Driver type determination method and device, data processing method

Abstract

The specification provides a driver type determination method and device, and a data processing method. Based on the above method, before implementation, a preset training set containing labeled positive sample data and labeled negative sample data can be constructed by processing initial sample data using a preset scoring model. Then, a preset classification model is trained using the preset training set. During implementation, the key factor parameters of a target driver to be predicted can be obtained. Then, the driver type of the target driver is determined by processing the key factor parameters of the target driver using the preset classification model. Thus, the initial sample data without accurate labeling can be fully utilized to train a preset classification model with high precision and good effect. Then, the driver type of the target driver can be accurately determined based on the key factor parameters of the target driver using the preset classification model, effectively reducing the classification error.

Description

Technical Field

[0001] This manual belongs to the field of intelligent mobility technology, and in particular relates to methods and devices for determining driver types and data processing methods. Background Technology

[0002] With the development and popularization of online ride-hailing services, more and more drivers are joining these services. These drivers can be further divided into full-time drivers and part-time drivers.

[0003] However, based on existing methods, online ride-hailing service platforms often find it difficult to accurately identify which drivers are full-time and which are part-time.

[0004] There is currently no effective solution to the above problems. Summary of the Invention

[0005] This specification provides a method and apparatus for determining driver type, as well as a data processing method. By utilizing a preset classification model based on the driver's key factor parameters, the driver type can be accurately determined, effectively reducing classification errors.

[0006] This specification provides a method for determining driver type, including: obtaining key factor parameters of a target driver; processing the key factor parameters of the target driver using a preset classification model to obtain a corresponding target classification result; wherein the preset classification model includes a model trained using a preset training set; the preset training set is constructed using a preset scoring model; and determining the driver type of the target driver based on the target classification result; wherein the driver type includes: full-time driver and part-time driver.

[0007] This specification also provides a data processing method, comprising: responding to an operation by a target driver by sending an order acceptance request to a server; wherein, the server responds to the order acceptance request by filtering matching passenger orders based on the target driver's driver type; the target driver's driver type is determined based on key factor parameters of the target driver; and displaying matching passenger orders to the target driver.

[0008] This specification also provides a data processing method, comprising: acquiring multiple initial sample data; wherein the initial sample data includes key factor parameters of sample drivers; determining a rating result of the initial sample data regarding the professionalism of the corresponding sample drivers through weighted calculation based on a preset rating model and the key factor parameters of the sample drivers; wherein the preset rating model includes weight values ​​of the key factor parameters; selecting positive sample data from the initial sample data for labeling based on the rating result, obtaining labeled positive sample data; selecting negative sample data from the initial sample data for labeling, obtaining labeled negative sample data; combining the labeled positive sample data and the labeled negative sample data to obtain a preset training set; and training the initial classification model using the preset training set to obtain a preset classification model.

[0009] This specification also provides a driver type determination device, comprising: an acquisition module for acquiring key factor parameters of a target driver; a processing module for processing the key factor parameters of the target driver using a preset classification model to obtain a corresponding target classification result; wherein the preset classification model includes a model trained using a preset training set; the preset training set is constructed using a preset scoring model; and a determination module for determining the driver type of the target driver based on the target classification result; wherein the driver type includes: full-time driver and part-time driver.

[0010] This specification also provides a server, including a processor and a memory for storing processor-executable instructions, wherein the processor executes the instructions to implement the relevant steps of the above-described method.

[0011] This specification also provides a computer storage medium storing computer instructions that, when executed by a processor, implement the relevant steps of the above-described method.

[0012] Based on the driver type determination method, apparatus, and data processing method provided in this specification, before implementation, a preset scoring model can be used to process the initial sample data to construct a preset training set containing labeled positive and negative sample data. Then, the preset training set is used to train the initial classification model to obtain a preset classification model that meets the requirements. During implementation, key factor parameters of the target driver can be obtained first. Then, the key factor parameters of the target driver are processed using the preset classification model to determine the driver type. This allows for full utilization of the initially unlabeled sample data to train a preset classification model with high accuracy and good performance. Furthermore, this preset classification model can be used to accurately determine the driver type of the target driver based on the key factor parameters, effectively reducing errors during classification and recognition. Attached Figure Description

[0013] To more clearly illustrate the embodiments of this specification, the accompanying drawings used in the embodiments will be briefly introduced below. The drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a flowchart illustrating a method for determining driver type provided in one embodiment of this specification;

[0015] Figure 2 This is a schematic diagram illustrating one embodiment of the driver type determination method provided in the embodiments of this specification, applied in a scenario example.

[0016] Figure 3 This is a schematic diagram illustrating one embodiment of the driver type determination method provided in the embodiments of this specification, applied in a scenario example.

[0017] Figure 4 This is a flowchart illustrating a data processing method provided in one embodiment of this specification;

[0018] Figure 5 This is a flowchart illustrating a data processing method provided in one embodiment of this specification;

[0019] Figure 6 This is a schematic diagram of the structural composition of a server provided in one embodiment of this specification;

[0020] Figure 7 This is a schematic diagram of the structural composition of a driver type determination device provided in one embodiment of this specification. Detailed Implementation

[0021] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0022] See Figure 1 As shown in the embodiments of this specification, a method for determining driver type is provided. In specific implementation, this method may include the following:

[0023] S101: Obtain key factor parameters for the target driver;

[0024] S102: Process the key factor parameters of the target driver using a preset classification model to obtain the corresponding target classification result; wherein, the preset classification model includes a model trained using a preset training set; the preset training set is constructed using a preset scoring model;

[0025] S103: Based on the target classification results, determine the driver type of the target driver; wherein, the driver type includes: full-time driver and part-time driver.

[0026] In some embodiments, the method for determining the driver type described above can be specifically applied to the server side.

[0027] For details, please refer to Figure 2 As shown, the server may specifically include a backend server applied to one side of a ride-hailing service platform (e.g., an online ride-hailing service platform), capable of data transmission, data processing, and other functions. Specifically, the server may be, for example, an electronic device with data processing, storage, and network interaction capabilities. Alternatively, the server may be a software program running on the electronic device, providing support for data processing, storage, and network interaction. In this embodiment, the number of servers is not specifically limited. The server may be a single server, several servers, or a server cluster formed by several servers. The aforementioned online ride-hailing service may specifically include: ride-hailing services, carpooling services, etc.

[0028] In this embodiment, the server can be connected to the database of the ride-hailing service platform. Specifically, the database can store user information of drivers registered on the ride-hailing service platform, as well as drivers' historical order records and other data.

[0029] It should be noted that all user information related to drivers and passengers mentioned in this manual was obtained and used with the user's knowledge and consent. Furthermore, the acquisition, storage, use, and processing of the aforementioned user information in this manual comply with relevant national laws and regulations.

[0030] In some embodiments, the target driver may specifically be understood as a driver whose driver type is to be determined.

[0031] In some specific embodiments, see [reference needed]. Figure 2 As shown, when a driver is ready to drive, they can generate an order acceptance request through their terminal device and send it to the server. This order acceptance request must at least include the driver's user identifier, such as their name, ID number, or username.

[0032] Specifically, the aforementioned terminal device may include a front-end applied to the driver's side, capable of data collection, data transmission, and other functions. Specifically, the terminal device may be, for example, an electronic device such as a tablet computer, laptop computer, smartphone, or car speaker. Alternatively, the terminal device may also be a software application that can run on the aforementioned electronic device. For example, it could be a ride-hailing app running on a smartphone.

[0033] After receiving an order request, the server can query the driver's user information in the database based on the user identifier to determine the driver's type. Then, based on the driver's type, it can filter passenger orders that match the driver from the passenger order pool and push them to the terminal device for the driver to choose from.

[0034] Specifically, the types of drivers mentioned above can include two types: full-time drivers and part-time drivers (e.g., rideshare drivers).

[0035] Specifically, the aforementioned full-time drivers can be understood as those whose primary job is driving, who dedicate relatively more time to it (e.g., accepting orders at all times), are more accepting of passenger orders, and have higher income expectations for their driving career; they are relatively professional drivers. The aforementioned part-time drivers can be understood as those whose job is part-time, who dedicate relatively less time to it (e.g., only accepting orders from passengers going in the same direction), are less accepting of passenger orders (e.g., only accepting orders from passengers going in the same direction), and have lower income expectations for their driving career; they are relatively less professional drivers.

[0036] If the server queries the database and determines that there is no driver type tag for the driver stored in the database, or that the driver type tag for the driver stored in the database has exceeded its validity period, the server can identify the driver as the target driver and trigger the driver type determination method provided in the embodiments of this specification to determine the driver's driver type.

[0037] In some embodiments, the aforementioned key factor parameters can be understood as indicative data that can directly or indirectly reflect the driver's work behavior characteristics within a preset time period. Specifically, the preset time period can be the most recent week, or the most recent month, etc.

[0038] In some embodiments, see Figure 3 As shown, the key factor parameters may specifically include at least one of the following: order acceptance capability parameter, order acceptance preference parameter, route tolerance parameter, and profitability parameter, etc.

[0039] Of course, it should be noted that the key factors listed above are only illustrative. In actual implementation, depending on the specific application scenario and processing requirements, the key factors may include other types of parameters besides those listed above, such as driver-defined parameters (including: driver-defined order acceptance time, driver-defined passenger types for which the driver is willing to accept orders, driver-defined areas for which the driver is willing to accept orders, etc.). This manual does not limit this.

[0040] In some embodiments, see Figure 3 As shown, the order-taking capacity parameters may specifically include: monthly contribution parameters, daily average workload parameters, and order-taking time saturation parameters. The monthly contribution parameters may further include: monthly order-taking contribution parameters (e.g., monthly order volume) and monthly order completion contribution parameters (e.g., monthly order volume). The daily average workload parameters may further include: daily average order completion volume (e.g., the ratio of monthly order completion volume to the number of days with completed orders), daily average order-taking volume (e.g., the ratio of monthly order-taking volume to the number of days with completed orders), and daily average order completion mileage (e.g., the ratio of monthly order completion mileage to the number of days with completed orders). The order-taking time saturation parameters may further include: monthly attendance rate (e.g., the ratio of the number of days with completed orders to the total number of days in the month), the percentage of departure times with appointments (the ratio of appointment departure times to the total daily time), and the percentage of order volume taken during working hours (e.g., the ratio of order volume taken during working hours to the daily order volume). The working hours specifically refer to 9:00 AM to 5:00 PM on weekdays. Conversely, non-working hours can specifically include: the entire day on non-working days (e.g., weekends, holidays, etc.), and 0:00 to 9:00 and 17:00 to 24:00 on working days.

[0041] The order acceptance preference parameters may specifically include: ride-sharing parameters, order channel preference parameters, and instant order acceptance parameters. Ride-sharing parameters may further include: ride-sharing volume (e.g., the mileage of completed ride-sharing orders) and ride-sharing order percentage (e.g., the ratio of daily received ride-sharing orders to daily order volume). Order channel preference parameters may further include: order volume from neighboring areas (e.g., order volume originating from surrounding areas) and order volume from across areas (e.g., order volume originating from across city districts). Instant order acceptance parameters may further include: instant order volume (e.g., the number of orders with a time difference of less than 10 minutes between departure and acceptance time) and instant order percentage (e.g., the ratio of daily instant order volume to daily order volume).

[0042] The route tolerance parameter may specifically include: route condition parameters and detour condition parameters, etc. The route condition parameter may further include: average route compatibility of monthly orders (e.g., average route compatibility of monthly orders), average route compatibility of monthly completed orders (e.g., average route compatibility of monthly completed orders), 1 / 4 route compatibility of monthly orders (e.g., the route compatibility value of passenger orders in the first 1 / 4 of monthly orders sorted from low to high), 1 / 4 route compatibility of monthly completed orders, etc. (e.g., the route compatibility value of passenger orders in the first 1 / 4 of monthly completed orders sorted from low to high). The detour condition parameter may further include: average detour mileage within a region (e.g., the average distance between the end location and the passenger's starting location for passenger orders picked up by the driver within the region during the month), average detour mileage across regions (e.g., the average distance between the end location and the passenger's starting location for passenger orders picked up by the driver across regions during the month), etc. Specifically, the aforementioned route compatibility can be understood as the degree of matching between the passenger's order-indicated travel route and the driver's original planned driving route.

[0043] The profitability parameters may specifically include: revenue parameters, reward parameters, etc. The revenue parameters may further include: monthly revenue (e.g., the sum of driver commissions, platform subsidies, and passenger tips for the month), monthly profitability rate (e.g., the ratio of monthly revenue to monthly completed mileage), etc. The reward parameters may specifically include: platform rewards (e.g., additional rewards from the platform to drivers), passenger tips (e.g., additional rewards from passengers to drivers), etc.

[0044] In some embodiments, the server can obtain key factor parameters of the target driver by querying the driver's user information stored in the database, as well as the driver's historical order records, based on the target driver's user identifier.

[0045] In some embodiments, the aforementioned preset classification model can be understood as a model capable of predicting the probability that a driver is a professional driver based on key driver factor parameters. Specifically, the preset classification model can be a model trained in advance using a preset training set. The preset training set can be understood as a model training set with good performance and strong targeting, constructed in advance by processing initial sample data using a preset scoring model.

[0046] In some embodiments, the initial sample data mentioned above can be understood as key factor parameters of sample drivers obtained directly. The initial sample data includes at least the key factor parameters of the sample drivers.

[0047] It should be noted that, on the one hand, the amount of initial sample data that can be directly collected in this field is often relatively small, requiring full and effective utilization of this limited initial sample data. On the other hand, most of the collected initial sample data does not come with driver type labels, and it is usually difficult to accurately determine whether the driver corresponding to the sample data is a full-time or part-time driver. Based on existing methods, most of the time it is necessary to rely on technicians to manually determine whether the driver is a full-time or part-time driver and to manually label it. This not only increases processing costs, but the labeling results are also prone to errors due to the subjective influence of technicians, resulting in less than ideal training effects of the sample data used for model training.

[0048] In view of the above problems, this manual introduces and utilizes a preset scoring model to process the initial sample data, so as to automatically and efficiently construct a preset training set that is effective, targeted, and suitable for training a classification model to determine driver type.

[0049] In some embodiments, in specific implementations, a preset training set can be obtained by processing the initial sample data using a preset scoring model in the following manner:

[0050] S1: Obtain multiple initial sample data; wherein, the initial sample data includes key factor parameters of the sample drivers;

[0051] S2: Based on the preset scoring model and the key factor parameters of the sample drivers, the initial sample data is used to determine the scoring result of the professionalism of the corresponding sample drivers through weighted calculation; wherein, the preset scoring model includes the weight values ​​of the key factor parameters.

[0052] S3: Based on the scoring results, select positive sample data from the initial sample data for labeling to obtain labeled positive sample data; select negative sample data from the initial sample data for labeling to obtain labeled negative sample data.

[0053] S4: Combine the labeled positive sample data and the labeled negative sample data to obtain the preset training set.

[0054] In some embodiments, the aforementioned preset scoring model can be a model constructed based on the Analytic Hierarchy Process (AHP) that can quantitatively evaluate the importance of key factor parameters in determining whether a driver is a professional driver.

[0055] In some embodiments, the preset scoring model can be constructed in the following manner:

[0056] S1: Determine the relative importance of key factor parameters;

[0057] S2: Establish a judgment matrix based on the relative importance of key factor parameters;

[0058] S3: Perform a consistency check on the judgment matrix based on the judgment logic;

[0059] S4: If the consistency check of the judgment matrix based on the judgment logic is passed, the weight values ​​of the key factor parameters are obtained by solving the normalized eigenvector corresponding to the largest eigenvalue of the judgment matrix.

[0060] S5: Construct a preset scoring model based on the weight values ​​of the key factor parameters.

[0061] In some embodiments, the aforementioned preset scoring model may specifically include weight values ​​for each key factor parameter. The weight values ​​of the key factors are determined based on their relative importance.

[0062] In some embodiments, determining the relative importance of key factor parameters as described above may include: collecting historical sample data; wherein the historical sample data package contains key factor parameters of historical sample users who have been identified as full-time or part-time drivers; and performing cluster analysis on the historical sample data to determine the relative importance of key factor parameters for driver type determination.

[0063] In practice, based on the cluster analysis results, the importance of each key factor parameter individually for driver type judgment can be determined. Furthermore, by combining the importance of each key factor parameter individually for driver type judgment with the cluster analysis results, the relative importance of any two key factor parameter combinations for driver type judgment can be determined.

[0064] Specifically, for example, if based on historical sample data, key factor parameter A is determined to be extremely important relative to key factor parameter B, then the quantified relative importance of A relative to B can be set to 9; conversely, the quantified relative importance of B relative to A can be set to 1 / 9. As another example, if based on historical sample data, key factor parameter A is determined to be strongly important relative to key factor parameter B, then the quantified relative importance of A relative to B can be set to 7; conversely, the quantified relative importance of B relative to A can be set to 1 / 7. As yet another example, if based on historical sample data, key factor parameter A is determined to be slightly important relative to key factor parameter B, then the quantified relative importance of A relative to B can be set to 3; conversely, the quantified relative importance of B relative to A can be set to 1 / 3. Finally, if based on historical sample data, key factor parameter A is determined to be equally important relative to key factor parameter B, then the quantified relative importance of A relative to B can be set to 1; conversely, the quantified relative importance of B relative to A can be set to 1.

[0065] In some embodiments, each key factor parameter includes at least two layers of sub-parameters. For example, profitability parameters include: revenue parameters, reward parameters, etc. (first-layer sub-parameters). The revenue parameter may further include: monthly revenue, monthly profit margin (second-layer sub-parameters). Therefore, in practical implementation, a hierarchical analysis algorithm can be used to analyze each layer based on the relative importance of the key factor parameters to determine the weight values ​​corresponding to each layer of sub-parameters; subsequently, a corresponding judgment matrix can be constructed.

[0066] Specifically, for example, order acceptance capability parameters, order acceptance preference parameters, route convenience tolerance parameters, and profitability parameters can be used as the first layer. Based on the relative importance of these parameters, their weight values ​​are determined. Taking profitability parameters as another example, revenue parameters and reward parameters can be used as the second layer. Based on the weight values ​​of profitability parameters, the relative importance of revenue and reward parameters is determined to obtain their weight values ​​in the second layer. Further, taking revenue parameters as another example, monthly revenue and monthly profit margin can be used as the third layer. Based on the weight values ​​of revenue parameters and the relative importance of monthly revenue and monthly profit margin, their weight values ​​in the third layer are obtained.

[0067] In some embodiments, after performing a consistency check based on judgment logic on the judgment matrix, the method may further include the following: adjusting the relative importance of key factor parameters to re-establish the judgment matrix. Then, based on the re-established judgment matrix, a new consistency check based on logical judgment can be performed.

[0068] In some embodiments, when implementing the project, the weight values ​​of the key factor parameters contained in the preset scoring model and the key factor parameters of the sample drivers can be used to obtain the initial sample data score on the professionalism of the sample drivers through weighted budgeting, which is then used as the scoring result of the initial sample data.

[0069] In some embodiments, the above-mentioned selection of positive sample data from the initial sample data based on the scoring results for labeling, resulting in labeled positive sample data; and the selection of negative sample data from the initial sample data for labeling, resulting in labeled negative sample data, may specifically include the following:

[0070] S1: Based on the scoring results, select the initial sample data whose scoring results are greater than the preset first score threshold from the initial sample data as positive sample data; and label the positive sample data to obtain labeled positive sample data;

[0071] S2: Based on the scoring results, select initial sample data whose scoring results are less than the preset second score threshold from the initial sample data as negative sample data; and label the negative sample data to obtain labeled negative sample data.

[0072] In practice, negative sample data can be selected from the initial sample data whose scores are less than a preset second score threshold and greater than a preset third score threshold.

[0073] The specific values ​​of the aforementioned preset first score threshold, preset second score threshold, and preset third score threshold can be determined by learning from historical sample data.

[0074] Specifically, for example, the aforementioned preset first score threshold can be 90. The aforementioned preset second score threshold can be 30. The aforementioned preset third score threshold can be 20. Of course, the preset first, second, and third score thresholds listed above are only illustrative. In actual implementation, other suitable values ​​can be flexibly set as the preset first, second, and third score thresholds according to specific circumstances and accuracy requirements.

[0075] The initial sample data also includes some non-positive samples. This portion of the initial sample data lacks distinct features and specificity, and because it is unlabeled, it's impossible to accurately determine the true driver type corresponding to this portion of the initial sample data. Therefore, to guide the model to focus on learning features that significantly influence driver type identification during subsequent model training, and to avoid using inaccurately labeled sample data to train the model and affect its accuracy, this portion of the initial sample data is not used as training data for subsequent model training.

[0076] Based on the above embodiments, positive samples belonging to full-time drivers with more obvious features and stronger targeting can be effectively and precisely screened from the initial sample data based on a preset first score threshold; at the same time, negative samples belonging to part-time drivers with more obvious features and stronger targeting can be effectively screened from the initial sample data based on a preset second score threshold.

[0077] Furthermore, in specific labeling, positive labels can be set for positive samples to represent positive samples, and negative labels can be set for negative samples to represent negative samples.

[0078] Then, the initial sample data that cannot accurately determine the driver type, has unclear features, or is not targeted based on the preset scoring model is filtered out. Only the labeled positive samples and labeled negative samples are combined. In this way, the initial sample data can be effectively used to construct a preset training set that is more targeted and has better results.

[0079] In some embodiments, the preset classification model can be trained using a preset training set in the following manner:

[0080] S1: Construct the initial classification model;

[0081] S2: Train the initial classification model using a preset training set to obtain the preset classification model.

[0082] In some embodiments, the initial classification model includes at least one of the following: a model based on the LogisticRegression structure, a model based on the GBDT structure, or a model based on the XGBoost structure.

[0083] In this embodiment, to avoid poor model accuracy and errors during use due to over-reliance on features corresponding to a single key factor parameter during model training, it is necessary to train not only the single-dimensional features corresponding to a single key factor parameter individually, but also the multi-dimensional features corresponding to multiple key factors in association. Based on the above considerations, a model based on the Logistic Regression structure was chosen as the initial classification model.

[0084] Among them, the Logistic Regression structure performs well in training correlations between multi-dimensional features. Using a model based on the Logistic Regression structure as the initial classification model for training not only accurately learns the features reflected by the four key factor parameters—order acceptance ability, order preference, route tolerance, and profitability—but also comprehensively and effectively trains the correlations between the various features reflected by these four key factor parameters. This avoids the model over-relying on one or two features, resulting in higher accuracy, better generalization ability, and robustness.

[0085] Of course, in specific implementation, depending on the specific circumstances and needs, models based on structures such as GBDT and XGBoost can also be used as the initial classification model.

[0086] In some embodiments, the target classification result includes a probability value that the driver is a full-time driver. Accordingly, determining the driver type of the target driver based on the target classification result can, in specific implementation, include: determining the target driver's driver type as a full-time driver if the probability value is greater than or equal to a preset probability threshold; and determining the target driver's driver type as a part-time driver if the probability value is less than the preset probability threshold. The preset probability threshold can be specifically set to 50%, etc.

[0087] In some embodiments, after determining the driver type of the target driver based on the target classification results, the method may further include the following:

[0088] S1: Determine the target order push strategy that matches the target driver based on the driver type of the target driver;

[0089] S2: Push passenger orders to target drivers according to the target order push strategy.

[0090] In this embodiment, prior to implementation, preset order push strategies can be configured and assigned to different types of drivers, such as full-time and part-time drivers, based on their specific needs. These preset order push strategies are then stored in the platform database. Each preset order push strategy also carries a driver type tag corresponding to its respective driver type.

[0091] In practice, the server can query the preset order push strategy carrying the corresponding driver type tag based on the determined driver type of the target driver, and use it as the target order push strategy that matches the target driver.

[0092] In some embodiments, when the target driver is a professional driver, passenger orders are pushed to the target driver according to the target order push strategy. Specifically, this may include: according to the target order push strategy, selecting passenger orders from the passenger order pool that have professional requirements, and / or have a route affinity of less than a preset route affinity threshold, and / or have a travel time during working hours, and / or have an income greater than a preset income threshold, and / or are package orders, as matching passenger orders; and pushing the matching passenger orders to the target driver.

[0093] In some embodiments, when the target driver is a part-time driver, passenger orders are pushed to the target driver according to a target order push strategy. Specifically, this may include: filtering passenger orders from the passenger order pool that do not require professional skills, and / or have a route affinity greater than or equal to a preset route affinity threshold, and / or have travel times outside of working hours, as matching passenger orders; and pushing these matching passenger orders to the target driver. Furthermore, relevant coupons may be pushed to part-time drivers on non-working days (e.g., holidays, weekends, etc.) to encourage them to take more orders.

[0094] This allows for the differentiation of driver types among target drivers, enabling the targeted delivery of matching passenger orders. This can increase the order acceptance rate for target drivers and shorten the waiting time for passengers to accept orders. Furthermore, it makes it easier for target drivers to receive passenger orders that meet their needs and expectations, thus improving their user experience.

[0095] In some embodiments, after determining the driver type of the target driver based on the target classification result, the method may further include the following: pushing business services that match the driver type of the target driver to the target driver based on the driver type of the target driver.

[0096] Specifically, for example, a smart filtering service for ride-sharing orders can be pushed to part-time drivers. Another example is a priority push service for full-time drivers, prioritizing high-paying, high-requirement orders.

[0097] This allows for the precise differentiation of different types of drivers, and based on the actual needs and expectations of these drivers, more targeted and differentiated business services can be designed and provided, further enhancing the driver's interactive experience.

[0098] As can be seen from the above, based on the driver type determination method provided in the embodiments of this specification, before specific implementation, a preset scoring model can be used to process the initial sample data to construct a preset training set containing labeled positive sample data and labeled negative sample data; then, the preset training set is used to train the initial classification model to obtain the preset classification model; in specific implementation, the key factor parameters of the target driver can be obtained first; then, the key factor parameters of the target driver can be processed by the preset classification model to determine the driver type of the target driver. Thus, the initial sample data without accurate labeling can be fully utilized to train a preset classification model with high accuracy and good performance; subsequently, the preset classification model can be used to accurately determine the driver type of the target driver based on the key factor parameters of the target driver, effectively reducing classification and recognition errors.

[0099] This specification also provides another method for determining driver type, which may include the following steps in its implementation:

[0100] S1: Obtain key factor parameters for the target driver;

[0101] S2: Based on the preset scoring model and the key factor parameters of the target driver, the scoring result of the target driver is determined through weighted calculation; wherein, the preset scoring model includes the weight values ​​of the key factor parameters;

[0102] S3: Based on the scoring results, if the target driver's score is greater than the preset first score threshold, the target driver can be identified as a full-time driver; if the target driver's score is less than the preset second score threshold, the target driver can be identified as a part-time driver; if the target driver's score is greater than or equal to the preset second score threshold and less than or equal to the preset first score threshold, the key factor parameters of the target driver are processed using a preset classification model to determine the driver type of the target driver.

[0103] Through the above embodiments, preset scoring models and preset classification models can be used in combination to determine the driver type of the target driver more efficiently and accurately.

[0104] See Figure 4 As shown in the embodiments of this specification, a data processing method is also provided. Specifically, this method can be applied to a terminal device deployed on the driver's side. In specific implementation, this method may include the following:

[0105] S401: In response to the target driver's operation, send an order acceptance request to the server; wherein, the server responds to the order acceptance request and filters out matching passenger orders based on the target driver's driver type; the target driver's driver type is determined based on the target driver's key factor parameters;

[0106] S402: Show the target driver a matching passenger order.

[0107] In some embodiments, the driver type of the target driver can be determined by the server calling a preset classification model based on the key factor parameters of the target driver. The preset classification model can be trained in advance using a preset training set.

[0108] In some embodiments, the preset training set may specifically include a training set constructed using a preset scoring model. The preset scoring model may specifically include a scoring model constructed based on the Analytic Hierarchy Process (AHP).

[0109] See Figure 5 As shown in the embodiments of this specification, a data processing method is also provided. In specific implementation, this method may include the following:

[0110] S501: Acquire multiple initial sample data; wherein, the initial sample data includes key factor parameters of the sample drivers;

[0111] S502: Based on the preset scoring model and the key factor parameters of the sample drivers, the initial sample data is used to determine the scoring result of the professionalism of the corresponding sample drivers through weighted calculation; wherein, the preset scoring model includes the weight values ​​of the key factor parameters.

[0112] S503: Based on the scoring results, select positive sample data from the initial sample data for labeling to obtain labeled positive sample data; select negative sample data from the initial sample data for labeling to obtain labeled negative sample data.

[0113] S504: Combine the labeled positive sample data and the labeled negative sample data to obtain a preset training set;

[0114] S505: Train the initial classification model using a preset training set to obtain the preset classification model.

[0115] Through the above embodiments, the initial sample data that is not accurately labeled can be fully utilized to construct a preset classification model with high accuracy and good performance.

[0116] This specification also provides a server, including a processor and a memory for storing processor-executable instructions. Specifically, the processor can perform the following steps according to the instructions: acquiring key factor parameters of a target driver; processing the key factor parameters of the target driver using a preset classification model to obtain a corresponding target classification result; wherein the preset classification model includes a model trained using a preset training set; the preset training set is constructed using a preset scoring model; and determining the driver type of the target driver based on the target classification result; wherein the driver type includes: full-time driver and part-time driver.

[0117] To execute the above instructions more accurately, please refer to... Figure 6 As shown in the embodiments of this specification, another specific server is also provided, wherein the server includes a network communication port 601, a processor 602 and a memory 603, and the above structures are connected by internal cables so that the various structures can perform specific data interaction.

[0118] Specifically, the network communication port 601 can be used to obtain key factor parameters of the target driver.

[0119] The processor 602 can specifically be used to process the key factor parameters of the target driver using a preset classification model to obtain the corresponding target classification result; wherein, the preset classification model includes a model trained using a preset training set; the preset training set is constructed using a preset scoring model; and the driver type of the target driver is determined based on the target classification result; wherein, the driver type includes: full-time driver and part-time driver.

[0120] The memory 603 can be used to store the corresponding instruction program.

[0121] In this embodiment, the network communication port 601 can be a virtual port bound to different communication protocols, thereby enabling the sending or receiving of different data. For example, the network communication port can be a port responsible for web data communication, a port responsible for FTP data communication, or a port responsible for email data communication. Furthermore, the network communication port can also be a physical communication interface or communication chip. For example, it can be a wireless mobile network communication chip, such as GSM or CDMA; it can also be a Wi-Fi chip; or it can be a Bluetooth chip.

[0122] In this embodiment, the processor 602 can be implemented in any suitable manner. For example, the processor can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers, etc. This specification is not limiting.

[0123] In this embodiment, the memory 603 may include multiple layers. In a digital system, anything that can store binary data can be a memory. In an integrated circuit, a circuit with storage function but no physical form is also called a memory, such as RAM, FIFO, etc. In a system, a storage device with a physical form is also called a memory, such as a memory stick, TF card, etc.

[0124] This specification also provides a terminal device, including a processor and a memory for storing processor-executable instructions. In a specific implementation, the processor can perform the following steps according to the instructions: responding to the operation of the target driver and sending an order acceptance request to the server; wherein, the server responds to the order acceptance request and filters out matching passenger orders based on the driver type of the target driver; the driver type of the target driver is determined based on the key factor parameters of the target driver; and displays the matching passenger orders to the target driver.

[0125] This specification also provides a computer storage medium based on the above-described method for determining driver type. The computer storage medium stores computer program instructions that, when executed, perform the following: acquiring key factor parameters of the target driver; processing the key factor parameters of the target driver using a preset classification model to obtain a corresponding target classification result; wherein the preset classification model includes a model trained using a preset training set; the preset training set is constructed using a preset scoring model; and determining the driver type of the target driver based on the target classification result; wherein the driver type includes: full-time driver and part-time driver.

[0126] This specification also provides a computer storage medium based on the above data processing method. The computer storage medium stores computer program instructions, which, when executed, perform the following: responding to the operation of the target driver, sending an order acceptance request to the server; wherein, the server responds to the order acceptance request and filters out matching passenger orders based on the target driver's driver type; the target driver's driver type is determined based on the target driver's key factor parameters; and displays the matching passenger orders to the target driver.

[0127] This specification also provides a computer storage medium based on the above data processing method. The computer storage medium stores computer program instructions, which, when executed, perform the following: acquiring multiple initial sample data sets; wherein the initial sample data sets include key factor parameters of sample drivers; determining a score for the professionalism of the corresponding sample drivers based on a weighted calculation using a preset scoring model and the key factor parameters of the sample drivers; wherein the preset scoring model includes weight values ​​for the key factor parameters; selecting positive sample data from the initial sample data and labeling it based on the scoring results to obtain labeled positive sample data; selecting negative sample data from the initial sample data and labeling it to obtain labeled negative sample data; combining the labeled positive sample data and the labeled negative sample data to obtain a preset training set; and training the initial classification model using the preset training set to obtain a preset classification model.

[0128] In this embodiment, the storage medium includes, but is not limited to, Random Access Memory (RAM), Read-Only Memory (ROM), cache, hard disk drive (HDD), or memory card. The memory can be used to store computer program instructions. The network communication unit can be an interface configured according to standards specified in the communication protocol for network connection communication.

[0129] In this embodiment, the specific functions and effects implemented by the program instructions stored in the computer storage medium can be explained in comparison with other implementation methods, and will not be repeated here.

[0130] See Figure 7 As shown, at the software level, this specification also provides a driver type determination device, which may specifically include the following structural modules:

[0131] The acquisition module 701 can be used to acquire key factor parameters of the target driver.

[0132] The processing module 702 is specifically used to process the key factor parameters of the target driver using a preset classification model to obtain the corresponding target classification result; wherein, the preset classification model includes a model trained using a preset training set; the preset training set is constructed using a preset scoring model;

[0133] The determination module 703 can be used to determine the driver type of the target driver based on the target classification results; wherein, the driver type includes: full-time driver and part-time driver.

[0134] In some embodiments, the key factor parameters may specifically include at least one of the following: order acceptance capability parameter, order acceptance preference parameter, route tolerance parameter, and profitability parameter, etc.

[0135] In some embodiments, the apparatus further includes a construction module. Accordingly, when the construction module is specifically implemented, it can construct a preset training set in the following manner: acquiring multiple initial sample data; wherein the initial sample data includes key factor parameters of sample drivers; determining the score result of the initial sample data regarding the professionalism of the corresponding sample drivers through weighted calculation based on a preset scoring model and the key factor parameters of the sample drivers; wherein the preset scoring model includes weight values ​​of the key factor parameters; selecting positive sample data from the initial sample data for labeling based on the scoring results, obtaining labeled positive sample data; selecting negative sample data from the initial sample data for labeling, obtaining labeled negative sample data; combining the labeled positive sample data and the labeled negative sample data to obtain the preset training set.

[0136] In some embodiments, when the construction module is specifically implemented, a preset scoring model can be constructed in the following manner: determining the relative importance of key factor parameters; establishing a judgment matrix based on the relative importance of key factor parameters; performing a consistency check on the judgment matrix based on judgment logic; if the consistency check of the judgment matrix based on judgment logic is passed, obtaining the weight values ​​of key factor parameters by solving the normalized eigenvector corresponding to the largest eigenvalue of the judgment matrix; and constructing a preset scoring model based on the weight values ​​of the key factor parameters.

[0137] In some embodiments, when the construction module is specifically implemented, it can select positive sample data from the initial sample data and label them according to the scoring results to obtain labeled positive sample data; and select negative sample data from the initial sample data and label them to obtain labeled negative sample data: according to the scoring results, select initial sample data whose scoring results are greater than a preset first score threshold as positive sample data; and label the positive sample data to obtain labeled positive sample data; according to the scoring results, select initial sample data whose scoring results are less than a preset second score threshold as negative sample data; and label the negative sample data to obtain labeled negative sample data.

[0138] In some embodiments, the apparatus may further include a training module. Accordingly, when the training module is specifically implemented, it may train a preset classification model in the following manner: construct an initial classification model; train the initial classification model using a preset training set to obtain the preset classification model.

[0139] In some embodiments, the initial classification model may specifically include a model based on the Logistic Regression structure, etc.

[0140] In some embodiments, the device may further include a push module, which may be used to determine a target order push strategy that matches the target driver based on the driver type of the target driver; and to push passenger orders to the target driver according to the target order push strategy.

[0141] In some embodiments, when the target driver is a professional driver, the above-mentioned push module can push passenger orders to the target driver according to the target order push strategy in the following manner: according to the target order push strategy, select passenger orders from the passenger order pool that have professional requirements, and / or have a route affinity of less than a preset route affinity threshold, and / or have a ride time that falls within working hours, and / or have an income greater than a preset income threshold, and / or are package orders, as matching passenger orders; and push the matching passenger orders to the target driver.

[0142] In some embodiments, when the target driver is a part-time driver, the above-mentioned push module can push passenger orders to the target driver according to the target order push strategy in the following manner: according to the target order push strategy, select passenger orders from the passenger order pool that have no professional requirements, and / or have a route affinity greater than or equal to a preset route affinity threshold, and / or have a ride time outside of working hours, as matching passenger orders; and push the matching passenger orders to the target driver.

[0143] This embodiment also provides a data processing apparatus, which may specifically include the following structural modules:

[0144] The sending module can be used to respond to the operation of the target driver and send an order acceptance request to the server; wherein, the server responds to the order acceptance request and filters out matching passenger orders based on the driver type of the target driver; the driver type of the target driver is determined based on the key factor parameters of the target driver;

[0145] The display module can be used to show matching passenger orders to target drivers.

[0146] It should be noted that the units, devices, or modules described in the above embodiments can be implemented by computer chips or physical entities, or by products with certain functions. For ease of description, the above devices are described by dividing them into various modules according to their functions. Of course, in implementing this specification, the functions of each module can be implemented in one or more software and / or hardware, or the module that implements the same function can be implemented by a combination of multiple sub-modules or sub-units, etc. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection between the devices or units shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0147] As can be seen from the above, the driver type determination device provided in the embodiments of this specification can make full use of the initial sample data that is not accurately labeled to train a preset classification model with high accuracy and good performance; then, the preset classification model can be used to accurately determine the driver type of the target driver based on the key factor parameters of the target driver, effectively reducing the classification and recognition error.

[0148] While this specification provides the steps of operation for the methods described in the embodiments or flowcharts, more or fewer steps may be included based on conventional or non-inventive means. The order of steps listed in the embodiments is merely one possible order of execution among many steps and does not represent the only possible order. In actual device or client product execution, the methods shown in the embodiments or drawings may be executed sequentially or in parallel (e.g., in a parallel processor or multi-threaded processing environment, or even a distributed data processing environment). The terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, product, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, product, or apparatus. Without further limitations, the presence of other identical or equivalent elements in a process, method, product, or apparatus that includes said elements is not excluded. The terms "first," "second," etc., are used to denote names and do not indicate any particular order.

[0149] Those skilled in the art will also know that, besides implementing the controller using purely computer-readable program code, the same functions can be achieved by logically programming the method steps, making the controller function as logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers (PLCs), and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the devices within it used to implement various functions can also be considered structures within that hardware component. Alternatively, the devices used to implement various functions can be considered as both software modules implementing the method and structures within a hardware component.

[0150] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, classes, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0151] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this specification can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solutions of this specification can essentially be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, mobile terminal, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments of this specification.

[0152] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on its differences from other embodiments. This specification can be used in numerous general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable electronic devices, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices, etc.

[0153] Although this specification has been described by way of examples, those skilled in the art will recognize that many variations and modifications are possible without departing from the spirit of this specification, and it is intended that the appended claims cover such variations and modifications without departing from the spirit of this specification.

Claims

1. A method of determining a driver type, characterized by, The method comprises the following steps: obtaining key factor parameters of a target driver; the key factor parameters comprise at least one of the following: order acceptance capability parameters, order acceptance preference parameters, route tolerance parameters, and profitability parameters; processing the key factor parameters of the target driver by using a preset classification model to obtain a corresponding target classification result; wherein the preset classification model comprises a model trained by using a preset training set; the preset training set is obtained by using a preset scoring model; determining a driver type of the target driver according to the target classification result; wherein the driver type comprises a full-time driver and a part-time driver; wherein the preset training set is obtained in the following manner: obtaining a plurality of initial sample data; wherein the initial sample data comprises key factor parameters of a sample driver, and the initial sample data does not comprise a sample driver type label; determining a score result of the initial sample data on the full-time degree of the corresponding sample driver by using a preset scoring model and the key factor parameters of the sample driver through weighted operation; wherein the preset scoring model comprises weight values of the key factor parameters; selecting positive sample data from the initial sample data according to the score result to obtain labeled positive sample data; selecting negative sample data from the initial sample data to obtain labeled negative sample data; and filtering out initial sample data with unobvious features and weak pertinence; combining the labeled positive sample data and the labeled negative sample data to obtain the preset training set; determining a target order pushing strategy matched with the target driver according to the driver type of the target driver; in the case that the driver type of the target driver is a full-time driver, selecting, according to the target order pushing strategy, a passenger order with professional requirements, and / or a passenger order with a route tolerance less than a preset route tolerance threshold, and / or a passenger order with a ride time belonging to a working time period, and / or a passenger order with an income greater than a preset income threshold, and / or a packed order, as a matched passenger order from a passenger order pool; and pushing the matched passenger order to the target driver.

2. The method of claim 1, wherein, The preset scoring model is obtained in the following manner: determining the relative importance degree of the key factor parameters; establishing a judgment matrix according to the relative importance degree of the key factor parameters; performing a consistency check based on judgment logic on the judgment matrix; in the case that the consistency check based on judgment logic on the judgment matrix is passed, obtaining the weight values of the key factor parameters by solving a normalized eigenvector corresponding to a maximum eigenvalue of the judgment matrix; constructing the preset scoring model according to the weight values of the key factor parameters.

3. The method of claim 1, wherein, The method for labeling the initial sample data comprises the following steps: selecting positive sample data from the initial sample data according to the score result to obtain labeled positive sample data; and selecting negative sample data from the initial sample data to obtain labeled negative sample data. According to the score result, the initial sample data with a score result greater than a preset first score threshold is selected as positive sample data from the initial sample data; and the positive sample data is labeled to obtain labeled positive sample data. According to the score result, the initial sample data with a score result less than a preset second score threshold is selected as negative sample data from the initial sample data; and the negative sample data is labeled to obtain labeled negative sample data.

4. The method of claim 1, wherein, The method further comprises: In a case where the driver type of the target driver is a part-time driver, according to a target order pushing strategy, a passenger order without a professional requirement, and / or a passenger order with a ride-along degree greater than or equal to a preset ride-along degree threshold, and / or a passenger order with a ride time belonging to a non-working time period is selected as a matched passenger order from a passenger order pool; and the matched passenger order is pushed to the target driver.

5. A data processing method, characterized by, Comprise: In response to an operation of the target driver, a server is sent an order accepting request; wherein the server responds to the order accepting request, and according to the driver type of the target driver, a matched passenger order is selected; the driver type of the target driver is determined by a preset classification model according to key factor parameters of the target driver; the preset classification model comprises a model trained by a preset training set; the preset training set is constructed by a preset scoring model; the key factor parameters comprise at least one of the following: order accepting ability parameters, order accepting preference parameters, ride-along tolerance parameters, and profitability parameters; The target driver is displayed with the matched passenger order; The preset training set is constructed in the following manner: a plurality of initial sample data is obtained; wherein the initial sample data comprises key factor parameters of a sample driver, and the initial sample data comprises sample data without a driver type label; according to a preset scoring model and the key factor parameters of the sample driver, a score result of the initial sample data on the full-time degree of the corresponding sample driver is determined by weighted operation; wherein the preset scoring model contains weight values of the key factor parameters; according to the score result, positive sample data is selected from the initial sample data for labeling to obtain labeled positive sample data; negative sample data is selected from the initial sample data for labeling to obtain labeled negative sample data; and initial sample data with unobvious features and weak pertinence is filtered out; the labeled positive sample data and the labeled negative sample data are combined to obtain the preset training set; the server is further configured to determine a target order pushing strategy matched with the target driver according to the driver type of the target driver; in a case where the driver type of the target driver is a full-time driver, a passenger order with a professional requirement, and / or a passenger order with a ride-along degree less than a preset ride-along degree threshold, and / or a passenger order with a ride time belonging to a working time period, and / or a passenger order with an income greater than a preset income threshold, and / or a packaged order is selected as a matched passenger order from a passenger order pool; and the matched passenger order is pushed to the target driver.

6. A data processing method, characterized by, Comprise: Obtaining a plurality of initial sample data; wherein, the initial sample data includes key factor parameters of sample drivers, the initial sample data includes sample data without driver type labels; the key factor parameters include at least one of the following: order receiving ability parameters, order receiving preference parameters, on-the-way tolerance parameters and profitability parameters; According to a preset scoring model and the key factor parameters of the sample driver, the scoring result of the initial sample data about the full-time degree of the corresponding sample driver is determined by weighted operation; wherein, the preset scoring model contains the weight value of the key factor parameters; According to the scoring result, the positive sample data is selected from the initial sample data for labeling to obtain labeled positive sample data; the negative sample data is selected from the initial sample data for labeling to obtain labeled negative sample data; and the initial sample data with unobvious features and weak pertinence is filtered out; Combining the labeled positive sample data and the labeled negative sample data, a preset training set is obtained; Using the preset training set to train an initial classification model to obtain a preset classification model; Wherein, the preset classification model is used to determine the driver type of the target driver by processing the key factor parameters of the target driver, so as to determine the target order pushing strategy matched with the target driver according to the driver type of the target driver; and in the case that the driver type of the target driver is a full-time driver, the target order pushing strategy is used to select the passenger order from the passenger order pool, which has professional requirements, and / or, the on-the-way degree is less than a preset on-the-way degree threshold, and / or, the ride time belongs to a working time period, and / or, the income is greater than a preset income threshold, and / or, the order is a packed order, as a matched passenger order; and the matched passenger order is pushed to the target driver.

7. A driver type determination apparatus characterized by comprising: Comprise: An acquisition module is configured to acquire key factor parameters of a target driver; the key factor parameters include at least one of the following: order receiving ability parameters, order receiving preference parameters, on-the-way tolerance parameters and profitability parameters; A processing module is configured to process the key factor parameters of the target driver by using a preset classification model to obtain a corresponding target classification result; wherein, the preset classification model includes a model trained by using a preset training set; the preset training set is obtained by using a preset scoring model; A determination module is configured to determine a driver type of the target driver according to the target classification result; wherein, the driver type includes a full-time driver and a part-time driver; The preset training set is constructed in the following manner: a plurality of initial sample data are obtained; the initial sample data include key factor parameters of sample drivers, and the initial sample data include sample data without driver type labels; a score result of the full-time degree of the corresponding sample driver of the initial sample data is determined through weighted operation according to a preset scoring model and the key factor parameters of the sample driver; the preset scoring model includes weight values of the key factor parameters; positive sample data are selected from the initial sample data according to the score result, and the positive sample data are labeled to obtain labeled positive sample data; negative sample data are selected from the initial sample data, and the negative sample data are labeled to obtain labeled negative sample data; and initial sample data with unobvious features and weak pertinence are filtered out; the labeled positive sample data and the labeled negative sample data are combined to obtain the preset training set. The device is further configured to: determine a target order pushing strategy matched with the target driver according to the driver type of the target driver; in a case where the driver type of the target driver is a full-time driver, select, according to the target order pushing strategy, a passenger order having professional requirements, and / or a pick-up degree less than a preset pick-up degree threshold, and / or a pick-up time belonging to a working time period, and / or an income greater than a preset income threshold, and / or a packed order, from a passenger order pool as a matched passenger order; and push the matched passenger order to the target driver.

8. A server, characterized by A computer program product including a processor and a memory storing processor-executable instructions that, when executed by the processor, implement the steps of the method of any one of claims 1-4 or 6.

9. A computer storage medium, characterized in that A computer program product including a processor and a memory storing processor-executable instructions that, when executed by the processor, implement the steps of the method of any one of claims 1-6.

Images