Method, device and system for activating transportation card

Abstract

The invention discloses a traffic card activation method, device and system, and belongs to the technical field of smart cards, and the method comprises the steps: obtaining the behavior data of a target user; according to the behavior data of the target user and a pre-trained machine learning model, iteratively calculating a card activating probability value of the target user in a first preset time period in the future, and outputting a pre-downloading instruction; downloading a transportation card basic data packet based on the pre-downloading instruction; and acquiring the personal key of the target user, and activating the card. According to the technical scheme, the active response of the downloading service is realized by predicting the card activating probability of the user; a traffic card data packet is divided into a basic packet and a dynamic key, and the dynamic key is used for activating the pre-downloaded basic packet, so that rapid activation of a traffic card is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of smart cards, in particular to a traffic card activation method. BACKGROUND

[0002] The mobile phone traffic card NFC technology is based on near field communication, and integrates the virtual entity traffic card in the device security chip (eSE) or software scheme (HCE). The user can open and recharge online through the card package application, and can directly complete payment by approaching the gate induction area with the mobile phone in the bus, subway and other scenes. This technology completely replicates the function of the entity traffic card, supports out-of-town recharge and multi-card management, and realizes the convenient experience of the mobile phone as a traffic card.

[0003] Currently, the process of opening a mobile phone NFC traffic card by a user is usually a passive response mode of "user initiates request-server response". The user end processing process is tedious and time-consuming, and the card opening system cannot predict user demand, and the server is under pressure from concentrated requests during peak periods. However, the existing optimization scheme only focuses on improving the server response speed or compressing the data packet size, but does not change the fundamental mode of passive response. Therefore, the defect of the prior art is that it cannot actively predict the user's card opening behavior to make pre-resource deployment, and realize the rapid activation of the traffic card. SUMMARY

[0004] The present application provides a traffic card activation method, device and system, which can realize the pre-download and rapid activation of the traffic card.

[0005] The present application provides a traffic card activation method suitable for a server, comprising: obtaining target user behavior data; According to the target user behavior data and the pre-trained machine learning model, the opening card probability value of the target user in the first preset time period in the future is iteratively calculated until the opening card probability value is greater than the first preset threshold, and a pre-download instruction is output; wherein, the target user behavior data is updated each time; wherein, the pre-trained machine learning model is obtained by training historical user behavior data; The pre-download instruction is sent to the user end, so that the user end downloads the traffic card basic data packet according to the pre-download instruction, and completes the card opening activation according to the traffic card basic data packet.

[0006] The application obtains a user card opening probability by inputting target user behavior data into a machine learning model, actively acquires a user state, and pre-downloads a traffic card basic data package, which can avoid a download peak period, smooth system load, avoid network congestion, and realize fast calling of the basic data package in a subsequent card opening step. By training a user card opening probability calculation model, accurate prediction of a user card opening behavior is realized, and a judgment basis for whether to pre-download a traffic card basic data package is provided. Compared with the prior art, the application can convert a traffic card download service from passive response to active download according to a machine learning prediction technology, greatly compresses a card opening time, balances user experience and system resource consumption, and realizes the technical effect of quickly activating a traffic card.

[0007] Further, the pre-trained machine learning model is obtained by training historical user behavior data, comprising: According to the historical user behavior data, a plurality of observation days are established; According to the historical user behavior data and the plurality of observation days, sample feature vectors and sample user labels are calculated until all observation days are calculated, and the sample feature vectors and the sample user labels of each observation day are output as a model training data set; wherein each observation day corresponds to a plurality of sample users; wherein the sample feature vector is quantitatively obtained according to the historical user behavior data of each sample user in a second preset time period before the corresponding observation day; wherein the sample user label is obtained according to whether each sample user opens a traffic card in a third preset time period after the corresponding observation day; The pre-trained machine learning model is trained by the model training data set.

[0008] In this way, by setting observation days, analyzing historical user behavior data in a specific time range before and after the observation days, forming sample feature vectors and sample user labels under different observation days, and obtaining a user data set required for machine learning model training, the machine learning model is trained by a large amount of data in the user data set. The model automatically learns the weight influence of different behavior data on the final prediction result, captures different feature weights, realizes prediction of the card opening probability value of the user in a specific time period in the future, and provides a solid data basis for subsequent intelligent pre-download decisions.

[0009] Further, the opening probability value of the target user in the first preset time period in the future is iteratively calculated according to the target user behavior data and the pre-trained machine learning model, specifically: The target user behavior data is quantified to extract features of each data source of the target user; The features of each data source of the target user are spliced into a unified high-dimensional feature vector; input the unified high-dimensional feature vector into the pre-trained machine learning model to obtain a card opening probability value of the target user in a future first preset time period.

[0010] In this way, by extracting the behavior features of the target user, quantifying different behavior features, and merging them into a high-dimensional feature vector, different dimensions of user behavior are converted into vector data that can be recognized by a machine learning model, thereby providing input data basis for model prediction. By inputting the high-dimensional feature vector, the user card opening probability is predicted, thereby providing a judgment basis for the subsequent traffic card data package pre-download.

[0011] The application provides a traffic card activation method suitable for a user end, including: uploading target user behavior data to a server, so that the server iteratively calculates a card opening probability value of the target user in a future first preset time period according to the target user behavior data and a pre-trained machine learning model, and outputs a pre-download instruction when the card opening probability value is greater than a first preset threshold; wherein the target user behavior data is updated each time. receiving the pre-download instruction output by the server, and downloading a traffic card basic data package based on the pre-download instruction; obtaining a target user personal key, and activating a card according to the target user personal key and the traffic card basic data package.

[0012] In this way, by uploading the data obtained by the user end to the cloud for processing and downloading through the server instruction, the memory occupation of the data processing and pre-download judgment in the local user end is reduced. By dividing the traffic card data download into a pre-downloadable basic data package and a real-time obtainable personal key, when the user actually initiates the card opening, only the key needs to be downloaded and the pre-downloaded basic data package needs to be activated, so that the traffic card can be quickly activated, the user card opening time is greatly shortened, the operation cost of the user is reduced, and the user experience is improved.

[0013] Further, the obtaining of the target user personal key and the activation of the card according to the target user personal key and the traffic card basic data package are specifically: sensing the range of a gate induction area, and applying for a target user personal key from a server when it is sensed that the gate induction area is entered and it is detected that there is a traffic card basic data package in the local; injecting the target user personal key into the traffic card basic data package to complete the card activation.

[0014] In this way, the user end judges whether to enter the gate induction area, and the user card opening behavior is judged in combination with the local basic data packet, so that the accurate scene triggering of the user card opening is realized. By dividing the traffic card data packet into the pre-downloaded traffic card basic data packet and the real-time applied personal key, when the user actually uses, only two lightweight steps of key application and injection are needed to realize the rapid activation of the traffic card, reduce the operation threshold, and improve the user experience.

[0015] Further, the traffic card basic data packet is downloaded based on the pre-download instruction, including: According to the pre-download instruction, the intelligent scheduler is enabled; Based on the intelligent scheduler and the first preset condition, the optimal download strategy is determined; According to the optimal download strategy, the traffic card basic data packet is downloaded; wherein the traffic card basic data includes application file structure, interface resource and rate table.

[0016] In this way, the pre-download instruction is converted into a specific download strategy by setting the intelligent scheduler, so that the user end can perform specific operations. By pre-downloading the traffic card basic data packet, the system resource utilization is optimized, the data download process is dispersed from the peak period of opening to the idle period of ordinary time, the server load is smoothed, network congestion is avoided, and the overall stability and efficiency of the system are improved.

[0017] Further, the optimal download strategy is determined based on the intelligent scheduler and the first preset condition, specifically: The first preset condition includes the card opening probability value in the pre-download instruction, the current network type, the device power and the load state; The device power and the load state are judged, when the device power is greater than or equal to the second preset threshold value and the load state is not in the high load state, the second preset condition is used for downloading; if the device power is less than the second preset threshold value, the task is suspended, and the downloading is performed when the device power is higher than the second preset threshold value; if the load state is in the high load state, the task is suspended, and the downloading is performed when the load state enters the idle state; The second preset condition is specifically: if the current network type is Wi-Fi environment, the downloading is performed; if the current network type is cellular network, if the card opening probability value in the pre-download instruction is greater than the third preset threshold value, the downloading is performed, otherwise, the task enters the waiting queue, and the downloading is performed when the current network type is Wi-Fi environment.

[0018] Thus, by formulating the optimal download strategy for the card opening probability value, the current network type, the device power and the load state, when the device power load is in a healthy state, the card opening probability value is judged, if the card opening probability is high, the data packet is immediately downloaded, otherwise, the download is carried out in the Wi-Fi environment, the current device state and the user card opening demand are balanced, and the optimal and reasonable download of resources is realized.

[0019] The application provides a traffic card activation device suitable for a server, comprising a data acquisition module, a probability prediction module and an instruction sending module. The data acquisition module is configured to acquire target user behavior data. The probability prediction module is configured to iteratively calculate a card opening probability value of a target user in a first preset time period in the future according to the target user behavior data and a pre-trained machine learning model, and output a pre-download instruction when the card opening probability value is greater than a first preset threshold; wherein, the target user behavior data is updated each time the iteration is performed; and the pre-trained machine learning model is trained according to historical user behavior data. The instruction sending module is configured to send the pre-download instruction to a user terminal, so that the user terminal downloads a traffic card basic data packet according to the pre-download instruction, and completes card opening activation according to the traffic card basic data packet.

[0020] The application provides a traffic card activation device suitable for a user terminal, comprising a data uploading module, a data downloading module and a card opening activation module. The data uploading module is configured to upload target user behavior data to a server, so that the server iteratively calculates a card opening probability value of a target user in a first preset time period in the future according to the target user behavior data and a pre-trained machine learning model, and outputs a pre-download instruction when the card opening probability value is greater than a first preset threshold; wherein, the target user behavior data is updated each time the iteration is performed. The data downloading module is configured to receive the pre-download instruction output by the server, and download a traffic card basic data packet based on the pre-download instruction. The card opening activation module is configured to acquire a target user personal key, and perform card opening activation according to the target user personal key and the traffic card basic data packet.

[0021] The application provides a traffic card activation system suitable for a server, comprising a user terminal and a server; wherein, the user terminal is configured to execute a traffic card activation method suitable for the user terminal; and the server is configured to execute a traffic card activation method suitable for the server. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. Obviously, the drawings described below only relate to some of the embodiments of the present application, and other drawings can also be obtained by those of ordinary skill in the art without any creative effort.

[0023] Figure 1 is a flowchart of an embodiment of the traffic card activation method for a server provided by the present application.

[0024] Figure 2 is a flowchart of an embodiment of the traffic card activation method for a client provided by the present application.

[0025] Figure 3 is a structural diagram of an embodiment of the traffic card activation device for a server provided by the present application.

[0026] Figure 4 is a structural diagram of an embodiment of the traffic card activation device for a client provided by the present application. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without any creative effort fall within the scope of protection of the present application.

[0028] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "include" and "have" and any variations thereof in the specification and claims of the present application and the above description of drawings are intended to cover non-exclusive inclusion.

[0029] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "a plurality of" is two or more, unless otherwise explicitly and specifically limited.

[0030] Reference to“an embodiment” herein means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase“in one embodiment” in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily all referring to a common set of embodiments, although they can. It will be explicitly understood that the application can be combined with one or more other applications, some of which will be known or developed by the applicant and some of which will be presently unknown or developed by others but later associated with the applicant.

[0031] Embodiment one Reference to Figure 1 To solve the problem of traffic card activation in the prior art, an embodiment of the application provides a traffic card activation method, which comprises steps 11 to 13, and the specific steps are as follows: Step 11, obtaining target user behavior data.

[0032] Further, the target user behavior data is quantified, and the features of each data source of the target user are extracted; and the features of each data source of the target user are spliced into a unified high-dimensional feature vector.

[0033] In one embodiment, obtaining target user behavior data comprises steps 1101 to 1102, and each step is specifically as follows: Step 1101, quantitatively processing the target user behavior data to extract data features.

[0034] Among them, the target user behavior data contains multiple data sources, and each data source is quantitatively processed independently in parallel, such as quantifying“geographic positioning data” into“daily average stay time in the target city”, and creating high-value features such as“daily average number of map navigation within the geographic fence in the target city” through data cross. In this way, by quantifying abstract user behavior to generate high-value features, data value deep mining is realized, high-discriminative user behavior features are created through data cross, and accurate prediction of user card opening behavior is realized based on multiple data sources.

[0035] Step 1102, splicing the feature data after quantitatively processing to obtain a unified high-dimensional feature vector.

[0036] In this way, by cross-fusing multiple data features, a unified high-dimensional feature vector for model input and realizing traffic card pre-opening behavior prediction is constructed, which provides a judgment basis for subsequent traffic card data package pre-download, and enhances the expression ability and accuracy of the prediction model.

[0037] Step 12, according to the target user behavior data and the pre-trained machine learning model, iteratively calculate the card opening probability value of the target user in the future first preset time period, until the card opening probability value is greater than the first preset threshold, output the pre-download instruction; wherein, each time iteration, update the target user behavior data; wherein, the pre-trained machine learning model is obtained by training according to the historical user behavior data.

[0038] Further, input the unified high-dimensional feature vector into the pre-trained machine learning model to obtain the card opening probability value of the target user in the future first preset time period.

[0039] Further, according to the historical user behavior data, establish a number of observation days; according to the historical user behavior data and the number of observation days, calculate the sample feature vector and the sample user label until all observation days are calculated, output the sample feature vector and the sample user label of each observation day as the model training data set; wherein, each observation day corresponds to a number of sample users; wherein, the sample feature vector is quantitatively obtained according to the historical user behavior data of each sample user in the second preset time period before the corresponding observation day; wherein, the sample user label is obtained according to whether each sample user opens the traffic card in the third preset time period after the corresponding observation day; train the pre-trained machine learning model through the model training data set.

[0040] In one embodiment, according to the target user behavior data and the pre-trained machine learning model, iteratively calculate the card opening probability value of the target user in the future first preset time period, until the card opening probability value is greater than the first preset threshold, output the pre-download instruction, including steps 1201 to 1203, each step is as follows: Step 1201, generate a model training data set according to historical user behavior data.

[0041] The model uses historical data accumulated and stored automatically during long-term operation for supervised learning training. First, a past date is selected from the historical timeline as an observation day to construct the baseline of the sample. Each user existing on the observation day is regarded as a sample user. The user behavior data of the sample user in a review period (e.g., 30 days) before the observation day is traced back and processed into a high-dimensional feature vector defined by the cost application to construct a sample feature vector X. Whether the sample user opens a traffic card in a city in a time window (e.g., 7 days) after the observation day is viewed. The time window is regarded as a third preset time period. If the sample user opens the traffic card, the sample user label Y of the sample user is defined as 1 (positive sample). Otherwise, the sample user label Y of the sample user is defined as 0 (negative sample). A training sample is constructed by sliding the time window. The observation day is slid backward on the time axis (e.g., from a certain day one year ago, sliding once a day until the recent period). The sample feature vector and the sample user label of different sample users are repeatedly constructed to generate millions or even tens of millions of training samples to form a complete model training data set.

[0042] The time window (e.g., 7 days) for predicting the future card opening behavior is set based on business logic and should meet the short-term travel decision cycle. The time window for predicting the future card opening behavior is regarded as the first preset time period.

[0043] The review period (e.g., 30 days) of the sample user behavior data is selected through experiments to balance the capture of effective behavior patterns and the filtering of noise. The review period of the sample user behavior data is regarded as the second preset time period.

[0044] The review period should capture the stable and periodic behavior habit patterns of the user. For example, a user may commute 5 days a week. A 30-day review period can capture about 20 commuting instances, which is sufficient for the model to learn that the user “moves between fixed locations in A city during morning and evening peak hours on weekdays”. If the review period is too short (e.g., 3 days), the model may not be able to distinguish between a casual visit and a long-term stable behavior, such as misjudging a short business trip as a strong card opening signal, thereby introducing noise.

[0045] At the same time, the review period also needs to ensure the timeliness of the user portrait and filter out obsolete and irrelevant historical behaviors. A long review period (e.g., 180 days) may include a large number of invalid behavior patterns. For example, a user worked in B city half a year ago but has returned to A city after leaving the job. A long review period will cause the user's behavior data in B city half a year ago to continue to interfere with the model, as if the user still has a strong association with B city, which introduces historical noise and reduces the accuracy of the model in predicting the current state of the user.

[0046] Therefore, to meet the above requirements for the review period, the embodiment uses different review periods (such as 14 days, 30 days, 60 days, and 90 days) to construct multiple sets of features in the model development stage, and uses the same validation data to evaluate the performance of these models, and finally selects the review period with the best performance on the validation set as the calculation parameter, determines the optimal time range of the review period, and improves the scientificity and accuracy of the prediction results.

[0047] In step 1202, the model is trained according to the obtained model training data set.

[0048] The machine learning model (such as gradient boosting decision tree GBDT) is trained according to the model training data set. During the model training process, the advanced machine learning model such as GBDT automatically learns the importance of each feature by analyzing a large amount of training data, and discovers which features are more critical to predicting the card opening behavior. For example, it may learn that the contribution degree (i.e., “weight”) of the “core circle target city card opening status” social feature is the highest, and the weight of the “device connection data” is relatively low. At the same time, the machine learning model can automatically evaluate the importance of the features and capture the nonlinear interaction between the features, to replace the traditional method of manually setting weights or rules. For example, the model may find that when the “stay time in the target city” (geographic feature) and “core circle card opening ratio” (social feature) are both at a high level, the prediction effect (probability increase) generated is much higher than the simple addition of the effects of the two, achieving an interaction that cannot be achieved by manually setting fixed weights. By training the machine learning model with a large amount of data in the user data set, the model automatically learns the weight influence of different behavior data on the final prediction result, captures different feature weights, and realizes the prediction of the user's card opening probability value in a specific time period in the future, providing a solid data foundation for subsequent intelligent pre-download decisions.

[0049] In step 1203, the target user behavior data is input into the trained machine learning model, the target user card opening probability value is iteratively calculated, and the pre-download instruction is output.

[0050] The feature vector is input into the pre-trained machine learning model as a whole, and the probability value of the target user opening a certain city transportation card in a specific time period (such as 7 days) in the future is calculated. When the opening probability value for a certain target city exceeds a first preset threshold (for example, 80%), the pre-download scheduler of the cloud server generates a pre-download instruction, otherwise the target user behavior data acquisition is returned to continue.

[0051] Step 13, sending the pre-download instruction to the user terminal, so that the user terminal downloads the traffic card basic data package according to the pre-download instruction, and completes the card opening activation according to the traffic card basic data package.

[0052] In one embodiment, sending the pre-download instruction to the user terminal, so that the user terminal downloads the traffic card basic data package according to the pre-download instruction, and completes the card opening activation according to the traffic card basic data package, comprises step 1301, and the specific steps are as follows: Step 1301, sending the pre-download instruction to the user terminal by the cloud server through the push service.

[0053] In this way, by dividing the traffic card opening activation into two parts of cloud server and user terminal for processing, a large amount of user historical behavior data is stored in the cloud server, the prediction model is continuously updated and improved, centralized and efficient operation of data is realized, the user terminal is used for data receiving and final activation operation of the traffic card, and a collaborative processing architecture of cloud intelligence and terminal agility is constructed.

[0054] Referring to Figure 2 An embodiment of the present application provides a traffic card activation method, which comprises steps 21 to 23, and the specific steps are as follows: Step 21, uploading target user behavior data to a server, so that the server iteratively calculates a card opening probability value of a target user in a first preset time period in the future according to the target user behavior data and a pre-trained machine learning model, and outputs a pre-download instruction when the card opening probability value is greater than a first preset threshold; wherein the target user behavior data is updated each time.

[0055] In one embodiment, uploading target user behavior data to a server, so that the server iteratively calculates a card opening probability value of a target user in a first preset time period in the future according to the target user behavior data and a pre-trained machine learning model, and outputs a pre-download instruction when the card opening probability value is greater than a first preset threshold, comprises steps 2101 to 2102, and the specific steps are as follows: Step 2101, obtaining user behavior data.

[0056] Among them, the mobile terminal (such as a mobile phone) of the user continuously collects anonymous user behavior data through its data collection module after obtaining the authorization of the user. These data include: geographic positioning data, application usage data, device connection data, and social association data. In this way, multi-dimensional user behavior information is obtained from multi-source data such as the user's geography, social relationships, and software usage habits, which is used as the input of the machine learning model to provide basic data for predicting the user's card opening probability. Through the geographic positioning data of the user terminal, the city information of the target user can be collected, and it can be determined whether the user enters a new city or frequently activities in a city. By obtaining the records in the map software of the user terminal, the user's travel habit information can be obtained. By obtaining the user's social association data, the social relationship network can be used to significantly improve the prediction accuracy.

[0057] Specifically, the geographic positioning data includes GPS and base station information, which is used to determine whether the user enters a new city or frequently activities in a city. The quantified features include, for example, "the average daily stay time in the target city". The application usage data includes the search and navigation records of bus and subway lines in the map App. The quantified features include, for example, "the average number of map navigation per day within the geographic fence in the target city". The device connection data includes the Wi-Fi SSID (such as the dedicated Wi-Fi of the subway and airport) connected by the mobile phone. The social association data includes the transportation card information of the user's family members or frequently used contacts obtained through secure encryption.

[0058] Step 2102, upload the obtained user data to the cloud server.

[0059] Specifically, in the process of obtaining social association data, the user terminal performs irreversible encryption and hash processing on the contact identifier locally, and only uploads the anonymous hash value to the cloud server for secure matching with the card opening user library. On this basis, the cloud server quantifies features such as "the number of associated transportation cards in the target city" and "the proportion of card opening in the family group", which are used as the input of the machine learning model. In this way, under the premise of ensuring user privacy and security, the data collection, transmission and processing scheme can utilize the strong social relationship network of the user to significantly improve the prediction accuracy.

[0060] Step 22, receiving the pre-download instruction output by the server, and downloading the transportation card basic data package based on the pre-download instruction.

[0061] Further, the range of the gate induction area is sensed. When it is sensed that the gate induction area is entered and it is detected that there is a local transportation card basic data package, the server is applied for a target user personal key. The target user personal key is injected into the transportation card basic data package to complete the card opening activation.

[0062] Further, according to the pre-download instruction, an intelligent scheduler is enabled; based on the intelligent scheduler and a first preset condition, an optimal download strategy is determined; and according to the optimal download strategy, a traffic card basic data packet is downloaded; wherein the traffic card basic data includes an application file structure, interface resources and a rate table.

[0063] Further, the first preset condition includes an open card probability value in the pre-download instruction, a current network type, a device power and a load state; the device power and the load state are judged; when the device power is greater than or equal to a second preset threshold value and the load state is not a high load state, downloading is performed according to a second preset condition; if the device power is less than the second preset threshold value, the task is temporarily suspended and downloading is performed when the device power is higher than the second preset threshold value; if the load state is in a high load state, the task is temporarily suspended and downloading is performed when the load state enters an idle state; wherein the second preset condition is specifically: if the current network type is a Wi-Fi environment, downloading is performed; if the current network type is a cellular network, if the open card probability value in the pre-download instruction is greater than a third preset threshold value, downloading is performed, otherwise, the task enters a waiting queue and downloading is performed when the current network type is a Wi-Fi environment.

[0064] In one embodiment, the pre-download instruction output by the server is received, and based on the pre-download instruction, a traffic card basic data packet is downloaded, including steps 2201 to 2202, each of which is as follows: Step 2201, according to the pre-download instruction, starting an intelligent scheduler to determine an optimal download strategy.

[0065] Wherein, the intelligent scheduler makes a comprehensive decision according to the current network type, signal strength, device power, load state and open card probability value in the pre-download instruction to determine the optimal download strategy. In this way, through the trigger and control mechanism of the cloud server prediction and user end execution cooperation, the pre-download scheduling strategy is obtained which fully considers the open card probability threshold, network and device state, and fully balances the current device state and user card opening demand, and realizes the optimal and reasonable download of resources. By separating the user end data acquisition and cloud server data processing, only downloading according to the server instruction, the memory occupation of the user end local is reduced in the data processing and pre-download judgment process.

[0066] The core principle of the optimal download strategy is: first, the device power and the load state are judged; if the device power is lower than a second preset threshold value (such as 30%) or in a high load state, the task is temporarily suspended and downloading is performed when the device is charged or in an idle state.

[0067] If the device power and load state are in a healthy state, the current network type and card opening probability value of the device are judged: when connecting Wi-Fi, it is considered as the highest priority channel, and the basic data package is immediately downloaded in the background in silence; when using a cellular network, a secondary judgment is made, if the predicted card opening probability is greater than a third preset threshold (such as 90%), the download is immediately initiated, or downloaded after the user's consent is obtained; if the probability is lower than the threshold, the task enters the waiting queue, and the download is delayed to the Wi-Fi environment.

[0068] Step 2202, performing download according to the download strategy.

[0069] After the download strategy is confirmed, the traffic card basic data package of the target city is downloaded in silence from the content distribution network (CDN), which contains application file structure, interface resources, fare table, etc., but does not contain personalized keys bound to user identity, and the data package is stored in the reserved space of the user end security area (SE / eSE). In this way, by pre-downloading the traffic card basic data package, the system resource utilization is optimized, the data download process is dispersed from the peak period of opening to the idle period of ordinary time, the server load is smoothed, network congestion is avoided, and the overall stability and efficiency of the system are improved Step 23, obtaining the target user personal key, and performing card opening activation according to the target user personal key and the traffic card basic data package.

[0070] In one embodiment, obtaining the target user personal key and performing card opening activation according to the target user personal key and the traffic card basic data package includes step 2301, and the specific steps are as follows: Step 2301, judging the user end scene, and applying the key for activation.

[0071] When the user end approaches the scene such as the gate through the user end App interface or through multi-source sensors, the traffic card opening behavior is formally triggered, the card management application detects the locally downloaded traffic card basic data package, if it is detected that there is a pre-downloaded data package locally, the user personal key is applied to the key management system (KMS) of the cloud server and the pre-downloaded basic data package is injected, and the final activation of the traffic card is completed.

[0072] The determination of the user terminal approaching the gate and the like is performed by comprehensively judging whether the user terminal enters the gate sensing area defined by the low-power Bluetooth beacon, and can be combined with the user's card swiping posture recognized by the motion sensor or the precise ranging by the UWB to achieve the determination. In this way, the user's approach to the gate is determined by multi-sensor fusion, and the precise scenario-based fast triggering of the traffic card activation is realized. By dividing the traffic card data packet into the pre-downloaded traffic card basic data packet and the real-time application of the personal key to the cloud, when the user actually uses it, only two lightweight steps of key application and injection are required to realize the fast activation of the traffic card, which reduces the operation threshold of the traffic card activation, greatly shortens the user's card opening time, and improves the user experience.

[0073] Referring to Figure 3 Another embodiment of the present application also provides a traffic card activation device suitable for a server, comprising a data acquisition module 301, a probability prediction module 302 and an instruction sending module 303.

[0074] The data acquisition module 301 is configured to acquire target user behavior data.

[0075] The probability prediction module 302 is configured to iteratively calculate a card opening probability value of a target user in a first preset time period in the future according to the target user behavior data and a pre-trained machine learning model, and output a pre-download instruction when the card opening probability value is greater than a first preset threshold; wherein the target user behavior data is updated each time the iteration is performed; and the pre-trained machine learning model is trained according to historical user behavior data.

[0076] The instruction sending module 303 is configured to send the pre-download instruction to a user terminal, so that the user terminal downloads a traffic card basic data packet according to the pre-download instruction and completes the card opening activation according to the traffic card basic data packet.

[0077] In this way, by inputting the target user behavior data into the machine learning model, the user's card opening probability is obtained, the user's state is actively acquired, and the traffic card basic data packet is pre-downloaded, which can avoid the download peak period, smooth the system load, avoid network congestion, and realize the fast calling of the basic data packet in the subsequent card opening step. By training the user card opening probability calculation model, the user's card opening behavior is accurately predicted, and a basis for determining whether to pre-download the traffic card basic data packet is provided.

[0078] Referring to Figure 4 Another embodiment of the present application also provides a traffic card activation device suitable for a user terminal, comprising a data uploading module 401, a data downloading module 402 and a card opening activation module 403.

[0079] The data uploading module 401 is configured to upload target user behavior data to a server, so that the server iteratively calculates a card opening probability value of a target user in a future first preset time period according to the target user behavior data and a pre-trained machine learning model, and outputs a pre-download instruction when the card opening probability value is greater than a first preset threshold; wherein the target user behavior data is updated each time.

[0080] The data downloading module 402 is configured to receive the pre-download instruction output by the server, and download a traffic card basic data package based on the pre-download instruction.

[0081] The card opening activation module 403 is configured to obtain a target user personal key, and perform card opening activation according to the target user personal key and the traffic card basic data package.

[0082] In this way, by uploading the data obtained by the user end to the cloud end for processing and downloading through the server instruction, the memory occupation of the data processing and pre-download judgment in the local of the user end is reduced. By dividing the traffic card data download into a basic data package that can be pre-downloaded and a personal key that needs to be obtained in real time, when the user actually initiates card opening, only the key needs to be downloaded and the pre-downloaded basic data package needs to be activated, so that the traffic card can be quickly activated, the card opening time of the user is greatly shortened, the operation cost of the user is reduced, and the user experience is improved.

[0083] Another embodiment of the application also provides a traffic card activation system suitable for a server, comprising a user end and a server; wherein the user end is configured to execute a traffic card activation method suitable for the user end; and the server is configured to execute a traffic card activation method suitable for the server.

[0084] It can be understood that the system embodiments described above correspond to the method embodiments of the application, and can implement the traffic card activation method provided by any one of the method embodiments of the application.

[0085] It should be noted that the system embodiments described above are only schematic, and part or all of the modules thereof can be selected to achieve the purpose of the embodiment. In addition, in the system embodiments provided by the application, the connection relationship between the modules indicates that there is a communication connection therebetween, which can be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement it without creative labor.

[0086] The above describes the preferred embodiments of the application. It should be noted that for those skilled in the art, without departing from the principles of the application, a number of improvements and refinements can be made, which are also considered within the protection scope of the application.

Claims

1. A method for activating a transportation card, characterized in that, For servers, the activation method includes: Obtain target user behavior data; Based on the target user behavior data and the pre-trained machine learning model, the probability of the target user opening a card within a first preset time period is iteratively calculated until the probability of opening a card is greater than a first preset threshold, at which point a pre-download instruction is output; wherein, the target user behavior data is updated in each iteration; wherein, the pre-trained machine learning model is trained based on historical user behavior data; The pre-download instruction is sent to the user terminal so that the user terminal can download the basic data package of the transportation card according to the pre-download instruction and complete the card activation according to the basic data package of the transportation card.

2. The method for activating a transportation card according to claim 1, characterized in that, The pre-trained machine learning model is trained based on historical user behavior data and includes: Based on the historical user behavior data, establish several observation days; Based on the historical user behavior data and the several observation days, sample feature vectors and sample user labels are calculated until all observation days are calculated. The sample feature vectors and sample user labels for each observation day are then output as the model training dataset. Each observation day corresponds to several sample users. The sample feature vectors are quantified based on the historical user behavior data of each sample user within a second preset time period before the corresponding observation day. The sample user labels are obtained based on whether each sample user has activated a transportation card within a third preset time period after the corresponding observation day. The pre-trained machine learning model is trained using the model training dataset.

3. The method for activating a transportation card according to claim 1, characterized in that, The step of iteratively calculating the probability of a target user opening a membership card within a first preset time period based on the target user behavior data and a pre-trained machine learning model is as follows: The target user behavior data is quantified, and the features of each data source of the target user are extracted. The features of each data source of the target user are concatenated into a unified high-dimensional feature vector; The unified high-dimensional feature vector is input into the pre-trained machine learning model to obtain the probability value of the target user opening a card in the first preset time period in the future.

4. A method for activating a transportation card, characterized in that, For user terminals, the activation method includes: The system uploads target user behavior data to the server, so that the server iteratively calculates the probability value of the target user opening a card in the future within a first preset time period based on the target user behavior data and a pre-trained machine learning model, until the card opening probability value is greater than a first preset threshold, at which point a pre-download instruction is output; wherein, the target user behavior data is updated in each iteration; Receive the pre-download instruction output by the server, and download the basic data package of the transportation card based on the pre-download instruction; Obtain the target user's personal key, and activate the card based on the target user's personal key and the basic data packet of the transportation card.

5. The method for activating a transportation card according to claim 4, characterized in that, The step of obtaining the target user's personal key and activating the card based on the target user's personal key and the basic data packet of the transportation card specifically involves: The system senses the range of the gate's sensing area. When it senses that someone has entered the gate's sensing area and detects that there is already a basic data packet of the local transportation card, it requests the target user's personal key from the server. The target user's personal key is injected into the basic data packet of the transportation card to complete the card activation.

6. The method for activating a transportation card according to claim 4, characterized in that, The process of downloading the basic data package for the transportation card based on the pre-download instruction includes: The intelligent scheduler is activated according to the pre-download instruction; Based on the intelligent scheduler and the first preset conditions, the optimal download strategy is determined; According to the optimal download strategy, download the basic data package of the transportation card; wherein, the basic data package of the transportation card includes the application file structure, interface resources and rate table.

7. The method for activating a transportation card according to claim 6, characterized in that, The step of determining the optimal download strategy based on the intelligent scheduler and the first preset conditions specifically involves: The first preset condition includes the card activation probability value, current network type, device battery level, and load status in the pre-download instruction; The system determines the device's battery level and load status. When the device's battery level is greater than or equal to a second preset threshold and the load status is not high, the download is performed according to the second preset condition. If the device's battery level is less than the second preset threshold, the task is paused and the download is performed when the device's battery level is higher than the second preset threshold. If the load status is high, the task is paused and the download is performed when the load status enters an idle state. Specifically, the second preset condition is as follows: if the current network type is a Wi-Fi environment, download is performed; if the current network type is a cellular network, and if the SIM card activation probability value in the pre-download instruction is greater than a third preset threshold, download is performed; otherwise, the task enters a waiting queue and is postponed until the current network type is a Wi-Fi environment to perform the download.

8. An activation device for a transportation card, characterized in that, Applicable to servers, the activation device includes: a data acquisition module, a probability prediction module, and an instruction sending module; The data acquisition module is used to acquire target user behavior data; The probability prediction module iteratively calculates the probability of a target user opening a card within a first preset time period in the future, based on the target user behavior data and a pre-trained machine learning model, until the card opening probability value is greater than a first preset threshold, at which point a pre-download instruction is output; wherein, the target user behavior data is updated during each iteration; wherein, the pre-trained machine learning model is trained based on historical user behavior data; The instruction sending module is used to send the pre-download instruction to the user terminal, so that the user terminal can download the basic data package of the transportation card according to the pre-download instruction and complete the card activation according to the basic data package of the transportation card.

9. An activation device for a transportation card, characterized in that, For user terminals, the activation device includes: a data upload module, a data download module, and a card activation module; The data upload module is used to upload target user behavior data to the server, so that the server can iteratively calculate the probability value of the target user opening a card in the future within a first preset time period based on the target user behavior data and a pre-trained machine learning model, until the card opening probability value is greater than a first preset threshold, and then output a pre-download instruction; wherein, the target user behavior data is updated in each iteration; The data download module is used to receive the pre-download instruction output by the server and download the basic data package of the transportation card based on the pre-download instruction; The card activation module is used to obtain the target user's personal key and activate the card based on the target user's personal key and the basic data packet of the transportation card.

10. A transportation card activation system, characterized in that, It includes a user terminal and a server; wherein the user terminal is used to execute the transportation card activation method as described in any one of claims 1 to 3; and the server is used to execute the transportation card activation method as described in any one of claims 4 to 7.

Images