Medical Device Order Processing Method and Device

Abstract

This application provides a method and apparatus for processing medical device orders. The method includes: determining the data distribution of normal orders based on historical order data in a medical device order management system; identifying abnormal orders in the current order data of the medical device order management system that deviate from the data distribution of normal orders based on a set anomaly detection model; setting the abnormal orders as available orders; outputting available order-grabbing customer service representatives through the output layer of an integrated model; assigning the available orders to the pending work page of the available order-grabbing customer service representatives; and changing the processing permission of the available orders to the available order-grabbing customer service representatives after receiving a list request from them. This application can effectively improve the efficiency and accuracy of order processing in the medical device field.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence, specifically to a method and apparatus for processing medical device orders. Background Technology

[0002] In the medical device industry, an order management system (OMS) is crucial for ensuring the smooth operation of orders from receipt to delivery. An efficient OMS can significantly improve operational efficiency and enhance customer satisfaction, giving companies a competitive edge in a fiercely competitive market. However, with the rapid development of the medical device industry and the increasing market demand, existing OMS systems have also revealed many challenges and limitations.

[0003] Currently, many medical device companies use internal operating systems based on the "whoever saves the data processes it" principle. This means that once an employee processes and stores order information, other employees cannot access or modify the order unless they use the same account. This design severely limits the team's flexibility and responsiveness. Especially when employees take leave, leave, or encounter other emergencies, it can lead to delays in order processing, impacting overall service quality and customer satisfaction.

[0004] Furthermore, order processing in the medical device industry typically needs to adapt to rapidly changing environments, including changes in customer needs, payment delays, or product supply issues. Current systems are not flexible enough in handling these special circumstances, still requiring operation through the original account holder's account, increasing complexity and time delays. This reliance on a single account not only reduces efficiency but can also impact timely order adjustments and responses to urgent customer needs at critical moments.

[0005] Therefore, existing internal operating systems are significantly inadequate in improving work efficiency and responding to emergencies. A more flexible and efficient system is needed to address these challenges, ensuring that medical device companies can maintain efficient operations while responding quickly to changes in market and customer needs. Summary of the Invention

[0006] To address the problems in the prior art, this application provides a medical device order processing method and apparatus, which can effectively improve the efficiency and accuracy of order processing in the medical device field.

[0007] To solve at least one of the above problems, this application provides the following technical solution:

[0008] In a first aspect, this application provides a method for processing medical device orders, including:

[0009] The distribution of normal orders is determined based on historical order data in the medical device order management system. Abnormal orders that deviate from the distribution of normal orders in the current order data of the medical device order management system are identified based on the set anomaly detection model. These abnormal orders are then set as available orders.

[0010] Based on the business parameters, service standards and time requirements in the historical order data, input the preset natural language processing model to determine the corresponding order features, use the order features as the model training set to train the preset time series analysis model, and obtain the order change prediction model.

[0011] The professional qualifications, processing efficiency, and processing success rate of customer service personnel in the medical device order management system are input into a preset natural language processing model to determine the corresponding customer service personnel characteristics. The customer service personnel characteristics are used as the model training set to train a preset recurrent neural network model to obtain a customer service personnel workload prediction model.

[0012] The order features extracted from the order change prediction model and the customer service personnel features extracted from the customer service personnel workload prediction model are input into the feature fusion layer of a preset integrated model for feature fusion. The result of the feature fusion is input into the input layer of the integrated model, and the customer service personnel who can grab orders are output through the output layer of the integrated model. The order features are the order change patterns over time extracted from the order change prediction model, and the customer service personnel features are the output of the hidden layer of the recurrent neural network and the learned temporal relationship extracted from the customer service personnel workload prediction model.

[0013] The available orders are assigned to the pending work page of the customer service representative who is eligible to accept them, and upon receiving a list request from the customer service representative who is eligible to accept them, the processing authority for the available orders is changed to that customer service representative.

[0014] Further, the step of determining the data distribution of normal orders based on historical order data in the medical device order management system, identifying abnormal orders in the current order data of the medical device order management system that deviate from the data distribution of normal orders based on a set anomaly detection model, and setting the abnormal orders as available-for-order orders includes:

[0015] Statistical analysis was performed on historical order data in the medical device order management system to determine the data distribution of normal orders that meet the normal judgment rules for order amount, order time, and order type.

[0016] Calculate the Mahalanobis distance between the current order data and the data distribution of normal orders in the medical device order management system. Determine the current order whose calculated Mahalanobis distance value exceeds the threshold as an abnormal order and set the abnormal order as an order that can be bid on in the medical device order management system.

[0017] Further, the step of determining corresponding order features by inputting business parameters, service standards, and time requirements from the historical order data into a preset natural language processing model, and using the order features as a model training set to train a preset time series analysis model to obtain an order change prediction model includes:

[0018] The business parameters, service standards, and time requirements in the historical order data are converted into structured order features using a preset natural language processing model, and order feature time series data is constructed based on the order features.

[0019] Order feature time series data are input into a preset time series analysis model for model training. The model parameters and hyperparameters are iteratively adjusted according to the model training results until they meet the expected model output, thus obtaining an order change prediction model.

[0020] Further, the step of determining corresponding customer service personnel characteristics by inputting the professional qualifications, processing efficiency, and processing success rate of customer service personnel in the medical device order management system into a preset natural language processing model, and using the customer service personnel characteristics as a model training set to train a preset recurrent neural network model to obtain a customer service personnel workload prediction model, includes:

[0021] The professional qualifications, processing efficiency, and processing success rate of customer service personnel in the medical device order management system are converted into structured customer service personnel features through a preset natural language processing model, and a model training set is constructed based on the customer service personnel.

[0022] The training set of the model is input into a preset recurrent neural network model, and the recurrent neural network model is iteratively trained by setting a mean squared error loss function to obtain a customer service staff workload prediction model.

[0023] Further, the step of inputting the order features extracted from the order change prediction model and the customer service personnel features extracted from the customer service personnel workload prediction model into the feature fusion layer of the preset integrated model for feature fusion includes:

[0024] The order change prediction model extracts the pattern of order changes over time and uses it as order features; the customer service staff workload prediction model extracts the output of the hidden layer of the recurrent neural network and the learned temporal relationship as customer service staff work features.

[0025] The order features and the customer service personnel work features are input into the feature fusion layer of the reinforcement learning model for feature fusion operation.

[0026] Furthermore, the output layer of the integrated model that provides customer service representatives capable of accepting orders includes:

[0027] The order characteristics and customer service personnel work characteristics are set as the state space of the integrated model, multiple preset order grabbing strategies are set as the action space of the integrated model, and the success of order grabbing post-processing and the failure of order grabbing post-processing are set as the reward function of the integrated model, thus constructing a reinforcement learning model.

[0028] The reinforcement learning model outputs a customer service representative who can grab orders through its output layer, wherein the reinforcement learning model iteratively optimizes the order-grabbing strategy in the action space through the reward function.

[0029] Furthermore, after changing the processing authority for the order-grabbing service to the order-grabbing customer service representative, the following steps are included:

[0030] Continuously monitor the real-time workload of the customer service representatives who are eligible to accept orders;

[0031] When the real-time workload exceeds a threshold, any unprocessed orders will be automatically transferred to other customer service representatives with lower real-time workloads.

[0032] Secondly, this application provides a medical device order processing device, comprising:

[0033] The order anomaly determination module is used to determine the data distribution of normal orders based on historical order data in the medical device order management system, identify abnormal orders in the current order data of the medical device order management system that deviate from the data distribution of normal orders based on a set anomaly detection model, and set the abnormal orders as available orders.

[0034] The order change prediction module is used to input business parameters, service standards and time requirements in the historical order data into a preset natural language processing model to determine the corresponding order features, and use the order features as a model training set to train a preset time series analysis model to obtain the order change prediction model.

[0035] The customer service staff workload prediction module is used to determine the corresponding customer service staff characteristics based on the professional qualifications, processing efficiency and processing success rate of customer service staff in the medical device order management system, and to train the preset recurrent neural network model using the customer service staff characteristics as the model training set to obtain the customer service staff workload prediction model.

[0036] The order-grabbing customer service matching module is used to input the order features extracted from the order change prediction model and the customer service personnel features extracted from the customer service personnel workload prediction model into the feature fusion layer of a preset integrated model for feature fusion. The result of the feature fusion is input into the input layer of the integrated model, and the order-grabbing customer service personnel are output through the output layer of the integrated model. The order features are the order change patterns over time extracted from the order change prediction model, and the customer service personnel features are the output of the hidden layer of the recurrent neural network and the learned temporal relationship extracted from the customer service personnel workload prediction model.

[0037] The intelligent order transfer module is used to assign the available orders to the pending work page of the customer service representative who can grab the orders, and after receiving the list request sent by the customer service representative who can grab the orders, change the processing permission of the available orders to that customer service representative.

[0038] Thirdly, this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the medical device order processing method.

[0039] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the medical device order processing method described above.

[0040] Fifthly, this application provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of the medical device order processing method described above.

[0041] As can be seen from the above technical solution, this application provides a medical device order processing method and apparatus. It determines the data distribution of normal orders through historical order data in a medical device order management system, identifies abnormal orders in the current order data of the medical device order management system that deviate from the data distribution of normal orders based on a set anomaly detection model, and sets these abnormal orders as available for processing. The available customer service representatives are output through the output layer of the integrated model. These available orders are assigned to the pending work page of the available customer service representatives, and upon receiving a list request from the available customer service representatives, the processing authority for the available orders is changed to that representative. This effectively improves the efficiency and accuracy of order processing in the medical device field. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is one of the flowcharts illustrating the medical device order processing method in the embodiments of this application;

[0044] Figure 2 This is a second flowchart illustrating the medical device order processing method in the embodiments of this application;

[0045] Figure 3 This is the third flowchart illustrating the medical device order processing method in the embodiments of this application;

[0046] Figure 4 This is the fourth flowchart illustrating the medical device order processing method in the embodiments of this application;

[0047] Figure 5 This is the fifth flowchart illustrating the medical device order processing method in the embodiments of this application;

[0048] Figure 6 This is the sixth flowchart illustrating the medical device order processing method in the embodiments of this application;

[0049] Figure 7 This is the seventh flowchart illustrating the medical device order processing method in the embodiments of this application;

[0050] Figure 8 This is a structural diagram of the medical device order processing device in the embodiments of this application;

[0051] Figure 9 This is a schematic diagram of the structure of the electronic device in the embodiments of this application.

[0052] Figure label:

[0053] Electronic device 9600, central processing unit 9100, memory 9140, communication module 9110, input unit 9120, audio processor 9130, display 9160, power supply 9170, buffer memory 9141, application / function storage unit 9142, data storage unit 9143, driver storage unit 9144, antenna 9111, speaker 9131, microphone 9132. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0055] The acquisition, storage, use, and processing of data in this application all comply with the relevant provisions of national laws and regulations.

[0056] Considering that order processing in the medical device industry often needs to adapt to rapidly changing environments, including changes in customer needs, payment delays, or product supply issues, current systems lack flexible mechanisms for handling these special situations. They still require operation through the original account holder's account, increasing complexity and time delays. This reliance on a single account not only reduces efficiency but can also affect timely order adjustments and responses to urgent customer needs at critical moments. This application provides a medical device order processing method and apparatus. It determines the data distribution of normal orders using historical order data in a medical device order management system, identifies abnormal orders in the current order data that deviate from the normal order data distribution based on a pre-defined anomaly detection model, and sets these abnormal orders as available for processing. The method outputs available customer service representatives through the integrated model's output layer, assigns these available orders to the pending work page of each representative, and changes the processing permission to that representative upon receiving a list request. This effectively improves the efficiency and accuracy of order processing in the medical device field.

[0057] To effectively improve the efficiency and accuracy of order processing in the medical device field, this application provides an embodiment of a medical device order processing method, see [link to embodiment]. Figure 1 The medical device order processing method specifically includes the following:

[0058] Step S101: Determine the data distribution of normal orders based on historical order data in the medical device order management system, identify abnormal orders in the current order data of the medical device order management system that deviate from the data distribution of normal orders based on the set anomaly detection model, and set the abnormal orders as orders that can be snatched.

[0059] Optionally, in this embodiment, the technical principles, working process, and technical effects of step S101 are described in detail. This step addresses the abnormal order detection and processing strategy in a medical device order management system, aiming to improve the system's order processing efficiency and security.

[0060] First, the technical principle relies on constructing a normal order data distribution model based on historical order data. This application's embodiments analyze historical order data, including multi-dimensional data such as order amount, time, content, and customer information, and utilize statistical methods or machine learning algorithms (e.g., Gaussian distribution model, cluster analysis, etc.) to establish a data model describing the characteristics of normal orders, serving as a benchmark for subsequent order normality judgments.

[0061] Next, the deployment of the anomaly detection model is the core of step S101. The model used in this embodiment utilizes a pre-defined data distribution model as a reference, identifying abnormal orders by comparing the deviation between the current order data and the data distribution of normal orders. This deviation may involve unusual order amounts, times, purchasing behavior, or anomalies in customer information and order combinations. Therefore, anomaly detection can employ various techniques, such as threshold-based detection, probability-based anomaly scoring systems, and machine learning anomaly detection algorithms (such as support vector machines and isolated forests).

[0062] In the operational process, this embodiment first extracts the order data to be processed from the database, and then applies an anomaly detection model to analyze each order data. The model calculates an "anomaly score" for each order by comparing it with the normal data model. When the anomaly score of an order exceeds a preset threshold, the order is marked as an abnormal order. This process is automated and can be performed in real time, ensuring that all orders entering the system are evaluated promptly and accurately.

[0063] Furthermore, in this embodiment, marking identified abnormal orders as "orders available for grabbing" is an innovative measure. In practice, this means that these orders will be flagged and prioritized for processing in the system. The purpose of marking them as orders available for grabbing is to attract more review resources, such as manual review or more detailed customer background investigations, to ensure that potential risks are not overlooked due to the limitations of automated system processing.

[0064] In terms of technical effectiveness, the embodiments of this application significantly improve the intelligence and security of the medical device order management system. Through automated anomaly detection, the system can quickly identify potentially risky orders, reducing human error and operational delays. Simultaneously, setting abnormal orders as available for preemption strengthens the management and control of these orders, improves customer service quality, and protects the interests of both the company and its customers. Furthermore, the implementation of the system can also serve as part of enterprise risk management, helping companies better comply with industry regulations and enhance their market competitiveness.

[0065] In summary, the embodiments of this application provide an efficient and secure solution for the medical device order management system through the technical means and workflows employed, effectively improving the accuracy of order processing and the overall responsiveness of the system, while reducing potential risks caused by order anomalies.

[0066] Step S102: Based on the business parameters, service standards and time requirements in the historical order data, input the data into a preset natural language processing model to determine the corresponding order features. Use the order features as the model training set to train the preset time series analysis model to obtain the order change prediction model.

[0067] Optionally, in this embodiment, step S102 involves using historical order data to develop a model for predicting order changes through natural language processing and time series analysis models. This step aims to improve the predictive capabilities of the order management system, thereby more effectively responding to market demands and optimizing resource allocation.

[0068] First, this embodiment of the application uses a Natural Language Processing (NLP) model to parse business parameters, service standards, and time requirements from historical order data. In this process, the NLP model is primarily responsible for extracting key information from the order text data, such as product type, customer-required service level, and delivery time; these are crucial features for subsequent modeling.

[0069] The extracted order features are then used as the model training set and input into the time series analysis model. Time series analysis is a statistical technique used to analyze data points ordered over time to identify long-term trends, periodic fluctuations, etc. In this embodiment, a suitable time series model (such as ARIMA, seasonal ARIMA, or a complex machine learning model such as an LSTM network) is selected to process the data, aiming to build a model capable of predicting order changes.

[0070] In the process, this embodiment first extracts historical order data from a database or order management system, and then processes this data using a preset NLP model to extract relevant features. The extracted features include, but are not limited to, order volume, delivery cycle, and customer satisfaction, which reflect the business parameters and service standards of the orders. These feature data are then used as training data and input into a time series analysis model for training, thereby constructing a model capable of predicting future order changes.

[0071] Once trained, this model can be used to predict future order trends, such as increases or decreases in order volume and changes in customer demand. By predicting these changes, the order management system can prepare in advance, such as adjusting inventory, optimizing production plans, and adjusting human resources accordingly, thereby responding more effectively to market changes.

[0072] In terms of technical effectiveness, the embodiments of this application, by combining natural language processing and time series analysis technologies, not only improve the accuracy of order data analysis but also enhance the predictive capabilities of the order management system. This enables the system to respond more flexibly to changes in market demand, improve resource utilization efficiency, and reduce risks caused by market fluctuations. Furthermore, more accurate order forecasting also helps improve customer satisfaction because the timeliness and accuracy of service are significantly enhanced.

[0073] Step S103: Based on the professional qualifications, processing efficiency, and processing success rate of customer service personnel in the medical device order management system, input the corresponding customer service personnel characteristics into a preset natural language processing model, use the customer service personnel characteristics as a model training set to train the preset recurrent neural network model, and obtain a customer service personnel workload prediction model.

[0074] Optionally, in this embodiment, step S103 describes how to use Natural Language Processing (NLP) technology and Recurrent Neural Network (RNN) models to predict the workload saturation of customer service personnel in a medical device order management system. Implementing this step helps to rationally allocate customer service resources and improve overall service efficiency and quality.

[0075] First, this embodiment of the application analyzes data such as customer service personnel's professional qualifications, processing efficiency, and success rate using a natural language processing (NLP) model. In practice, this data is typically contained in employee work records, evaluation reports, and feedback, and may exist in the form of natural language. The task of the NLP model is to extract key information from this textual data, such as the level of qualification certification, average order processing time, and customer satisfaction rating. This information constitutes the characteristics of the customer service personnel.

[0076] The extracted customer service personnel features were then used as training data and input into a pre-defined recurrent neural network model. A recurrent neural network is a type of neural network model suitable for processing sequential data, capable of handling time-series data, and is particularly well-suited for predicting behaviors that change over time, such as employee workload and work efficiency.

[0077] During model training, the RNN model learns how to predict the future workload of customer service representatives based on their historical performance. This includes analyzing the frequency of peak work periods, changes in order processing speed, and fluctuations in success rates. Once the model is trained, it can predict the workload of each customer service representative over a future period, i.e., their potential workload.

[0078] The effectiveness of this predictive model is evident in several aspects: First, by predicting the workload of customer service staff, managers can conduct more scientific human resource planning, such as increasing manpower when high workloads are predicted, or providing training and rest when workloads are low. Second, this prediction also helps balance the workload of various customer service staff, preventing some employees from becoming overworked and affecting service quality. Finally, the predictive model can also serve as a basis for customer service performance evaluation and training needs analysis, thereby continuously improving the overall performance of the service team.

[0079] In summary, step S103, by combining NLP and RNN technologies, not only optimized the allocation of customer service resources but also improved the efficiency and quality of customer service work, effectively supporting the operation of the medical device order management system.

[0080] Step S104: Input the order features extracted from the order change prediction model and the customer service personnel features extracted from the customer service personnel workload prediction model into the feature fusion layer of the preset integrated model for feature fusion. Input the result of the feature fusion into the input layer of the integrated model and output the customer service personnel who can grab orders through the output layer of the integrated model. The order features are the order change patterns over time extracted from the order change prediction model, and the customer service personnel features are the output of the hidden layer of the recurrent neural network and the learned temporal relationship extracted from the customer service personnel workload prediction model.

[0081] Optionally, in step S104, the embodiment describes how to integrate the features extracted from the order change prediction model and the customer service staff workload prediction model, and use the integrated model to predict the customer service staff suitable for bidding. The key to this step lies in the application of feature fusion technology, as well as the construction and application of the integrated model, with the aim of achieving more accurate service staff matching and resource allocation.

[0082] First, the order features extracted from the order change prediction model include patterns in order changes over time, such as periodic fluctuations and seasonal variations in order volume. These features reflect the dynamic changes in order demand and are crucial for predicting future order trends.

[0083] Meanwhile, the features extracted from the customer service staff workload prediction model are mainly based on the outputs of the hidden layers of the recurrent neural network. These outputs capture the efficiency and success rate of customer service staff in processing orders, as well as their changes over time. These features reflect the workload and efficiency of customer service staff and play an important role in predicting their future workload.

[0084] Next, these extracted features are fed into the feature fusion layer of the pre-defined ensemble model. The role of the feature fusion layer is to integrate features from different sources to form a comprehensive feature set, which helps the model to understand the situation more comprehensively and make decisions. After feature fusion, the comprehensive feature data is fed into the input layer of the ensemble model.

[0085] Ensemble models may employ various machine learning algorithms, such as deep learning, decision tree ensemble methods (e.g., random forests), and gradient boosting machines. The specific choice depends on the specific needs of the problem and the characteristics of the data. The model learns through training how to predict the most suitable customer service representative to accept orders based on current order demand and the workload of existing customer service staff.

[0086] Finally, the model outputs a list of customer service representatives who can accept orders. This output is based on the model's assessment of the current and future working status of each customer service representative, aiming to optimize the allocation of customer service resources and improve response efficiency and customer satisfaction.

[0087] Overall, step S104 achieves dynamic matching between order requirements and customer service resources through efficient feature fusion and intelligent integration models. This not only improves the operational efficiency of the medical device order management system but also optimizes the customer service experience.

[0088] Step S105: Assign the available orders to the pending work page of the customer service representative who can accept orders, and after receiving the list request sent by the customer service representative who can accept orders, change the processing permission of the available orders to that customer service representative.

[0089] Optionally, in step S105, the entire process revolves around effectively and efficiently allocating available orders to suitable customer service personnel and ensuring the correct transfer of order processing authority. This step is a crucial link in the service allocation system, directly affecting the timeliness of service response and customer satisfaction. The following is a detailed description of the execution process and internal logic of this step.

[0090] First, in step S104, the system needs to determine which customer service representatives are "order-grabbing representatives" using an integrated model. This process involves complex data analysis and predictive models, including predicting order change trends and assessing the workload of customer service representatives. Once the model identifies the most suitable customer service representatives, the system will automatically assign the order-grabbing representatives to their pending work pages.

[0091] This automated allocation process requires the backend system to have real-time data processing and update capabilities. The system monitors order status and customer service availability, ensuring that each new available order is quickly displayed on the corresponding customer service representative's work interface. This real-time update mechanism not only improves work efficiency but also reduces customer waiting time and enhances the overall quality of service.

[0092] After orders automatically appear on the customer service representative's pending task page, the representative can view the order details and decide whether to accept the order based on their workload and expertise. If the representative decides to accept the order, they need to send a list request through the system interface. This is usually done by clicking the "Request Processing" or "Accept Order" button. This action is the customer service representative's formal application to accept the order, triggering a series of permission reviews and update processes in the system backend.

[0093] Upon receiving a list request, the system first verifies its validity, including checking the order's current status (ensuring the order hasn't been processed) and verifying the customer service representative's qualifications (ensuring the representative has the authority to handle this type of order). Once verification is successful, the system changes the order's processing authority from a pending state to a state exclusively held by that customer service representative. This authority change is achieved by modifying the order record in the database, including updating the current handler and status fields, thus ensuring the order's status in the system accurately reflects its actual processing status.

[0094] In addition, the system needs to synchronize status updates after permission changes to ensure that all relevant personnel (including other members of the customer service team, the order submitter, etc.) can see the changes in order status in real time. This transparency can reduce communication costs and improve team collaboration efficiency.

[0095] The implementation of the entire S105 process not only relies on strong technical support, including real-time data processing, database management, and user interface design, but also requires rigorous process management to ensure that each step is executed accurately. This design and implementation guarantees high service efficiency and customer satisfaction, making it an indispensable part of modern service business management systems.

[0096] Through the detailed description above, this embodiment demonstrates the importance and complexity of step S105 in the entire order processing flow. This is not merely a matter of technical implementation, but also a challenge in process design and user experience. Correctly executing this step can significantly improve service efficiency and customer satisfaction, and is key to enhancing a company's competitiveness.

[0097] As described above, the medical device order processing method provided in this application can determine the data distribution of normal orders through historical order data in the medical device order management system, identify abnormal orders in the current order data of the medical device order management system that deviate from the data distribution of normal orders according to a set anomaly detection model, and set the abnormal orders as available orders; output available order-grabbing customer service through the output layer of the integrated model; assign the available orders to the pending work page of the available order-grabbing customer service, and change the processing permission of the available orders to the available order-grabbing customer service after receiving the list request sent by the available order-grabbing customer service, thereby effectively improving the efficiency and accuracy of order processing in the medical device field.

[0098] In one embodiment of the medical device order processing method of this application, see [link to embodiment]. Figure 2 It can also specifically include the following:

[0099] Step S201: Perform data statistics on historical order data in the medical device order management system to determine the data distribution of normal orders that meet the normal judgment rules for order amount, order time, and order type;

[0100] Step S202: Calculate the Mahalanobis distance between the data distribution of the current order data and the normal order data in the medical device order management system. Determine the current order whose calculated Mahalanobis distance value exceeds the threshold as an abnormal order, and set the abnormal order as an order that can be snatched in the medical device order management system.

[0101] Optionally, in steps S201 and S202, this embodiment addresses a complex data analysis and anomaly detection process for a medical device order management system. This is a crucial step in ensuring the effectiveness and security of the order processing flow. Specifically, step S201 involves statistical analysis of historical order data to establish a data distribution model for normal orders. Step S202, based on this model, identifies abnormal orders by calculating the Mahalanobis distance between the current order and the normal data distribution, and marks these abnormal orders as eligible for bidding for special handling. This process is not only essential for maintaining the normal operation of the system but also plays a vital role in preventing potential fraudulent activities.

[0102] First, step S201 requires this embodiment to perform in-depth data statistics on historical order data in the medical device order management system. This includes, but is not limited to, parameters such as order amount, order time, and order type. By analyzing these parameters, this embodiment can determine which orders conform to normal business behavior patterns. For example, the normal determination rule for order amount may be based on the distribution of amount in past data, such as the mean and standard deviation; the normal determination rule for order time may take into account the cyclical characteristics of business peaks and troughs; and the normal determination of order type may be based on the proportion and frequency of various types of orders in historical orders. The key to this step is to establish a statistically significant data model that can represent normal business activities. This model will serve as the benchmark for anomaly detection in subsequent steps.

[0103] Next, in step S202, this embodiment uses Mahalanobis distance to measure the similarity between the current order data point and the historical normal order data distribution. Mahalanobis distance is a distance metric that considers the covariance of the dataset and is well-suited for outlier detection when there is correlation between different dimensions of the data. In the medical device order management system, by calculating the Mahalanobis distance between each current order and the normal order data distribution, this embodiment can quantify the degree of deviation of each order from the normal behavior pattern. If the Mahalanobis distance of an order exceeds a preset threshold, this indicates that the order differs significantly from the normal order pattern and can therefore be identified as an abnormal order.

[0104] Furthermore, marking these identified abnormal orders as available for bidding is based on preventative and investigative needs. Abnormal orders may involve errors or fraud and need to be prioritized to avoid potential financial losses or legal risks. In practice, these orders are pushed to the work interface of specific customer service or review teams, who will conduct a more detailed manual review or contact the customer directly to confirm the authenticity and accuracy of the order.

[0105] The steps S201 and S202 not only demonstrate the application of data science in modern medical device order management, but also reflect how statistical and machine learning techniques can be used to enhance the security and efficiency of business processes in the era of big data. Through these technological means, enterprises can better control and manage business risks, while also improving the quality and responsiveness of customer service. The implementation of this process requires close cross-departmental collaboration, including the joint efforts of IT, data analytics, customer service, and risk management teams, to ensure accuracy and real-time performance at every stage. Furthermore, regular model evaluation and updates are necessary to adapt to the ever-changing market environment and business models.

[0106] From a technical implementation perspective, the data statistics and analysis in step S201 can be accomplished using various statistical software or programming languages ​​such as Python and R. These tools provide powerful data processing and analysis libraries to help analysts perform complex data calculations and visualizations. However, the Mahalanobis distance calculation in step S202 requires knowledge of linear algebra and statistics, involving mathematical operations such as the inverse of the covariance matrix. These operations can be efficiently implemented using scientific computing libraries such as NumPy.

[0107] In one embodiment of the medical device order processing method of this application, see [link to embodiment]. Figure 3 It can also specifically include the following:

[0108] Step S301: Convert the business parameters, service standards, and time requirements in the historical order data into structured order features using a preset natural language processing model, and construct order feature time series data based on the order features;

[0109] Step S302: Input the order feature time series data into the preset time series analysis model for model training. Iteratively adjust the model parameters and hyperparameters according to the model training results until they meet the expected model output, and obtain the order change prediction model.

[0110] Optionally, in step S301, this embodiment first processes historical order data, which includes extracting information from diverse business parameters, service standards, and time requirements. To achieve this, this embodiment employs advanced Natural Language Processing (NLP) techniques. These techniques can understand and parse key information in text data, converting it into a machine-readable structured format. Specifically, this embodiment can use pre-trained language models such as BERT or GPT, which are trained on large-scale corpora and can capture subtle differences in language, thereby accurately extracting key business parameters. This step is fundamental to the entire data processing flow, ensuring data quality and the accuracy of subsequent analysis.

[0111] After the data is structured, this embodiment further transforms this structured data into feature time series. Time series data reflects the changing trends of order characteristics over time, which is crucial for understanding order dynamics. For example, this embodiment may track changes in order amounts or seasonal fluctuations in order processing times. Constructing such time series data not only helps this embodiment capture the time-dependent characteristics of the business but also provides a foundation for subsequent predictive models.

[0112] Next, in step S302, this embodiment inputs these order feature time series data into a preset time series analysis model for training. Time series analysis is a statistical method used to analyze time series data to extract meaningful statistical and other features. This embodiment may employ models such as ARIMA (Autoregressive Integral Moving Average) or more modern methods such as LSTM (Long Short-Term Memory) models, which are capable of handling and predicting complex patterns in time series data.

[0113] During model training, this embodiment iteratively adjusts the model's parameters and hyperparameters. Parameter tuning is an experimental and optimization process that needs to be based on the model's predictive performance. For example, this embodiment may need to adjust the number of neural layers or the number of neurons per layer in the LSTM model to achieve the best learning effect. In addition, hyperparameter tuning, such as learning rate and batch size, is equally important, as these directly affect the model's training speed and effectiveness.

[0114] This series of parameter and hyperparameter adjustments is to ensure that the model output meets expectations, namely, its ability to accurately predict future order changes. Through continuous model evaluation and parameter tuning, this embodiment ultimately yields a highly efficient order change prediction model. This model can not only predict future order trends but also help companies make more data-driven decisions, such as inventory management and resource allocation.

[0115] Through these technologies and methods, the functionality of the medical device order management system has been greatly enhanced. Furthermore, the entire implementation process demonstrates the value of data science in practical business operations. By utilizing these advanced analytical tools and methods, companies can better understand market dynamics and prepare in advance to meet future challenges.

[0116] In one embodiment of the medical device order processing method of this application, see [link to embodiment]. Figure 4 It can also specifically include the following:

[0117] Step S401: Convert the professional qualifications, processing efficiency, and processing success rate of customer service personnel in the medical device order management system into structured customer service personnel features through a preset natural language processing model, and construct a model training set based on the customer service personnel.

[0118] Step S402: Input the model training set into a preset recurrent neural network model, and iteratively train the recurrent neural network model by setting the mean squared error loss function to obtain a customer service staff workload prediction model.

[0119] Optionally, in step S401, the focus is on using Natural Language Processing (NLP) technology to transform unstructured data such as the professional qualifications, processing efficiency, and success rate of customer service personnel in the medical device order management system into structured customer service personnel features. This step is crucial because it directly affects the training quality and prediction accuracy of subsequent models.

[0120] First, it's necessary to summarize and organize the evaluations and performance records of customer service personnel. These records may exist in text form, such as work reports, performance evaluations, and customer feedback. Using a pre-defined natural language processing model, this text data is parsed and transformed into structured data. For example, a customer service representative's processing efficiency can be quantified by their average processing time and the number of orders processed per day, while the success rate can be measured by the ratio of completed orders to total orders.

[0121] After the data is structured, this information is organized into a model training set. This training set includes not only various performance indicators of customer service personnel, but also other relevant information, such as the number of years of work experience and educational background of customer service personnel, which are important factors that may affect their work performance.

[0122] Next, in step S402, the model training set is input into a pre-defined recurrent neural network (RNN) model. RNNs are particularly well-suited for processing sequential data, making them ideal for analyzing and predicting behaviors such as customer service staff workload over time. Here, each customer service staff member's historical performance data is treated as a time series, and the model predicts future workload by learning patterns within these time series.

[0123] To train this recurrent neural network, this embodiment uses mean squared error (MSE) as the loss function. MSE is the average of the squared differences between predicted and true values, and it is a commonly used performance metric in regression tasks. By minimizing the MSE during training, the model can fit the training data more accurately, thereby improving the accuracy of predictions for future data.

[0124] During the iterative training of the model, this embodiment continuously adjusts network parameters, such as weights and biases. It may also require adjustments to the network architecture or hyperparameters, such as the learning rate, batch size, and number of iterations. This process requires significant computational resources and time, but with appropriate techniques and strategies, such as early stopping and regularization, overfitting can be effectively avoided, and the model's generalization ability can be improved.

[0125] Ultimately, through these meticulous design and training steps, this embodiment yields a neural network model capable of predicting customer service staff workload. This model can not only help management allocate customer service resources more effectively and prevent staff from overworking, but it can also be used to develop more reasonable work plans, thereby improving the overall efficiency of the customer service department and customer satisfaction.

[0126] In summary, through this highly structured data processing and advanced machine learning technology, the functionality of the medical device order management system has been substantially enhanced. This not only optimizes the allocation of customer service resources and improves work efficiency, but also leverages technology to improve the overall service quality and customer satisfaction of the enterprise. Such technological application is particularly important because it directly relates to the company's operational efficiency and customer relationship management. In practice, this technology helps companies better understand the workload distribution of their customer service teams, predict peak periods, and thus prepare human resources and scheduling in advance.

[0127] Furthermore, by accurately predicting the workload of customer service staff, management can allocate tasks more rationally, preventing a decline in service quality due to staff overload. This plays a crucial role in maintaining customer satisfaction and loyalty. For example, if increased customer service pressure is anticipated during a certain period, management can make advance adjustments to staffing or use technological means such as automation tools and AI-assisted systems to alleviate the workload, ensuring that customer issues are resolved promptly and effectively.

[0128] Furthermore, the establishment and application of such predictive models also provide enterprises with data-driven decision-making support. Enterprises can adjust their business strategies, optimize service processes, and even strategically reposition the role and function of their customer service centers based on the information output by the model. For example, if data shows that certain types of problems frequently lead to high customer service saturation, the enterprise may consider conducting a more in-depth analysis of these problems to identify the root causes and reduce their occurrence by improving products or services.

[0129] Furthermore, data-driven customer service management models can help companies make more targeted decisions regarding employee training and development. By analyzing the performance data of individual customer service staff, companies can identify which employees may need additional training or support, and which high-performing employees should be given more development opportunities. This not only improves the job satisfaction and professional achievement of individual employees but also helps improve the overall service quality and efficiency of the team.

[0130] In terms of technical implementation, recurrent neural network models provide a powerful tool for processing time series data, but they also bring challenges, such as model complexity and computational resource requirements. To address these challenges, enterprises need to invest corresponding technical resources and professional personnel. At the same time, as these models are increasingly integrated into daily enterprise operations, higher demands are being placed on data quality and the management of processing workflows.

[0131] In summary, by implementing such a customer service staff workload prediction model in a medical device order management system, companies can not only improve the efficiency and effectiveness of customer service operations, but also optimize human resource management and customer service strategies on a broader level. This will enable companies to maintain a competitive edge in a highly competitive market, while simultaneously improving customer satisfaction and the company's long-term sustainable development capabilities.

[0132] In one embodiment of the medical device order processing method of this application, see [link to embodiment]. Figure 5 It can also specifically include the following:

[0133] Step S501: Extract the pattern of order changes over time from the order change prediction model and use it as order features; extract the output of the hidden layer of the recurrent neural network and the learned temporal relationship from the customer service staff workload prediction model and use it as customer service staff work features.

[0134] Step S502: Input the order features and the customer service personnel work features into the feature fusion layer of the reinforcement learning model for feature fusion operation.

[0135] Optionally, in step S501, this embodiment extracts important features from two key models. First, from the order change prediction model, this embodiment extracts patterns regarding order changes over time. These patterns reflect trends in order quantity, type, and frequency, which are crucial for understanding market demand fluctuations and seasonal patterns. For example, if the prediction model shows a sharp increase in order quantity before and after holidays, this time-series pattern can be transformed into order features for subsequent analysis and decision-making.

[0136] Next, this embodiment extracts the output of the hidden layer of the recurrent neural network and the temporal relationships learned by the model from the customer service staff workload prediction model. The output of the hidden layer can capture the internal representation of the customer service staff's workload, while the temporal relationships reveal how the customer service staff's workload changes over time. These features not only include the current working state of the customer service staff, but also the trend of their working state, which is extremely important for predicting future workload.

[0137] In step S502, these features extracted from different sources—order features and customer service personnel job features—are input into a feature fusion layer of a reinforcement learning model. The main task of the feature fusion layer is to integrate features from different models to generate a comprehensive feature representation, which is the basis for the reinforcement learning model to make decisions.

[0138] The purpose of applying reinforcement learning models here is to optimize customer service resource allocation and order processing strategies. Through feature fusion, the model can comprehensively understand the changing patterns of order volume and the working status of customer service personnel, thereby formulating the optimal resource allocation plan. For example, if the model predicts that orders will increase in the near future, while the workload of customer service personnel is already close to saturation, the reinforcement learning model can suggest increasing the number of customer service personnel or adjusting their working hours to cope with the upcoming peak period.

[0139] Reinforcement learning models learn optimal strategies through interaction with their environment. In this scenario, each decision made by the model not only affects current customer service efficiency but may also impact future order processing capacity and customer satisfaction. The design of the feature fusion layer is crucial because it ensures the model receives all relevant information, making the decision-making process more accurate and efficient.

[0140] In practice, feature fusion may involve various techniques, such as principal component analysis (PCA), autoencoders, or attention mechanisms in deep learning. These techniques can effectively integrate high-dimensional data from different sources. Furthermore, to improve the performance of reinforcement learning models, it may be necessary to standardize the fused features to eliminate the dimensionality differences between them and ensure the stability of model training.

[0141] In summary, through steps S501 and S502, this embodiment effectively integrates key business insights and data-driven forecasting into the practical operations of customer service and order management. This reinforcement learning-based approach not only improves the data-dependent and intelligent level of decision-making but also provides enterprises with a powerful tool for dynamically adjusting and optimizing resource allocation to adapt to ever-changing market conditions.

[0142] In one embodiment of the medical device order processing method of this application, see [link to embodiment]. Figure 6 It can also specifically include the following:

[0143] Step S601: Set the order features and customer service personnel work features as the state space of the integrated model, set multiple preset order grabbing strategies as the action space of the integrated model, and set successful order grabbing and failed order grabbing as the reward function of the integrated model to construct a reinforcement learning model;

[0144] Step S602: Output order-grabbing customer service through the output layer of the reinforcement learning model, wherein the reinforcement learning model iteratively optimizes the order-grabbing strategy in the action space through the reward function.

[0145] Optionally, in step S601, this embodiment designs and constructs an integrated reinforcement learning model to optimize customer service order-grabbing strategies. First, this embodiment defines the order features and customer service personnel work features extracted from step S501 as the model's state space. These features represent all the environmental information the model needs to consider when making decisions, such as order quantity, order type, current workload of customer service personnel, and historical workload.

[0146] The action space consists of multiple preset order-grabbing strategies, which may include, but are not limited to, order-grabbing based on the fastest response time, order-grabbing based on customer service load balancing, or order-grabbing based on customer satisfaction optimization. The choice of which order-grabbing strategy to use depends on the current state space and the evaluation of historical performance.

[0147] To effectively train and optimize these order-grabbing strategies, this embodiment defines a reward function based on the processing outcome after order grabbing—success or failure. Successful processing may result in a positive reward, while failure may result in a negative reward or a lower positive reward. This reward mechanism incentivizes the model to learn how to select the optimal action based on the current state to maximize success rate and customer satisfaction.

[0148] Next, in step S602, the reinforcement learning model determines the customer service representative who should accept the order through its output layer. At each decision point, the model receives the current environmental state (i.e., order characteristics and customer service job characteristics) and calculates the expected utility of each possible order-accepting strategy. In this way, the model can dynamically adjust its strategy selection to respond to changes in the environment and the results of previous decisions.

[0149] The model iteratively optimizes through feedback from the reward function. In practice, this typically involves an exploration-exploitation tradeoff, where the model not only utilizes the known best policy to obtain immediate rewards but also explores less frequently used but potentially effective policies. This exploration mechanism helps the model find better decision-making strategies in the long run.

[0150] In practical deployments, this reinforcement learning model can be integrated into customer service systems to receive order and customer service status information in real time and dynamically allocate orders to customer service personnel. This not only optimizes the use of customer service resources and improves processing efficiency but also enhances overall customer satisfaction, as the model learns how to adopt the most suitable order-grabbing strategy in different situations.

[0151] Through such a system, enterprises can manage massive customer service demands more intelligently and efficiently, while continuously improving service quality and customer experience through iterative optimization of machine learning models.

[0152] In one embodiment of the medical device order processing method of this application, see [link to embodiment]. Figure 7 It can also specifically include the following:

[0153] Step S701: Continuously monitor the real-time workload of the customer service representatives who can accept orders;

[0154] Step S702: When the real-time workload exceeds the threshold, the unprocessed orders will be automatically transferred to other customer service representatives with lower real-time workloads.

[0155] Optionally, in step S701, the system needs to implement a mechanism to continuously monitor the real-time workload of each customer service representative who can accept orders. Workload refers to the combined effect of the number and complexity of orders currently being processed by the customer service representative. This metric is crucial for ensuring that customer service representatives are not overloaded and thus affect service quality. Real-time monitoring can be achieved by integrating a real-time data tracking system, which can collect and analyze information such as the current activity status of customer service representatives, order processing speed, and order waiting queues.

[0156] Step S702 involves an automated workload balancing mechanism. When the system detects that a customer service representative's real-time workload exceeds a preset threshold, this usually means that the number or complexity of orders being processed by that representative has reached or exceeded their optimal capacity. To prevent a decline in service quality and excessive customer service fatigue, the system will automatically transfer some unprocessed orders to other customer service representatives with lower workloads.

[0157] The implementation of this automatic transfer mechanism requires consideration of the following key points:

[0158] 1. Threshold Setting: Setting a reasonable threshold for workload saturation is key to ensuring customer service staff are not overloaded while maintaining high efficiency. Threshold settings can be based on historical data analysis, customer service feedback, and actual work efficiency testing.

[0159] 2. Order Priority and Complexity Assessment: When automatically transferring orders, the system needs to consider the urgency and processing complexity of the orders. Prioritizing orders with longer processing times or requiring specific skills can better balance the workload of each customer service representative.

[0160] 3. Dynamic Adjustment: Workload thresholds and order transfer rules should not be static, but need to be dynamically adjusted according to actual conditions. For example, during peak periods, thresholds may need to be adjusted to accommodate sudden increases in order volume and changes in customer service processing capacity.

[0161] 4. System Feedback and Optimization: The system should be able to collect feedback after the operation, such as the processing efficiency of customer service and customer satisfaction after the transfer, in order to further optimize threshold settings and workload balancing strategies.

[0162] This system design ensures that the customer service team can maintain efficient and high-quality service even when faced with a large volume of orders. Automatic workload balancing not only improves the responsiveness and flexibility of the entire customer service system but also helps increase customer service staff satisfaction, preventing fatigue and decreased efficiency caused by overwork. Implementing this strategy is a crucial step in achieving intelligent customer service management, significantly improving a company's customer service levels and internal operational efficiency.

[0163] To effectively improve the efficiency and accuracy of order processing in the medical device field, this application provides an embodiment of a medical device order processing apparatus for implementing all or part of the aforementioned medical device order processing method. See [link to embodiment]. Figure 8 The medical device order processing device specifically includes the following components:

[0164] The order anomaly determination module 10 is used to determine the data distribution of normal orders based on historical order data in the medical device order management system, determine abnormal orders in the current order data of the medical device order management system that deviate from the data distribution of normal orders based on the set anomaly detection model, and set the abnormal orders as orders that can be snatched.

[0165] The order change prediction module 20 is used to input the business parameters, service standards and time requirements in the historical order data into a preset natural language processing model to determine the corresponding order features, and use the order features as a model training set to train the preset time series analysis model to obtain the order change prediction model.

[0166] The customer service staff workload prediction module 30 is used to determine the corresponding customer service staff characteristics based on the professional qualifications, processing efficiency and processing success rate of the customer service staff in the medical device order management system by inputting them into a preset natural language processing model. The customer service staff characteristics are used as the model training set to train the preset recurrent neural network model to obtain the customer service staff workload prediction model.

[0167] The order-grabbing customer service matching module 40 is used to input the order features extracted from the order change prediction model and the customer service personnel features extracted from the customer service personnel workload prediction model into the feature fusion layer of a preset integrated model for feature fusion, input the result of the feature fusion into the input layer of the integrated model, and output the order-grabbing customer service personnel through the output layer of the integrated model. The order features are the order change patterns over time extracted from the order change prediction model, and the customer service personnel features are the output of the hidden layer of the recurrent neural network and the learned temporal relationship extracted from the customer service personnel workload prediction model.

[0168] The intelligent order transfer module 50 is used to assign the available orders to the pending work page of the customer service representative who can grab the orders, and after receiving the list request sent by the customer service representative who can grab the orders, change the processing permission of the available orders to that customer service representative.

[0169] As described above, the medical device order processing device provided in this application embodiment can determine the data distribution of normal orders through historical order data in the medical device order management system, identify abnormal orders in the current order data of the medical device order management system that deviate from the data distribution of normal orders according to a set anomaly detection model, and set the abnormal orders as available orders; output available order-grabbing customer service through the output layer of the integrated model; assign the available orders to the pending work page of the available order-grabbing customer service, and change the processing permission of the available orders to the available order-grabbing customer service after receiving the list request sent by the available order-grabbing customer service, thereby effectively improving the efficiency and accuracy of order processing in the medical device field.

[0170] From a hardware perspective, in order to effectively improve the efficiency and accuracy of order processing in the medical device field, this application provides an embodiment of an electronic device for implementing all or part of the aforementioned medical device order processing method. The electronic device specifically includes the following components:

[0171] The system comprises a processor, memory, a communications interface, and a bus; wherein the processor, memory, and communications interface communicate with each other via the bus; the communications interface is used to realize information transmission between the medical device order processing device and core business systems, user terminals, and related databases and other related devices; the logic controller can be a desktop computer, tablet computer, or mobile terminal, etc., and this embodiment is not limited to these. In this embodiment, the logic controller can be implemented with reference to the embodiments of the medical device order processing method and the medical device order processing device in the embodiments, the content of which is incorporated herein, and repeated details will not be described again.

[0172] It is understood that the user terminal may include smartphones, tablet computers, network set-top boxes, portable computers, desktop computers, personal digital assistants (PDAs), in-vehicle devices, smart wearable devices, etc. Among these, the smart wearable devices may include smart glasses, smartwatches, smart bracelets, etc.

[0173] In practical applications, the medical device order processing method can be partially executed on the electronic device side as described above, or all operations can be completed on the client device. The choice can be made based on the processing capabilities of the client device and the limitations of the user's usage scenario. This application does not impose any limitations on this. If all operations are completed on the client device, the client device may further include a processor.

[0174] The aforementioned client device may have a communication module (i.e., a communication unit) that can communicate with a remote server to achieve data transmission. The server may include a server on the task scheduling center side; in other implementation scenarios, it may also include a server on an intermediate platform, such as a server on a third-party server platform that has a communication link with the task scheduling center server. The server may include a single computer device, a server cluster consisting of multiple servers, or a distributed server structure.

[0175] Figure 9 This is a schematic block diagram illustrating the system configuration of the electronic device 9600 according to an embodiment of this application. Figure 9 As shown, the electronic device 9600 may include a central processing unit 9100 and a memory 9140; the memory 9140 is coupled to the central processing unit 9100. It is worth noting that... Figure 9 This is an example; other types of structures can also be used to supplement or replace this structure to achieve telecommunications functions or other functions.

[0176] In one embodiment, the medical device order processing method function can be integrated into the central processing unit 9100.

[0177] The central processing unit 9100 can be configured to perform the following controls:

[0178] Step S101: Determine the data distribution of normal orders based on historical order data in the medical device order management system, identify abnormal orders in the current order data of the medical device order management system that deviate from the data distribution of normal orders based on the set anomaly detection model, and set the abnormal orders as orders that can be snatched.

[0179] Step S102: Based on the business parameters, service standards and time requirements in the historical order data, input the data into a preset natural language processing model to determine the corresponding order features. Use the order features as the model training set to train the preset time series analysis model to obtain the order change prediction model.

[0180] Step S103: Based on the professional qualifications, processing efficiency, and processing success rate of customer service personnel in the medical device order management system, input the corresponding customer service personnel characteristics into a preset natural language processing model, use the customer service personnel characteristics as a model training set to train the preset recurrent neural network model, and obtain a customer service personnel workload prediction model.

[0181] Step S104: Input the order features extracted from the order change prediction model and the customer service personnel features extracted from the customer service personnel workload prediction model into the feature fusion layer of the preset integrated model for feature fusion. Input the result of the feature fusion into the input layer of the integrated model and output the customer service personnel who can grab orders through the output layer of the integrated model. The order features are the order change patterns over time extracted from the order change prediction model, and the customer service personnel features are the output of the hidden layer of the recurrent neural network and the learned temporal relationship extracted from the customer service personnel workload prediction model.

[0182] Step S105: Assign the available orders to the pending work page of the customer service representative who can accept orders, and after receiving the list request sent by the customer service representative who can accept orders, change the processing permission of the available orders to that customer service representative.

[0183] As described above, the electronic device provided in this application embodiment determines the data distribution of normal orders through historical order data in the medical device order management system, identifies abnormal orders in the current order data of the medical device order management system that deviate from the data distribution of normal orders according to a set anomaly detection model, and sets the abnormal orders as available orders; outputs available order-grabbing customer service through the output layer of the integrated model; assigns the available orders to the pending work page of the available order-grabbing customer service, and changes the processing permission of the available orders to the available order-grabbing customer service after receiving the list request sent by the available order-grabbing customer service, thereby effectively improving the efficiency and accuracy of order processing in the medical device field.

[0184] In another embodiment, the medical device order processing device can be configured separately from the central processing unit 9100. For example, the medical device order processing device can be configured as a chip connected to the central processing unit 9100, and the medical device order processing method function can be implemented through the control of the central processing unit.

[0185] like Figure 9As shown, the electronic device 9600 may further include: a communication module 9110, an input unit 9120, an audio processor 9130, a display 9160, and a power supply 9170. It is worth noting that the electronic device 9600 does not necessarily need to include these components. Figure 9 All components shown; in addition, the electronic device 9600 may also include Figure 9 For components not shown, please refer to existing technologies.

[0186] like Figure 9 As shown, the central processing unit 9100, sometimes also referred to as a controller or operating control, may include a microprocessor or other processor device and / or logic device, which receives inputs and controls the operation of various components of the electronic device 9600.

[0187] The memory 9140 may be, for example, one or more of a cache, flash memory, hard drive, removable media, volatile memory, non-volatile memory, or other suitable devices. It may store the aforementioned failure-related information, and also store a program for executing that information. The central processing unit 9100 may execute the program stored in the memory 9140 to perform information storage or processing, etc.

[0188] Input unit 9120 provides input to central processing unit 9100. Input unit 9120 may be, for example, a keypad or touch input device. Power supply 9170 provides power to electronic device 9600. Display 9160 displays images and text. Display may be, for example, an LCD display, but is not limited thereto.

[0189] The memory 9140 can be a solid-state memory, such as a read-only memory (ROM), random access memory (RAM), a SIM card, etc. It can also be a memory that retains information even when power is off, can be selectively erased, and contains more data; examples of this type of memory are sometimes referred to as EPROMs. The memory 9140 can also be some other type of device. The memory 9140 includes a buffer memory 9141 (sometimes referred to as a buffer). The memory 9140 may include an application / function storage unit 9142 for storing application programs and function programs or processes for executing the operation of the electronic device 9600 via the central processing unit 9100.

[0190] The memory 9140 may also include a data storage unit 9143 for storing data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit 9144 of the memory 9140 may include various drivers for the electronic device's communication functions and / or for performing other functions of the electronic device (such as messaging applications, address book applications, etc.).

[0191] The communication module 9110 is a transmitter / receiver that sends and receives signals via the antenna 9111. The communication module 9110 (transmitter / receiver) is coupled to the central processing unit 9100 to provide input signals and receive output signals, which is the same as in a conventional mobile communication terminal.

[0192] Based on different communication technologies, multiple communication modules 9110 can be configured in the same electronic device, such as cellular network modules, Bluetooth modules, and / or wireless LAN modules. The communication module 9110 (transmitter / receiver) is also coupled to a speaker 9131 and a microphone 9132 via an audio processor 9130 to provide audio output via the speaker 9131 and receive audio input from the microphone 9132, thereby realizing typical telecommunications functions. The audio processor 9130 may include any suitable buffer, decoder, amplifier, etc. Additionally, the audio processor 9130 is coupled to a central processing unit 9100, enabling on-device recording via the microphone 9132 and on-device playback of stored audio via the speaker 9131.

[0193] Specific embodiments have been used to illustrate the principles and implementation methods of this invention. The descriptions of the embodiments above are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A method for processing medical device orders, characterized in that, Applied to a medical device order management system, the method includes: The distribution of normal orders is determined based on historical order data in the medical device order management system. Abnormal orders that deviate from the distribution of normal orders in the current order data of the medical device order management system are identified based on the set anomaly detection model. These abnormal orders are then set as available orders. Based on the business parameters, service standards and time requirements in the historical order data, input the preset natural language processing model to determine the corresponding order features, and use the order features as the model training set to train the preset time series analysis model to obtain the order change prediction model. The professional qualifications, processing efficiency, and processing success rate of customer service personnel in the medical device order management system are input into a preset natural language processing model to determine the corresponding customer service personnel characteristics. The customer service personnel characteristics are used as the model training set to train a preset recurrent neural network model to obtain a customer service personnel workload prediction model. The order features extracted from the order change prediction model and the customer service personnel features extracted from the customer service personnel workload prediction model are input into the feature fusion layer of a preset integrated model for feature fusion. The result of the feature fusion is input into the input layer of the integrated model, and the customer service personnel who can grab orders are output through the output layer of the integrated model. The order features are the order change patterns over time extracted from the order change prediction model, and the customer service personnel features are the output of the hidden layer of the recurrent neural network and the learned temporal relationship extracted from the customer service personnel workload prediction model. The available orders are assigned to the pending work page of the customer service representative who is eligible to accept them, and upon receiving a list request from the customer service representative who is eligible to accept them, the processing authority for the available orders is changed to that customer service representative.

2. The medical device order processing method according to claim 1, characterized in that, The process of determining the data distribution of normal orders based on historical order data in the medical device order management system, identifying abnormal orders in the current order data of the medical device order management system that deviate from the data distribution of normal orders based on a set anomaly detection model, and setting the abnormal orders as available for purchase includes: Statistical analysis was performed on historical order data in the medical device order management system to determine the data distribution of normal orders that meet the normal judgment rules for order amount, order time, and order type. Calculate the Mahalanobis distance between the current order data and the data distribution of normal orders in the medical device order management system. Determine the current order whose calculated Mahalanobis distance value exceeds the threshold as an abnormal order and set the abnormal order as an order that can be bid on in the medical device order management system.

3. The medical device order processing method according to claim 1, characterized in that, The process involves inputting business parameters, service standards, and time requirements from the historical order data into a preset natural language processing model to determine the corresponding order features, and then using these order features as a model training set to train a preset time series analysis model to obtain an order change prediction model. This includes: The business parameters, service standards, and time requirements in the historical order data are converted into structured order features using a preset natural language processing model, and order feature time series data is constructed based on the order features. The order feature time series data is input into a preset time series analysis model for model training. The model parameters and hyperparameters are iteratively adjusted according to the model training results until they meet the expected output of the model, thus obtaining the order change prediction model.

4. The medical device order processing method according to claim 1, characterized in that, The process involves inputting the professional qualifications, processing efficiency, and success rate of customer service personnel in the medical device order management system into a preset natural language processing model to determine corresponding customer service personnel characteristics. These characteristics are then used as a model training set to train a preset recurrent neural network model, resulting in a customer service personnel workload prediction model. This process includes: The professional qualifications, processing efficiency, and processing success rate of customer service personnel in the medical device order management system are converted into structured customer service personnel features through a preset natural language processing model, and a model training set is constructed based on the customer service personnel. The training set of the model is input into a preset recurrent neural network model, and the recurrent neural network model is iteratively trained by setting a mean squared error loss function to obtain a customer service staff workload prediction model.

5. The medical device order processing method according to claim 1, characterized in that, The step of inputting the order features extracted from the order change prediction model and the customer service personnel features extracted from the customer service personnel workload prediction model into the feature fusion layer of the preset integrated model for feature fusion includes: The order change prediction model extracts the pattern of order changes over time and uses it as order features; the customer service staff workload prediction model extracts the output of the hidden layer of the recurrent neural network and the learned temporal relationship as customer service staff work features. The order features and the customer service personnel work features are input into the feature fusion layer of the reinforcement learning model for feature fusion operation.

6. The medical device order processing method according to claim 5, characterized in that, The output layer of the integrated model that provides customer service representatives capable of accepting orders includes: The order characteristics and customer service personnel work characteristics are set as the state space of the integrated model, multiple preset order grabbing strategies are set as the action space of the integrated model, and the success of order grabbing post-processing and the failure of order grabbing post-processing are set as the reward function of the integrated model, thus constructing a reinforcement learning model. The reinforcement learning model outputs a customer service representative who can grab orders through its output layer, wherein the reinforcement learning model iteratively optimizes the order-grabbing strategy in the action space through the reward function.

7. The medical device order processing method according to claim 1, characterized in that, After changing the processing authority for the order-grabbing service to the order-grabbing customer service representative, the following is included: Continuously monitor the real-time workload of the customer service representatives who are eligible to accept orders; When the real-time workload exceeds a threshold, any unprocessed orders will be automatically transferred to other customer service representatives with lower real-time workloads.

8. A medical device order processing device, characterized in that, The device includes: The order anomaly determination module is used to determine the data distribution of normal orders based on historical order data in the medical device order management system, identify abnormal orders in the current order data of the medical device order management system that deviate from the data distribution of normal orders based on a set anomaly detection model, and set the abnormal orders as available orders. The order change prediction module is used to input business parameters, service standards and time requirements in the historical order data into a preset natural language processing model to determine the corresponding order features, and use the order features as a model training set to train a preset time series analysis model to obtain the order change prediction model. The customer service staff workload prediction module is used to determine the corresponding customer service staff characteristics based on the professional qualifications, processing efficiency and processing success rate of customer service staff in the medical device order management system, and to train the preset recurrent neural network model using the customer service staff characteristics as the model training set to obtain the customer service staff workload prediction model. The order-grabbing customer service matching module is used to input the order features extracted from the order change prediction model and the customer service personnel features extracted from the customer service personnel workload prediction model into the feature fusion layer of a preset integrated model for feature fusion. The result of the feature fusion is input into the input layer of the integrated model, and the order-grabbing customer service personnel are output through the output layer of the integrated model. The order features are the order change patterns over time extracted from the order change prediction model, and the customer service personnel features are the output of the hidden layer of the recurrent neural network and the learned temporal relationship extracted from the customer service personnel workload prediction model. The intelligent order transfer module is used to assign the available orders to the pending work page of the customer service representative who can grab the orders, and after receiving the list request sent by the customer service representative who can grab the orders, change the processing permission of the available orders to that customer service representative.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the medical device order processing method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the medical device order processing method according to any one of claims 1 to 7.

Images