Information pushing method and system based on smart medical big data

Abstract

The invention relates to an information pushing method and system based on smart medical big data, and the method comprises the steps: calculating a single-classification project enabling rate for each user classification and project classification combination, and carrying out the calculation of a plurality of single-classification project enabling rates through employing a weighted average method, and obtaining a comprehensive project enabling rate; setting a project active factor according to the value of the single-classification project enabling rate; setting user active factors according to the number of item types involved by a single user; generating an optimization scheme by combining the project active factor and the user active factor according to the single-classification project enabling rate; by calculating the starting rate of the single-classification item, the user and the smart medical item can be more accurately matched, the interest and the demand of the user are considered, the characteristics of the item and the target user group are combined, the characteristics and the preferences of different user groups are more met, and thus the pushing accuracy and the user experience are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of smart medical big data, and in particular to an information push method and system based on smart medical big data. Background Art

[0002] With the continuous development of computer science and information technology, big data has been widely involved in all walks of life. For example, in the smart medical industry, industry applications based on big data have also emerged like mushrooms after rain. In a typical smart medical industry application scenario, information push based on smart medical big data has been widely used. However, most common information push methods push corresponding information to the push object based on established information push rules, which leads to poor push effect and user experience.

[0003] The Chinese invention patent with application number CN202111349320.4 provides an information push method and a smart medical cloud server based on smart medical big data. According to the session feature data of the smart medical project without a preset session tag corresponding to the target push business scope and the target smart medical session, and the push business information statistics corresponding to the target push business scope, the network is optimized based on the information push strategy to obtain a push strategy optimization plan for the target smart medical session, and then information is pushed to the target smart medical session through the optimized information push strategy according to the push business information statistics.

[0004] In the existing patent technology, a conversation project feature data set without preset conversation tags is used to perform a conversation project activation frequency analysis, and a push strategy optimization solution for the target smart medical conversation is obtained based on the conversation project activation frequency analysis data. However, the conversation project activation frequency used is analyzed and optimized based on all projects and users, and the conversation project activation frequency is not calculated based on different classifications of users and projects, resulting in inaccurate and non-targeted calculation of the conversation project activation frequency, which in turn reduces the accuracy of information push. Summary of the invention

[0005] In view of the technical problems existing in the prior art, the present invention provides an information push method and system based on smart medical big data.

[0006] The technical solution of the present invention to solve the above technical problems is as follows: an information push method based on smart medical big data, the method comprising: S100: Obtain smart medical project information and corresponding session feature data, and simultaneously obtain push service information statistics, and classify users and projects according to the smart medical project information and corresponding session feature data; S200: for each user classification and project classification combination, calculating the single classification project activation rate, and calculating the activation rates of multiple single classification projects using a weighted average method to obtain a comprehensive project activation rate; S300: setting a project activity factor according to the value of the activation rate of the single-category project; setting a user activity factor according to the number of project types involved by a single user; S400: generating an optimization plan according to the activation rate of the single-category project, combined with the project activity factor and the user activity factor; optimizing the information push strategy according to the optimization plan, pushing information according to the optimized information push strategy, and monitoring and acquiring push feedback data in real time; S500: Evaluate the push optimization effect according to the comprehensive project activation rate, and generate a user trend change report for data analysis and real-time adjustment of the optimization strategy.

[0007] Furthermore, the optimization plan is to divide users into highly active, medium active and low active users according to the activation rate of the single-category project, the project activity factor and the user activity factor; and divide the projects into three priorities: high, medium and low; the optimization plan sets a basic push strategy and a cross-pre-test push strategy, and the basic push strategy adopts information push of different frequencies for different active users and different priority projects; the cross-pre-test push strategy is to push high-priority projects to low-active users, and push low-priority projects to high-active users, set a time limit to collect user feedback result data, analyze the feedback results collected by the cross-pre-test push strategy, determine the push frequency and push range in combination with the basic push strategy, and generate the final information push optimization strategy.

[0008] Furthermore, the optimization scheme also includes related project recommendation rules, which select complementary projects that are strongly related to the user's current most popular projects based on the project association matrix, push information based on the complementary projects, and guide the optimization of user access paths.

[0009] Furthermore, the project association matrix is ​​constructed by analyzing the association between smart medical projects using an association rule learning algorithm, and based on the calculated association, the project association matrix is ​​constructed; the strongly associated complementary projects refer to projects whose association is greater than an association threshold.

[0010] Furthermore, the user access path optimization refers to identifying key nodes in the user access path and inserting complementary items at the key nodes to achieve a natural transition; the key nodes refer to project pages that users visit and stay on, and complementary items are inserted during the time they stay on the project page.

[0011] Furthermore, the push feedback data is obtained, the user push service results are integrated, the matching value of each push item received by the user is calculated, the matching values ​​of all push items of each user are counted and the median is calculated, and the initial item matching value range is set with the median as the origin; the initial item matching value range is debugged according to the debugging rules to obtain the accurate item matching value range.

[0012] Furthermore, the debugging rule is to pre-test the lower limit of the initial project matching value range before and after, until the maximum number of pre-tests is reached, observe user feedback and push effects, and adjust the project matching value range based on the pre-test results; the maximum number of pre-tests is pre-set.

[0013] Furthermore, the precise project matching value range is obtained, and the project matching value range is dynamically adjusted in combination with a random exploration mechanism to form a jumping project matching value range, and information push optimization is performed based on the jumping project matching value range.

[0014] Furthermore, the random exploration mechanism randomly extracts a rated number of untouched items to fill into the precise item matching value range, sets a time interval, periodically replaces items, and removes invalid items in the precise item matching value range; the rated number is the number of untouched items to be added based on the number of items in the precise item matching value range and the random disturbance factor.

[0015] The information push system based on smart medical big data is equipped with the information push method based on smart medical big data as described above, and the system includes: a data acquisition module for acquiring smart medical project information, session feature data and push business information statistics; The classification calculation module is used to receive data information from the data acquisition module, classify users and projects, and calculate the activation rate of single-classification projects; An active factor setting module, used to receive the activation rate of a single-category project in the classification calculation module, and set the project active factor and the user active factor; An optimization scheme module, used to receive the activation rate of single-category items, item activity factor and user activity factor to generate an optimization scheme; The feedback adjustment module is used to monitor and obtain push feedback data in real time and adjust the optimization strategy in real time.

[0016] The beneficial effects of the present invention are as follows: by classifying users and projects, calculating the activation rate of single-category projects for each combination of user classification and project classification, it is possible to more accurately match users with smart medical projects, taking into account the interests and needs of users, combining the characteristics of the project and the target user groups, and more in line with the characteristics and preferences of different user groups, thereby improving the accuracy of push and user experience; by setting user activity factors and project activity factors, it is possible to identify and actively push information to users who may be interested in the project but are currently less active, thereby improving overall user activity. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of the information push method based on smart medical big data of the present invention; Figure 2 This is an architecture diagram of the information push system based on smart medical big data of the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0019] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, "plurality" means two or more, unless otherwise clearly and specifically defined.

[0020] In the description of the present application, the term "for example" is used to mean "used as an example, illustration or description". Any embodiment described as "for example" in the present application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid unnecessary details to obscure the description of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present application.

[0021] Example 1: Figure 1As shown, the present application provides an information push method based on smart medical big data, the method comprising: S100: Obtain smart medical project information and corresponding session feature data, and simultaneously obtain push service information statistics, and classify users and projects according to the smart medical project information and corresponding session feature data; Specifically, the session feature data includes consultation content, consultation projects, consultation types and number of sessions, etc.; the smart medical project information includes the smart medical project name, sub-project name, project description, target user group, project type, project service content and project feedback; the project type represents whether the project is a general project or a special project; the push business information statistical results include: push project statistics, the name and number of pushed projects; user feedback statistics, user click-through rate, conversion rate and satisfaction with push information; push time statistics, push information sending time, frequency, etc.; user behavior points, user behavior after receiving push information, such as whether to view, whether to consult, etc.; push effect analysis, a comprehensive analysis of the above content is performed to obtain the push effect, expressed as a specific score or percentage, without specific limitation.

[0022] Specifically, the user classification is based on the user's historical session feature data, extracts key features from the session data, such as the user's interests, health needs, health status, etc., uses a clustering algorithm to group users, and classifies them based on factors such as users' interests, needs, health status, etc., such as "cardiovascular disease patients", "diabetic patients", "health consultants", etc.; the project classification is divided into two categories: general projects and special projects. The general projects are suitable for most users, such as health consultations, routine physical examinations, etc.; special projects are aimed at specific user groups or specific health problems, such as cardiovascular disease treatment, diabetes management, etc., and the target user group, project type, and project service content of each project are analyzed and classified.

[0023] S200: for each user classification and project classification combination, calculating the single classification project activation rate, and calculating the activation rates of multiple single classification projects using a weighted average method to obtain a comprehensive project activation rate; Specifically, each user classification is combined with each project classification to form a user-project classification combination. For each user-project classification combination, the number of sessions of the user under the project classification is counted, and the single-classification project activation rate is calculated, where the single-classification project activation rate = the number of sessions of the user under the project classification / the total number of sessions of the user. The weighted average method is used to calculate the activation rates of multiple single-classification projects to obtain the comprehensive project activation rate. The weights of different classification combinations are set in combination with the user classification and the project classification. The weighted average method is used to calculate the comprehensive project activation rate. The trend changes of the project openings can be clearly identified based on the comprehensive project activation rate. The single-classification project activation rate can be used to accurately match users and projects, thereby improving the push effect.

[0024] S300: setting a project activity factor according to the value of the activation rate of the single-category project; setting a user activity factor according to the number of project types involved by a single user; Specifically, according to the value of the activation rate of the single-classification project, the specific steps of setting the project activity factor are as follows: S301: According to business needs or historical data, determine a series of threshold ranges of the activation rate of the single-classification project, and set different grades according to the threshold range; S302: Compare the activation rate of each single-classification project with the set threshold to determine the grade to which the activation rate of each single-classification project belongs; S303: Pre-assign a project activity factor to each different grade, and match the corresponding project activity factor according to the actual grade. For example, the threshold is set to three grades of low (0%-20%), medium (21%-50%), and high (51%-100%). The project activity factor of the low grade is 1, the project activity factor of the medium grade is 2, and the project activity factor of the high grade is 3. For example, there are the following three single-classification project activation rate data: the activation rate of project A: 45%; the activation rate of project B: 75%; the activation rate of project C: 15%. According to step 1, we set the following thresholds: low threshold: 0%-20%; medium threshold: 21%-50%; high threshold: 51%-100%. Next, we categorize the project activation rates according to step 2: Project A’s activation rate is medium (21%-50%); Project B’s activation rate is high (51%-100%); Project C’s activation rate is low (0%-20%). Finally, according to step 3, we set the project activity factor for each project: Project A’s project activity factor: 2; Project B’s project activity factor: 3; Project C’s project activity factor: 1.

[0025] Specifically, the user activity factor is set based on the number of project types involved by a single user, reflecting the user's enthusiasm for participating in smart medical projects. For each user, the number of different project types involved is counted, and a user activity factor is assigned to the user based on the number of project types involved. For example, a baseline value (such as 1) is set, and then the user activity factor increases by a certain value (such as 0.5) for each additional project type. Projects with high project activity factors indicate that users have a high degree of interest in the project, and users with high user activity factors indicate that they are highly motivated to participate in smart medical projects.

[0026] S400: generating an optimization plan according to the activation rate of the single-category project, combined with the project activity factor and the user activity factor; optimizing the information push strategy according to the optimization plan, pushing information according to the optimized information push strategy, and monitoring and acquiring push feedback data in real time; Specifically, the optimization plan is to divide users into high-active, medium-active and low-active users according to the activation rate of the single-category project, the project activity factor and the user activity factor; divide the projects into three priorities: high, medium and low; set a basic push strategy and a cross-pre-test push strategy in the optimization plan, the basic push strategy is to adopt information push of different frequencies for different active users and different priority projects, the cross-pre-test push strategy is to push high-priority projects to low-active users, and push low-priority projects to high-active users, set a time limit to collect user feedback result data, analyze the feedback results collected by the cross-pre-test push strategy, determine the push frequency and push range in combination with the basic push strategy, and generate the final information push optimization strategy.

[0027] Specifically, users are subdivided into highly active, medium active and low active users based on user activity factors. Based on the project activity factors and the activation rate of single-category projects, projects are divided into three priorities: high, medium and low. The basic push strategy is to implement high-frequency, personalized information push for high-priority projects and highly active users; to implement a moderate information push strategy for medium-priority projects and medium-active users; and to adopt low-frequency but accurate information push for low-priority projects and low-active users. The push frequency refers to the frequency of push. The high and low frequencies are set according to the actual situation and adjusted according to the specific habits of the users.

[0028] The cross-pre-test push strategy is to conduct cross-pre-tests according to the priorities and activity levels, and analyze the results to form a final more accurate optimization push strategy. Specifically, high-priority projects are pushed to some low-activity users and their feedback is collected, aiming to understand the reactions of low-activity users to high-priority projects, and further analyze the reasons for the inactivity of low-activity users and possible problems with the projects. At the same time, low-priority projects are pushed to some high-activity users, and their feedback is collected to identify possible deficiencies and room for improvement of low-priority projects. The time limit is the time limit for cross-pre-test pushes, which can be set according to actual conditions. Preferably, the time limit should not be less than one week.

[0029] Analyze the feedback data collected from cross-pre-test pushes to identify the specific reasons why users are not interested in the project or are not active. Adjust the project content, presentation method or push strategy based on the feedback to increase the attractiveness of the project and the activity of users. Regularly review the entire push strategy, including the effects of basic push and cross-pre-test push, and perform necessary iterations and optimizations. Continuously update user segmentation and project classification methods to adapt to changing market needs and user preferences.

[0030] Specifically, for example, there are two smart medical projects A and B, and two user groups C and D. The following results are obtained through data calculation: Project A has a high activation frequency of single-category projects and a high project activity factor, indicating that the project is popular and users are interested in it. Project B has a medium activation frequency of single-category projects and a low project activity factor, indicating that the project is used in general and users are not interested in it.

[0031] The user activity factor of user group C is high, indicating that this group actively participates in the smart medical project. The user activity factor of user group D is low, indicating that this group is not very enthusiastic about participating in the smart medical project.

[0032] Based on the above data, the following push strategy can be formulated: For project A, push it to user groups C and D at the same time. Because the project is popular and user group C is highly active, the push effect is expected to be better. At the same time, push it to user group D to form a comparison group with user group C, analyze the subsequent push result feedback, and further determine the reason why the activity factor of user group D is low.

[0033] For project B, push it to user group C. Although the project is not very interesting, the activity of user group C may stimulate their interest in the project. At the same time, push it appropriately to highly active users in user group D to expand the influence of the project.

[0034] Set the appropriate push frequency and time to ensure that information reaches users at the right time. The forms of push information can be diverse, including SMS, email, APP notification, etc. The specific form is selected according to user needs and platform characteristics.

[0035] S500: Evaluate the push optimization effect according to the comprehensive project activation rate, and generate a user trend change report for data analysis and real-time adjustment of the optimization strategy; Specifically, the push optimization effect is evaluated through the changes in the comprehensive project activation rate before optimization and the comprehensive project optimization rate after optimization, and a user trend change report is generated based on the changes in the comprehensive project activation rate. The user trend change report is analyzed to understand the current trends that users are concerned about, which plays a decisive role in the addition and reduction of subsequent smart medical projects. At the same time, user feedback data after push is collected, including indicators such as click-through rate, conversion rate, and user satisfaction, to further analyze the push effect and evaluate the effectiveness of the optimization strategy. According to the evaluation results, the optimization strategy is adjusted to continuously iterate and improve the information push effect.

[0036] In the present application, preferably, the smart medical project can be a single project or a comprehensive project including multiple sub-projects. If it includes multiple sub-projects, the categories of the sub-projects must be consistent.

[0037] In this embodiment, by classifying users and projects, and calculating the activation rate of single-category projects for each combination of user classification and project classification, users and smart medical projects can be matched more accurately, taking into account the interests and needs of users, combining the characteristics of the project and the target user group, and more in line with the characteristics and preferences of different user groups, thereby improving the accuracy of push and user experience; evaluating the push optimization effect through the comprehensive project activation rate helps to understand the effectiveness of the push strategy and provide data support for subsequent project development; by setting user activity factors and project activity factors, information can be identified and actively pushed to users who may be interested in the project but are currently less active, thereby improving overall user activity.

[0038] Embodiment 2: In the above embodiment, users and projects are classified in detail and different priorities are set for different users' activities and projects. An optimization plan is set to optimize and adjust the push strategy. This embodiment makes further limitations.

[0039] The optimization scheme also includes a related project recommendation rule, which selects complementary projects that are strongly related to the project with the highest current attention of the user according to the project association matrix, pushes information according to the complementary projects, and guides the optimization of the user's access path; The project association matrix is ​​constructed by analyzing the association between smart medical projects using an association rule learning algorithm, and based on the calculated association, the project association matrix is ​​constructed; the strongly associated complementary projects refer to projects with an association greater than an association threshold; Specifically, association rule learning algorithms (such as Apriori or FP-Growth) are used to find frequent item sets. Association rule learning algorithms are common and well-known existing technologies in this field. Combinations of items that often appear together are found. The correlation between items is analyzed based on the frequent item sets. Based on the calculated correlation, a project correlation matrix is ​​constructed. The project correlation matrix is ​​a two-dimensional matrix, in which each cell identifies the correlation strength between two items. The association rule learning algorithm is used to find the association rules between items, and the association matrix is ​​filled according to the found association rules. A correlation threshold is set, and the correlation between two items is calculated. When the correlation exceeds the correlation threshold, it is determined that there is a strong correlation between the two items. When information is pushed to users, when a user visits a certain pushed item, other items in the correlation matrix whose correlation with the item exceeds the threshold are found, and these items are pushed as the second recommended items.

[0040] The user access path optimization refers to identifying key nodes in the user access path and inserting complementary items at the key nodes to achieve a natural transition; the key nodes refer to the project pages that users visit and stay on, and complementary items are inserted during the time they stay on the project pages.

[0041] Specifically, user access path optimization refers to inserting guidance and prompts for complementary projects so that users can transition more naturally to other related projects during the access process, thereby improving the overall activation frequency of projects and user retention rate; identifying key nodes in the user access path, that is, pages or projects that users frequently visit or stay on, and inserting guidance and prompts for other projects that are strongly related to them at the key nodes, such as pop-up recommendations, sidebar prompts or bottom navigation bar recommendations, etc. At the same time, according to the user's historical behavior and preferences, the inserted related guidance is personalized to improve the pertinence and effectiveness of the guidance.

[0042] In this embodiment, project correlation analysis is introduced, and a project correlation matrix is ​​constructed, so that the internal connection between projects can be quantified and clarified, which deepens the understanding of the relationship between projects and provides a more accurate basis for subsequent information push and user guidance. By identifying the strong correlation rules between projects, when a user visits a project, it can intelligently recommend complementary projects closely related to it, achieving a highly accurate second recommendation, thereby greatly improving the user experience; the guidance and prompts of complementary projects are inserted into the user access path, so that users can transition to other related projects more naturally, and effectively improve the overall activation frequency and user retention rate of the project. By guiding users to contact and learn more about smart medical projects, the stickiness and satisfaction of users on the platform are enhanced.

[0043] Embodiment 3 introduces project correlation analysis in the above embodiment, constructs a project correlation matrix, and implements secondary push in the process of information push, which improves the user experience and increases the activation rate of the project. This embodiment makes further limitations on the above content.

[0044] Acquire the push feedback data, integrate the user push service results, calculate the matching value of each push item received by the user, count the matching values ​​of all push items of each user and calculate the median, set the initial item matching value range with the median as the origin; debug the initial item matching value range according to the debugging rules to obtain the accurate item matching value range; Specifically, the push feedback data includes the number of user clicks, user dwell time, user conversion rate, user feedback, whether the user clicks on the pushed message, and tracks the user's dwell time on the push content page. The user's satisfaction with the pushed content is collected through questionnaires, rating systems or direct feedback channels. If the user consults, makes an appointment, or undergoes an examination because of the push, there is a conversion rate. The results of the integrated user push service are to classify the data according to the dimensions of push items, user groups, time, etc., and calculate the average, median, standard deviation and other statistics of various indicators to reflect the overall situation.

[0045] The matching value is an indicator used to measure the degree of fit between the pushed project and the user's interests and needs. A standard value is set for each indicator in the push feedback data (number of user clicks, user dwell time, user conversion rate, etc.). The standard value is based on the average value of each indicator in a certain project of the user. Each indicator of the project is compared with the average value to obtain the ratio of each indicator, and the matching value is generated based on the ratio of each indicator.

[0046] For example, the average value of all indicators of a certain project for all users is 50, and user A's indicators for the project are 80. The ratio of each item is 1.6, and the combined value is 4.8, which is the final matching value.

[0047] Specifically, the matching values ​​of all pushed items of each user are counted and the median is calculated, and the initial item matching value range is set with the median as the origin. For example, the lower limit of the initial item matching value range is set with the median as the origin and floating downward by 30%, forming the initial item matching value range with the highest value.

[0048] The debugging rule is to pre-test the lower limit of the initial project matching value range before and after, until the maximum number of pre-tests is reached, observe user feedback and push effects, and adjust the project matching value range based on the pre-test results; the maximum number of pre-tests is pre-set.

[0049] Specifically, the debugging rule is based on the initial project matching value range, and pre-tests are performed on the lower limit. The pre-test range can be pre-set according to actual conditions. Two groups of test items are created, one group with a matching value slightly lower than the lower limit (such as 5% lower), and the other group with a matching value slightly higher than the lower limit (such as 5% higher). Other variables (such as push time, content type, etc.) are kept consistent, and push is performed to users, and user feedback on the pushed items is recorded, including click-through rate, conversion rate, dwell time, etc.; at the same time, other users similar to the user are selected, and pre-tested in the same way and recorded; user feedback data of the two groups of test items are compared, and different user reactions to items with slightly lower and slightly higher matching values ​​are observed, as well as different feedback effects between similar users are observed. According to the preliminary test results of the above two groups of experiments, if multiple users respond positively to items with slightly lower matching values, the lower limit of the range can be appropriately lowered; if multiple users respond better to items with slightly higher matching values, the lower limit of the range can be appropriately raised; select items within the new matching value range for push, and repeat the above steps until the maximum number of preliminary tests (such as set to 5 times) is reached. Based on the results of the five preliminary tests, the final item matching value range, that is, the precise item matching value range, is determined.

[0050] In this embodiment, the matching values ​​of all projects involved in a certain user are calculated through analysis of the project results. At the same time, after debugging, the most accurate project matching value range is obtained. The projects within this matching value range are all project contents that are most interesting to the user and highly matched with his needs. The user is recommended projects within the accurate project matching value range. The user will be more likely to receive projects that match their interests and needs, thereby improving user satisfaction and acceptance, and further enhancing user stickiness and activity.

[0051] Embodiment 4: In the above embodiment, the most accurate item matching value range is obtained, and the items within the accurate item matching value range are recommended to the user. This embodiment makes further limitations on the above.

[0052] Obtain the precise project matching value range, dynamically adjust the project matching value range in combination with the random exploration mechanism to form a fluctuating project matching value range, and optimize information push according to the fluctuating project matching value range; The random exploration mechanism randomly extracts a rated number of untouched items to fill into the precise item matching value range, sets a time interval, periodically replaces items, and removes invalid items in the precise item matching value range; the rated number is the number of untouched items to be added based on the number of items in the precise item matching value range and the random disturbance factor.

[0053] Specifically, the rated number in the random exploration mechanism is determined by combining historical data and user feedback, specifically based on the number of items within the precise item matching value range and the random perturbation factor, to ensure that while maintaining a certain recommendation accuracy, diversity is increased. The setting of the random perturbation factor should be based on historical data of user behavior, actual test results, and business goals; the initial value can be set to a smaller proportion, such as 5% or 10%, and then adjusted according to the actual effect. If the recommendation accuracy is too high, causing users to always see similar content, the value of the random perturbation factor can be appropriately increased. The number of untouched items to be added is jointly determined by the random perturbation factor and the number of items within the precise item matching value range. For example, if the random perturbation factor is 5% and the number of items within the precise item matching value range is 100, then the rated number of newly added untouched items is 5.

[0054] Specifically, a reasonable time interval is set, which can be daily, weekly or monthly, depending on the timeliness of the content and the frequency of user behavior. If the user interacts frequently, the update cycle can be shortened. In each update cycle, some recommended content is replaced based on user behavior and feedback data to ensure that users can continue to be exposed to fresh and popular content. The invalid project is determined based on the user's real-time feedback on the project, the timeliness of the project and the user behavior data to comprehensively judge whether the project is invalid. Specifically, the user's negative feedback on the project, such as low score, negative comments or quick skip; the timeliness of the project expires, such as a seasonal science popularization or temporary medical lecture; according to user behavior data, if a project has not been interacted by users for a long time; according to the above three situations, a comprehensive judgment is made, and if the above situation exists, it is determined to be an invalid project. Preferably, the user behavior data can be judged according to the time threshold, and a threshold is set. For example, if a project has not been clicked by the user for a continuous period of time (such as a month) or the score is lower than a certain value (such as 2 points), it is determined to be invalid.

[0055] This embodiment introduces a random exploration mechanism based on the above embodiment, which can increase the diversity of recommendations while ensuring the accuracy of recommendations. This diversity is not only reflected in the types of recommended content, but also in the recommendation of new content, thereby helping users discover more content of interest; by dynamically adjusting the project content within the project matching value range, it can continuously adapt to changes in user interests and content updates. This dynamic adaptability can always keep pace with user needs; setting time intervals can ensure that users are exposed to the latest smart medical projects in a timely manner.

[0056] Example 5: This example provides an information push system based on smart medical big data. Figure 2 As shown, the system includes: a data acquisition module for acquiring smart medical project information, session feature data and push business information statistics; The classification calculation module is used to receive data information from the data acquisition module, classify users and projects, and calculate the activation rate of single-classification projects; An active factor setting module, used to receive the activation rate of a single-category project in the classification calculation module, and set the project active factor and the user active factor; An optimization scheme module, used to receive the activation rate of single-category items, item activity factor and user activity factor to generate an optimization scheme; The feedback adjustment module is used to monitor and obtain push feedback data in real time and adjust the optimization strategy in real time.

[0057] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0058] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0059] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0060] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0061] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0062] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0063] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. Information push method based on smart medical big data, It is characterized in that The method comprises: S100: Obtain smart medical project information and corresponding session feature data, and simultaneously obtain push service information statistics, and classify users and projects according to the smart medical project information and corresponding session feature data; S200: for each user classification and project classification combination, calculating the single classification project activation rate, and calculating the activation rates of multiple single classification projects using a weighted average method to obtain a comprehensive project activation rate; S300: setting a project activity factor according to the value of the activation rate of the single-category project; setting a user activity factor according to the number of project types involved by a single user; S400: generating an optimization plan according to the activation rate of the single-category project, combined with the project activity factor and the user activity factor; optimizing the information push strategy according to the optimization plan, pushing information according to the optimized information push strategy, and monitoring and acquiring push feedback data in real time; S500: Evaluate the push optimization effect according to the comprehensive project activation rate, and generate a user trend change report for data analysis and real-time adjustment of the optimization strategy.

2. The information push method based on smart medical big data according to claim 1, It is characterized in that The optimization scheme is to classify users into highly active, medium active and low active users according to the activation rate of the single-category project, the project activity factor and the user activity factor; to classify projects into three priorities: high, medium and low; in the optimization scheme, a basic push strategy and a cross-pre-test push strategy are set, and the basic push strategy adopts information push of different frequencies for different active users and different priority projects; the cross-pre-test push strategy is to push high-priority projects to low-active users, and push low-priority projects to high-active users, set a time limit to collect user feedback result data, analyze the feedback results collected by the cross-pre-test push strategy, determine the push frequency and push range in combination with the basic push strategy, and generate the final information push optimization strategy.

3. The information push method based on smart medical big data according to claim 2, It is characterized in that The optimization scheme also includes a related project recommendation rule, which selects complementary projects that are strongly related to the user's current most popular project based on the project association matrix, pushes information based on the complementary projects, and guides the optimization of user access paths.

4. The information push method based on smart medical big data according to claim 3, It is characterized in that The project association matrix is ​​constructed by analyzing the association between smart medical projects using an association rule learning algorithm, and based on the calculated association, the project association matrix is ​​constructed; the strongly associated complementary projects refer to projects with an association greater than an association threshold.

5. The information push method based on smart medical big data according to claim 3, It is characterized in that The user access path optimization refers to identifying key nodes in the user access path and inserting complementary items at the key nodes to achieve a natural transition; the key nodes refer to the project pages that users visit and stay on, and complementary items are inserted during the time they stay on the project pages.

6. The information push method based on smart medical big data according to claim 1, It is characterized in that Acquire the push feedback data, integrate the user push service results, calculate the matching value of each push item received by the user, count the matching values ​​of all push items of each user and calculate the median, set the initial item matching value range with the median as the origin; debug the initial item matching value range according to the debugging rules to obtain the accurate item matching value range.

7. The information push method based on smart medical big data according to claim 6, It is characterized in that The debugging rule is to pre-test the lower limit of the initial project matching value range before and after, until the maximum number of pre-tests is reached, observe user feedback and push effects, and adjust the project matching value range based on the pre-test results; the maximum number of pre-tests is pre-set.

8. The information push method based on smart medical big data according to claim 6, It is characterized in that The precise project matching value range is obtained, and the project matching value range is dynamically adjusted in combination with a random exploration mechanism to form a jumping project matching value range, and information push optimization is performed according to the jumping project matching value range.

9. The information push method based on smart medical big data according to claim 8, It is characterized in that The random exploration mechanism randomly extracts a rated number of untouched items to fill into the precise item matching value range, sets a time interval, periodically replaces items, and removes invalid items in the precise item matching value range; the rated number is the number of untouched items to be added based on the number of items in the precise item matching value range and the random disturbance factor.

10. An information push system based on smart medical big data, supporting an information push method based on smart medical big data as described in any one of claims 1 to 9, It is characterized in that The system includes: a data acquisition module for acquiring smart medical project information, session feature data and push business information statistics; The classification calculation module is used to receive data information from the data acquisition module, classify users and projects, and calculate the activation rate of single-classification projects; An active factor setting module, used to receive the activation rate of a single-category project in the classification calculation module, and set the project active factor and the user active factor; An optimization scheme module, used to receive the activation rate of single-category items, item activity factor and user activity factor to generate an optimization scheme; The feedback adjustment module is used to monitor and obtain push feedback data in real time and adjust the optimization strategy in real time.

Images