Cross-border e-commerce content pushing method and system based on artificial intelligence

Abstract

The invention discloses a cross-border e-commerce content pushing method and system based on artificial intelligence, relates to the technical field of cross-border e-commerce, and solves the specific problem that the network quality of a user side is not fully considered during cross-border e-commerce pushing. Adjusting the frame rate and the resolution of the pushed content according to the frame rate and the resolution; it is ensured that the pushed content can be stably and smoothly transmitted to the user in different network environments, and the situations of pushing failure, slow content loading or packet loss and the like caused by network problems are avoided; the pushing problem caused by network differences in different regions is solved for cross-border e-commerce, the user coverage range is widened, and the user experience under various network conditions is improved; for example, in a region with relatively low network flow, the resolution and frame rate of the pushed content are reduced, and the content can be normally pushed to the user side.

Description

Technical Field

[0001] The present invention relates to the technical field of cross-border e-commerce, and specifically to a cross-border e-commerce content push method and system based on artificial intelligence. Background Art

[0002] In the traditional cross-border e-commerce content push method, if the user-side network conditions are not fully considered, it will cause a series of serious problems, greatly affecting user experience and business development.

[0003] First, in terms of push content transmission, due to the huge differences in network environment in different regions, if the push is not adjusted according to the user's characteristic traffic range, the high-resolution and high-frame-rate push content will be easily loaded slowly or even unable to load when transmitted to the user end with low network traffic. For example, in some remote areas with relatively weak network infrastructure, or during network congestion, users may face the loading icon on the screen for a long time and cannot obtain the pushed product information in time. This will cause the user's patience with the platform to quickly wear out, thereby reducing their favorability and willingness to use the platform.

[0004] Secondly, push failures due to network traffic overlap or insufficiency occur frequently. When the transmission traffic required to push content exceeds the user's actual network carrying capacity, packet loss will occur, making it impossible to fully present the push content to the user. This will not only cause users to miss important product promotion information, but also for e-commerce platforms, it means that the invested push resources are wasted and the expected marketing effect cannot be achieved.

[0005] In addition, a unified push method that does not consider the user's network conditions will also cause irrational use of network resources. Users with better network traffic could have received higher-quality push content, but they cannot get the best experience due to unified settings. For users with poor network conditions, forced push of high-specification content will cause network congestion, further worsening the usage environment. In the long run, the platform user activity and retention rate will be seriously affected, hindering the continued expansion and in-depth development of cross-border e-commerce business. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention provides a cross-border e-commerce content push method and system based on artificial intelligence, which solves the specific problem of not fully considering the network quality of the user end when performing cross-border e-commerce push.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a cross-border e-commerce content push method based on artificial intelligence, comprising the following steps: Step 1: Limit the association period, record the user behavior data of different users within the association period, and determine the push content associated with the corresponding user from the recorded user behavior data. The specific method is as follows: S11, limiting the association period, where the association period is a preset period, and recording the user behavior data of different users using the e-commerce platform within the association period; S12, confirming the number of times the corresponding user clicks on the relevant product and the viewing time of the relevant product from the recorded user behavior data, recording the click behavior with viewing time ≤ Y1 as invalid behavior, and eliminating such invalid behavior, otherwise, recording the corresponding click behavior as valid behavior, where Y1 is a preset value; S13. From the recorded valid behaviors, confirm the category of the clicked related products. The related products have been marked in advance in the e-commerce platform. The related products belonging to the same category are recorded as similar products. The total number of times and total viewing time associated with similar products are locked. The total number of times the same products are associated within the association period is marked as CS. i , the total viewing time is marked as SC i , where i represents different categories; Using PD i =CS i ×C1+SC i ×C2 confirms the evaluation value PD generated by the corresponding category within this associated period i , where C1 and C2 are preset fixed coefficient factors; From the rating values ​​PD associated with several categories i In the PD i The category associated with max is recorded as the pushed category, and relevant content is randomly selected from the pushed category and recorded as the pushed content; Step 2: After the push content associated with the corresponding user is determined, the push period of the corresponding user is identified from the user behavior data recorded in the past historical data, and the best push period is selected from the identified several groups of push periods. The past historical data is the period data associated with each associated period. The specific method is as follows: S21. Identify the user behavior data associated with each different association period from the past historical data, and confirm the relevant time period of the user behavior from the different user behavior data; S22. Among the different relevant time periods confirmed by the behavior data of different users, the existing intersecting time periods are locked, and the number of intersecting times of the corresponding intersecting time periods is confirmed. The group of intersecting time periods with the largest number of intersecting times is recorded as the selected time period. If there is only one group of selected time periods, the selected time period is directly recorded as the optimal push time period for this user. If there are multiple groups of selected time periods, the group of selected time periods with the longest total duration is determined to be recorded as the optimal push time period for this user. Step 3: From the user behavior data recorded in the association period, confirm the network traffic data generated during the data interaction process, process several groups of network traffic data in a unified manner, and lock the characteristic traffic interval of this user. The specific sub-steps are: S31. Based on the network traffic data generated during the data interaction process, different network traffic data associated at different times are confirmed, and a network traffic data fluctuation curve corresponding to the data interaction process is generated. The network traffic data fluctuation curves at different time periods are all placed in the same two-dimensional coordinate system, and the horizontal coordinate of the two-dimensional coordinate system is the timeline, and the vertical coordinate is the network traffic; S32, confirming the maximum network traffic value Lmax and the minimum network traffic value Lmin of the network traffic data fluctuation curve, determining the associated point on the vertical coordinate axis based on the confirmed Lmax and Lmin, and constructing a parallel line perpendicular to the associated point and parallel to the horizontal coordinate axis, recording the parallel line associated with Lmax as the upper script line, and recording the parallel line associated with Lmin as the lower script line; S33, moving the upper and lower marking lines toward each other, preferably keeping the upper marking line unchanged, and gradually moving the lower marking line upward, confirming a number of moving processes, and stopping when the lower marking line moves to a group of network flow data below the upper marking line; Then move the upper line downward to a group of network flow data, and then gradually move the lower line upward, confirm several moving processes, and stop when the lower line moves to a group of network flow data below the upper line, and so on, execute several groups of moving processes, during the moving process, the upper line and the lower line cannot overlap; Record each different mobile process, process each mobile process, and confirm the characteristic value TZ of the mobile process k , where k represents different mobile processes, and the confirmation method is: S331, for a single group of mobile processes, record the difference CZ between the upper and lower marking lines, where the difference CZ=the network flow corresponding to the upper marking line-the network flow corresponding to the lower marking line, and then confirm the partial curve segments located at the network flow corresponding to the upper marking line and the network flow corresponding to the lower marking line from the network flow data fluctuation curve, identify the line length of the curve segment, and record it as XC; S332, using: XC÷CZ=TZ to determine the characteristic value TZ associated with the current moving process; S333: For other mobile processes, the same method as steps S331-S332 is used to confirm the characteristic value TZ associated with the corresponding mobile process. k , where k represents different mobile processes; S34, different characteristic values ​​TZ associated with different mobile processes k In the menu, select TZ k The mobile process associated with max records the selected mobile process as the standard process: If there is only one set of standard processes, the characteristic flow range of this user is determined based on the network flow data of the upper and lower lines of the standard process; If there are multiple groups of standard processes, the flow interval associated with each group of standard processes is confirmed first, and then the group of intervals with the largest flow interval value range is used as the characteristic flow interval of this user; Step 4: Based on the determined push content and optimal push period, adjust the frame rate and resolution associated with the push content, keep the transmission traffic used by the corresponding push content within the characteristic traffic range, and simultaneously ensure that the push of the push content is completed within the optimal push period. The specific method is: S41, confirm the specific duration of the optimal push period and mark it as Ts, then determine the specific capacity R of the push content, and use R÷Ts=V to confirm the minimum transmission flow V required for the corresponding push content; S42, checking the confirmed V with the determined characteristic flow interval: if V∈the characteristic flow interval or V<the minimum value of the characteristic flow interval, no processing is required, and when the best push period arrives, the push content is directly sent to the designated user; If V>the maximum value of the characteristic traffic interval, when the best push period arrives, the push content will be sent directly to the designated user, and the frame rate and resolution associated with the push content will be reduced in real time, and the push rate generated during the push process will be averaged to confirm the real-time average rate. When the confirmed average rate∈the characteristic traffic interval, stop reducing the frame rate and resolution associated with the push content. If the subsequent average rate is greater than the characteristic traffic interval, reduce the frame rate and resolution again until the average rate∈the characteristic traffic interval, and complete the push of this type of push content to this type of user.

[0008] Preferably, the cross-border e-commerce content push system based on artificial intelligence includes: The user behavior data analysis end defines the association period, records the user behavior data of different users within the association period, and determines the push content associated with the corresponding user from the recorded user behavior data; The best time period confirmation end, after the push content associated with the corresponding user is determined, identifies the push period of the corresponding user from the user behavior data recorded in the past historical data, and selects the best push period from the identified several groups of push periods, whose past historical data is the period data associated with each associated period; The traffic interval confirmation end confirms the network traffic data generated during the data interaction process from the user behavior data recorded in the association period, processes several groups of network traffic data in a unified manner, and locks the characteristic traffic interval of this user; The content push processing end adjusts the frame rate and resolution associated with the push content based on the determined push content and the optimal push time period, keeps the transmission traffic used by the corresponding push content within the characteristic traffic range, and simultaneously ensures that the relevant push of the push content is completed within the optimal push time period.

[0009] The present invention provides a cross-border e-commerce content push method and system based on artificial intelligence. Compared with the prior art, it has the following beneficial effects: The present invention records user behavior data by limiting the association period, screening effective behaviors, determining push categories and randomly selecting push content, accurately grasping user interests, and greatly improving the matching degree between push content and user needs. It changes the traditional blind push mode, reduces the frequency of users receiving irrelevant information, improves users' favorability and usage experience of e-commerce platforms, increases users' attention to push content and click-through rate, and thus improves the sales conversion rate of goods; Identify user behavior time periods from historical data, and select the best push time period by determining the overlapping time periods, frequency, duration, etc.; ensure that push information is delivered during the time period when users are most active and most likely to pay attention, thereby improving the effect of information reach. Avoid push during time periods when users are busy or not using the platform, reduce the probability of information being ignored, improve push efficiency, and make more reasonable use of push resources; Conduct in-depth analysis of network traffic data, lock in the user's characteristic traffic range, and adjust the frame rate and resolution of the pushed content accordingly; ensure that the pushed content can be transmitted to users stably and smoothly in different network environments, and avoid push failures, slow content loading or packet loss due to network problems; solve the push problems caused by network differences in different regions for cross-border e-commerce, expand user coverage, and improve user experience under various network conditions; for example, in areas with low network traffic, reduce the resolution and frame rate of the pushed content to ensure that the content can be pushed to the user end normally. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 It is a schematic diagram of the process of the present invention; Figure 2 It is a schematic diagram of the principle framework of the present invention. DETAILED DESCRIPTION

[0011] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0012] First embodiment

[0013] See also Figure 1 , this application provides a cross-border e-commerce content push method based on artificial intelligence, comprising the following steps: Step 1: Limit the association period, record the user behavior data of different users within the association period, and determine the push content associated with the corresponding user (that is, the relevant push content with the most clicks and viewing time) from the recorded user behavior data. The user behavior data is recorded only with the user's permission. When the corresponding user shops or views on the corresponding e-commerce platform, relevant behavior data will be generated. The specific method of confirming the push content associated with the corresponding user is as follows: S11. Limit the association period, which is a preset period, which is determined by relevant operators based on experience, and is generally set at 24 hours, and records the user behavior data of different users using this e-commerce platform within the association period; S12. From the recorded user behavior data, confirm the number of times the corresponding user clicks on the relevant product and the viewing time of the relevant product, record the click behavior with a viewing time ≤ Y1 as invalid behavior, and remove such invalid behavior. Otherwise, record the corresponding click behavior as valid behavior, where Y1 is a preset value, which is formulated by the relevant operator based on experience, and is generally set to 3 seconds. It may be a case of a wrong click, and such click viewing behavior is not considered at all; S13. From the recorded valid behaviors, confirm the category of the clicked related product. The related products have been marked in advance in the e-commerce platform. The related products belonging to the same category are recorded as similar products. The total number of times and the total viewing time associated with the similar products are locked (the total number of times is the sum of the click times of several related products belonging to the same category, and the total viewing time is the sum of the viewing time of several related products). The total number of times the similar products are associated within the association period is marked as CS. i , the total viewing time is marked as SC i , where i represents different categories; Using PD i =CS i ×C1+SC i×C2 confirms the evaluation value PD generated by the corresponding category within this associated period i , where C1 and C2 are preset fixed coefficient factors, and their specific values ​​are determined by the operator based on experience; From the rating values ​​PD associated with several categories i In the PD i The category associated with max is recorded as the push category, and relevant content is randomly selected from the push category and recorded as the push content. The relevant content is preset in advance and operated by relevant personnel; Specifically, when certain users use the e-commerce platform within a certain period, relevant usage records will be generated, and the usage records will include the click behavior and viewing time of the corresponding products. Therefore, the relevant content that the corresponding users are interested in can be confirmed from the user records generated by the corresponding users, so as to confirm the relevant push content; Step 2: After the push content associated with the corresponding user is determined, the push time period of the corresponding user is identified from the user behavior data recorded in the past historical data, and the best push time period is selected from the identified several groups of push time periods. The past historical data is the period data associated with each associated period, and the specific method of selecting the best push time period is: S21. Identify the user behavior data associated with each different association period from the past historical data, and confirm the relevant time period of the user behavior from the different user behavior data (that is, the relevant time period of the corresponding user using the corresponding e-commerce platform, and the relevant time period adopts a 24-hour system); S22. In different relevant time periods confirmed by different user behavior data, the existing cross time periods are locked, and the number of cross times of the corresponding cross time periods is confirmed (if a group of cross time periods is obtained by cross confirmation of five groups of relevant time periods, then the number of cross times is 5), and the group of cross time periods with the largest number of cross times is recorded as the selected time period. If there is only one group of selected time periods, then this selected time period is directly recorded as the best push time period for this user. If there are multiple groups of selected time periods, then the group of selected time periods with the longest total duration is determined as the best push time period for this user (the so-called longest total duration means the longest internal duration. For example, if the selected time periods are 9-10, 10-11 and 11-13, then the duration of 11-13 is the longest, so the time period of 11-13 is the best push time period for this user. In general, the selected time period is generally short, which may be a few minutes); Step 3: From the user behavior data recorded in the association period, confirm the network traffic data generated during the data interaction process, process several groups of network traffic data in a unified manner, and lock the characteristic traffic interval of this user. Specifically, due to different regions, the network ports used by the corresponding users are also different, so the network traffic characteristics generated by each port are also relatively different. If push content of the same parameter type is used for push, users with better network traffic can normally receive such push content, but users with crossed network traffic may not be able to normally receive such push content. Therefore, it is necessary to confirm the characteristic traffic interval of each different user to facilitate the adjustment of the resolution and frame rate of the push content to ensure that the corresponding user can normally receive such push content; The specific sub-steps for locking the characteristic flow range of this user are: S31. Based on the network traffic data generated during the data interaction process (that is, when the corresponding user uses this e-commerce platform, there will be traffic interaction), different network traffic data associated at different times are confirmed, and a network traffic data fluctuation curve for the corresponding data interaction process is generated. The network traffic data fluctuation curves for different time periods are all placed in the same two-dimensional coordinate system, and the horizontal coordinate of the two-dimensional coordinate system is the timeline, and the vertical coordinate is the network traffic; S32, confirming the maximum network traffic value Lmax and the minimum network traffic value Lmin of the network traffic data fluctuation curve, determining the associated point on the vertical coordinate axis based on the confirmed Lmax and Lmin, and constructing a parallel line perpendicular to the associated point and parallel to the horizontal coordinate axis, recording the parallel line associated with Lmax as the upper script line, and recording the parallel line associated with Lmin as the lower script line; S33, moving the upper and lower marking lines toward each other, preferably keeping the upper marking line unchanged, and gradually moving the lower marking line upward, confirming a number of moving processes, and stopping when the lower marking line moves to a group of network flow data below the upper marking line; Then move the upper line downward to a group of network flow data, and then gradually move the lower line upward, confirm several moving processes, and stop when the lower line moves to a group of network flow data below the upper line, and so on, execute several groups of moving processes. During the moving process, the upper line and the lower line cannot overlap (the value range generated by the overlap is 0, and there is no consideration of any meaning). For example: the network flow data corresponding to the proposed upper line is 10, and the network flow data corresponding to the lower line is 2. The lower line is moved upward gradually first, and the moving process includes: 3, 4, 5, 6, 7, 8, 9, and the first group of moving methods stops; Move the upper line downward to the position where the network traffic data is 9, and then move the lower line upward. The movement process includes: 3, 4, 5, 6, 7, 8, and the second group of movement methods stops. Similarly, several groups of value ranges associated with this fluctuation curve can be confirmed in turn to confirm several groups of movement processes; Record each different mobile process, process each mobile process, and confirm the mobile process characteristic value: S331, for a single group of mobile processes, record the difference CZ between the upper and lower marking lines, where the difference CZ=the network flow corresponding to the upper marking line-the network flow corresponding to the lower marking line, and then confirm the partial curve segments located at the network flow corresponding to the upper marking line and the network flow corresponding to the lower marking line from the network flow data fluctuation curve, identify the line length of the curve segment, and record it as XC; S332, using: XC÷CZ=TZ to determine the characteristic value TZ associated with this mobile process. The larger the TZ, the longer the line segment associated with the network traffic data within the corresponding value range, and the smaller the associated difference value. In this case, such data is more clustered and its characteristics are more obvious. The smaller the TZ, the less the corresponding data is clustered and its characteristics are not obvious. In order to confirm the characteristic traffic of the corresponding user, it is necessary to find the most clustered network traffic to lock the characteristic traffic interval of the corresponding user. S333: For other mobile processes, the same method as steps S331-S332 is used to confirm the characteristic value TZ associated with the corresponding mobile process. k , where k represents different mobile processes; S34, different characteristic values ​​TZ associated with different mobile processes k In the menu, select TZ k The mobile process associated with max records the selected mobile process as the standard process: If there is only one set of standard processes, the characteristic flow range of this user is determined based on the network flow data of the upper and lower lines of the standard process; If there are multiple groups of standard processes, the flow intervals associated with each group of standard processes are confirmed first, and then the group of intervals with the largest flow interval value range is used as the characteristic flow interval of this user (that is, the interval with the largest difference between the maximum and minimum values ​​in the corresponding interval belongs to the characteristic flow interval of the corresponding user); Step 4: Based on the determined push content and optimal push period, adjust the frame rate and resolution associated with the push content, keep the transmission traffic used for the corresponding push content within the characteristic traffic range, and simultaneously ensure that the push of the push content is completed within the optimal push period; The specific methods of adjustment are as follows: S41, confirm the specific duration of the best push period and mark it as Ts, then determine the specific capacity R of the push content, and use R÷Ts=V to confirm the minimum transmission flow V (that is, the transmission rate per unit time) required for the corresponding push content; S42, check the confirmed V with the determined characteristic traffic interval: if V∈characteristic traffic interval or V<the minimum value of characteristic traffic interval, no processing is required, and when the best push period arrives, the push content can be directly sent to the designated user; when the corresponding V is less than the minimum value of the corresponding interval, it means that the network traffic associated with the corresponding user terminal is sufficient to receive the push content within the corresponding period, because the corresponding characteristic traffic interval is higher than the corresponding minimum transmission traffic V, so the content can be directly pushed within the specific period; If V> the maximum value of the characteristic traffic interval, when the optimal push period arrives, the push content is directly sent to the designated user, and the frame rate and resolution associated with the push content are reduced in real time, and the push rate generated during the push process is simultaneously averaged to confirm the real-time average rate. When the confirmed average rate∈the characteristic traffic interval, the frame rate and resolution associated with the push content are stopped from being reduced. If the subsequent average rate is greater than the characteristic traffic interval, the frame rate and resolution are reduced again to stop when the average rate∈the characteristic traffic interval, and the relevant push of such push content to such users is completed. Specifically, in order to enable the corresponding push content to be transmitted to the corresponding user party completely and without packet loss, it is necessary to determine the transmission rate associated with the corresponding time period, and then based on the confirmed specific transmission, adjust the relevant parameters of the relevant push content to reduce the relevant frame rate and the corresponding resolution, so that the corresponding push content can be completely and effectively transmitted to the designated user end, which can not only ensure the normal push of the corresponding push content, but also ensure that the corresponding push process can be sufficiently applicable to different network bandwidths to achieve a more comprehensive push effect.

[0014] Second embodiment

[0015] Combination Figure 2 , a cross-border e-commerce content push system based on artificial intelligence, including: The user behavior data analysis end defines the association period, records the user behavior data of different users within the association period, and determines the push content associated with the corresponding user from the recorded user behavior data; The best time period confirmation end, after the push content associated with the corresponding user is determined, identifies the push period of the corresponding user from the user behavior data recorded in the past historical data, and selects the best push period from the identified several groups of push periods, whose past historical data is the period data associated with each associated period; The traffic interval confirmation end confirms the network traffic data generated during the data interaction process from the user behavior data recorded in the association period, processes several groups of network traffic data in a unified manner, and locks the characteristic traffic interval of this user; The content push processing end adjusts the frame rate and resolution associated with the push content based on the determined push content and the optimal push time period, keeps the transmission traffic used by the corresponding push content within the characteristic traffic range, and simultaneously ensures that the relevant push of the push content is completed within the optimal push time period.

[0016] Some of the data in the above formulas are numerically calculated by removing their dimensions, and the contents not described in detail in this specification belong to the prior art known to those skilled in the art; It should be noted that all user data collected in this application is collected with the user's consent and authorization, and the use of user data is legal and compliant, and the use and processing of user data complies with relevant laws, regulations and standards in relevant regions.

[0017] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A cross-border e-commerce content push method based on artificial intelligence, characterized in that: The following steps are involved: Step 1: limit the association period, record the user behavior data of different users within the association period, and determine the push content associated with the corresponding user from the recorded user behavior data; Step 2: After the push content associated with the corresponding user is determined, the push period of the corresponding user is identified from the user behavior data recorded in the past historical data, and the best push period is selected from the identified several groups of push periods, and the past historical data is the period data associated with each associated period; Step 3: From the user behavior data recorded in the association period, confirm the network traffic data generated during the data interaction process, process several groups of network traffic data in a unified manner, and lock the characteristic traffic interval of this user; Step 4: Based on the determined push content and optimal push period, adjust the frame rate and resolution associated with the push content, keep the transmission traffic used by the corresponding push content within the characteristic traffic range, and simultaneously ensure that the push content is pushed within the optimal push period.

2. The cross-border e-commerce content push method based on artificial intelligence according to claim 1 is characterized in that: In step 1, the specific method of confirming the pushed content is: S11, limiting the association period, where the association period is a preset period, and recording the user behavior data of different users using the e-commerce platform within the association period; S12, confirming the number of times the corresponding user clicks on the relevant product and the viewing time of the relevant product from the recorded user behavior data, recording the click behavior with viewing time ≤ Y1 as invalid behavior, and eliminating such invalid behavior, otherwise, recording the corresponding click behavior as valid behavior, where Y1 is a preset value; S13. From the recorded valid behaviors, confirm the category of the clicked related products. The related products have been marked in advance in the e-commerce platform. The related products belonging to the same category are recorded as similar products. The total number of times and total viewing time associated with similar products are locked. The total number of times the same products are associated within the association period is marked as CS. i , the total viewing time is marked as SC i , where i represents different categories; Using PD i =CS i ×C1+SC i ×C2 confirms the evaluation value PD generated by the corresponding category within this associated period i , where C1 and C2 are preset fixed coefficient factors; From the rating values ​​PD associated with several categories i In the PD i The category associated with max is recorded as the pushed category, and relevant content is randomly selected from the pushed category and recorded as the pushed content.

3. The cross-border e-commerce content push method based on artificial intelligence according to claim 1 is characterized in that: In step 2, the specific method of selecting the best push time period is: S21. Identify the user behavior data associated with each different association period from the past historical data, and confirm the relevant time period of the user behavior from the different user behavior data; S22. Lock the existing crossover time periods in different relevant time periods confirmed by different user behavior data, and confirm the number of crossovers of the corresponding crossover time periods. Record the group of crossover time periods with the largest number of crossovers as the selected time period. If there is only one group of selected time periods, directly record this selected time period as the best push time period for this user.

4. The cross-border e-commerce content push method based on artificial intelligence according to claim 3 is characterized in that: In step S22, if there are multiple groups of selected time periods, a group of selected time periods with the longest total duration is determined and recorded as the optimal push time period for this user.

5. The cross-border e-commerce content push method based on artificial intelligence according to claim 1 is characterized in that: In step 3, the specific sub-steps for locking the characteristic flow interval of this user are: S31. Based on the network traffic data generated during the data interaction process, different network traffic data associated at different times are confirmed, and a network traffic data fluctuation curve corresponding to the data interaction process is generated. The network traffic data fluctuation curves at different time periods are all placed in the same two-dimensional coordinate system, and the horizontal coordinate of the two-dimensional coordinate system is the timeline, and the vertical coordinate is the network traffic; S32, confirming the maximum network traffic value Lmax and the minimum network traffic value Lmin of the network traffic data fluctuation curve, determining the associated point on the vertical coordinate axis based on the confirmed Lmax and Lmin, and constructing a parallel line perpendicular to the associated point and parallel to the horizontal coordinate axis, recording the parallel line associated with Lmax as the upper script line, and recording the parallel line associated with Lmin as the lower script line; S33, moving the upper and lower marking lines toward each other, preferably keeping the upper marking line unchanged, and gradually moving the lower marking line upward, confirming a number of moving processes, and stopping when the lower marking line moves to a group of network flow data below the upper marking line; Then move the upper line downward to a group of network flow data, and then gradually move the lower line upward, confirm several moving processes, and stop when the lower line moves to a group of network flow data below the upper line, and so on, execute several groups of moving processes, during the moving process, the upper line and the lower line cannot overlap; Record each different mobile process, process each mobile process, and confirm the characteristic value TZ of the mobile process k , where k represents different mobile processes; S34, different characteristic values ​​TZ associated with different mobile processes k In the menu, select TZ k The mobile process associated with max records the selected mobile process as the standard process: If there is only one set of standard processes, the characteristic flow range of this user is determined based on the network flow data of the upper and lower lines of the standard process; If there are multiple groups of standard processes, the flow interval associated with each group of standard processes is confirmed first, and then the group of intervals with the largest flow interval value range is used as the characteristic flow interval of this user.

6. The cross-border e-commerce content push method based on artificial intelligence according to claim 5 is characterized in that: In step S33, the mobile process characteristic value is confirmed in the following manner: S331, for a single group of mobile processes, record the difference CZ between the upper and lower marking lines, where the difference CZ=the network flow corresponding to the upper marking line-the network flow corresponding to the lower marking line, and then confirm the partial curve segments located at the network flow corresponding to the upper marking line and the network flow corresponding to the lower marking line from the network flow data fluctuation curve, identify the line length of the curve segment, and record it as XC; S332, using: XC÷CZ=TZ to determine the characteristic value TZ associated with the current moving process; S333: For other mobile processes, the same method as steps S331-S332 is used to confirm the characteristic value TZ associated with the corresponding mobile process. k , where k represents different movement processes.

7. The cross-border e-commerce content push method based on artificial intelligence according to claim 1 is characterized in that: In step 4, the specific method of adjusting the pushed content is: S41, confirm the specific duration of the optimal push period and mark it as Ts, then determine the specific capacity R of the push content, and use R÷Ts=V to confirm the minimum transmission flow V required for the corresponding push content; S42, checking the confirmed V with the determined characteristic flow interval: if V∈the characteristic flow interval or V<the minimum value of the characteristic flow interval, no processing is required, and when the best push period arrives, the push content is directly sent to the designated user; If V>the maximum value of the characteristic traffic interval, when the best push period arrives, the push content will be sent directly to the designated user, and the frame rate and resolution associated with the push content will be reduced in real time, and the push rate generated during the push process will be averaged to confirm the real-time average rate. When the confirmed average rate∈the characteristic traffic interval, stop reducing the frame rate and resolution associated with the push content. If the subsequent average rate is greater than the characteristic traffic interval, reduce the frame rate and resolution again until the average rate∈the characteristic traffic interval, and complete the push of this type of push content to this type of user.

8. A cross-border e-commerce content push system based on artificial intelligence, the system is operated according to the cross-border e-commerce content push method based on artificial intelligence according to any one of claims 1 to 7, characterized in that: include: The user behavior data analysis end defines the association period, records the user behavior data of different users within the association period, and determines the push content associated with the corresponding user from the recorded user behavior data; The best time period confirmation end, after the push content associated with the corresponding user is determined, identifies the push period of the corresponding user from the user behavior data recorded in the past historical data, and selects the best push period from the identified several groups of push periods, whose past historical data is the period data associated with each associated period; The traffic interval confirmation end confirms the network traffic data generated during the data interaction process from the user behavior data recorded in the association period, processes several groups of network traffic data in a unified manner, and locks the characteristic traffic interval of this user; The content push processing end adjusts the frame rate and resolution associated with the push content based on the determined push content and the optimal push time period, keeps the transmission traffic used by the corresponding push content within the characteristic traffic range, and simultaneously ensures that the relevant push of the push content is completed within the optimal push time period.

Images