indicator sets further performs the following processing: extracting multiple slopes of the multiple fitted lines; determining whether the multiple slopes are less than or equal to a preset slope threshold; if so, authentication is successful; if not, authentication fails, a risk perception instruction is generated, and the risk perception instruction is sent to the staff.

[0010] In a possible implementation, the first risk identification result determination module further performs the following processing: calculating demand differences based on the multiple sets of storage environment requirements and the multiple sets of storage monitoring indicators to obtain multiple sets of storage demand differences; determining whether each of the multiple sets of storage demand differences is greater than or equal to a multiple set of preset tolerance differences; if so, obtaining multiple first-risk drugs; calculating the difference between the multiple sets of storage demand differences corresponding to the multiple first-risk drugs and the multiple set of preset tolerance differences, and comparing the calculation result with the multiple set of preset tolerance differences to obtain the multiple first-risk coefficients; and using the multiple first-risk drugs and the multiple first-risk coefficients as the first risk identification result.

[0011] In a possible implementation, the second risk identification result determination module further performs the following processing: randomly extracting one drug from each of the multiple drug sets in the multiple drug storage areas as a first nearest-neighbor central drug, obtaining multiple first nearest-neighbor central drugs; extracting multiple first drug neighborhoods of the multiple first nearest-neighbor central drugs based on the multiple drug storage location identifiers; performing similarity identification on the multiple first nearest-neighbor central drugs and the multiple first drug neighborhoods according to the multiple drug name sets and the multiple outer packaging image sets, obtaining multiple first nearest-neighbor similarity coefficients; performing nearest-neighbor similarity identification on the multiple drug sets, obtaining multiple nearest-neighbor similarity coefficient sets; and using a preset similarity threshold to perform regional confusion risk identification on the multiple nearest-neighbor similarity coefficient sets, determining the second risk identification result.

[0012] In a possible implementation, a similarity recognition function is constructed, wherein the similarity recognition function is: ; in, The first nearest neighbor similarity coefficient, Let n be the number of drugs in the first drug neighborhood, where n is a positive integer greater than or equal to 1. The name of the drug at the first nearest neighbor center. Let be the name of the i-th drug in the first drug neighborhood. Image of the outer packaging of the drug at the first nearest neighbor center. This is the outer packaging image of the i-th drug in the neighborhood of the first drug. This represents the weight given to the similarity of drug names when performing nearest neighbor similarity identification. The similarity of the outer packaging images is weighted in the nearest neighbor similarity recognition process. Based on the multiple sets of drug names and multiple sets of outer packaging images, the multiple first nearest neighbor center drugs and the multiple first drug neighborhoods are mapped and matched to obtain multiple first nearest neighbor center drug names, multiple first nearest neighbor center drug outer packaging images, multiple first drug neighborhood name sets, and multiple first drug neighborhood outer packaging image sets. The multiple first nearest neighbor center drug names, multiple first nearest neighbor center drug outer packaging images, multiple first drug neighborhood name sets, and multiple first drug neighborhood outer packaging image sets are respectively input into the similarity recognition function to obtain the multiple first nearest neighbor similarity coefficients.

[0013] This application proposes a multi-point collaborative analysis-based safety risk management platform to interact with the storage system of a target drug storage warehouse and extract drug storage distribution data. It then extracts features from multiple drug sets according to preset risk characteristics, generating multiple drug feature sets. Furthermore, it interacts with multiple intelligent sensing arrays in multiple drug storage areas to obtain multiple sets of storage monitoring indicators. Based on multiple sets of storage environment requirements, it identifies risks in these multiple sets of storage monitoring indicators to determine a first risk identification result. Finally, it identifies regional confusion risks based on multiple sets of drug names, multiple sets of outer packaging images, multiple drug sets, and multiple drug storage location identifiers to determine a second risk identification result. The platform uses multiple first-risk drugs, multiple first-risk coefficients, multiple risk areas, and multiple second-risk coefficients as the drug storage risk identification result. This addresses the technical problems in existing technologies, such as single-dimensional monitoring of drug storage safety risks, delayed risk identification, and difficulty in achieving multi-source data collaborative analysis, which leads to inaccurate prediction of potential risks during drug storage and insufficient drug storage safety. The platform achieves the technical effects of real-time dynamic monitoring of drug storage risks, multi-dimensional collaborative early warning, and intelligent and precise management. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings of the embodiments of this disclosure will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the platform according to the embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0015] Figure 1 This is a schematic diagram of the security risk management platform structure for multi-point collaborative analysis provided in an embodiment of this application; Figure 2 This is a schematic diagram illustrating the execution process of the module for obtaining the set of monitoring indicators in the security risk management platform for multi-point collaborative analysis provided in this application embodiment.

[0016] Explanation of reference numerals in the attached diagram: Drug storage and distribution data extraction module 10, drug feature set generation module 20, storage monitoring indicator set acquisition module 30, first risk identification result determination module 40, second risk identification result determination module 50, safety risk management module 60. Detailed Implementation

[0017] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application.

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application will be provided in conjunction with the accompanying drawings. The described embodiments should not be considered as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] In the following description, references to "some embodiments" describe a subset of all possible embodiments. However, it is understood that "some embodiments" can be the same or different subsets of all possible embodiments and can be combined with each other without conflict. The terms "first" and "second" are used merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, platform, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only.

[0020] This application provides a security risk management platform for multi-point collaborative analysis, such as... Figure 1 As shown, the platform includes: The drug storage distribution data extraction module 10 is used to interact with the storage system of the target drug storage warehouse to extract drug storage distribution data. This data includes multiple drug sets corresponding to multiple drug storage areas and multiple drug storage location identifiers. The storage system of the target drug storage warehouse is a comprehensive system integrating modern information technology, automation technology, and warehouse management technology. It achieves automatic drug storage and retrieval, inventory monitoring, and data analysis through rational planning of drug storage areas, precise identification of drug storage locations, and the use of intelligent methods. The drug storage distribution data mainly refers to detailed information about the distribution of drugs within the storage warehouse, including multiple drug sets corresponding to multiple drug storage areas and multiple drug storage location identifiers. Specifically, a storage area refers to a different spatial area within the storage warehouse divided according to drug type, purpose, and management requirements. Each area may correspond to different environmental conditions (e.g., ...). Temperature, humidity, security level, or access permissions; a drug collection refers to a group of drugs with similar attributes or uses stored in a specific area. For example, drugs can be categorized by type (prescription drugs, over-the-counter drugs, traditional Chinese medicine, Western medicine, etc.) or by use (antibiotics, analgesics, cardiovascular drugs, etc.), as well as psychotropic drugs, narcotics, and other drugs requiring special management. Storage location identifiers are information used to uniquely identify the specific location of each drug within the storage warehouse, including physical location such as shelf number, layer number, column number, and compartment number, and electronic identifiers such as RFID tags and QR codes, which can be quickly read using scanning devices. Below is an example table of drug storage data:

[0021] The drug feature set generation module 20 is used to extract features from the multiple drug sets according to preset risk characteristics, generating multiple drug feature sets. These multiple drug feature sets include multiple sets of drug names, multiple sets of outer packaging images, and multiple sets of storage environment requirements. In the storage system of the target drug storage warehouse, features are extracted from the multiple drug sets according to preset risk characteristics, generating multiple drug feature sets. Specifically, relevant information for each drug set is extracted from the storage system, including drug names, outer packaging images, and storage environment requirements. The collected information is classified, organized, and summarized to form a standardized data structure. Based on preset risk characteristics (such as drug names, outer packaging images, and storage environment requirements), key features are extracted from the organized information, and corresponding feature sets are generated. The generated drug feature sets are stored in a database and managed and queried uniformly through a management system. This results in multiple drug feature sets including multiple sets of drug names, multiple sets of outer packaging images, and multiple sets of storage environment requirements. The drug name set refers to the set of drug names extracted from all drug sets, containing the standard names of all drugs in the storage warehouse. Each drug name is unique and used to identify different drug types. The drug image set helps to quickly identify drug types, facilitating accurate operations during storage, retrieval, and inventory. The outer packaging image set refers to the collection of drug outer packaging images extracted from all drug sets, encompassing images of all drug outer packaging within the storage compartment. These images can be in the form of photographs, scans, or digital models, showcasing the characteristics of the drug's outer packaging. The outer packaging image set helps to quickly identify drug types and conditions through visual recognition, especially when dealing with drugs that look similar or have similar names, significantly reducing the risk of confusion. The storage environment requirement set refers to the collection of drug storage environment requirements extracted from all drug sets, including the specific requirements of each drug for the storage environment, such as temperature range, humidity requirements, light conditions, and whether light protection is necessary. The storage environment requirement set is crucial for ensuring the safety and effectiveness of drugs during storage. By monitoring and adjusting the storage environment in real time, the storage requirements of different drugs can be met, preventing drugs from becoming ineffective or deteriorating due to improper storage.

[0022] The storage monitoring indicator set acquisition module 30 is used to interact with multiple intelligent sensing arrays in the multiple drug storage areas to acquire multiple storage monitoring indicator sets, wherein the multiple storage monitoring indicator sets include multiple storage temperatures, multiple storage humidity levels, and multiple storage radiation levels. An intelligent sensing array is a device or system integrating multiple sensors and data processing technologies, capable of real-time monitoring and recording various environmental parameters within the drug storage area. These sensors may include temperature sensors, humidity sensors, radiation sensors, etc., used to measure key indicators such as temperature, humidity, and radiation levels in the storage area. In storage warehouses, drugs are typically stored in different areas, which may be divided according to the type, purpose, and management requirements of the drugs. Each area is equipped with a corresponding intelligent sensing array to ensure comprehensive monitoring of the drug storage environment within that area. Specifically, the intelligent sensing array collects and records key indicators of the drug storage environment in real time to acquire multiple storage monitoring indicator sets. The monitoring indicators include multiple storage temperatures, multiple storage humidity levels, and multiple storage radiation levels. Multiple storage temperatures refer to the real-time temperature values ​​measured in each storage area, reflecting the current temperature conditions of the storage area, which is crucial for drugs requiring specific temperature storage conditions. Multiple storage humidity levels refer to the real-time humidity values ​​measured in each storage area; humidity is an important factor affecting drug stability and efficacy. For most drugs, radiation levels may not be the primary monitoring indicator, but in certain special cases (such as radiopharmaceuticals or drugs requiring light-protected storage), radiation monitoring is necessary. Radiation levels may refer to light radiation or other forms of radiation, forming multiple storage radiation levels.

[0023] The first risk identification result determination module 40 is used to identify risks in the multiple storage monitoring indicator sets based on the multiple storage environment requirement sets, and determine the first risk identification result, wherein the first risk identification result includes multiple first risk drugs and multiple first risk coefficients. By comparing and analyzing storage environment requirements with real-time monitoring data, drugs with potential risks and their corresponding risk coefficients are identified. Specifically, the risk identification process involves comparing and analyzing the set of storage environment requirements with the set of storage monitoring indicators. The system compares the storage environment requirements of each drug with the actual monitoring data one by one according to a preset algorithm or rule. If a monitoring indicator is found to exceed the range of the drug's storage environment requirements, the drug is considered to have potential risks, resulting in a first risk identification result. This includes multiple first-risk drugs and multiple first-risk coefficients. Multiple first-risk drugs refer to drugs identified as having potential risks during the risk identification process. These drugs may have their quality deteriorated or become ineffective because one or more indicators of the storage environment exceed the range of their storage environment requirements. Multiple first-risk coefficients correspond to each first-risk drug and are used to quantify the degree of risk present by the drug. The calculation of the risk coefficient may be based on multiple factors, such as the degree to which the monitoring indicator exceeds the range, the duration of the exceedance, and the sensitivity of the drug itself. The higher the risk coefficient, the greater the risk of the drug, and the more urgent the measures required.

[0024] The second risk identification result determination module 50 is used to perform regional confusion risk identification based on the multiple drug name sets, multiple outer packaging image sets, multiple drug sets, and multiple drug storage location identifiers, and determine a second risk identification result. The second risk identification result includes multiple risk areas and multiple second risk coefficients. Regional confusion risk identification based on multiple drug name sets, multiple outer packaging image sets, multiple drug sets, and multiple drug storage location identifiers identifies potential confusion risks within storage areas due to similar drug names, similar outer packaging, or improper storage locations, quantifies the degree of these risks, and determines the second risk identification result. Specifically, regional confusion risk identification includes name similarity analysis, outer packaging similarity analysis, and storage location analysis. Name similarity analysis uses text analysis technology to compare names in the drug name sets, identify drugs with similar names or pronunciations, and analyze whether these similarly named drugs are stored in adjacent or easily confused areas. Outer packaging similarity analysis uses image recognition technology to analyze images in the outer packaging image sets, identify drugs with similar or easily confused outer packaging, and, combined with storage location identifiers, assesses whether these drugs with similar outer packaging are stored in easily confused locations. Storage location analysis involves examining drug storage location markings, analyzing the rationality of the storage layout, and identifying any confusion risks due to improper location. Particular attention is paid to the storage locations of frequently accessed, highly similar, or high-risk drugs. Then, the results of name similarity analysis, outer packaging similarity analysis, and storage location analysis are comprehensively evaluated to identify drug combinations and storage areas with confusion risks. This serves as a secondary risk identification result, including multiple risk areas and multiple secondary risk coefficients. Multiple risk areas indicate specific areas or locations within the storage warehouse where confusion risks exist, including multiple drugs with similar names, similar outer packaging, or improper storage locations. Multiple secondary risk coefficients quantify the degree of confusion risk in that area based on the number of drugs with similar names, the degree of similarity in outer packaging, and the rationality of the storage location. Corresponding to each risk area, a higher risk coefficient indicates a greater confusion risk in that area, and the more urgent the measures required.

[0025] The safety risk management module 60 is used to manage safety risks based on the multiple first-risk drugs, multiple first-risk coefficients, multiple risk areas, and multiple second-risk coefficients. The multiple first-risk drugs, multiple first-risk coefficients, multiple risk areas, and multiple second-risk coefficients are used as the drug storage risk identification results. These results provide a comprehensive and quantitative assessment of drug storage risks, encompassing various types of risk information. This helps managers clearly understand which drugs in the storage warehouse pose risks, the degree of risk, and the specific location of the risks. Based on this information, targeted risk response measures can be developed, such as adjusting the storage environment, redesigning the storage layout, and strengthening personnel training, to minimize drug storage risks, ensure patient medication safety, and promptly identify new risk points or risk trends, thereby continuously optimizing storage management strategies and improving storage efficiency and security.

[0026] The multi-point collaborative analysis security risk management platform according to embodiments of the present invention addresses the technical problem of insufficient drug storage security in existing drug storage risk identification methods. This is because the diverse types of drugs and varying storage conditions make it difficult to accurately identify risks at multiple storage locations within a drug storage facility, leading to an inability to predict potential risks during drug storage. The platform achieves intelligent monitoring and risk identification throughout the entire drug storage process, thereby improving drug storage security. The multi-point collaborative analysis security risk management platform includes: a drug storage distribution data extraction module 10, a drug feature set generation module 20, a storage monitoring indicator set acquisition module 30, a first risk identification result determination module 40, a second risk identification result determination module 50, and a security risk management module 60.

[0027] The specific configuration of the drug feature set generation module 20 will be described in detail below. The drug feature set generation module 20 may further include: the preset risk features include drug name, outer packaging image, and storage environment requirements. Pre-defined risk characteristics refer to standard parameters closely related to drug characteristics set in advance to identify and prevent potential risks. These include drug name, outer packaging image, and storage environment requirements. The drug name is the basic identifier of a drug, including the generic name and brand name. When multiple drug names are similar, confusion can easily occur during drug dispensing, preparation, or administration, increasing the risk to patients. The outer packaging image consists of visual elements on the drug packaging, including patterns, colors, and text. Similarity in outer packaging images may lead to misidentification or misuse of drugs during storage and distribution, especially when multiple drugs have similar outer packaging. Storage environment requirements refer to the specific conditions that a drug must meet during storage, such as temperature, humidity, and light. An unsuitable storage environment can directly affect the quality and stability of the drug, leading to drug inactivation or the generation of harmful substances. This not only reduces the drug's efficacy but may also increase the risk to patients. By establishing detailed storage environment standards and strengthening environmental monitoring and regulation, it is possible to ensure that drugs are stored under suitable conditions, reducing the risks caused by unsuitable environments.

[0028] The following will describe in detail the specific configuration of the module 30 for storing and acquiring monitoring indicator sets. For example... Figure 2As shown, the monitoring indicator set acquisition module 30 may further include: collecting data from the multiple intelligent sensing arrays within a preset verification window, generating multiple sets of verification temperature sequences, multiple sets of verification humidity sequences, and multiple sets of verification radiation sequences. The preset verification window is a predefined time period based on the periodicity of environmental changes, the stability of sensor responses, and the needs of data processing, specifically for data collection and verification. Specifically, during the preset verification window, sensor arrays deployed in the drug storage area or related environment continuously monitor and record changes in environmental parameters, including key indicators such as temperature, humidity, and radiation, at a predetermined frequency and accuracy, to reflect the environmental conditions of the storage area. Each array may contain multiple sensors to acquire environmental data from different angles or locations, generating multiple sets of verification temperature sequences, multiple sets of verification humidity sequences, and multiple sets of verification radiation sequences. The verification temperature sequences... The set refers to the sequence of temperature data collected from all intelligent sensing arrays, which is then processed and formed. Each sequence represents the temperature change trend monitored by the corresponding array during the verification window. The verification humidity sequence set is similar to the temperature sequence set, which is the sequence of humidity data collected from all intelligent sensing arrays, which is then processed and formed. Each sequence reflects the humidity change monitored by the corresponding array during the verification window. The verification radiation sequence set is the sequence of radiation data (which may include different types of radiation such as ultraviolet and infrared) collected from all intelligent sensing arrays, which is then processed and formed. It provides detailed information about the radiation environment of the storage area.

[0029] The monitoring indicator set acquisition module 30 further includes normalizing the multiple verification temperature sequence sets, multiple verification humidity sequence sets, and multiple verification radiation sequence sets to obtain multiple normalized verification temperature sequence sets, multiple normalized verification humidity sequence sets, and multiple normalized verification radiation sequence sets. Linear normalization (also known as min-max normalization) and Z-score normalization (also known as standard deviation normalization) are used to normalize the multiple verification temperature sequence sets, multiple verification humidity sequence sets, and multiple verification radiation sequence sets respectively. This involves scaling the original data to the [0, 1] interval or converting the data to a distribution with a mean of 0 and a standard deviation of 1, thereby eliminating dimensional and distributional differences between different data sets. This allows data from different sources or of different types to be compared and analyzed within the same framework, resulting in multiple normalized verification temperature sequence sets, multiple normalized verification humidity sequence sets, and multiple normalized verification radiation sequence sets.

[0030] The monitoring indicator set acquisition module 30 further includes window alignment of the multiple normalized verification temperature sequence sets, multiple normalized verification humidity sequence sets, and multiple normalized verification radiation sequence sets, and synchronously distributing them to multiple verification fitting branches to generate multiple verification fitting spaces, wherein the multiple verification fitting spaces include multiple verification fitting particle sets. Window alignment refers to adjusting the multiple normalized verification temperature sequence sets, multiple normalized verification humidity sequence sets, and multiple normalized verification radiation sequence sets to the same time start and end point, or at least ensuring that they overlap within the analysis time period, in order to eliminate minor time differences (such as network latency, different device startup times, etc.). Specifically, a common time window is determined, containing the effective portion of all sequence data, i.e., deleting data outside the window range, while padding may involve using interpolation or other methods to estimate missing data points. Then, the window-aligned dataset is simultaneously sent to multiple parallel processing units (i.e., verification fitting branches) for further analysis and processing, i.e., using a distributed computing system or parallel processing. The framework manages multiple validation fitting branches, taking the window-aligned dataset as input and sending it to all validation fitting branches simultaneously. Each validation fitting branch independently processes the input data and generates corresponding outputs or results, ultimately producing multiple validation fitting spaces. A validation fitting space is a virtual space used to store and process all data and information related to a specific validation task, providing a clear and orderly environment for data analysis and helping to reduce the complexity and error rate of data processing. Multiple validation fitting spaces include multiple sets of validation fitting particles, which are a set of elements within the validation fitting space used to represent data points, hypotheses, or model parameters. These sets may represent different subsets of data, combinations of model parameters, or fitting results.

[0031] The storage monitoring index set acquisition module 30 further includes performing linear fitting on the multiple verification fitting particle sets within the multiple verification fitting spaces, performing feasibility verification on the multiple intelligent sensing arrays based on the linear fitting results, and when the verification is successful, using the multiple intelligent sensing arrays to monitor the multiple drug storage areas respectively, generating the multiple storage monitoring index sets, wherein the linear fitting results include multiple fitted lines. Linear fitting is used to find the best-fitting straight line from a set of data points (in this scenario, the data points in the validation fit particle set). This line minimizes the sum of squared distances from the data points to corresponding points on the line, i.e., it uses the least squares method for fitting. Specifically, within each validation fit space, the validation fit particle set is preprocessed. Using the least squares method, a linear fitting operation is performed on the validation fit particle set within the selected validation fit space. This minimizes the squared error between the data points and the fitted line to find the best line. For example, statistical measures such as the mean and variance of the data points are calculated, and the parameters of the best line (such as slope and intercept) are solved according to the fitting method. The goodness of fit is evaluated by calculating indicators such as fitting error and goodness of fit. Finally, the feasibility of multiple intelligent sensing arrays is assessed based on the linear fitting results. The certification process specifically involves setting certification standards for the intelligent sensing array based on business needs and data characteristics. These standards include fitting error thresholds and data consistency requirements. The linear fitting results are compared with the certification standards to determine if the intelligent sensing array meets the requirements. If the fitting results of the intelligent sensing array meet the certification standards, the certification is passed. Multiple intelligent sensing arrays are used to monitor indicators in multiple drug storage areas. Based on the management needs of the drug storage areas, the indicators to be monitored are set, such as temperature, humidity, and light intensity. The intelligent sensing arrays collect various indicator data from the storage areas in real time and perform preliminary processing (such as filtering and noise reduction). The processed data is then organized according to time sequence or regional division to form multiple sets of storage monitoring indicators, which contain various indicator data of the drug storage areas at different times or under different conditions.

[0032] The monitoring indicator set acquisition module 30 also includes, as will be described in detail below, the specific configuration of the monitoring indicator set acquisition module 30. The monitoring indicator set acquisition module 30 may further include: randomly generating multiple initial straight lines within the multiple verification fitting spaces, counting the number of verification fitting particles whose distance to the multiple initial straight lines is a preset bandwidth, and generating multiple initial fitting values. Within each verification fitting space, multiple initial straight lines are randomly generated, and a preset bandwidth is set, defining the range of verification fitting particles sufficiently close to the line. Specifically, if the vertical distance of a verification fitting particle to the line is less than or equal to the preset bandwidth, the particle is considered to be within the bandwidth. Then, the distances of all particles in the verification fitting space to the line are calculated, and the number of all verification fitting particles whose distance to the initial straight line is within the preset bandwidth is counted, reflecting the proximity of the initial straight line to the data point set. For each initial straight line, an initial fitting value is generated based on its corresponding number of verification fitting particles (or the transformed number), used in subsequent iterative update processes to find a better fitting line.

[0033] The module 30 for storing and acquiring monitoring index sets further includes iteratively updating the multiple initial straight lines within the multiple verification fitting spaces according to the preset bandwidth and the multiple initial fitting values ​​to obtain multiple iterative straight lines. Using gradient descent (for continuous space optimization) and genetic algorithms (for discrete or complex spaces), multiple initial straight lines are iteratively updated across multiple validation fitting spaces according to a preset bandwidth and multiple initial fitting values. Specifically, based on the preset bandwidth and initial fitting values, an error function (or loss function) is defined to measure the degree of fit between the current straight line and the set of data points. In each iteration, the parameters of the straight line (such as slope, intercept, etc.) are updated according to the current error function value and the rules of the selected optimization algorithm. After each iteration, the distance (or error) from all data points in the validation fitting space to the updated straight line is recalculated, and the iteration effect is evaluated. For example, new fitting values ​​(such as the number of particles falling within the preset bandwidth, the value of the error function, etc.) are calculated and compared with the previous iteration results. If the iteration effect meets the preset stopping conditions (such as the error being less than a certain threshold, the number of iterations reaching the upper limit, etc.), the iteration stops; otherwise, the iteration continues until the stopping conditions are met, ultimately obtaining multiple iterative straight lines that can more accurately describe the linear relationship of the set of data points.

[0034] The module 30 for storing and acquiring monitoring index sets further includes iterating the multiple iterative lines within the multiple verification fitting spaces according to a preset angle step size to obtain multiple angle iterative lines. The preset angle step size is a fixed angle value used to adjust the direction of the lines during the iteration process. A smaller step size means a more refined search, but may require more iterations; a larger step size may cause the search process to skip the optimal solution. Specifically, one iterative line is selected from each verification fitting space as a reference line, serving as the starting point for subsequent angle iterations. Using the reference line as a benchmark, new lines are generated within the verification fitting space according to the preset angle step size. The newly generated lines are evaluated to determine whether their fitting effect is better than that of the reference line. If the fitting effect of the new line is better than that of the reference line, it is replaced with the new reference line; otherwise, the original reference line remains unchanged, and multiple iterations are performed until a preset stopping condition is met (such as reaching the maximum number of iterations, or the error being less than a certain threshold). During the iteration process, all generated and evaluated new lines are collected to form a set of angle iterative lines, representing the fitting results obtained after fine-tuning the reference line in different directions.

[0035] The module 30 for storing the monitoring index set acquisition also includes determining whether the difference in fitting values ​​between the multiple angle iteration lines and the multiple iteration lines meets a preset fitting value difference. If so, the multiple angle iteration lines are used as multiple fitting lines. The preset fitting value difference is a threshold used to determine whether angle iteration has brought sufficient improvement in fitting effect. Specifically, for each angle iteration line, its corresponding fitting value (such as the number of particles falling within a preset bandwidth, the value of the error function, etc.) is calculated. For each original iteration line, its corresponding fitting value is also calculated. The difference in fitting values ​​between each angle iteration line and its corresponding original iteration line is calculated, reflecting the degree of improvement in fitting effect after angle iteration. If the difference in fitting values ​​between multiple angle iteration lines and their corresponding original iteration lines is greater than or equal to the preset fitting value difference, then these angle iteration lines are considered to have significantly improved fitting effect and are selected as the final fitting lines, representing the best fitting result of the original data point set in different directions.

[0036] The following will describe in detail the specific configuration of the monitoring indicator set acquisition module 30. The monitoring indicator set acquisition module 30 may further include: extracting multiple slopes from the multiple fitted lines. Slopes are extracted from the multiple lines that have been determined as fitted lines; the smaller and higher the slope, the more stable the intelligent sensing array is, the more reliable the monitoring data is, and it can be put into use.

[0037] The module 30 for storing and acquiring monitoring index sets also includes determining whether the multiple slopes are less than or equal to a preset slope threshold. If so, the authentication is successful. The extracted slopes are compared with the preset slope threshold. If the slopes of all fitted lines are less than or equal to the preset slope threshold, the inclination of these lines is considered to meet the requirements, and therefore the authentication is successful. The preset slope threshold is a set value used to determine whether the inclination of the fitted lines is within an acceptable range.

[0038] The monitoring indicator set acquisition module 30 also includes: if not, authentication fails, a risk perception instruction is generated, and the risk perception instruction is sent to the staff. If the slope of any fitted straight line is greater than a preset slope threshold, the inclination of these lines is considered to exceed the acceptable range, which may indicate some abnormality or risk situation. Authentication fails, a risk perception instruction is generated, containing specific information about why the authentication failed, and sent to the staff, such as which line's slope exceeds the threshold and the degree of exceedance.

[0039] The specific configuration of the first risk identification result determination module 40 will be described in detail below. The first risk identification result determination module 40 may further include: calculating demand differences based on the multiple sets of storage environment requirements and the multiple sets of storage monitoring indicators to obtain multiple sets of storage demand differences. By comparing each storage environment requirement with its corresponding monitoring indicator and calculating the difference between them, multiple sets of storage demand differences are obtained, reflecting the gap between the actual storage environment and the ideal storage environment.

[0040] The first risk identification result determination module 40 further includes determining whether each of the multiple storage requirement difference sets is greater than or equal to a multiple preset tolerance difference set. If so, multiple first-risk drugs are obtained. These storage requirement differences are compared with the preset tolerance difference set. If a storage requirement difference is greater than or equal to the corresponding preset tolerance difference, it indicates that the environmental requirement has not been met, which may lead to drug quality damage or safety risks. Multiple first-risk drugs are obtained. The preset tolerance difference set defines the maximum acceptable deviation range for each environmental requirement.

[0041] The first risk identification result determination module 40 further includes calculating the difference between the multiple sets of storage requirement differences corresponding to the multiple first-risk drugs and the multiple sets of preset tolerance difference values, and comparing the calculation result with the multiple sets of preset tolerance difference values ​​to obtain the multiple first risk coefficients. For drugs marked as first-risk drugs, their risk level is further calculated, that is, the difference between their corresponding storage requirement difference and the preset tolerance difference value is calculated, and the difference is divided by the preset tolerance difference value to obtain multiple first risk coefficients, which reflect the degree to which environmental requirements are not met. The larger the coefficient, the higher the risk.

[0042] The first risk identification result determination module 40 further includes using the plurality of first-risk drugs and the plurality of first-risk coefficients as the first risk identification result. By outputting all identified first-risk drugs and their corresponding first-risk coefficients as the first risk identification result, it is possible to understand which drugs pose storage risks and the specific degree of those risks, thereby enabling corresponding measures to be taken to reduce the risks.

[0043] The specific configuration of the second risk identification result determination module 50 will be described in detail below. The second risk identification result determination module 50 may further include: randomly extracting one drug from each of the multiple drug sets in the multiple drug storage areas as the first nearest neighbor central drug, thereby obtaining multiple first nearest neighbor central drugs.

[0044] The second risk identification result determination module 50 further includes extracting multiple first drug neighborhoods of the multiple first nearest-neighbor central drugs based on the multiple drug storage location identifiers. A reasonable neighborhood range is defined, such as an area within a certain physical distance centered on the central drug, or other drugs on the same shelf, on the same floor, or within the same area. Based on the defined neighborhood range, all drugs located within that range are extracted from the storage area to constitute the first drug neighborhood of the central drug. For each central drug, a first drug neighborhood set containing all drugs within its neighborhood is formed, resulting in multiple first drug neighborhood sets corresponding to multiple first nearest-neighbor central drugs.

[0045] The second risk identification result determination module 50 further includes performing similarity identification on the multiple first nearest neighbor central drugs and the multiple first drug neighborhoods based on the multiple drug name sets and the multiple outer packaging image sets, to obtain multiple first nearest neighbor similarity coefficients. Using the multiple drug name sets and the multiple outer packaging image sets as data sources, similarity identification is performed on each first nearest neighbor central drug and the drugs in its corresponding first drug neighborhood. This includes text similarity comparison (such as string matching or fuzzy matching of drug names) and image similarity comparison (such as feature extraction and comparison of outer packaging images). Based on the similarity identification results, multiple first nearest neighbor similarity coefficients are obtained, reflecting the degree of similarity between the central drug and the drugs in its neighborhood.

[0046] The second risk identification result determination module 50 further includes performing nearest neighbor similarity identification on the multiple drug sets to obtain multiple nearest neighbor similarity coefficient sets. More broadly, nearest neighbor similarity identification is performed on all drugs within each drug set, and multiple nearest neighbor similarity coefficient sets are calculated using the same similarity identification method. Each set contains the similarity coefficients between all drug pairs within that region.

[0047] The second risk identification result determination module 50 further includes using a preset similarity threshold to identify regional confusion risks in the multiple nearest neighbor similarity coefficient sets, and determining the second risk identification result. Using the preset similarity threshold, regional confusion risk is identified for each nearest neighbor similarity coefficient set. If the similarity coefficient between a drug pair exceeds the threshold, the two drugs are considered to have a risk of confusion. By analyzing the similarity coefficients of all drug pairs, it is possible to determine which regions or drug combinations have a high risk of confusion. Based on the regional confusion risk identification result, the second risk identification result is determined, i.e., which regions or drug combinations require special attention to avoid confusion errors.

[0048] The specific configuration of the second risk identification result determination module 50 will be described in detail below. The second risk identification result determination module 50 may further include: constructing a similarity identification function, wherein the similarity identification function is: ; in, The first nearest neighbor similarity coefficient, Let n be the number of drugs in the first drug neighborhood, where n is a positive integer greater than or equal to 1. The name of the drug at the first nearest neighbor center. Let be the name of the i-th drug in the first drug neighborhood. Image of the outer packaging of the drug at the first nearest neighbor center. This is the outer packaging image of the i-th drug in the neighborhood of the first drug. This represents the weight given to the similarity of drug names when performing nearest neighbor similarity identification. This represents the weight of the similarity between the outer packaging images when performing nearest neighbor similarity recognition.

[0049] The second risk identification result determination module 50 further includes mapping and matching the multiple first nearest-neighbor central drugs and the multiple first drug neighborhoods based on the multiple drug name sets and the multiple outer packaging image sets, to obtain multiple first nearest-neighbor central drug names, multiple first nearest-neighbor central drug outer packaging images, multiple first drug neighborhood name sets, and multiple first drug neighborhood outer packaging image sets. For each first nearest neighbor central drug, the corresponding name is found in the drug name set by matching the drug's unique identifier (such as barcode, SKU number, or drug code), forming multiple first nearest neighbor central drug names. Similarly, the outer packaging image corresponding to each first nearest neighbor central drug is found in the outer packaging image set, forming multiple first nearest neighbor central drug outer packaging images. Based on the location of each first nearest neighbor central drug in the storage area, its first drug neighborhood is determined. Usually, other drugs that are physically adjacent or close to the central drug are selected as neighborhood members. For each first drug neighborhood, each drug in it is traversed, and the corresponding name is found in the drug name set, obtaining multiple first drug neighborhood name sets. Similarly, the outer packaging image corresponding to each drug in the neighborhood is found in the outer packaging image set, forming multiple first drug neighborhood outer packaging image sets.

[0050] The second risk identification result determination module 50 further includes inputting multiple first nearest neighbor central drug names, multiple first nearest neighbor central drug packaging images, multiple first drug neighborhood name sets, and multiple first drug neighborhood packaging image sets into the similarity identification function to obtain the multiple first nearest neighbor similarity coefficients. For each first nearest neighbor central drug, its name and packaging image are input into the similarity identification function as a set of data. For the name of the central drug and the name of each neighboring drug, the text similarity algorithm in the similarity identification function is used to calculate the similarity coefficient between them. Similarly, for the packaging image of the central drug and the packaging image of each neighboring drug, the image similarity algorithm (such as histogram comparison, feature matching, deep learning model, etc.) in the similarity identification function is used to calculate the similarity coefficient between them. For the similarity calculation of the name and image of the central drug and each drug in the neighborhood, a similarity coefficient will be obtained. Finally, the similarity identification function will output multiple first nearest neighbor similarity coefficients, reflecting the degree of similarity between each central drug and each drug in its neighborhood.

[0051] Although this application makes various references to certain modules in the platform according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0052] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A security risk management platform for multi-point collaborative analysis, characterized in that, The platform includes: The drug storage distribution data extraction module is used to interact with the storage system of the target drug storage warehouse to extract drug storage distribution data, wherein the drug storage distribution data includes multiple drug sets corresponding to multiple drug storage areas and multiple drug storage location identifiers. The drug feature set generation module is used to extract features from the multiple drug sets according to preset risk characteristics to generate multiple drug feature sets, wherein the multiple drug feature sets include multiple drug name sets, multiple outer packaging image sets, and multiple storage environment requirement sets. The storage monitoring index set acquisition module is used to interact with multiple intelligent sensing arrays in the multiple drug storage areas to acquire multiple storage monitoring index sets, wherein the multiple storage monitoring index sets include multiple storage temperatures, multiple storage humidity levels and multiple storage radiation levels. The first risk identification result determination module is used to identify risks in the multiple storage monitoring indicator sets based on the multiple storage environment requirement sets, and determine the first risk identification result, wherein the first risk identification result includes multiple first risk drugs and multiple first risk coefficients; The second risk identification result determination module is used to identify regional confusion risk based on the multiple drug name sets, multiple outer packaging image sets, the multiple drug sets and the multiple drug storage location identifiers, and determine the second risk identification result, wherein the second risk identification result includes multiple risk areas and multiple second risk coefficients; The safety risk management module is used to manage safety risks based on the multiple first-risk drugs, multiple first-risk coefficients, multiple risk areas, and multiple second-risk coefficients.

2. The security risk management platform for multi-point collaborative analysis as described in claim 1, characterized in that, The preset risk characteristics include drug name, outer packaging image, and storage environment requirements.

3. The security risk management platform for multi-point collaborative analysis as described in claim 1, characterized in that, include: Data is collected from the multiple intelligent sensing arrays within a preset verification window to generate multiple sets of verification temperature sequences, multiple sets of verification humidity sequences, and multiple sets of verification radiation sequences. The multiple sets of verified temperature sequences, multiple sets of verified humidity sequences, and multiple sets of verified radiation sequences are normalized to obtain multiple sets of normalized verified temperature sequences, multiple sets of normalized verified humidity sequences, and multiple sets of normalized verified radiation sequences. The multiple sets of normalized verification temperature sequences, multiple sets of normalized verification humidity sequences, and multiple sets of normalized verification radiation sequences are window-aligned and synchronously distributed to multiple verification fitting branches to generate multiple verification fitting spaces, wherein the multiple verification fitting spaces include multiple sets of verification fitting particles. Linear fitting is performed on the multiple sets of verification fitting particles within the multiple verification fitting spaces. The feasibility of the multiple intelligent sensing arrays is verified based on the linear fitting results. When the verification is successful, the multiple intelligent sensing arrays are used to monitor the multiple drug storage areas respectively, generating the multiple sets of storage monitoring indicators. The linear fitting results include multiple fitted lines.

4. The security risk management platform for multi-point collaborative analysis as described in claim 3, characterized in that, include: Multiple initial straight lines are randomly generated within the multiple verification fitting spaces, and the distance to the multiple initial straight lines is counted as the number of verification fitting particles with a preset bandwidth, thereby generating multiple initial fitting quantities; According to the preset bandwidth and the multiple initial fitting values, the multiple initial straight lines are iteratively updated in the multiple verification fitting spaces to obtain multiple iterative straight lines; The multiple iterative lines are iterated in the multiple verification fitting spaces according to a preset angle step size to obtain multiple angle iterative lines; Determine whether the difference in fitting values ​​between the multiple angle iteration lines and the multiple iteration lines meets the preset fitting value difference. If so, then the multiple angle iteration lines are used as multiple fitting lines.

5. The security risk management platform for multi-point collaborative analysis as described in claim 3, characterized in that, include: Extract multiple slopes from the fitted lines; Determine whether the multiple slopes are less than or equal to a preset slope threshold; if so, authentication is successful. If not, authentication fails, a risk awareness instruction is generated, and the risk awareness instruction is sent to the staff.

6. The security risk management platform for multi-point collaborative analysis as described in claim 1, characterized in that, include: Based on the multiple sets of storage environment requirements and the multiple sets of storage monitoring indicators, the demand difference is calculated to obtain multiple sets of storage demand difference; Determine whether the multiple sets of storage demand differences are greater than or equal to multiple preset tolerance difference sets. If so, obtain multiple first-risk drugs. Calculate the difference between the multiple storage requirement difference sets corresponding to the multiple first-risk drugs and the multiple preset tolerance difference sets, and compare the calculation result with the multiple preset tolerance difference sets to obtain the multiple first-risk coefficients; The plurality of first-risk drugs and the plurality of first-risk coefficients are used as the first risk identification results.

7. The security risk management platform for multi-point collaborative analysis as described in claim 1, characterized in that, include: One drug is randomly selected from multiple drug sets in the multiple drug storage areas as the first nearest neighbor drug, thereby obtaining multiple first nearest neighbor drugs; Based on the multiple drug storage location identifiers, extract multiple first drug neighborhoods of the multiple first nearest neighbor center drugs; Based on the multiple sets of drug names and multiple sets of outer packaging images, similarity identification is performed on the multiple first nearest neighbor center drugs and the multiple first drug neighborhoods to obtain multiple first nearest neighbor similarity coefficients; Perform nearest neighbor similarity identification on the multiple drug sets to obtain multiple sets of nearest neighbor similarity coefficients; The region confusion risk is identified by using a preset similarity threshold to identify the multiple nearest neighbor similarity coefficient sets, and the second risk identification result is determined.

8. The security risk management platform for multi-point collaborative analysis as described in claim 7, characterized in that, include: Construct a similarity recognition function, wherein the similarity recognition function is: ； in, The first nearest neighbor similarity coefficient, Let n be the number of drugs in the first drug neighborhood, where n is a positive integer greater than or equal to 1. The name of the drug at the first nearest neighbor center. Let be the name of the i-th drug in the first drug neighborhood. Image of the outer packaging of the drug at the first nearest neighbor center. This is the outer packaging image of the i-th drug in the neighborhood of the first drug. This represents the weight given to the similarity of drug names when performing nearest neighbor similarity identification. The weight of the similarity between outer packaging images in nearest neighbor similarity recognition; Based on the multiple drug name sets and multiple outer packaging image sets, the multiple first nearest neighbor center drugs and the multiple first drug neighborhoods are mapped and matched to obtain multiple first nearest neighbor center drug names, multiple first nearest neighbor center drug outer packaging images, multiple first drug neighborhood name sets, and multiple first drug neighborhood outer packaging image sets. The drug names of multiple first nearest neighbors, the outer packaging images of multiple first nearest neighbors, the sets of first drug neighborhood names, and the sets of first drug neighborhood images are respectively input into the similarity recognition function to obtain the multiple first nearest neighbor similarity coefficients.

Images