Information management system and method based on computer network

Abstract

The invention relates to the field of information management, and particularly discloses an information management system and method based on a computer network, and the method comprises the steps: firstly obtaining patient prescription information and pharmacy drug inventory information, then carrying out the feature extraction and correlation analysis of the patient prescription information and the pharmacy drug inventory information through a deep learning technology, and finally generating a drug delivery audit report through a generator. Therefore, automatic monitoring and auditing of the medicine ex-warehouse process are achieved, a pharmacy is helped to better formulate a purchasing strategy, the capacity of inventory management and patient service is optimized, and human errors and potential medicine management risks are reduced.

Description

Technical Field

[0001] This application relates to the field of information management, and more specifically, to an information management system and method based on computer networks. Background Technology

[0002] Healthcare information management involves collecting, storing, managing, and transmitting data and information in the healthcare field through efficient information technology systems. Modern hospital operations rely heavily on healthcare information management systems as crucial technological support and infrastructure. These systems aim to strengthen hospital management, improve efficiency, and optimize healthcare quality through modern, scientific, and standardized methods, thereby shaping a new image for modern hospitals.

[0003] However, in the existing technology, in the process of medical information management, since the consultation and pharmacy dispensing are in two separate systems, it is impossible to reasonably control the patient consultation process and the medication dispensing system, thereby affecting the overall operational efficiency and management effectiveness.

[0004] Therefore, an information management system and method based on computer networks is desired. Summary of the Invention

[0005] To address the aforementioned technical problems, this application is proposed. Embodiments of this application provide a computer network-based information management system and method. It first acquires patient prescription information and pharmacy drug inventory information, then utilizes deep learning technology to extract features and perform correlation analysis on both. Finally, a generator is used to generate a drug dispensing audit report, thereby achieving automated monitoring and auditing of the drug dispensing process. This helps pharmacies better formulate procurement strategies, optimize inventory management and patient service capabilities, and reduce human error and potential drug management risks.

[0006] According to one aspect of this application, a computer network-based information management system is provided, comprising:

[0007] The medical information data acquisition module is used to acquire patient prescription information and pharmacy drug inventory information;

[0008] The medical information data extraction module is used to extract the patient prescription information-related semantic understanding feature vector and the drug inventory multimodal fusion feature vector from the patient prescription information and the pharmacy drug inventory information.

[0009] The drug delivery audit report generation module is used to generate a drug delivery audit report based on the semantic understanding feature vector associated with the patient prescription information and the multimodal fusion feature vector of the drug inventory.

[0010] According to another aspect of this application, a computer network-based information management method is provided, comprising:

[0011] Obtain patient prescription information and pharmacy drug inventory information;

[0012] Extract the semantic understanding feature vector associated with the patient prescription information and the multimodal fusion feature vector of the drug inventory from the patient prescription information and the pharmacy drug inventory information;

[0013] Based on the semantic understanding feature vector associated with the patient's prescription information and the multimodal fusion feature vector of the drug inventory, a drug outbound audit report is generated.

[0014] Compared with existing technologies, this application provides an information management system and method based on computer networks. It first acquires patient prescription information and pharmacy drug inventory information, then uses deep learning technology to extract features and perform correlation analysis on the two, and finally uses a generator to generate a drug delivery audit report. This enables automated monitoring and auditing of the drug delivery process, thereby helping pharmacies to better formulate procurement strategies, optimize inventory management and patient service capabilities, and reduce human error and potential drug management risks. Attached Figure Description

[0015] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0016] Figure 1 This is a block diagram of a computer network-based information management system according to an embodiment of this application.

[0017] Figure 2 This is a block diagram of a medical information data extraction module in a computer network-based information management system according to an embodiment of this application.

[0018] Figure 3 This is a block diagram of a drug inventory information feature extraction unit in a computer network-based information management system according to an embodiment of this application.

[0019] Figure 4 This is a block diagram of the drug outbound audit report generation module in a computer network-based information management system according to an embodiment of this application.

[0020] Figure 5 This is a flowchart of a computer network-based information management method according to an embodiment of this application.

[0021] Figure 6 This is a block diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0022] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0023] Figure 1 This is a block diagram of a computer network-based information management system according to an embodiment of this application. Figure 1 As shown, the computer network-based information management system 100 according to an embodiment of this application includes: a medical information data acquisition module 110, used to acquire patient prescription information and pharmacy drug inventory information; a medical information data extraction module 120, used to extract patient prescription information-related semantic understanding feature vectors and drug inventory multimodal fusion feature vectors from the patient prescription information and the pharmacy drug inventory information; and a drug delivery audit report generation module 130, used to generate a drug delivery audit report based on the patient prescription information-related semantic understanding feature vectors and the drug inventory multimodal fusion feature vectors.

[0024] In the aforementioned computer network-based information management system 100, the medical information data acquisition module 110 is used to acquire patient prescription information and pharmacy drug inventory information. It should be understood that medical information management systems play a crucial role in modern hospital operations. They collect, store, manage, and transmit medical data and information through efficient information technology systems, serving as a key infrastructure for improving hospital management efficiency and the quality of medical services. However, a critical problem currently exists: the separation of the patient visit process from the pharmacy dispensing system within the medical information management system. This separation leads to a lack of effective integration and control between the patient visit process and the medication dispensing system. Because patient visit and pharmacy dispensing are in two independent systems, hospitals face challenges such as poor flow of patient information and incomplete information sharing. For example, after a patient's visit, they need to pick up their medication at the pharmacy, but due to the system separation, medical staff may not be able to obtain the patient's medication prescription and dispensing progress in a timely manner, resulting in reduced dispensing efficiency and a decreased patient service experience. This disconnect can also lead to problems such as data entry errors, dispensing delays, or duplicate dispensing, further affecting the overall efficiency of hospital operations and the effectiveness of management. Therefore, the technical solution in this application, by acquiring patient prescription information and pharmacy drug inventory information, and combining it with deep learning technology, generates drug dispensing audit reports to achieve automated monitoring and auditing of the drug dispensing process, helping pharmacies improve operational efficiency and management quality. This automated system can help pharmacy managers accurately analyze the details and data of each drug dispensing, thereby formulating more effective procurement strategies and optimizing inventory management. Simultaneously, by reducing human error and potential drug management risks, the system can improve the safety and accuracy of drug services, enhance patient service capabilities, and further ensure the compliance and operational efficiency of medical institutions.

[0025] Specifically, obtaining patient prescription information and pharmacy drug inventory information is crucial for efficient healthcare information management and precise drug supply chain control. Patient prescription information enables healthcare providers to accurately record and manage patient medication use, ensuring the effectiveness and safety of treatment. Simultaneously, pharmacy drug inventory information helps supply chain managers understand drug availability and flow in real time, effectively predicting and allocating drug demand to ensure timely medication supply to patients. The collection and utilization of this information not only improves the quality and efficiency of healthcare services but also reduces healthcare risks caused by drug shortages or mismatches, creating a safer and more reliable healthcare environment for patients and healthcare institutions.

[0026] In the aforementioned computer network-based information management system 100, the medical information data extraction module 120 is used to extract semantic understanding feature vectors related to patient prescription information and multimodal fusion feature vectors of drug inventory from the patient prescription information and the pharmacy drug inventory information. It should be understood that by extracting these feature vectors, the medical information management system can achieve a closed loop from data to decision-making, effectively improving the quality and efficiency of medical services. This intelligent data analysis and processing method not only helps improve the utilization rate of medical resources but also reduces medical errors and drug waste caused by information asymmetry, creating a safer and more reliable treatment environment for patients and medical institutions.

[0027] Figure 2 This is a block diagram of a medical information data extraction module in a computer network-based information management system according to an embodiment of this application. Figure 2 As shown, in a specific embodiment of this application, the medical information data extraction module 120 includes: a patient prescription information feature extraction unit 121, used to extract features from the patient prescription information to obtain a semantic understanding feature vector associated with the patient prescription information; and a drug inventory information feature extraction unit 122, used to extract features from the pharmacy drug inventory information to obtain a multimodal fusion feature vector of the drug inventory.

[0028] It is understandable that patient prescription information encompasses crucial medical treatment information, including drug names, dosages, and frequency of use, which are essential for providing healthcare services. Through feature extraction, the system can extract semantically relevant features from raw text data, such as disease diagnosis, treatment plans, and allergic reactions, thereby establishing a deep understanding and comprehensive analysis of the patient's medical history. The feature extraction process typically employs deep learning models, such as convolutional neural networks (CNNs) or recurrent neural networks (RNNs), which effectively capture the semantic structure and relationships within prescription information. By analyzing and processing the text data, the system can generate representative and information-rich feature vectors. These vectors not only support doctors' accurate understanding of the patient's condition but also provide scientific evidence and data support for subsequent medical decisions and treatment plan formulation.

[0029] Furthermore, the purpose of feature extraction from pharmacy drug inventory information is to optimize drug supply chain management and improve the accuracy and efficiency of drug inventory. Pharmacy drug inventory information encompasses a large amount of key data, including drug storage status, expiration dates, and supplier information. This information directly impacts drug supply to medical institutions and patient treatment arrangements. Through feature extraction, the system can obtain key inventory characteristics from multiple dimensions, such as drug quantity, variety, and storage conditions, thereby establishing a comprehensive and integrated inventory feature vector.

[0030] In a specific embodiment of this application, the patient prescription information feature extraction unit 121 includes: passing the patient prescription information through a patient prescription information semantic understanding model to obtain multiple patient prescription information semantic feature vectors; arranging the multiple patient prescription information semantic feature vectors according to the sample dimension to obtain a two-dimensional feature matrix of patient prescription information; and passing the two-dimensional feature matrix of patient prescription information through a patient prescription information text convolutional neural network to obtain the patient prescription information associated semantic understanding feature vector.

[0031] It is understandable that each patient prescription contains unique medical information, such as diagnosis, medication regimen, and allergies. This information is crucial for doctors to assess the patient's health and develop personalized treatment plans. Through a patient prescription information semantic understanding model, the system can extract key semantic features from each prescription, such as drug name, dosage, frequency, and important indicators related to the patient's health status. Semantic understanding models are typically based on deep learning techniques, such as Natural Language Processing (NLP) models, which can automatically identify and understand the semantic structure and information associations in prescription text, thereby generating representative and information-rich feature vectors. Specifically, the patient prescription information is segmented to obtain a sequence of prescription information words; the embedding layer of the patient prescription information semantic understanding model maps each prescription information word in the sequence to a prescription information word embedding vector to obtain a sequence of prescription information word embedding vectors; the BERT model based on a converter in the patient prescription information semantic understanding model is used to perform global contextual semantic encoding on the sequence of prescription information word embedding vectors to obtain multiple patient prescription information semantic feature vectors.

[0032] Furthermore, arranging the semantic feature vectors of multiple patient prescription information according to the sample dimension aims to effectively organize and analyze medical data for systematic research and application. Each patient's prescription information semantic feature vector represents their individualized medical data, including important information such as disease diagnosis, treatment plan, and medication history. By arranging these feature vectors according to the sample (i.e., patient) dimension, the feature vector set of each patient can be clearly represented in a two-dimensional feature matrix, with each row corresponding to one patient and each column corresponding to one feature. This construction of a two-dimensional feature matrix makes medical data analysis more systematic and operational. First, each element in the matrix represents a specific semantic feature value, such as drug name, dosage, and frequency, which collectively demonstrate the treatment trends and medication patterns of the patient group. Second, through statistical analysis and data mining of the matrix, commonalities and differences between different patient groups can be revealed, helping medical researchers discover potential disease patterns and key factors for evaluating treatment effectiveness.

[0033] Furthermore, the purpose of using a patient prescription information text convolutional neural network (CNN) to process the two-dimensional feature matrix of patient prescription information is to further mine and utilize the deep semantic information in medical data, achieving automated processing and semantic understanding of prescription data. Patient prescription information contains a large amount of medical text data, such as drug names, dosages, and frequency of use. This information needs to be analyzed efficiently and accurately to support medical decisions and treatment plan formulation. A text convolutional neural network is a deep learning model specifically designed for processing text data, effectively capturing local features and semantic relationships within the text. By inputting the two-dimensional feature matrix of patient prescription information into the CNN model, the system can extract key features from the text data at different levels and scales, such as semantic representations of words, phrases, or entire sentences. These feature vectors not only have good representational capabilities but also capture potential patterns and correlations in the prescription information, such as common disease patterns and treatment recommendations. Through processing by the CNN model, the semantically understood feature vectors associated with patient prescription information can provide more detailed and comprehensive medical data analysis results. These vectors can be used to automatically identify and classify patient disease types, assess treatment effectiveness, and predict patient health trends. Specifically, each layer of the patient prescription information text convolutional neural network performs convolution processing, mean pooling based on the local feature matrix, and nonlinear activation processing on the input data during the forward propagation of the layer, so that the last layer of the patient prescription information text convolutional neural network outputs the patient prescription information associated semantic understanding feature vector, wherein the input of the patient prescription information text convolutional neural network is the two-dimensional feature matrix of the patient prescription information.

[0034] Figure 3 This is a block diagram of a drug inventory information feature extraction unit in a computer network-based information management system according to an embodiment of this application. Figure 3 As shown, in a specific embodiment of this application, the drug inventory information feature extraction unit 122 includes: a drug inventory information text part feature encoding subunit 1221, used to perform feature encoding on the text part of the pharmacy drug inventory information to obtain a drug inventory semantic text understanding feature vector; a drug inventory information image part feature encoding subunit 1222, used to perform feature encoding on the image part of the pharmacy drug inventory information to obtain a drug inventory enhanced comprehensive feature vector; and a drug inventory information multimodal feature fusion subunit 1223, used to fuse the drug inventory semantic text understanding feature vector and the drug inventory enhanced comprehensive feature vector to obtain the drug inventory multimodal fusion feature vector.

[0035] It is understandable that drug inventory information includes key data such as drug names, specifications, and supplier information, which are crucial for the daily operations of pharmacies and medical services. Through feature encoding, the system can transform this unstructured text data into structured, quantifiable feature vectors, thereby achieving intelligent understanding and analysis of drug inventory information. Feature encoding typically involves word embedding and text vectorization methods in Natural Language Processing (NLP). Word embedding maps each drug name, specification, and other textual information to a vector representation in a high-dimensional space, preserving semantic relationships and contextual information between words. Through text vectorization, the system can effectively capture drug characteristics, supplier features, and important drug-related attributes such as storage conditions and expiration dates, thus forming representative and information-rich feature vectors.

[0036] Furthermore, the purpose of feature encoding on the image portion of pharmacy drug inventory information is to fully utilize the visual information in the image data to improve the accuracy and efficiency of drug inventory management. Image data in drug inventory typically includes visual features such as the appearance, packaging, and labels of drugs; this information is crucial for rapid identification and accurate classification of drugs. Through feature encoding, the system can convert these visual features into numerical, comprehensive feature vectors to achieve a deep understanding and intelligent management of drug inventory. Feature encoding of image data typically involves techniques in the field of computer vision, such as feature extraction, image descriptors, and deep learning models. The feature extraction process captures key attributes of drug appearance by extracting crucial visual feature points, edges, textures, and other information from images. Image descriptors then convert these extracted features into numerical vector representations, typically high-dimensional numerical vectors, which retain important visual features and semantic information from the image.

[0037] Furthermore, fusing the semantic text understanding feature vector and the enhanced comprehensive feature vector of drug inventory aims to comprehensively utilize information from different data modalities, thereby improving the overall understanding and precise management capabilities of drug inventory. Drug inventory information encompasses two different data formats: textual descriptions and visual images. The text portion provides detailed descriptions of drug attributes and characteristics, while the image portion presents the visual features of the drug's appearance and packaging. By fusing these two feature vectors, the system can achieve a more comprehensive and in-depth understanding and analysis of drug inventory. Here, the multimodal fusion feature vector of drug inventory combines semantic information from textual data and visual features from image data, providing a more accurate and comprehensive basis for drug classification, identification, and attribute analysis.

[0038] In a specific embodiment of this application, the feature encoding subunit 1221 for the text portion of the drug inventory information includes: extracting pharmacy drug inventory text data from the pharmacy drug inventory information; passing the pharmacy drug inventory text data through a drug inventory text information semantic encoder based on a context convolutional neural network to obtain multiple drug inventory semantic feature vectors; arranging the multiple drug inventory semantic feature vectors into a drug inventory semantic feature vector; and passing the drug inventory semantic feature vector through a drug inventory semantic multi-scale neighborhood feature extraction module to obtain the drug inventory semantic text understanding feature vector.

[0039] It is understandable that pharmaceutical inventory text data typically includes the name, specifications, quantity, supplier information, and related storage and usage instructions for each drug. This information is crucial for pharmacy management, helping pharmacy managers understand the current inventory status, effectively predict and plan drug procurement and sales activities, and optimize inventory layout and management strategies. By extracting pharmaceutical inventory text data, pharmacies can achieve real-time monitoring and management of their inventory. This data can be used to establish an inventory database or management system, automatically recording drug inbound and outbound transactions, ensuring drug updates and timely replenishment, thereby avoiding business interruptions or waste due to insufficient or excessive inventory. Specifically, the embedding layer of the aforementioned context-based convolutional neural network-based pharmaceutical inventory text information semantic encoder is used to transform the pharmacy pharmaceutical inventory text data into embedding vectors to obtain a sequence of pharmaceutical inventory embedding vectors; the converter-based BERT model of the aforementioned context-based convolutional neural network-based pharmaceutical inventory text information semantic encoder is used to perform global contextual semantic encoding on the sequence of pharmaceutical inventory embedding vectors to obtain multiple pharmaceutical inventory semantic feature vectors.

[0040] Furthermore, processing pharmacy drug inventory text data using a context-based convolutional neural network (CNN) semantic encoder aims to improve the semantic understanding and feature representation capabilities of the text data. Drug inventory text data contains a large amount of descriptive information about drugs, such as names, specifications, and suppliers. This information may have different meanings and importance in different contexts. Traditional text processing methods may not be able to effectively capture these complex semantic relationships and contextual information, while the context-based CNN semantic encoder can improve semantic representation by learning local and global features of the text data. The working principle of context-based CNN is to extract features from the input text data and preserve the structural information of the text through multiple layers of convolution and pooling operations. For pharmacy inventory text data, this method can help the system identify and understand the semantic and functional attributes of drug names, specifications, and other key information. By establishing multiple drug inventory semantic feature vectors, the system can more comprehensively describe the characteristics and uses of each drug, including important information such as therapeutic effects, indications, and contraindications.

[0041] Furthermore, arranging multiple drug inventory semantic feature vectors into a single drug inventory semantic feature vector aims to integrate the semantic features of individual drugs into a unified vector representation, enabling more effective drug management and analysis. Each drug inventory semantic feature vector represents a detailed semantic description and characteristic of a particular drug, including its name, specifications, function, indications, and other important information. By arranging these feature vectors into a unified drug inventory semantic feature vector, the system can achieve comprehensive analysis and management of the overall drug inventory. This vectorized representation not only facilitates rapid drug retrieval and classification but also supports similarity comparison and correlation analysis between drugs.

[0042] Specifically, processing the semantic feature vectors of drug inventory through the multi-scale neighborhood feature extraction module aims to further enrich and optimize the semantic understanding capabilities of drug inventory. The semantic feature vectors of drug inventory contain basic semantic information for each drug, such as its name, specifications, and functions. However, a single feature vector may not fully capture the diversity and complexity of drug characteristics. Therefore, by introducing a multi-scale neighborhood feature extraction module, the system can understand and analyze the semantic information of drugs from a broader and deeper perspective. The multi-scale neighborhood feature extraction module utilizes information windows and convolutional kernels of different scales to capture the diversity of drug semantic features. This method not only considers the overall characteristics of drugs but also pays attention to subtle changes and differences in local drug characteristics. For example, for a drug, its name and basic attributes may be one scale, while more specific information such as its therapeutic effects, indications, and side effects may require a finer-grained scale for processing and understanding. Through this multi-scale feature extraction, the system can generate richer and more accurate semantic text understanding feature vectors for drug inventory. Specifically, the drug inventory semantic multi-scale neighborhood feature extraction module includes: a first convolutional layer, a second convolutional layer parallel to the first convolutional layer, and a cascaded layer connected to the first convolutional layer and the second convolutional layer, wherein the first convolutional layer uses a one-dimensional convolutional kernel with a first scale, and the second convolutional layer uses a one-dimensional convolutional kernel with a second scale.

[0043] In a specific embodiment of this application, the feature encoding subunit 1222 of the drug inventory information image includes: extracting image data of multiple regions of the pharmacy drug inventory from the pharmacy drug inventory information; passing the image data of multiple regions of the pharmacy drug inventory through a drug inventory image denoising model to obtain multiple regions of drug inventory denoising enhanced images; and passing the multiple regions of drug inventory denoising enhanced images through a region drug inventory denoising enhanced feature extractor based on a convolutional neural network to obtain the drug inventory enhanced comprehensive feature vector.

[0044] It is understandable that multi-area image data of drug inventory provides an intuitive understanding of the actual layout and arrangement of drug storage. By capturing and analyzing images of drug shelves, warehouses, or specific areas, the system can accurately record the storage location, stacking method, and surrounding environment of drugs. This visual description not only helps pharmacy managers better understand the organizational structure of the inventory but also helps optimize inventory space utilization and layout design, improving warehouse operational efficiency and workflow optimization. Multi-area image data of drug inventory provides crucial support for real-time monitoring and management of drugs. By regularly capturing and updating images of drug storage areas, the system can achieve real-time monitoring of inventory status. This real-time monitoring not only helps to quickly identify and resolve potential problems in drug storage, such as expired drugs and misplaced items, but also effectively prevents and reduces the risk of drug theft or damage, ensuring the safety and integrity of drugs.

[0045] Furthermore, processing image data from multiple areas of the pharmacy's drug inventory using a drug inventory image denoising model aims to optimize and enhance image quality, thereby improving the visual recognition and analysis capabilities of the drug inventory. Images of drug inventory areas may be affected by factors such as lighting conditions, shooting angle, and device resolution, resulting in noise, blurring, or other undesirable effects. These problems can hinder the accurate understanding and analysis of drug storage conditions. Through the drug inventory image denoising model, the system can effectively remove noise and enhance image quality for each drug inventory area. This model is typically based on advanced image processing techniques, such as convolutional neural networks (CNNs) or deep learning-based image enhancement algorithms, capable of identifying and repairing noise, blur, and distortion in images, making the images of drug storage areas clearer and more identifiable. Specifically, the image data of multiple regions of the pharmacy's drug inventory is input into the encoder of the drug inventory image denoising model, wherein the encoder uses a convolutional layer to explicitly spatially encode the image data of multiple regions of the pharmacy's drug inventory to obtain image features; and the image features are input into the decoder of the drug inventory image denoising model, wherein the decoder uses a deconvolutional layer to deconvolve the image features to obtain the denoised and enhanced image of the multiple regions of the drug inventory.

[0046] Furthermore, processing multiple areas of drug inventory denoising and enhancement images using a region-based drug inventory denoising and enhancement feature extractor based on a convolutional neural network (CNN) aims to further optimize image quality and extract richer and more accurate feature information to support the efficient operation and intelligent decision-making of the drug inventory management system. The CNN-based feature extractor can effectively learn and extract key features of drug inventory areas from images using deep learning methods. CNNs can automatically identify patterns and structures in images, thereby further extracting detailed features of drug storage areas, such as drug labels, container shapes, and stacking methods, on top of denoising and enhancement. This feature extraction process is not merely simple image processing, but a process of understanding and parsing images based on deep learning models. Features trained through CNN models possess high abstraction capabilities and discriminative power, enabling the system to more accurately identify and classify different types of drugs and their storage conditions. The convolutional neural network-based feature extractor provides strong support for generating comprehensive feature vectors for drug inventory enhancement. These feature vectors not only contain visual features extracted from images but also information such as the spatial layout, correlation, and possible usage patterns of drug storage areas. Specifically, each layer of the convolutional neural network-based regional drug inventory noise reduction and enhancement feature extractor processes the input data during the forward pass of the layer as follows: The convolutional units of each layer of the convolutional neural network-based regional drug inventory noise reduction and enhancement feature extractor perform kernel-based convolution processing on the input data to obtain a convolutional feature map; the pooling units of each layer of the convolutional neural network-based regional drug inventory noise reduction and enhancement feature extractor perform channel-based pooling processing on the convolutional feature map to obtain a pooled feature map; and the activation units of each layer of the convolutional neural network-based regional drug inventory noise reduction and enhancement feature extractor perform non-linear activation on the feature values ​​at each position in the pooled feature map to obtain an activation feature map; wherein, the output of the last layer of the convolutional neural network-based regional drug inventory noise reduction and enhancement feature extractor is the comprehensive feature vector for drug inventory enhancement.

[0047] In the aforementioned computer network-based information management system 100, the drug dispensing audit report generation module 130 is used to generate a drug dispensing audit report based on the semantic understanding feature vector associated with the patient prescription information and the multimodal fusion feature vector of the drug inventory. It should be understood that the semantic understanding feature vector associated with the patient prescription information contains important information such as the individual patient's medical needs, drug formulation, and related treatment background, while the multimodal fusion feature vector of the drug inventory integrates multi-dimensional features such as the visual characteristics, quantity, and storage location of the drugs. Combining these two types of feature vectors can effectively track and confirm whether each drug dispensing meets the patient's prescription needs and medical standards. By comparing the feature vectors of the patient's prescription information with those of the actual dispensing drugs, the system can quickly identify potential discrepancies or errors, such as the accuracy of the drug formulation and the matching of drug types, thereby reducing potential medical risks and legal liabilities caused by erroneous dispensing.

[0048] Figure 4 This is a block diagram of a drug outbound audit report generation module in a computer network-based information management system according to an embodiment of this application. Figure 4 As shown in a specific embodiment of this application, the drug delivery audit report generation module 130 includes: a medical information feature association unit 131, used to associate the patient prescription information association semantic understanding feature vector and the drug inventory multimodal fusion feature vector to obtain a patient prescription-drug inventory information association feature vector; a medical information feature optimization unit 132, used to perform category-based patient prescription-drug inventory information association metric learning compensation on the patient prescription-drug inventory information association feature vector based on metric data to obtain an optimized patient prescription-drug inventory information association feature vector; and a drug information report generation unit 133, used to generate a drug delivery audit report by passing the optimized patient prescription-drug inventory information association feature vector through a generator.

[0049] It is understandable that by comparing the semantic understanding feature vector of patient prescription information with the multimodal fusion feature vector of drug inventory, the system can intelligently verify whether the dispensed drugs meet the prescription requirements, thereby avoiding medical risks caused by errors or misdispensing. By accurately obtaining the correlation feature vector between patient prescription and drug inventory information, the system can optimize drug delivery plans and inventory management strategies, ensuring timely supply and proper storage of drugs, and improving the efficiency and responsiveness of pharmacy operations.

[0050] Specifically, in the technical solution of this application, individual differences in blood glucose values ​​due to the physiological characteristics and disease changes of different diabetic patients are considered, which may affect the distribution of blood glucose feature vectors. The accuracy and stability of portable blood glucose measuring devices may affect the accuracy of blood glucose data, and thus affect the feature vectors. Real-time dynamic changes in drug inventory may lead to rapid updates of inventory information; if the data is not updated in time, it may affect the accuracy of the feature vectors. The quality of image data of pharmacy drug inventory areas may be affected by factors such as shooting conditions and image clarity, affecting the effect of image noise reduction and feature extraction, which may lead to local abnormal distributions in the patient prescription-drug inventory information associated feature vectors. This may cause a category outlier shift relative to the generator's predetermined class label when the feature vectors are generated and judged, affecting the generation iteration effect. Specifically, local abnormal distributions may cause the generator to learn incorrect patterns during training, leading to a bias towards outliers when generating and judging, i.e., category outlier shift. Abnormal distributions may affect the generator's prediction accuracy and generalization ability, and may cause the generated reports to fail to accurately reflect the actual situation. To address this issue, the technical solution of this application performs category-based patient prescription-drug inventory information association metric learning compensation on the patient prescription-drug inventory information association feature vector based on metric data to obtain an optimized patient prescription-drug inventory information association feature vector.

[0051] The process of applying metric-based metric learning compensation to the patient prescription-drug inventory information association feature vector to obtain an optimized patient prescription-drug inventory information association feature vector includes: determining the decoding weight matrix of the generator; decoupling the decoding weight matrix in row vector units to obtain a set of decoding weight vectors; calculating the patient prescription-drug inventory information association metric learning loss information term between each decoding weight vector in the set of decoding weight vectors and the patient prescription-drug inventory information association feature vector to obtain a patient prescription-drug inventory information association metric learning compensation representation vector; inputting the patient prescription-drug inventory information association metric learning compensation representation vector into a Softmax activation function to obtain a probabilistic patient prescription-drug inventory information association metric learning compensation weight vector; calculating the positional dot product of the probabilistic patient prescription-drug inventory information association metric learning compensation weight vector and the patient prescription-drug inventory information association feature vector to obtain a dot product feature vector, and calculating the weighted sum of the dot product feature vector and the patient prescription-drug inventory information association feature vector to obtain the optimized patient prescription-drug inventory information association feature vector.

[0052] The patient prescription-drug inventory information association metric learning loss information term between each decoded weight vector in the set of decoded weight vectors and the patient prescription-drug inventory information association feature vector is calculated using the following optimization formula to obtain the patient prescription-drug inventory information association metric learning compensation representation vector;

[0053] The optimization formula is as follows:

[0054]

[0055] in, Let x represent the eigenvalue at the i-th position of the j-th decoded weight vector in the set of decoded weight vectors. i Let L represent the feature value at the i-th position of the patient prescription-drug inventory information association feature vector, and D represent the length of the feature vector. j The feature value at the j-th position of the patient prescription-drug inventory information association metric learning compensation representation vector is represented.

[0056] In other words, the patient prescription-drug inventory information association feature vector is compensated using metric learning based on metric data. First, the decoded weight matrix of the generator is determined to understand how the generator distinguishes features from different categories. Then, the decoded weight matrix is ​​decoupled into row vectors to obtain a set of decoded weight vectors, allowing the model to independently analyze the weights of each category, providing finer control for feature compensation. Next, the patient prescription-drug inventory information association metric learning loss information term is calculated. This process involves quantifying the interaction between each category weight vector and the patient prescription-drug inventory information association feature vector. A specific loss function is used to capture inconsistencies between feature vectors and category weights, providing a quantitative indicator for compensation. Next, the patient prescription-drug inventory information association metric learning compensation representation vector is input into an activation function and subjected to a nonlinear transformation, converting it into a probabilistic form. The application of the activation function introduces nonlinearity, enhancing the model's interpretability of the compensation weights and making the compensation weights more consistent with actual distribution characteristics. Subsequently, using the patient prescription-drug inventory information association metric learning compensation representation vector processed by the activation function, the probabilistic patient prescription-drug inventory information association metric learning compensation weight vector is calculated. This step, through a probabilistic approach, provides probabilistic guidance for feature vector adjustment, making the optimization process more statistically consistent and thus enhancing the model's accuracy in feature adjustment. Finally, by calculating the dot product of the probabilistic patient prescription-drug inventory information association metric learning compensation weight vector and the patient prescription-drug inventory information association feature vector, and then calculating a weighted sum with the original patient prescription-drug inventory information association feature vector, the optimized patient prescription-drug inventory information association feature vector is obtained. This step integrates the original feature information and compensation information to generate the adjusted feature vector, reducing the impact of outlier distributions and improving the accuracy of generation.

[0057] Furthermore, the optimized patient prescription-drug inventory information association feature vector contains detailed data features of patients' medical needs and actual dispensing of drugs. Processed by advanced data association and matching algorithms, this optimized feature vector accurately reflects whether each drug dispensing complies with medical prescription requirements, including key information such as drug type, dosage, and formulation. By generating drug dispensing audit reports, the system can automatically analyze and summarize large amounts of prescription-inventory association data, providing detailed audit results and conclusions. The audit report is not merely a simple data summary; it also includes a comprehensive assessment of healthcare service compliance, drug safety, and operational procedures, helping managers and regulatory agencies to fully understand and evaluate the quality and efficiency of the drug dispensing process.

[0058] In summary, this application first obtains patient prescription information and pharmacy drug inventory information, then uses deep learning technology to extract features and perform correlation analysis on the two, and finally uses a generator to generate a drug delivery audit report, thereby realizing automated monitoring and auditing of the drug delivery process. This helps pharmacies better formulate procurement strategies, optimize inventory management and patient service capabilities, and reduce human error and potential drug management risks.

[0059] As described above, the computer network-based information management system 100 according to the embodiments of this application can be implemented in various terminal devices. In one example, the computer network-based information management system 100 can be integrated into the terminal device as a software module and / or a hardware module. For example, the computer network-based information management system 100 can be a software module in the operating system of the terminal device, or it can be an application developed for the terminal device; of course, the computer network-based information management system 100 can also be one of many hardware modules of the terminal device.

[0060] Alternatively, in another example, the computer network-based information management system 100 and the terminal device can also be separate devices, and the computer network-based information management system 100 can be connected to the terminal device via wired and / or wireless networks, and transmit interactive information in accordance with an agreed data format.

[0061] Figure 5 This is a flowchart of a computer network-based information management method according to an embodiment of this application. Figure 5 As shown, the information management method based on a computer network according to an embodiment of this application includes: S110, acquiring patient prescription information and pharmacy drug inventory information; S120, extracting a patient prescription information-related semantic understanding feature vector and a drug inventory multimodal fusion feature vector from the patient prescription information and the pharmacy drug inventory information; S130, generating a drug outbound audit report based on the patient prescription information-related semantic understanding feature vector and the drug inventory multimodal fusion feature vector.

[0062] Here, those skilled in the art will understand that the specific operations of each step in the above-described computer network-based information management method have been referenced above. Figures 1 to 4 The description of the computer network-based information management system is detailed here, and therefore, its repeated description will be omitted.

[0063] Below, for reference Figure 6 This describes an electronic device according to embodiments of the present application.

[0064] Please see Figure 6 , Figure 6The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes:

[0065] The processor 11 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application; the memory 12 can be implemented using a read-only memory (ROM), static storage device, dynamic storage device, or random access memory (RAM), etc.

[0066] The memory 12 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program code is stored in the memory 12 and is called and executed by the processor 11 to execute the information management method based on computer network of the embodiments of this application.

[0067] Input / output interface 13 is used to implement information input and output;

[0068] The communication interface 14 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0069] Bus 15 transmits information between various components of the device (e.g., processor 11, memory 12, input / output interface 13, and communication interface 14);

[0070] The processor 11, memory 12, input / output interface 13 and communication interface 14 are connected to each other within the device via bus 15.

[0071] This application embodiment also provides a storage medium, which is a computer-readable storage medium, storing a computer program that, when executed by a processor, implements the above-described information management method based on a computer network.

[0072] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0073] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0074] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0075] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0076] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0077] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0078] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0079] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0080] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0081] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods in the various embodiments of this application. The aforementioned storage medium includes: USB flash drive, mobile hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk, and other media capable of storing programs. The preferred embodiments of this application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the embodiments of this application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of this application shall be within the scope of the claims of the embodiments of this application.

Claims

1. An information management system based on a computer network, characterized in that, include: The medical information data acquisition module is used to acquire patient prescription information and pharmacy drug inventory information; The medical information data extraction module is used to extract the patient prescription information-related semantic understanding feature vector and the drug inventory multimodal fusion feature vector from the patient prescription information and the pharmacy drug inventory information. The drug delivery audit report generation module is used to generate a drug delivery audit report based on the semantic understanding feature vector associated with the patient prescription information and the multimodal fusion feature vector of the drug inventory.

2. The information management system based on a computer network according to claim 1, characterized in that, The medical information data extraction module includes: The patient prescription information feature extraction unit is used to extract features from the patient prescription information to obtain the semantic understanding feature vector associated with the patient prescription information; The drug inventory information feature extraction unit is used to extract features from the pharmacy drug inventory information to obtain the drug inventory multimodal fusion feature vector.

3. The information management system based on a computer network according to claim 2, characterized in that, The patient prescription information feature extraction unit includes: The patient prescription information is processed through a patient prescription information semantic understanding model to obtain multiple patient prescription information semantic feature vectors; Arrange the semantic feature vectors of the multiple patient prescription information according to the sample dimension to obtain a two-dimensional feature matrix of patient prescription information; The two-dimensional feature matrix of the patient prescription information is passed through a convolutional neural network for the patient prescription information text to obtain the semantic understanding feature vector associated with the patient prescription information.

4. The information management system based on a computer network according to claim 3, characterized in that, The drug inventory information feature extraction unit includes: The feature encoding subunit for the text portion of drug inventory information is used to encode the features of the text portion of pharmacy drug inventory information to obtain the semantic text understanding feature vector of drug inventory. The image feature encoding subunit for drug inventory information is used to encode the features of the image portion of pharmacy drug inventory information to obtain a comprehensive feature vector for enhanced drug inventory. The multimodal feature fusion subunit for drug inventory information is used to fuse the semantic text understanding feature vector of drug inventory and the enhanced comprehensive feature vector of drug inventory to obtain the multimodal fusion feature vector of drug inventory.

5. The information management system based on a computer network according to claim 4, characterized in that, The feature encoding subunit for the text portion of the drug inventory information includes: Extract pharmacy drug inventory text data from the pharmacy drug inventory information; The pharmacy drug inventory text data is processed by a drug inventory text information semantic encoder based on a context convolutional neural network to obtain multiple drug inventory semantic feature vectors. Arrange the multiple drug inventory semantic feature vectors into a drug inventory semantic feature vector; The drug inventory semantic feature vector is processed through the drug inventory semantic multi-scale neighborhood feature extraction module to obtain the drug inventory semantic text understanding feature vector.

6. The information management system based on a computer network according to claim 5, characterized in that, The feature encoding subunit of the drug inventory information image includes: Extract image data of multiple areas of the pharmacy's drug inventory from the pharmacy's drug inventory information; The image data of multiple areas of the pharmacy's drug inventory are processed through a drug inventory image denoising model to obtain denoised and enhanced images of the drug inventory in multiple areas. The multiple regional drug inventory noise reduction and enhancement images are processed by a regional drug inventory noise reduction and enhancement feature extractor based on a convolutional neural network to obtain the comprehensive feature vector of the drug inventory enhancement.

7. The information management system based on a computer network according to claim 6, characterized in that, The drug outbound audit report generation module includes: The medical information feature association unit is used to associate the patient prescription information association semantic understanding feature vector and the drug inventory multimodal fusion feature vector to obtain the patient prescription-drug inventory information association feature vector; The medical information feature optimization unit is used to perform category-based patient prescription-drug inventory information association metric learning compensation on the patient prescription-drug inventory information association feature vector based on metric data to obtain an optimized patient prescription-drug inventory information association feature vector. The drug information report generation unit is used to generate a drug outbound audit report by passing the optimized patient prescription-drug inventory information association feature vector through a generator.

8. The information management system based on a computer network according to claim 7, characterized in that, The medical information feature optimization unit includes: Determine the decoding weight matrix of the generator; The decoding weight matrix is ​​decoupled in units of row vectors to obtain a set of decoding weight vectors; Calculate the patient prescription-drug inventory information association metric learning loss information term between each decoded weight vector in the set of decoded weight vectors and the patient prescription-drug inventory information association feature vector to obtain the patient prescription-drug inventory information association metric learning compensated representation vector; The patient prescription-drug inventory information association metric learning compensation representation vector is input into the activation function to obtain the probabilistic patient prescription-drug inventory information association metric learning compensation weight vector. The positional dot product of the probabilistic patient prescription-drug inventory information association metric learning compensation weight vector and the patient prescription-drug inventory information association feature vector is calculated to obtain the dot product feature vector. The weighted sum of the dot product feature vector and the patient prescription-drug inventory information association feature vector is then calculated to obtain the optimized patient prescription-drug inventory information association feature vector.

9. An information management method based on computer networks, characterized in that, include: Obtain patient prescription information and pharmacy drug inventory information; Extract the semantic understanding feature vector associated with the patient prescription information and the multimodal fusion feature vector of the drug inventory from the patient prescription information and the pharmacy drug inventory information; Based on the semantic understanding feature vector associated with the patient's prescription information and the multimodal fusion feature vector of the drug inventory, a drug outbound audit report is generated.

10. The information management method based on computer networks according to claim 9, characterized in that, Extracting the patient prescription information-related semantic understanding feature vector and the pharmacy drug inventory multimodal fusion feature vector from the patient prescription information and the pharmacy drug inventory information, including: Feature extraction is performed on the patient prescription information to obtain the semantic understanding feature vector associated with the patient prescription information; Feature extraction is performed on the pharmacy's drug inventory information to obtain a multimodal fusion feature vector of the drug inventory.

Images