A method for pushing information of home medical equipment based on user positioning

Abstract

The present invention discloses a method for pushing home medical device information based on user positioning, which relates to the technical field of home medical device information push, and includes the following steps: classifying a number of push messages generated by a home medical device information push system to establish an analysis set; for the current push message, establishing a comparison set with the previous push message; after establishing the comparison set, performing feature extraction on the information in each comparison set. The present invention predicts the expiration hidden dangers of the push information through a pre-trained convolutional neural network, screens out high-expiration hidden danger information, ensures that users obtain accurate and real-time medical device information in emergency situations, and reduces the risk of delays. For low-expiration hidden danger information, its reliability is comprehensively evaluated based on historical comparison analysis, and it is pushed in sequence to ensure that users get the most reliable information first, thereby improving the accuracy of information push, user trust, and overall experience.

Description

Technical Field

[0001] The present invention relates to the technical field of home medical device information push technology, and in particular to a home medical device information push method based on user positioning. Background Art

[0002] Home medical device information push based on user location is an intelligent medical service model that uses the user's geographic location and health data to accurately push relevant medical device information. By obtaining the location data of the user's area, combined with the user's health needs, medical history and other information, the most suitable home medical equipment for the user, such as blood pressure monitors, blood glucose meters, etc., is analyzed, and equipment purchase or rental information from nearby hospitals, pharmacies or equipment suppliers is pushed based on the user's current location. This method can improve the accuracy and timeliness of information, help users more conveniently obtain the required medical resources, and improve the convenience and efficiency of health management.

[0003] The prior art has the following deficiencies:

[0004] In the process of pushing medical equipment information, existing technologies usually rely on pre-stored data such as equipment supply, inventory and location. If this data is not updated in a timely manner, the system may push expired or invalid information to users. When users need to quickly obtain medical equipment in an emergency, the risk of pushing expired information is particularly serious, which may cause users to receive outdated supplier information, out-of-stock equipment or discontinued services, ultimately delaying the acquisition of medical equipment and endangering the safety of patients.

[0005] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0006] The purpose of the present invention is to provide a method for pushing information of home medical devices based on user positioning, which predicts the expiration risk of the pushed information through a pre-trained convolutional neural network, screens out high expiration risk information, ensures that users obtain accurate and real-time medical device information in emergency situations, reduces delays and potential risks, and significantly improves the timeliness and accuracy of the push. At the same time, by performing historical comparison and analysis on low expiration risk information, comprehensively evaluating its reliability and pushing it in sequence, it ensures that users get the most reliable information first, enhances the accuracy of the push and user trust, optimizes the overall user experience, and improves the stability and reliability of the push system to solve the problems in the above-mentioned background technology.

[0007] In order to achieve the above object, the present invention provides the following technical solution: a method for pushing information of home medical devices based on user positioning, comprising the following steps:

[0008] Classify a number of push messages generated by the home medical device information push system and establish an analysis set;

[0009] For the current push information, create a comparison set with the previous push information;

[0010] After the comparison set is established, feature extraction is performed on the information in each comparison set, and the extracted features are analyzed and input into the pre-trained convolutional neural network to make a preliminary prediction of the expiration risk of the pushed information;

[0011] Based on the prediction results of the convolutional neural network, the current push information is classified into high expiration risk and low expiration risk; the push information with high expiration risk is initially screened out to prevent it from being pushed to users;

[0012] For the push information of low expiration hidden dangers, the prediction results of several historical comparison sets are further called, and the information obtained from multiple analyses is comprehensively evaluated. Based on the prediction results of the historical comparison sets, the reliability of each push information of low expiration hidden dangers is comprehensively evaluated;

[0013] Sort all push information with low risk of expiration by reliability, and push them sequentially in the sorted order to ensure that the most reliable information is pushed to users first.

[0014] Preferably, the previous push information refers to the home medical device information that was most recently generated and pushed to the user by the system before the current push information was generated.

[0015] Preferably, after establishing the comparison set, feature extraction is performed on the information in each comparison set, wherein the extracted features include the frequency of supplier supply chain updates and the success rate of home medical devices from order placement to delivery. After analyzing the frequency of supplier supply chain updates and the success rate of home medical devices from order placement to delivery, a supply chain update quantitative reference value and an order completion quantitative reference value are generated respectively. The supply chain update quantitative reference value is used to measure the activity level of the supplier's supply chain, and the order completion quantitative reference value is used to measure the performance of home medical devices from order placement to successful delivery.

[0016] Preferably, after obtaining the supply chain update quantitative reference value and the order completion quantitative reference value generated by comparing the two push information in each comparison set, the supply chain update quantitative reference value and the order completion quantitative reference value are input into a pre-trained convolutional neural network to generate an expiration evaluation index, and the initial expiration risk of the push information is predicted by the expiration evaluation index.

[0017] Preferably, the expiration evaluation index generated after predicting the current push information through the convolutional neural network is compared and analyzed with the pre-set expiration evaluation index reference threshold, and the current push information is divided. The specific division process is: if the expiration evaluation index is greater than or equal to the expiration evaluation index reference threshold, the current push information is divided into high expiration risk, and the push information with high expiration risk is preliminarily screened out; if the expiration evaluation index is less than the expiration evaluation index reference threshold, the current push information is divided into low expiration risk, and further analysis is performed later.

[0018] Preferably, for the push information with low expiration hidden danger, the prediction results of several historical comparison sets are further called for comprehensive analysis to generate a reliability evaluation coefficient, and the reliability of each push information with low expiration hidden danger is comprehensively evaluated through the reliability evaluation coefficient;

[0019] The steps to obtain the reliability evaluation coefficient are as follows:

[0020] For each push message with low expiration risk, collect its expiration evaluation index in multiple historical comparison sets to establish a data set, and record the data set as E. Then , where m is the number of historical alignment sets, is the expiration evaluation index in the jth historical comparison set, reflecting the expiration risk of the pushed information in the past;

[0021] The standard deviation of the overdue evaluation index is calculated, and the coefficient of variation is introduced to measure the stability of the overdue evaluation index. The expression of the coefficient of variation is: ,in is the geometric mean of the overdue assessment index, , is the coefficient of variation, which reflects the degree of fluctuation of the overdue evaluation index. The smaller the coefficient of variation, the higher the stability;

[0022] The reliability evaluation coefficient of each low expiration risk push information is calculated by combining the expiration evaluation index, the expiration evaluation index reference threshold, and the coefficient of variation corresponding to the current push information. The calculation expression is: , where is the reliability assessment coefficient, and its value range is , The expiration evaluation index of the current push information. is the reference threshold of the overdue assessment index, , , are adjustment parameters, The value range is , The value range is , The value range is .

[0023] Preferably, the specific steps of generating a quantitative reference value for supply chain updates after analyzing the frequency of supplier supply chain updates are as follows:

[0024] Extract the supplier's update records from the comparison set, including the timestamp of each update, and set the timestamp to ,in, Indicates the timestamp of the supplier's i-th update, and records a total of N supplier update behaviors;

[0025] The supplier's stagnation time between two updates is weighted to calculate the fluctuation factor of the update frequency, which is used to measure the volatility of the supplier's update behavior. The calculation expression of the fluctuation factor is: , where is the fluctuation factor of the update frequency, which measures the fluctuation of the supplier's downtime between two updates. Indicates the i-th and Downtime of supplier updates, is the sensitivity factor used to adjust the dead time The impact on volatility factors, is the reference threshold of the dead time, representing the expected maximum reasonable update dead time;

[0026] The quantitative reference value of supply chain update is calculated by combining the volatility factor and the sum of stagnation time. The calculation expression is: , where A quantitative reference value for supply chain updates, used to measure the activity of supplier update behavior. In order to adjust the impact weight of volatility factors on the quantitative reference value of supply chain updates, To adjust the weight of the impact of the sum of time intervals on the supply chain update quantitative reference value.

[0027] Preferably, the specific steps of analyzing the success rate of home medical devices from order placement to delivery to generate a quantitative reference value for order completion are as follows:

[0028] Each order of home medical devices is divided into two categories: successful delivery and unsuccessful delivery. A marking function is defined to judge the uth order. Whether the delivery is successful, the expression of the marking function is: ;

[0029] Introducing fulfillment time distribution weights , used to measure the time difference from order placement to delivery, assuming order The fulfillment time is , and the ideal delivery time is , the expression defining the distribution weight of fulfillment time is: , where k is the steepness of the time distribution curve;

[0030] The order completion quantitative reference value is generated based on the order success mark and the fulfillment time distribution weight. The expression for generating the order completion quantitative reference value is: , where is the weight factor of each order, indicating the importance or priority of the order, It is a quantitative reference value for order completion, and n represents the total number of orders.

[0031] Preferably, all push information with low expiration risk is sorted by reliability, specifically:

[0032] A data set is established based on the reliability assessment coefficients generated by comprehensively evaluating the reliability of each low-expired hidden danger push information, and the reliability assessment coefficients in the data set are sorted in order. The sorted reliability assessment coefficients are associated one-to-one with their corresponding low-expired hidden danger push information, and the low-expired hidden danger push information is then pushed sequentially in the sorted push order to ensure that the most reliable information is pushed to users first.

[0033] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0034] The present invention uses a pre-trained convolutional neural network to make preliminary predictions on the potential expiration risks of pushed information, effectively classify pushed information with high expiration risks, and screen it out in a timely manner to avoid delivering invalid or expired information to users. This method significantly reduces the interference of expired information on user decision-making, ensuring that users can obtain accurate and real-time medical equipment information in emergency situations, thereby reducing delays and possible risks, and greatly improving the timeliness and accuracy of information push.

[0035] By analyzing multiple historical comparison sets of information pushed with low expiration risks, the present invention realizes a comprehensive reliability evaluation of each piece of information, and sorts and pushes it according to reliability, ensuring that users get the most reliable equipment information first. This not only enhances the accuracy of information push, but also improves the user's trust in the medical equipment procurement or leasing decision-making process, optimizes the overall user experience, and ensures that information push is more robust and reliable. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0037] Figure 1 The present invention is a method flow chart of a method for pushing home medical device information based on user positioning. DETAILED DESCRIPTION

[0038] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0039] The present invention provides Figure 1 A method for pushing home medical device information based on user positioning is shown, comprising the following steps:

[0040] Classify a number of push messages generated by the home medical device information push system and establish an analysis set;

[0041] This set contains information in multiple dimensions, such as the supplier information, inventory status, location information, and update time of the equipment. By establishing an analysis set, each push message can be managed and tracked as a whole, ensuring that all information is fully captured and recorded by the system. In the existing technology, the push method usually relies on static databases. The introduction of analysis sets makes the data more dynamic and can be updated in real time through data streams, laying the foundation for subsequent analysis.

[0042] For the current push information, create a comparison set with the previous push information;

[0043] The previous push information refers to the home medical device information that was most recently generated and pushed to the user by the system before the current push information was generated. Specifically, the previous push information usually contains similar content to the current push information, such as equipment supplier, inventory, location information, etc., but may have been pushed at an earlier point in time. By comparing the current push information with the previous push information, the system can identify changes between the information, such as whether the equipment inventory has decreased, whether the supplier has changed, etc., thereby evaluating the validity and accuracy of the information. This comparison can help determine whether the push information has expired or is still valid.

[0044] The purpose of this operation is to compare the current push information with the previous push information horizontally to find the differences between the two push information, such as inventory changes, price adjustments, or whether there is a new version of the device. This comparison method is convenient for effectively identifying potential outdated risks.

[0045] After the comparison set is established, feature extraction is performed on the information in each comparison set. The extracted features are analyzed and input into a pre-trained convolutional neural network (CNN) to make a preliminary prediction of the expiration risk of the pushed information.

[0046] After establishing the comparison set, feature extraction is performed on the information in each comparison set, wherein the extracted features include the frequency of supplier supply chain updates and the success rate of home medical devices from order placement to delivery. After analyzing the frequency of supplier supply chain updates and the success rate of home medical devices from order placement to delivery, a supply chain update quantitative reference value and an order completion quantitative reference value are generated respectively. The supply chain update quantitative reference value is used to measure the activity of the supplier's supply chain. By analyzing the update frequency of supplier information, it is reflected whether the supplier continues to provide medical devices and maintains the effectiveness of inventory. The order completion quantitative reference value is used to measure the performance of home medical devices from order placement to successful delivery, reflecting the stability of the supply chain of home medical devices and the probability that users can successfully obtain the devices.

[0047] After obtaining the supply chain update quantitative reference value and order completion quantitative reference value generated by comparing the two push information in each comparison set, the supply chain update quantitative reference value and the order completion quantitative reference value are input into the pre-trained convolutional neural network to generate an expiration assessment index, and the initial expiration risk of the push information is predicted through the expiration assessment index.

[0048] A pre-trained convolutional neural network (CNN) is a deep learning model that has been trained with a large amount of historical data. The model can automatically extract complex features from the data and make predictions and classifications based on these features. In this specific home medical device information push system, CNN uses the processed supply chain update quantitative reference value and order completion quantitative reference value as input data to determine the validity or expiration risk of device information. By learning from a large amount of historical push information data, the convolutional neural network has the ability to identify which supply chain features or order fulfillment information may cause the push information to expire, and can effectively associate the features in different push information to make accurate predictions.

[0049] The convolutional layer in the convolutional neural network processes the multi-dimensional features of the quantitative reference values ​​of supply chain updates and order completion, and extracts the correlation between information through different convolution kernels, such as whether the frequent data updates by suppliers are related to the successful delivery of equipment. During the training process, CNN continuously adjusts its internal weights to optimize its ability to identify expired information. After multiple iterations, the model will gradually improve its prediction accuracy of the hidden dangers of pushed information. When applied in practice, the pre-trained CNN can quickly process the input data and generate an expiration assessment index, which helps the system make preliminary predictions on the hidden dangers of pushed information and significantly improves the reliability and effectiveness of pushed information.

[0050] If the supplier supply chain updates less frequently between two push messages, it usually indicates that the current push message has a potential risk of expiration. This is because the supplier supply chain update frequency reflects the supplier's timely update of the device inventory and supply information. If the supply chain update frequency is low, it means that the supplier has not updated the device inventory or supply information for a long time, which may indicate that the supply of the device is no longer stable or has even stopped supplying, thereby increasing the risk of push information expiration.

[0051] The specific steps to generate a quantitative reference value for supply chain updates after analyzing the frequency of supplier supply chain updates are as follows:

[0052] Extract the supplier's update records from the comparison set (the current push information and the previous push information), including the timestamp of each update, and set the timestamp to ,in, Indicates the timestamp of the supplier's i-th update, and records a total of N supplier update behaviors;

[0053] The supplier's stagnation time between two updates is weighted to calculate the fluctuation factor of the update frequency, which is used to measure the volatility of the supplier's update behavior. The calculation expression of the fluctuation factor is: , where is the fluctuation factor of the update frequency, which measures the fluctuation of the supplier's downtime between two updates. Indicates the i-th and The downtime of supplier update is larger It may indicate low supply chain activity and expiration risk. It is the sensitivity coefficient. The specific value is not limited here. It can be determined according to the actual situation and needs and is used to adjust the dead time. The impact on volatility factors, is the reference threshold of the dead time, representing the expected maximum reasonable update dead time;

[0054] This step exploits a nonlinear function (the Sigmoid function), which aggravates its impact when the interval is large, thus highlighting the unevenness of the supply chain update frequency.

[0055] The quantitative reference value of supply chain update is calculated by combining the volatility factor and the sum of stagnation time. The calculation expression is: , where A quantitative reference value for supply chain updates, used to measure the activity of supplier update behavior. In order to adjust the impact weight of volatility factors on the quantitative reference value of supply chain updates, In order to adjust the influence weight of the sum of time intervals on the quantitative reference value of supply chain update, the setting of influence weight is not specifically limited here and can be adjusted according to actual conditions and needs;

[0056] This step generates a quantitative reference value for supply chain updates by balancing the fluctuations in supplier update frequency with the total stagnation time. A lower quantitative reference value for supply chain updates indicates that supply chain updates are less active, and suppliers may have stopped updating, which may lead to expiration risks. The nonlinear function used here ensures that the quantitative reference value for supply chain updates has higher sensitivity when fluctuation factors and time intervals change.

[0057] It can be seen from the quantitative reference value of supply chain updates that between two pieces of push information, the larger the performance value of the quantitative reference value of supply chain updates generated after analyzing the frequency of supplier supply chain updates, the lower the frequency of supplier supply chain updates and the inactive supply chain, which infers that the current push information may have a greater risk of expiration. Conversely, the smaller the performance value of the quantitative reference value of supply chain updates, the higher the frequency of supplier supply chain updates and the more active the supply chain, indicating that the supply information of home medical devices is more timely and the current push information has a smaller risk of expiration. Therefore, the quantitative reference value of supply chain updates can effectively measure the timeliness of push information and help identify whether there is a risk of expiration.

[0058] If the success rate of home medical device orders from order to delivery between two push messages is low, it usually indicates that the current push information is potentially out of date. A low success rate may reflect problems such as unstable supply chain, tight equipment inventory, or suppliers unable to fulfill their contracts on time. These situations may mean that the supply of the equipment has changed, but the pushed information has not been updated in time, increasing the risk of the push information being out of date or invalid. Therefore, a low success rate is a key indicator that the push information may no longer be accurate or valid.

[0059] The specific steps to generate a quantitative reference value for order completion after analyzing the success rate of home medical devices from order placement to delivery are as follows:

[0060] Each order of home medical devices is divided into two categories: successful delivery and unsuccessful delivery. A marking function is defined to judge the uth order. Whether the delivery is successful, the expression of the marking function is: ;

[0061] On this basis, the system marks the delivery status of each order as 1 (successful) or 0 (failed). This marking function provides basic data for subsequent calculations.

[0062] In order to reflect the timeliness of order delivery, a fulfillment time distribution weight is introduced , used to measure the time difference from order placement to delivery, assuming order The fulfillment time is , and the ideal delivery time is , the expression defining the distribution weight of fulfillment time is: , where k is the steepness of the time distribution curve;

[0063] This formula reflects the time when the order is delivered The closer When the performance time distribution weight The closer it is to 1, the better the performance. Far beyond , performance time distribution weight will drop rapidly, indicating that there is a risk of delay or failure in order fulfillment.

[0064] The order completion quantitative reference value is generated based on the order success mark and the fulfillment time distribution weight. The expression for generating the order completion quantitative reference value is: , where It is the weight factor of each order, indicating the importance or priority of the order. It can be set based on factors such as the order amount and the importance of the equipment. No specific restrictions are made here. It is the quantitative reference value for order completion, and n represents the total number of orders;

[0065] This step takes into account the success or failure of the order, the fulfillment of the time, and the importance of each order, thereby generating a quantitative reference value for order completion that reflects the overall fulfillment of home medical devices from order placement to successful delivery.

[0066] It can be seen from the order completion quantitative reference value that between the two push messages, the smaller the performance value of the order completion quantitative reference value generated after analyzing the success rate of home medical devices from order placement to delivery, the lower the success rate of home medical devices from order placement to delivery, which reflects that the supply chain may be unstable or there are problems with the delivery of equipment, which in turn indicates that the current push information has a greater risk of expiration. On the contrary, the larger the performance value of the order completion quantitative reference value, the better the order fulfillment, the stable supply chain of home medical devices, and the on-time delivery of home medical devices, which means that the current push information has a smaller risk of expiration. Therefore, the level of the order completion quantitative reference value can effectively predict the effectiveness and timeliness of push information.

[0067] The convolutional neural network is not specifically limited here, and can achieve the quantitative reference value of supply chain update and order completion quantitative reference value Conduct comprehensive analysis to generate an overdue valuation index In order to realize the technical solution of the present invention, the present invention provides a specific implementation method; the overdue evaluation index The resulting calculation formula is: , where , Update quantitative reference values ​​for the supply chain and order completion quantitative reference value The preset scaling factor of , Both are greater than 0.

[0068] It can be seen from the expiration evaluation index that between the two push messages, the larger the performance value of the supply chain update quantitative reference value generated after analyzing the frequency of supplier supply chain updates, the smaller the performance value of the order completion quantitative reference value generated after analyzing the success rate of home medical equipment from order placement to delivery. In other words, the larger the performance value of the expiration evaluation index generated after predicting the current push information through the convolutional neural network, the greater the risk of expiration of the current push information, and vice versa.

[0069] Based on the prediction results of the convolutional neural network, the current push information is classified into high expiration risk and low expiration risk; the push information with high expiration risk is initially screened out to prevent it from being pushed to users, thereby reducing the impact of expired or invalid information on users;

[0070] The expiration evaluation index generated by predicting the current push information through the convolutional neural network is compared and analyzed with the pre-set expiration evaluation index reference threshold, and the current push information is divided. The specific division process is: if the expiration evaluation index is greater than or equal to the expiration evaluation index reference threshold, the current push information is divided into high expiration risks, and the push information with high expiration risks is initially screened out. High expiration risks mean that the push information may no longer be accurate or timely, such as unstable equipment supply chain, outdated inventory data, and suppliers are no longer active. In this case, the pushed information may have expired, and the medical equipment obtained by the user may not be delivered on time or cannot be purchased. Therefore, the high expiration risk information needs to be further screened or automatically screened out to avoid pushing invalid information to the user and ensure that the information received by the user is the latest and reliable. If the expiration evaluation index is less than the expiration evaluation index reference threshold, the current push information is divided into low expiration risks. Low expiration risks mean that the pushed information is relatively reliable, the supply chain is stable, the inventory data is the latest, and the equipment can be delivered on time or is still in supply. The effectiveness of the equipment in the push information is high, and users can rely on this information to make decisions on purchasing or leasing medical equipment. Therefore, information with low risk of expiration can be pushed to users first to ensure the timeliness and accuracy of the information.

[0071] Existing systems often lack such a risk classification mechanism, so users may receive a large amount of outdated or useless equipment information. By introducing risk classification, the system can push information more intelligently and accurately.

[0072] For the push information of low expiration hidden dangers, the prediction results of several historical comparison sets are further called, and the information obtained from multiple analyses is comprehensively evaluated. Based on the prediction results of the historical comparison sets, the reliability of each push information of low expiration hidden dangers is comprehensively evaluated;

[0073] For the push information of low-expiration hidden dangers, the prediction results of several historical comparison sets are further called for comprehensive analysis to generate a reliability evaluation coefficient. The reliability of each push information of low-expiration hidden dangers is comprehensively evaluated through the reliability evaluation coefficient. The steps for obtaining the reliability evaluation coefficient are as follows:

[0074] For each push message with low expiration risk, collect its expiration evaluation index in multiple historical comparison sets to establish a data set, and record the data set as E. Then , where m is the number of historical alignment sets, is the expiration evaluation index in the jth historical comparison set, reflecting the expiration risk of the pushed information in the past;

[0075] The main function of this step is to extract the expiration evaluation index of the pushed information from multiple historical comparison sets. These historical data reflect the expiration risks of the pushed information at multiple time points in the past, providing a basis for changes in the time dimension for subsequent analysis. By collecting these historical data, the system can identify the long-term stability or risk trend of the pushed information.

[0076] The standard deviation of the overdue evaluation index is calculated, and the coefficient of variation is introduced to measure the stability of the overdue evaluation index. The expression of the coefficient of variation is: ,in is the geometric mean of the overdue assessment index, , is the coefficient of variation, which reflects the degree of fluctuation of the overdue evaluation index. The smaller the coefficient of variation, the higher the stability;

[0077] The purpose of this step is to quantify the historical fluctuations of the pushed information by calculating the stability coefficient of the historical expiration evaluation index. Specifically, the coefficient of variation is introduced to measure the fluctuation range of the expiration evaluation index. If the coefficient of variation is large, it means that the pushed information has shown great instability in history, which means that its supply chain or inventory information may change frequently, which will affect the reliability assessment of the pushed information. If the coefficient of variation is small, it means that the information has shown high stability in historical data and the pushed information is more reliable. Through this process, the system can provide a quantitative risk basis for subsequent reliability assessments.

[0078] The reliability evaluation coefficient of each low expiration risk push information is calculated by combining the expiration evaluation index, the expiration evaluation index reference threshold, and the coefficient of variation corresponding to the current push information. The calculation expression is: , where is the reliability assessment coefficient, and its value range is , the larger the value, the higher the reliability. The expiration evaluation index of the current push information. is the reference threshold of the overdue assessment index, , , are adjustment parameters, The value range is usually The specific value depends on the system's sensitivity setting to expiration risks. The influence of the overdue evaluation index in the control formula on the reliability evaluation coefficient. When the value increases, the influence of the expiration evaluation index on the reliability evaluation coefficient will be more significant, which means that even if the expiration evaluation index slightly exceeds the expiration evaluation index reference threshold, the reliability evaluation coefficient will drop rapidly. This makes the system more sensitive and can quickly filter out information with high expiration risks. On the contrary, a smaller A value of 0 will reduce the sensitivity, allowing the system to tolerate small expiration risks to a certain extent;

[0079] The value range is , usually used to adjust the expiration evaluation index corresponding to the current push information Nonlinear effects relative to the reference threshold T. The main function of is to adjust the degree of nonlinearity in the formula. For example, when When it is close to 1, the formula shows a linear growth, which means that the relationship between the expiration evaluation index and the reliability evaluation coefficient is linear. As the value increases, the formula becomes more nonlinear, which means that when the overdue evaluation index approaches or exceeds the overdue evaluation index reference threshold, the reliability evaluation coefficient will decrease more dramatically, thereby accelerating the screening of high-risk information;

[0080] The value range is , which is mainly used to adjust the influence of coefficient of variation on reliability assessment coefficient. The influence strength of stability (volatility) in the control formula. When the coefficient of variation is large, it means that the historical overdue evaluation index fluctuates greatly, which means that the stability of the information is low. The value will amplify the impact of this volatility on the reliability evaluation coefficient, making the system pay more attention to information stability and thus make a stricter evaluation of unstable push information. A low value weakens this effect, and the system becomes more tolerant to fluctuations in information.

[0081] The reliability of each push information is judged based on its reliability assessment coefficient. The closer the reliability assessment coefficient is to 1, the more reliable the information is. In this step, the reliability of each low-risk push information can be evaluated by comparing similar information obtained from historical push records. This multi-level comprehensive analysis enables the system to make full judgments even in low-risk situations, avoiding relying solely on single prediction results. The cumulative effect of historical data is used to evaluate factors such as the stability of equipment suppliers and the regularity of inventory fluctuations. In this way, the system can better judge which push information will remain valid and reliable in the future. Compared with existing technologies, this step ensures that the information is not only currently valid, but also has a certain future validity.

[0082] Sort all push information with low expiration risk by reliability, and push them in the sorted order to ensure that the most reliable information is pushed to users first;

[0083] Sort all push messages with low risk of expiration by reliability, specifically:

[0084] A data set is established based on the reliability assessment coefficients generated by comprehensively evaluating the reliability of each low-expired hidden danger push information, and the reliability assessment coefficients in the data set are sorted in order. The sorted reliability assessment coefficients are associated one-to-one with their corresponding low-expired hidden danger push information, and the low-expired hidden danger push information is then pushed sequentially in the sorted push order to ensure that the most reliable information is pushed to users first.

[0085] In an emergency, users need to get the latest and most effective device information as soon as possible, so the optimization of push order is crucial. In existing technologies, push is usually random or in chronological order, but by introducing a sorting mechanism, the system can ensure that users receive the most valuable information first based on reliability, thereby improving push efficiency and reducing the risk of delays.

[0086] The present invention uses a pre-trained convolutional neural network to make preliminary predictions on the potential expiration risks of pushed information, effectively classify pushed information with high expiration risks, and screen it out in a timely manner to avoid delivering invalid or expired information to users. This method significantly reduces the interference of expired information on user decision-making, ensuring that users can obtain accurate and real-time medical equipment information in emergency situations, thereby reducing delays and possible risks, and greatly improving the timeliness and accuracy of information push.

[0087] By analyzing multiple historical comparison sets of information pushed with low expiration risks, the present invention realizes a comprehensive reliability evaluation of each piece of information, and sorts and pushes it according to reliability, ensuring that users get the most reliable equipment information first. This not only enhances the accuracy of information push, but also improves the user's trust in the medical equipment procurement or leasing decision-making process, optimizes the overall user experience, and ensures that information push is more robust and reliable.

[0088] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0089] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0090] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A method for pushing home medical device information based on user positioning, characterized in that: The following steps are involved: Classify a number of push messages generated by the home medical device information push system and establish an analysis set; For the current push information, create a comparison set with the previous push information; After the comparison set is established, feature extraction is performed on the information in each comparison set, and the extracted features are analyzed and input into the pre-trained convolutional neural network to make a preliminary prediction of the expiration risk of the pushed information; Based on the prediction results of the convolutional neural network, the current push information is classified into high expiration risk and low expiration risk; the push information with high expiration risk is initially screened out to prevent it from being pushed to users; For the push information of low expiration hidden dangers, the prediction results of several historical comparison sets are further called, and the information obtained from multiple analyses is comprehensively evaluated. Based on the prediction results of the historical comparison sets, the reliability of each push information of low expiration hidden dangers is comprehensively evaluated; Sort all push information with low expiration risk by reliability, and push them in the sorted order to ensure that the most reliable information is pushed to users first; After establishing the comparison set, feature extraction is performed on the information in each comparison set, wherein the extracted features include the frequency of supplier supply chain updates and the success rate of home medical devices from order placement to delivery. After analyzing the frequency of supplier supply chain updates and the success rate of home medical devices from order placement to delivery, a supply chain update quantitative reference value and an order completion quantitative reference value are generated respectively. The activity of the supplier supply chain is measured by the supply chain update quantitative reference value, and the performance of home medical devices from order placement to successful delivery is measured by the order completion quantitative reference value. The specific steps to generate a quantitative reference value for supply chain updates after analyzing the frequency of supplier supply chain updates are as follows: Extract the supplier's update records from the comparison set, including the timestamp of each update, and set the timestamp to ,in, Indicates that the supplier timestamp of updates, total records N Sub-supplier update behavior; The supplier's stagnation time between two updates is weighted to calculate the fluctuation factor of the update frequency, which is used to measure the volatility of the supplier's update behavior. The calculation expression of the fluctuation factor is: In the formula, is the fluctuation factor of the update frequency, which measures the fluctuation of the supplier's downtime between two updates. Indicates and Downtime of supplier updates, is the sensitivity factor used to adjust the dead time The impact on volatility factors, is the reference threshold of the dead time, representing the expected maximum reasonable update dead time; The quantitative reference value of supply chain update is calculated by combining the volatility factor and the sum of stagnation time. The calculation expression is: In the formula, A quantitative reference value for supply chain updates, used to measure the activity of supplier update behavior. In order to adjust the impact weight of volatility factors on the quantitative reference value of supply chain updates, To adjust the weight of the impact of the sum of time intervals on the supply chain update quantitative reference value.

2. A method for pushing home medical device information based on user positioning according to claim 1, characterized in that: The previous push information refers to the home medical device information that was most recently generated and pushed to the user by the system before the current push information was generated.

3. A method for pushing home medical device information based on user positioning according to claim 1, characterized in that: After obtaining the supply chain update quantitative reference value and order completion quantitative reference value generated by comparing the two push information in each comparison set, the supply chain update quantitative reference value and the order completion quantitative reference value are input into the pre-trained convolutional neural network to generate an expiration assessment index, and the initial expiration risk of the push information is predicted through the expiration assessment index.

4. A method for pushing home medical device information based on user positioning according to claim 3, characterized in that: The expiration evaluation index generated after predicting the current push information through the convolutional neural network is compared and analyzed with the pre-set expiration evaluation index reference threshold, and the current push information is divided. The specific division process is: if the expiration evaluation index is greater than or equal to the expiration evaluation index reference threshold, the current push information is divided into high expiration risk, and the push information with high expiration risk is preliminarily screened out; if the expiration evaluation index is less than the expiration evaluation index reference threshold, the current push information is divided into low expiration risk, and further analysis is performed later.

5. A method for pushing home medical device information based on user positioning according to claim 4, characterized in that: For the push information of low-expiration hidden dangers, the prediction results of several historical comparison sets are further called for comprehensive analysis to generate a reliability evaluation coefficient, and the reliability of each push information of low-expiration hidden dangers is comprehensively evaluated through the reliability evaluation coefficient; The steps to obtain the reliability evaluation coefficient are as follows: For each push message with low expiration risk, collect its expiration evaluation index in multiple historical comparison sets to establish a data set, and record the data set as E ,but ,in m is the number of historical alignment sets, For the The expiration evaluation index in the historical comparison set reflects the expiration risk of the pushed information in the past; The standard deviation of the overdue evaluation index is calculated, and the coefficient of variation is introduced to measure the stability of the overdue evaluation index. The expression of the coefficient of variation is: in is the geometric mean of the overdue assessment index, , is the coefficient of variation, which reflects the degree of fluctuation of the overdue evaluation index. The smaller the coefficient of variation, the higher the stability; The reliability evaluation coefficient of each low expiration risk push information is calculated by combining the expiration evaluation index, the expiration evaluation index reference threshold, and the coefficient of variation corresponding to the current push information. The calculation expression is: In the formula, is the reliability assessment coefficient, and its value range is , The expiration evaluation index of the current push information. is the reference threshold of the overdue assessment index, are adjustment parameters, The value range is , The value range is , The value range is .

6. A method for pushing home medical device information based on user positioning according to claim 1, characterized in that: The specific steps to generate a quantitative reference value for order completion after analyzing the success rate of home medical devices from order placement to delivery are as follows: Each order of home medical equipment is divided into two categories: successful delivery and unsuccessful delivery. A marking function is defined to judge the u Orders Whether the delivery is successful, the expression of the marking function is: ; Introducing fulfillment time distribution weights , used to measure the time difference from order placement to delivery, assuming order The fulfillment time is , and the ideal delivery time is , the expression defining the distribution weight of fulfillment time is: , where k is the steepness of the time distribution curve; The order completion quantitative reference value is generated based on the order success mark and the fulfillment time distribution weight. The expression for generating the order completion quantitative reference value is: In the formula, is the weight factor of each order, indicating the importance or priority of the order, Quantitative reference value for order completion, n Indicates the total number of orders.

7. A method for pushing home medical device information based on user positioning according to claim 5, characterized in that: Sort all push messages with low risk of expiration by reliability, specifically: A data set is established based on the reliability assessment coefficients generated by comprehensively evaluating the reliability of each low-expired hidden danger push information, and the reliability assessment coefficients in the data set are sorted in order. The sorted reliability assessment coefficients are associated one-to-one with their corresponding low-expired hidden danger push information, and the low-expired hidden danger push information is then pushed sequentially in the sorted push order to ensure that the most reliable information is pushed to users first.

Images