Order cancellation probability prediction method, electronic equipment and readable storage medium

Abstract

The invention provides an order cancellation probability prediction method, electronic equipment and a readable storage medium, and belongs to the technical field of data processing. According to the method, through obtaining a first feature information set, a second feature information set and a third feature information set, multi-dimensional features related to a target goods source order, a first driver and an interaction relationship thereof can be used as input of a target prediction model; and the target prediction model outputs a cancellation probability that the first driver cancels the target goods source order. According to the method and the device, the historical behaviors of the target cargo owner to which the target cargo source order belongs, the attributes of the target cargo source and the configuration requirements of the target cargo source order are considered, and the transport capacity characteristics and the historical behavior characteristics of the driver and the cross characteristics between the driver and the target cargo owner as well as between the target cargo source and the target cargo source order are introduced; and the target prediction model comprehensively covers diversified factors which cause cancelling of the target goods source order, and accurate prediction of the cancelling probability of cancelling the target goods source order by the driver is realized.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method for predicting the probability of order cancellation, an electronic device, and a readable storage medium. Background Technology

[0002] In freight matching scenarios, matchmaking platforms collect and process information from both supply and demand sides, matching the freight demands posted by shippers with the transportation capacity resources provided by drivers to facilitate the completion of transportation orders. However, in actual operation, it is common for drivers to cancel orders after accepting them. Frequent order cancellations not only affect the user experience of shippers but also reduce the overall fulfillment efficiency of the matchmaking platform.

[0003] In related technologies, prediction models based on conversion rates are typically used to estimate the matching degree between drivers and cargo, thereby increasing the likelihood of a transaction during the recommendation process. However, these conversion rate-based prediction models primarily focus on the probability of a transaction and lack targeted prediction and control for subsequent fulfillment and cancellation processes. The main reasons drivers cancel orders after accepting them include: First, the driver's vehicle does not match the cargo's length, model, tonnage, etc., leading to cancellation after acceptance; second, there is a conflict between the cargo's loading / unloading time and the driver's transportation schedule, causing cancellation if the driver cannot arrive on time or wait; third, some drivers and cargo owners bypass the matching platform to conduct private transactions, resulting in online order cancellations; fourth, the cargo's remarks contain special requirements, such as the need for lifting equipment, tailgate vehicles, or truck-mounted assistance, which the accepting driver lacks, leading to cancellation. These diverse reasons can all contribute to a high order cancellation rate, thus affecting the fulfillment rate of the matching platform.

[0004] Therefore, in existing vehicle-cargo matching scenarios, the reasons for order cancellation are complex and diverse, and traditional prediction models based on completion rates are unable to accurately identify the risk of order cancellation, which has become an urgent problem to be solved. Summary of the Invention

[0005] This application provides a method, electronic device, and readable storage medium for predicting the probability of order cancellation, in order to solve the problem that in existing vehicle-cargo matching scenarios, the reasons for order cancellation are complex and diverse, and traditional prediction models based on completion rates are difficult to accurately identify the risk of order cancellation.

[0006] In a first aspect, this application provides a method for predicting the probability of order cancellation, the method comprising: Obtain a first set of feature information related to the target cargo order. The first set of feature information includes first feature information related to the historical behavior features of the target cargo owner to which the target cargo order belongs, second feature information related to the inherent attributes of the target cargo corresponding to the target cargo order, and third feature information related to the configuration of the target cargo order. Obtain a second set of feature information related to the first driver. The second set of feature information includes a fourth set of feature information related to the target vehicle corresponding to the first driver, and a fifth set of feature information related to the historical behavior features of the first driver. Obtain a third set of feature information related to the target cargo order and the first driver. The third set of feature information includes first cross-feature information between the first driver and the target cargo owner, second cross-feature information between the first driver and the target cargo, and third cross-feature information between the first driver and the target cargo order. The first feature information set, the second feature information set, and the third feature information set are processed using a target prediction model to obtain the probability that the first driver will cancel the target cargo order.

[0007] In one possible design, the first feature information includes the target cargo owner's transaction and cancellation information within a preset time period, as well as the target cargo owner's credit score; wherein, the transaction information includes the target cargo owner's number of transaction days and number of transactions; the cancellation information includes the target cargo owner's number of cancellation days and number of cancellations; the preset time period is located before the time when the target cargo order is published; The second feature information includes the type information, shape information, volume information, weight information, and sensitive attribute information of the target cargo, as well as the quality score of the target cargo; The third feature information includes the transportation route and associated conditions of the target cargo.

[0008] In one possible design, the fourth feature information includes the target vehicle's length information, vehicle type information, origin information, destination information, and load information; The fifth feature information includes the number of times the first driver is exposed, clicked, completed, canceled, and placed orders without fulfilling them within the preset time period; wherein, the number of exposures is the number of times the target platform shows the first cargo order to the first driver; the number of clicks is the number of times the first driver clicks to view the first cargo order; the number of completed transactions is the number of times the first driver accepts the first cargo order; the number of cancellations is the number of times the first driver accepts but does not fulfill the first cargo order; and the number of placed orders without fulfilling them is the number of times the first driver completes a transaction on the target platform but does not fulfill the first cargo order through the target platform; wherein, the first cargo order is any cargo order on the target platform.

[0009] In one possible design, the first cross-feature information includes transaction information and cancellation information between the first driver and the target cargo owner within the preset time period; The second cross-feature information includes the number of times the first driver exposes the second cargo order, the number of times it is clicked, the number of transactions completed, the number of cancellations, and the number of times the first driver cancels the order within the preset time period; wherein, the cargo corresponding to the second cargo order is the target cargo. The third cross-feature information includes the number of cargo orders completed by the first driver along the transportation route within the preset time period.

[0010] In one possible design, prior to acquiring the first feature information related to the target source order, the method further includes: A training sample set is constructed, comprising multiple training samples. Each training sample includes a set of historical feature information and label information corresponding to the set of historical feature information. The set of historical feature information includes a fourth set of feature information related to historical freight orders, a fifth set of feature information related to the second driver corresponding to the historical freight orders, and a sixth set of feature information related to both the historical freight orders and the second driver. The label information is used to indicate whether the second driver cancels the historical freight orders. The initial prediction model is trained using the training sample set to obtain the target prediction model.

[0011] In one possible design, the target prediction model is a tree model or a deep neural network model.

[0012] In one possible design, after processing the first feature information set, the second feature information set, and the third feature information set using a target prediction model to obtain the cancellation probability of the first driver canceling the target freight order, the method further includes: Determine whether the target cargo meets the tiered exposure conditions, wherein the tiered exposure conditions are used to determine whether the tiered exposure is effective in reducing the probability of the first driver canceling the order for the target cargo; If the target cargo meets the tiered exposure conditions, based on the cancellation probability of each first driver canceling the target cargo order, the multiple first drivers are respectively divided into high cancellation rate drivers, medium cancellation rate drivers and low cancellation rate drivers; For drivers with high cancellation rates, the exposure of the target cargo order is delayed by a first duration; for drivers with medium cancellation rates, the exposure of the target cargo order is delayed by a second duration; and for drivers with low cancellation rates, the exposure of the target cargo order is direct; wherein the first duration is longer than the second duration.

[0013] In one possible design, the step of classifying the multiple first drivers into high-cancellation-rate drivers, medium-cancellation-rate drivers, and low-cancellation-rate drivers based on the cancellation probability of each first driver canceling the target cargo order includes: Based on the number of the plurality of first drivers and a preset ratio, a target number is determined, wherein the target number is the total number of drivers with high cancellation rates and drivers with medium cancellation rates among the plurality of first drivers; A first cancellation probability threshold is determined based on the sum of the base cancellation probability and a first preset value; wherein, the base cancellation probability is the median of the cancellation probabilities of the plurality of first drivers when they are arranged in descending order. A second cancellation probability threshold is determined based on the sum of the basic cancellation probability and the second preset value; wherein the second preset value is less than the first preset value. If the probability of the first driver canceling the target cargo order is greater than the first cancellation probability threshold, and the number of drivers with high cancellation rates is less than the target number, the first driver is identified as the driver with high cancellation rates. If the first driver's probability of canceling the target cargo order is greater than the second cancellation probability threshold, the first driver's probability of canceling the target cargo order is less than or equal to the first cancellation probability, and the number of drivers with high cancellation rates and drivers with medium cancellation rates is less than the target number, then the first driver is identified as the driver with medium cancellation rates. The first driver, excluding the high cancellation rate driver and the medium cancellation rate driver, is identified as the low cancellation rate driver.

[0014] Secondly, this application provides an apparatus for predicting the probability of order cancellation, the apparatus comprising: The first acquisition module is used to acquire a first set of feature information related to the target cargo order. The first set of feature information includes first feature information related to the historical behavior features of the target cargo owner to which the target cargo order belongs, second feature information related to the inherent attributes of the target cargo corresponding to the target cargo order, and third feature information related to the configuration of the target cargo order. The second acquisition module is used to acquire a second set of feature information related to the first driver. The second set of feature information includes a fourth set of feature information related to the target vehicle corresponding to the first driver, and a fifth set of feature information related to the historical behavior features of the first driver. The third acquisition module is used to acquire a third set of feature information related to the target cargo order and the first driver. The third set of feature information includes the first cross-feature information between the first driver and the target cargo owner, the second cross-feature information between the first driver and the target cargo, and the third cross-feature information between the first driver and the target cargo order. The processing module is used to process the first feature information set, the second feature information set, and the third feature information set using a target prediction model to obtain the cancellation probability of the first driver canceling the target cargo order.

[0015] Thirdly, this application provides an electronic device, including: a memory and at least one processor; The memory stores computer-executed instructions; The at least one processor executes computer execution instructions stored in the memory, causing the at least one processor to perform the method described in the first aspect or various possible designs of the first aspect.

[0016] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed, implement the method described in the first aspect or various possible designs of the first aspect.

[0017] Fifthly, this application provides a computer program product, which includes computer program code that, when run on a computer, causes the computer to implement the method described in the first aspect or various possible designs of the first aspect.

[0018] In a sixth aspect, this application provides a chip, comprising: an interface circuit and a logic circuit, wherein the interface circuit is configured to receive signals from other chips outside the chip and transmit them to the logic circuit, or to send signals from the logic circuit to other chips outside the chip, and the logic circuit is configured to implement the method described in the first aspect or various possible designs of the first aspect.

[0019] This application provides a method, electronic device, and readable storage medium for predicting order cancellation probability. The method acquires a first set of feature information, a second set of feature information, and a third set of feature information to simultaneously obtain multi-dimensional features related to the target cargo order and the first driver, as well as the interaction between the target cargo order and the first driver. These three sets of feature information are then used as input to a target prediction model to output the probability of the first driver canceling the target cargo order. Compared to existing methods that only predict based on the completion rate, the method provided in this application considers not only the historical behavior of the target cargo owner to which the target cargo order belongs, the attributes of the target cargo itself, and the configuration requirements of the target cargo order, but also the driver's capacity characteristics, historical behavior characteristics, and cross-features between the driver and the target cargo owner, the target cargo, and the target cargo order. This allows the target prediction model to comprehensively cover the diverse factors leading to the cancellation of target cargo orders at the feature level. Therefore, this method can accurately predict the probability of a driver canceling a target cargo order, thus solving the problem that traditional prediction models based on completion rates cannot accurately identify order cancellation risks. Attached Figure Description

[0020] Figure 1 A flowchart illustrating a method for predicting order cancellation probability provided in an embodiment of this application; Figure 2 A flowchart illustrating another method for predicting order cancellation probability provided in this application embodiment; Figure 3 A flowchart illustrating another method for predicting order cancellation probability provided in an embodiment of this application; Figure 4 A schematic diagram of the structure of an order cancellation probability prediction device provided in an embodiment of this application; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

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

[0022] 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 this application pertains; the terminology used herein in the specification of the application is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims and drawings of this application are intended to cover non-exclusive inclusion.

[0023] The term "embodiment" as used herein means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of the phrase "embodiment" in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0024] In this article, the term "and / or" simply describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists, A and B can exist simultaneously, and B exists. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0025] Furthermore, the terms "first," "second," etc., in the specification and claims of this application or in the aforementioned drawings are used to distinguish different objects rather than to describe a specific order, and may explicitly or implicitly include one or more of the features.

[0026] In the description of this application, unless otherwise stated, "multiple" and "at least two" mean two or more (including two), and similarly, "multiple groups" and "at least two groups" mean two or more (including two groups).

[0027] In the description of this application, it should be noted that, unless otherwise explicitly specified and limited, the terms "connected" and "linked" should be interpreted broadly. For example, "connected" or "linked" can refer not only to a physical connection, but also to an electrical connection or a signal connection. For instance, it can be a direct connection, i.e., a physical connection, or an indirect connection through at least one intermediate component, as long as the circuit is connected. It can also refer to the internal connection between two components. A signal connection can refer not only to a signal connection through a circuit, but also to a signal connection through a medium, such as radio waves. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0028] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. It should be noted that, unless otherwise specified, different technical features in this application can be combined with each other.

[0029] In freight matching scenarios, matchmaking platforms collect and process information from both supply and demand sides, matching the freight demands posted by shippers with the transportation capacity resources provided by drivers to facilitate the completion of transportation orders. However, in actual operation, it is common for drivers to cancel orders after accepting them. Frequent order cancellations not only affect the user experience of shippers but also reduce the overall fulfillment efficiency of the matchmaking platform.

[0030] In related technologies, order cancellations between drivers and shippers are usually caused by a combination of factors, which can be summarized into the following categories: First, there's the issue of mismatched vehicle parameters. In the recommendation logic of long-haul freight, the matching platform doesn't solely rely on identical vehicle length and model when recommending cargo. Instead, it allows for a certain degree of difference between vehicles and cargo; for example, a large vehicle might transport small-volume cargo, or a small vehicle might transport moderately oversized cargo. In such cases, some drivers may misjudge their vehicle's carrying capacity when accepting an order, only to discover later, after communicating with the shipper, that they cannot complete the transport, ultimately leading to order cancellation.

[0031] Second, the pickup time doesn't match. Orders posted by shippers on the matching platform include both immediate orders requiring drivers to arrive within a specified time and next-day orders with less urgent delivery times. If a driver's empty journey is too long to arrive within the stipulated time, they may cancel an immediate order; if a driver primarily engages in short-distance transportation and prefers quick transactions, they may cancel a next-day order due to the mismatch in timing requirements.

[0032] Third, private transactions. In some cases, drivers who find the prices offered by the matching platform too low after accepting an order may contact the shipper directly through means outside the platform. If the shipper also wants to avoid the service fees charged by the matching platform, they may complete the transaction with the driver, bypassing the platform, which could lead to the cancellation of the online order.

[0033] Fourth, special requirements were not met. When listing cargo, shippers often specify special requirements in the order notes, such as the need for lifting equipment, the need for truck closures, or the requirement for vehicles with tailgates. If the driver accepting the order does not meet these requirements, the order will be cancelled.

[0034] This demonstrates the diversity and complexity of reasons for order cancellations. Related technologies typically employ prediction models based on completion rates to estimate the matching degree between drivers and freight, thereby increasing the likelihood of a successful transaction during the recommendation process. However, these completion rate-based prediction models primarily focus on the probability of a successful transaction and struggle to accurately identify cancellation risks.

[0035] To address the problems existing in related technologies, this application provides a method, electronic device, and readable storage medium for predicting order cancellation probability. In the method for predicting order cancellation probability, by acquiring a first set of feature information, a second set of feature information, and a third set of feature information, multi-dimensional features related to the target cargo order and the first driver, as well as the interaction between the target cargo order and the first driver, can be simultaneously obtained. These three sets of feature information are then used as input to a target prediction model for processing, thereby outputting the probability of the first driver canceling the target cargo order. Compared to existing schemes that only predict based on the completion rate, the method provided in this application not only considers the historical behavior of the target cargo owner to which the target cargo order belongs, the inherent attributes of the target cargo, and the configuration requirements of the target cargo order, but also introduces the driver's capacity characteristics, historical behavior characteristics, and cross-features between the driver and the target cargo owner, the target cargo, and the target cargo order. This allows the target prediction model to comprehensively cover the diverse factors leading to the cancellation of target cargo orders at the feature level. Therefore, this method can accurately predict the probability of drivers canceling orders for target cargo, thus solving the problem that traditional prediction models based on completion rates cannot accurately identify order cancellation risks.

[0036] Next, through some specific embodiments and accompanying drawings, we will describe in detail how this application solves the problem that in the above-mentioned vehicle-cargo matching scenario, the reasons for order cancellation are complex and diverse, and traditional prediction models based on the completion rate are difficult to accurately identify the risk of order cancellation.

[0037] Figure 1 This is a flowchart illustrating a method for predicting order cancellation probability provided in an embodiment of this application. Figure 1As shown, the method for predicting the order cancellation probability provided in this application embodiment specifically includes S101 to S104, and S101 to S104 will be described in detail below.

[0038] It should be noted that the execution subject of the order cancellation probability prediction method provided in this application embodiment can be an electronic device. The electronic device can be a mobile phone, a laptop computer, etc., and this embodiment does not specifically limit it.

[0039] S101. Obtain the first set of feature information related to the target source order.

[0040] The first set of feature information includes first feature information related to the historical behavior characteristics of the target cargo owner to which the target cargo order belongs, second feature information related to the inherent attributes of the target cargo corresponding to the target cargo order, and third feature information related to the configuration of the target cargo order.

[0041] It should be noted that the first set of feature information is used to describe the target source order from multiple dimensions. The first set of feature information includes first feature information, second feature information, and third feature information.

[0042] The first feature information is generated by the target platform based on historical transaction data. It is used to reflect the behavioral characteristics of the target cargo owner within a given time range, thereby reflecting the target cargo owner's transaction habits and stability during modeling.

[0043] The second characteristic information consists of the description, quantity, specifications, transportation requirements, etc. of the target cargo source filled in by the target cargo owner when publishing the target cargo source, which can reflect the basic situation of the target cargo source in terms of physical characteristics, usage conditions or business needs.

[0044] The third characteristic information mainly involves the operational constraints set for the target freight order on the target platform, which reflects the specific requirements for drivers and vehicles during the execution of the target freight order. The target platform is a matching platform.

[0045] The first feature information set does not rely on single-dimensional feature information, and can provide a more comprehensive characterization of the target cargo order, providing input basis for the subsequent target prediction model to calculate the probability prediction of the first driver canceling the target cargo order.

[0046] In one possible embodiment, the first feature information includes the target cargo owner's transaction and cancellation information within a preset time period, as well as the target cargo owner's credit score.

[0047] The preset time period is located before the time when the target source order is published.

[0048] It should be noted that the preset time period refers to the time range used for statistical behavioral data, which is located before the statistical time (the time when the target source order is released).

[0049] The preset time period can be set by the administrators of the target platform, and this embodiment does not impose specific limitations on it.

[0050] For example, the preset time period is the past 7 days, the past 14 days, or the past 30 days.

[0051] It should be noted that the target cargo owner's credit score is a numerical indicator that represents the target platform's comprehensive quantitative assessment of the target cargo owner's credit status based on factors such as the target cargo owner's historical behavior data, transaction performance, and default records.

[0052] The transaction information includes the number of days and the number of transactions for the target cargo owner.

[0053] It should be noted that the transaction information is used to indicate the situation where the freight orders posted by the target cargo owner within a preset time period are accepted by the driver.

[0054] The number of days of a transaction refers to the number of calendar days within a preset time period during which at least one of the cargo orders posted by the target cargo owner is accepted by a driver.

[0055] The number of completed orders refers to the total number of cargo orders posted by the target cargo owner that are accepted by the driver within a preset time period.

[0056] The cancellation information includes the number of days the target cargo owner cancelled and the number of cancelled orders.

[0057] It should be noted that cancellation information is used to indicate situations where a cargo owner's order posted within a preset time period is cancelled after being accepted by a driver.

[0058] Cancellation days refer to the number of calendar days within a preset time period during which at least one of the cargo orders posted by the target cargo owner is cancelled after the driver accepts the order.

[0059] The number of cancelled orders refers to the total number of cargo orders posted by the target cargo owner that were cancelled after being accepted by the driver within a preset time period.

[0060] The second feature information includes the target cargo's type, shape, volume, weight, and sensitive attribute information, as well as the target cargo's quality score.

[0061] It should be noted that the type information of the target goods is used to characterize the category to which the target goods belong, such as steel, building materials, electronic products, food, etc.

[0062] The morphological information of the target cargo is used to characterize the physical state of the target cargo during transportation, such as bulk, full truckload, pallet, or liquid.

[0063] The volume information of the target cargo is used to characterize the size of the target cargo in terms of space, and can be described by the length, width and height of the target cargo.

[0064] The weight information of the target cargo is used to characterize the weight or tonnage of the target cargo.

[0065] Sensitive attribute information of the target cargo is used to characterize whether the target cargo has special attributes, such as fragile, flammable, requiring refrigeration, or dangerous goods.

[0066] The quality score of the target source is a numerical indicator used to characterize the quality of the target source after the target platform conducts a comprehensive quantitative evaluation based on the completeness, standardization, and historical performance of the target source.

[0067] The third characteristic information includes the transportation route and associated conditions of the target cargo.

[0068] It should be noted that the transportation route of the target cargo is used to characterize the geographical path between the origin and destination of the target cargo, and may include the origin city, the destination city, and information on intermediate stops.

[0069] The attached conditions of the target cargo are used to characterize the additional requirements set by the target cargo owner in the target cargo order, such as whether lifting equipment is required, whether truck jacking is required, or whether a vehicle with a tailgate is required.

[0070] In this embodiment, by incorporating the target cargo owner's transaction information, cancellation information, and credit score into the first feature information, and combining the target cargo source's attribute features, quality score, transportation route, and target cargo source order remarks into the second and third feature information, the input data of the target prediction model can be made more comprehensive. It considers both the target cargo owner's historical transaction behavior and the target cargo source's own attributes and target cargo source order configuration requirements, thereby effectively improving the target prediction model's ability to identify and predict the cancellation risk of the first driver canceling the target cargo source order.

[0071] S102, Obtain a set of second feature information related to the first driver.

[0072] The second set of feature information includes fourth feature information related to the target vehicle corresponding to the first driver, and fifth feature information related to the historical behavior features of the first driver.

[0073] It should be noted that the second set of feature information is used to comprehensively describe the first driver's transportation capacity and historical behavioral characteristics. The second set of feature information includes the fourth and fifth feature information.

[0074] The fourth feature information is obtained by the first driver when registering the vehicle or updating the vehicle information on the target platform. It is used to characterize the basic situation of the target vehicle in terms of physical parameters and operating conditions, thereby reflecting the first driver's ability to adapt to different cargo orders.

[0075] The fifth feature information comes from the target platform's historical operation data records of the first driver within a predetermined time range. It is used to characterize the first driver's behavior patterns in the order acceptance, transportation and fulfillment stages, so as to reflect the first driver's transaction habits and fulfillment stability when modeling.

[0076] In one possible embodiment, the fourth feature information includes the target vehicle's length information, vehicle type information, origin information, destination information, and load information.

[0077] It should be noted that the target vehicle's length information is used to represent the numerical or range information of the target vehicle's length. The length information is recorded in meters, for example, 4.2m, 6.8m, 7.6m, 9.6m, 13m, 17.5m, etc.; it can also be discretized according to the target platform's preset gears.

[0078] The vehicle type information of the target vehicle is used to characterize the vehicle body type category, such as van, flatbed, dump, refrigerated, hazardous materials, etc. The vehicle type information can be stored using either an enumerated type or one-hot encoding.

[0079] The origin information of the target vehicle is used to characterize the origin or frequent departure location preference of the target vehicle, and can be represented by administrative division (province / city / district / county) identifier, geographic code or preset area code.

[0080] The destination information of the target vehicle is used to characterize the destination or frequently visited location preference of the target vehicle. The way the destination information of the target vehicle is represented is consistent with the way the departure information of the target vehicle is represented.

[0081] The load information of the target vehicle is used to characterize the load-bearing capacity of the target vehicle. It can be the rated load, and the unit can be tons or kilograms.

[0082] Specifically, the origin and destination information of the target vehicle can be configured by the first driver in the vehicle information on the target platform, or obtained through the statistical results of the first driver's historical completed orders or the route preferences of the first driver recorded by the target platform.

[0083] For example, the fourth characteristic information includes: the target vehicle is 9.6m long, is a van, frequently departs from City A1 in Province A, frequently arrives at City B2 in Province B, and has a load capacity of no more than 10 tons.

[0084] The fifth feature information includes the number of times the first driver is exposed, clicked, completed, canceled, and placed orders within a preset time period.

[0085] The number of exposures refers to the number of times the target platform displays the first source of goods order to the first driver within a preset time period.

[0086] The first source order is any source order on the target platform.

[0087] It should be noted that each time the target platform shows the first source order to the first driver, it is counted as one instance; if the same first source order is shown to the first driver again at different times or through different entry points, it will be counted separately.

[0088] The number of clicks refers to the number of times the first driver clicks to view the first source of goods orders within a preset time period.

[0089] It should be noted that each click by the first driver to "enter the order details page" for the first source of goods is counted as one event; multiple clicks by the first driver for the same first source of goods are counted separately.

[0090] Among them, the number of transactions refers to the number of times the first driver accepts the first source of goods order within a preset time period.

[0091] It should be noted that the "order accepted successfully" event recorded by the target platform shall prevail. If the same first source order is cancelled and then accepted again, it can be counted separately according to the target platform's rules or only the first valid order can be counted.

[0092] The cancellation count refers to the number of times the first driver accepts but fails to fulfill the first source of goods order within a preset time period.

[0093] Among them, the number of "flying orders" refers to the number of times the first driver fails to fulfill the first source of goods order after it has been completed on the target platform within a preset time period.

[0094] For example, within a preset time period of the past 30 days, the first driver's exposure count was 120, click count was 35, transaction count was 12, cancellation count was 3, and order cancellation count was 1. The above counts of exposure count, click count, transaction count, cancellation count, and cancellation count are based on the event logs of the target platform. Multiple displays and multiple clicks of the same first source order are counted separately.

[0095] In this embodiment, by introducing the target vehicle's length, model, origin, destination, and load capacity attributes into the second feature information set, and combining it with the first driver's behavioral statistics such as exposure count, click count, transaction count, cancellation count, and order cancellation count within a preset time period, the input features of the target prediction model on the driver side can be made more comprehensive. This can reflect the adaptability of the driver's vehicle in terms of carrying capacity and route selection, and also reflect the driver's historical performance in platform interaction and fulfillment processes. This improves the target prediction model's ability to distinguish and predict the cancellation risk of different drivers canceling target cargo orders.

[0096] S103. Obtain a set of third feature information related to the target cargo order and the first driver.

[0097] The third set of feature information includes the first cross-feature information between the first driver and the target cargo owner, the second cross-feature information between the first driver and the target cargo source, and the third cross-feature information between the first driver and the target cargo source order.

[0098] It should be noted that the third feature information set is used to characterize the features of the multi-dimensional interaction between the first driver and the target freight order. The third feature information set includes the first cross-feature information, the second cross-feature information, and the third cross-feature information.

[0099] The first cross-feature information is generated by the target platform based on the cooperation records between the first driver and the target cargo owner in historical orders. It is used to reflect the transaction frequency, cooperation stability and historical performance between the first driver and the target cargo owner.

[0100] The second cross-feature information is obtained by the target platform based on the interaction between the first driver and the target cargo source within a preset time period. It is used to reflect the first driver's familiarity with the target cargo source and his historical performance.

[0101] The third cross-feature information involves the matching between the first driver and the target cargo order in terms of transportation routes, time arrangements, or execution conditions, and is used to characterize the first driver's suitability in a specific order scenario.

[0102] By acquiring the third feature information set, the interaction characteristics between the first driver and the target cargo order can be characterized from different perspectives such as cooperative relationship, cargo preference and order execution conditions, providing a more comprehensive feature basis for the input of the cancellation probability prediction model.

[0103] In one possible embodiment, the first cross-feature information includes transaction information and cancellation information between the first driver and the target cargo owner within a preset time period.

[0104] It should be noted that the first cross-feature information is used to characterize the historical cooperation between the first driver and the target cargo owner within a preset time period, and may include at least transaction information and cancellation information.

[0105] The transaction information between the first driver and the target cargo owner within a preset time period is used to represent the situation where the first driver accepts the cargo orders posted by the target cargo owner within the preset time period.

[0106] The number of days during which the first driver and the target cargo owner have a transaction within a preset time period is used to indicate the number of natural days during which the target cargo owner and the first driver have at least one successful order.

[0107] The number of completed orders between the first driver and the target cargo owner within a preset time period indicates the total number of orders successfully accepted between the target cargo owner and the first driver within the preset time period.

[0108] Cancellation information between the first driver and the target cargo owner within a preset time period is used to indicate situations where a cargo order posted by the target cargo owner is accepted by the first driver and then cancelled within the preset time period.

[0109] The number of cancellation days between the first driver and the target cargo owner within a preset time period is used to indicate the number of natural days during which at least one order was cancelled after the target cargo owner and the first driver accepted the order within the preset time period.

[0110] The number of cancellations between the first driver and the target cargo owner within a preset time period indicates the total number of orders cancelled between the target cargo owner and the first driver after acceptance within the preset time period.

[0111] For example, within a preset time period of nearly 30 days, the number of days of transactions between the first driver and the target cargo owner is 3, the number of transactions is 5, the number of days of cancellation is 1, and the number of cancellations is 1.

[0112] The second set of cross-feature information includes the number of times the first driver exposes, clicks, completes, cancels, and skips orders for the second source of goods within a preset time period.

[0113] Among them, the source of goods corresponding to the second source of goods order is the target source of goods.

[0114] It should be noted that the second cross-feature information is used to characterize the first driver's interactive behavior with the corresponding order for the target cargo within a preset time period.

[0115] The number of times the first driver is exposed to the second source of goods order within a preset time period is the number of times the target platform shows the second source of goods order to the first driver within the preset time period.

[0116] The number of times the first driver clicks on the second source of goods order within the preset time period is the number of times the first driver clicks to view the details of the second source of goods order within the preset time period.

[0117] The number of times the first driver completes a second cargo order within a preset time period is the number of times the first driver confirms the acceptance of the second cargo order within the preset time period.

[0118] The number of times the first driver cancels a second source of goods order within a preset time period is the number of times the first driver fails to fulfill a second source of goods order that has been accepted within a preset time period and is recorded by the platform as a cancellation. The number of times the first driver accepts second-source orders within a preset time period is the number of times the first driver accepts second-source orders on the target platform within the preset time period but fails to fulfill the second-source orders through the target platform.

[0119] The third cross-feature information includes the number of cargo orders completed by the first driver along the transportation route within a preset time period.

[0120] The number of cargo orders completed by the first driver along the transportation route within a preset time period refers to the total number of orders completed by the first driver within the preset time period, and whose origin and destination match the transportation route.

[0121] The matching rules can be precise at the city level, or equivalent matching can be determined by regional coding, radius aggregation, etc. in the implementation, and are not limited to a single implementation method.

[0122] Specifically, when there is no record for any interaction among the first, second, and third cross-feature information, the count for that interaction can be recorded as 0.

[0123] In this embodiment of the application, by introducing historical transaction and cancellation data between the first driver and the target cargo owner, interactive data such as the first driver's exposure, clicks, transactions, cancellations, and unauthorized orders for the target cargo source, as well as the number of orders completed by the first driver along the transportation route within a preset time period into the third feature information set, the interaction characteristics of the driver and the target cargo source orders can be comprehensively characterized from three perspectives: cooperative relationship, cargo source preference, and route matching. This makes the input data of the target prediction model more comprehensive, thereby improving the target prediction model's ability to identify cancellation risks between the first driver and the target cargo source orders.

[0124] S104. The first feature information set, the second feature information set, and the third feature information set are processed using the target prediction model to obtain the cancellation probability of the first driver canceling the target cargo order.

[0125] The input to the target prediction model consists of a first set of feature information, a second set of feature information, and a third set of feature information. The output of the target prediction model is a probability value that represents the likelihood of the first driver canceling the target cargo order.

[0126] The probability of a first driver canceling a target cargo order typically falls within the range [0,1]. The closer the probability of a first driver canceling a target cargo order is to 1, the greater the likelihood that the first driver will cancel the target cargo order; the closer the probability of a first driver canceling a target cargo order is to 0, the smaller the likelihood that the first driver will cancel the target cargo order.

[0127] In this embodiment, by acquiring a first set of feature information, a second set of feature information, and a third set of feature information, multi-dimensional features related to the target cargo order and the first driver, as well as the interaction between the target cargo order and the first driver, can be simultaneously obtained. These three sets of feature information are then used as input to a target prediction model to output the probability of the first driver canceling the target cargo order. Compared to existing schemes that only predict based on the completion rate, the method provided in this application not only considers the historical behavior of the target cargo owner to which the target cargo order belongs, the inherent attributes of the target cargo, and the configuration requirements of the target cargo order, but also introduces the driver's capacity characteristics, historical behavior characteristics, and cross-features between the driver and the target cargo owner, the target cargo, and the target cargo order. This allows the target prediction model to comprehensively cover the diverse factors leading to the cancellation of target cargo orders at the feature level. Therefore, this method can achieve accurate prediction of the probability of a driver canceling a target cargo order, thereby solving the problem that traditional prediction models based on completion rates cannot accurately identify order cancellation risks.

[0128] In the above embodiments, the electronic device needs to process the first feature information set, the second feature information set, and the third feature information set using a target prediction model to obtain the probability of the first driver canceling the target freight order. The specific process by which the electronic device obtains the target prediction model will be described in detail below.

[0129] Figure 2 This is a flowchart illustrating another method for predicting order cancellation probability provided in an embodiment of this application. Figure 2 As shown, in one possible embodiment, before the method step shown in S101, the prediction method further includes implementations S201 and S202, which are described in detail below.

[0130] S201. Construct a training sample set.

[0131] The training sample set includes multiple training samples, and each training sample includes a set of historical feature information and the label information corresponding to the set of historical feature information.

[0132] It should be noted that the training sample set is used to support the offline / online training of the initial prediction model. The training sample set contains multiple training samples, each of which consists of a set of historical feature information and the label information corresponding to the set of historical feature information.

[0133] The historical feature information set includes a fourth feature information set related to historical freight orders, a fifth feature information set related to the second driver corresponding to the historical freight orders, and a sixth feature information set related to the historical freight orders and the second driver.

[0134] It should be noted that the feature information included in the fourth feature information set is similar to that included in the first feature information set, and will not be described again in this embodiment.

[0135] The fourth set of feature information includes feature information related to the historical behavior of the cargo owners to which the historical cargo orders belong, feature information related to the inherent attributes of the cargo corresponding to the historical cargo orders, and feature information related to the configuration of the historical cargo orders.

[0136] The feature information included in the fifth feature information set is similar to that included in the second feature information set, and will not be described again in this embodiment.

[0137] The fifth set of feature information includes feature information related to the vehicle of the third driver corresponding to historical freight orders, as well as feature information related to the historical behavioral characteristics of the third driver.

[0138] The feature information included in the sixth feature information set is similar to that included in the third feature information set, and will not be described again in this embodiment.

[0139] The third feature information set includes cross-feature information between the third driver and the cargo owner to which the historical freight orders belong, cross-feature information between the third driver and the cargo corresponding to the historical freight orders, and cross-feature information between the third driver and the historical freight orders.

[0140] The tag information is used to indicate whether the second driver has cancelled a previous cargo order.

[0141] It should be noted that the tag information is used to indicate whether the second driver cancels the historical freight order, and is denoted as a binary variable y∈{0,1}.

[0142] y=1 indicates that the historical order was cancelled after being accepted by the second driver.

[0143] y=0 indicates that the historical order was not cancelled after being accepted by the second driver (e.g., it is in a fulfillment completed or other non-cancelled state).

[0144] S202. Train the initial prediction model using the training sample set to obtain the target prediction model.

[0145] In one possible embodiment, the target prediction model is a tree model or a deep neural network model.

[0146] The initial prediction model can be a tree-based binary classification model (such as a gradient boosting model) or a deep neural network model. The specific network structure, feature interaction method and hyperparameter settings of the initial prediction model are not limited to a specific implementation.

[0147] It should be noted that the process of training the initial prediction model using the training sample set to obtain the target prediction model is similar to the existing model training process, and will not be described in detail in this embodiment.

[0148] In this embodiment, by constructing a training sample set that simultaneously includes a set of feature information of historical freight orders, a set of driver feature information corresponding to the historical freight orders, and a set of feature information of their interaction, and using whether the driver cancels the historical freight orders as label information, the target prediction model can fully learn the relationship between driver cancellation behavior and multi-dimensional features during the training phase. Furthermore, by training the initial prediction model and generating the target prediction model through the training sample set, the target prediction model can, during the inference phase, output the cancellation probability corresponding to the feature output of the target freight order and the first driver, thereby enabling the target prediction model to effectively identify and predict cancellation risks in practical applications.

[0149] In the above embodiment, the electronic device needs to obtain the probability that the first driver will cancel the target cargo order. Next, the method executed by the electronic device after obtaining the probability of the first driver canceling the target cargo order will be described in detail.

[0150] Figure 3 This is a flowchart illustrating another method for predicting order cancellation probability provided in an embodiment of this application. Figure 3 As shown, in one possible embodiment, after the method steps shown in S104, the prediction method further includes implementations in S105 to S107, which are described in detail below.

[0151] S105. Determine whether the target source of goods meets the conditions for layered exposure.

[0152] Among them, the layered exposure condition is used to determine whether layered exposure is effective in reducing the probability of the first driver canceling the target cargo order.

[0153] It should be noted that the tiered exposure condition is used to determine whether implementing a tiered exposure strategy for the first driver is an effective strategy in the current target cargo order scenario.

[0154] In one possible implementation, it can be determined whether the target source meets the stratified exposure conditions based on the control group and the intervention group.

[0155] During the offline training phase, the case where tiered exposure of goods is implemented is used as the intervention group, and the case where tiered exposure is not implemented is used as the control group. With the fulfillment of goods orders as the target result, an estimation model for improving the fulfillment rate is constructed.

[0156] Specifically, the difference between the intervention group and the control group can be estimated using an uplift model. The Uplift value is defined as: .

[0157] in, This indicates that layered exposure is being performed. This indicates that layered exposure was not performed. The Uplift value represents the net impact of layered exposure on the fulfillment rate of goods orders.

[0158] In the process of estimating the model, the training set and the test set can be divided, and the ratio of the intervention group to the control group in the two groups should be kept consistent to ensure the stability of the estimation results.

[0159] Furthermore, X-Learner can be selected as the learning method, trained with Uplift value as the optimization objective, and combined with Uplift tree to split the features in order to extract target feature patterns that meet the layered exposure conditions.

[0160] The target feature pattern refers to the combination of conditions extracted from the splitting results of the Uplift model after modeling and analyzing historical order data, which can characterize the relationship between the characteristics of the goods and the layered exposure effect.

[0161] During offline training, the target platform uses multi-dimensional features of historical freight orders as input, trains the Uplift model using the X-Learne method, and performs splitting using an Uplift tree. At the leaf nodes of the Uplift tree, a set of combinations of feature conditions can be formed, such as a freight quality score greater than a certain value and remarks including end-of-line requirements, or a transportation distance exceeding a certain mileage and the shipper's historical cancellation rate being lower than a certain percentage. Each combination of feature conditions corresponds to an Uplift value, which characterizes the net improvement in fulfillment rate achieved by implementing tiered exposure under that feature condition combination.

[0162] When the Uplift value of a certain combination of feature conditions exceeds the first preset threshold set by the target platform, and its sample size and confidence index meet the preset requirements, the combination of feature conditions is identified as the target feature pattern. The target feature pattern is essentially a set of rules that describes the effectiveness of tiered exposure under the feature conditions of this type of goods.

[0163] During the online determination phase, for a target source of goods, the feature vector corresponding to the target source of goods can be used to determine whether the feature vector of the target source of goods conforms to any target feature pattern generated in the offline training phase, and then determine whether the target source of goods meets the tiered exposure conditions.

[0164] Specifically, if the feature vector corresponding to the target source matches the target feature pattern, and the Uplift value of the target feature pattern that matches the feature vector corresponding to the target source is greater than a first preset threshold, then the target source is determined to meet the layered exposure conditions; otherwise, the target source is determined not to meet the layered exposure conditions.

[0165] In another possible implementation, when the target product is in a special category, special route, or special note set configured by the target platform, it is directly determined that the target product meets the tiered exposure conditions.

[0166] S106. If the target cargo meets the tiered exposure conditions, based on the cancellation probability of each first driver canceling the target cargo order, the multiple first drivers are divided into high cancellation rate drivers, medium cancellation rate drivers and low cancellation rate drivers.

[0167] If the target cargo meets the tiered exposure conditions, based on the cancellation probability of each of the multiple first drivers, the multiple first drivers are divided into three categories: high cancellation rate drivers, medium cancellation rate drivers, and low cancellation rate drivers.

[0168] In one possible implementation, multiple first drivers can be classified using a fixed threshold method based on their respective cancellation probabilities.

[0169] Set a threshold θh>θm≥0. When the first driver's cancellation probability pi≥θh, the first driver is classified as a high cancellation rate driver; when the first driver's cancellation probability θh>pi≥θm, the first driver is classified as a medium cancellation rate driver; when the first driver's cancellation probability pi<θm, the first driver is classified as a low cancellation rate driver.

[0170] In another possible implementation, a proportional constraint can be added to the fixed threshold method to control the size of different types of drivers, such as limiting the number of drivers with high cancellation rates to no more than α% of the total number of multiple first drivers, where α is a configurable parameter.

[0171] S107. For drivers with high cancellation rates, the exposure of target cargo orders is delayed by the first time; for drivers with medium cancellation rates, the exposure of target cargo orders is delayed by the second time; and for drivers with low cancellation rates, the exposure of target cargo orders is direct.

[0172] The first duration is longer than the second duration.

[0173] It should be noted that different exposure timing control parameters are configured for drivers of different levels, and the display of target cargo orders is scheduled.

[0174] Specifically, the exposure timing control parameter configured for drivers with high cancellation rates is the first duration T1, the exposure timing control parameter configured for drivers with medium cancellation rates is the second duration T2, and the exposure timing control parameter configured for drivers with low cancellation rates is the third duration T0; T1 > T2, T0 = 0.

[0175] For drivers with high cancellation rates, the target cargo order is added to the driver's visible queue after a first delay of T1; for drivers with medium cancellation rates, the target cargo order is added to the driver's visible queue after a second delay of T2; for drivers with low cancellation rates, the target cargo order is exposed directly.

[0176] The first duration T1 and the second duration T2 can be set by the target platform administrators themselves, and this embodiment does not impose specific limitations on this.

[0177] For example, the first duration T1 is 10 minutes, and the second duration T2 is 5 minutes.

[0178] It should be noted that the target platform maintains an exposure scheduling queue for each first driver. When the exposure delay time corresponding to the first driver is reached, the target cargo order is pushed to the front of the queue or inserted into an appropriate position according to a predetermined sorting factor (such as distance, timeliness, platform sorting score, etc.). If the status of the target cargo order changes during the exposure delay period (such as being accepted by another driver, canceled, or completed), the target cargo order will no longer be exposed to the first driver.

[0179] In this embodiment, by introducing a tiered exposure condition determination step, and when the target cargo meets the tiered exposure condition, the candidate drivers are divided into high cancellation rate drivers, medium cancellation rate drivers, and low cancellation rate drivers based on the cancellation probability of each candidate driver. Then, different delayed exposure durations are configured for drivers of different cancellation rate levels. Based on the output of the target prediction model, differentiated management of the driver group can be further realized, so that the exposure scheduling link can be graded according to the degree of cancellation risk. This will suppress high-risk drivers and give priority support to low-risk drivers in the order allocation process of the target platform.

[0180] In the above embodiments, the electronic device needs to classify the multiple first drivers into high-cancellation-rate drivers, medium-cancellation-rate drivers, and low-cancellation-rate drivers based on the cancellation probability of each first driver canceling the target cargo order. The specific process by which the electronic device classifies the multiple first drivers into these categories will be described in detail below.

[0181] In one possible embodiment, the method steps shown in S106 can be implemented by S1061 to S1066, which are described in detail below.

[0182] S1061. Determine the target quantity based on the number of multiple first drivers and the preset ratio.

[0183] The target number is the total number of drivers with high cancellation rates and drivers with medium cancellation rates among multiple first-time drivers.

[0184] It should be noted that the preset ratio can be configured by the administrators of the target platform, and this embodiment does not impose any specific limitations on it.

[0185] For example, the preset ratio is 20%.

[0186] If the number of multiple first drivers is N, and the preset ratio is 20%, then the target number M = N × 20%.

[0187] For example, if the total number of first drivers is 50 and the preset ratio is set to 20%, then the target number is 10, meaning that a maximum of 10 first drivers can be classified as high cancellation rate drivers or medium cancellation rate drivers.

[0188] S1062. Determine the first cancellation probability threshold based on the sum of the basic cancellation probability and the first preset value.

[0189] The base cancellation probability is the median of the cancellation probabilities of multiple first drivers when they are arranged in descending order.

[0190] It should be noted that if N is an even number, the base cancellation probability is the average of the cancellation probabilities of the multiple first drivers arranged in descending order, specifically the cancellation probability at the N / 2th position and the cancellation probability at the (N / 2+1th position).

[0191] The first preset value is a positive offset set by the administrator of the target platform, which is used to shift the first preset value upward based on the basic cancellation probability.

[0192] For example, if the base cancellation probability is 0.7 and the first preset value is 0.15, then the first cancellation probability threshold is 0.85.

[0193] S1063. Determine the second cancellation probability threshold based on the sum of the basic cancellation probability and the second preset value.

[0194] The second preset value is less than the first preset value.

[0195] The second preset value is a positive offset set by the administrator of the target platform, which is used to shift the second preset value upward based on the basic cancellation probability.

[0196] For example, if the base cancellation probability is 0.7 and the second preset value is set to 0.05, then the second cancellation probability threshold is 0.75.

[0197] S1064. If the first driver's cancellation probability of canceling the target cargo order is greater than the first cancellation probability threshold, and the number of drivers with high cancellation rates is less than the target number, the first driver is identified as a driver with a high cancellation rate.

[0198] If the probability of the first driver canceling the target cargo order is greater than the first cancellation probability threshold, and the number of drivers already classified as high cancellation rate has not reached the target number, the first driver will be identified as a high cancellation rate driver.

[0199] S1065. If the first driver's cancellation probability of canceling the target cargo order is greater than the second cancellation probability threshold, the first driver's cancellation probability of canceling the target cargo order is less than or equal to the first cancellation probability, and the number of drivers with high cancellation rates and medium cancellation rates is less than the target number, then the first driver is identified as a driver with a medium cancellation rate.

[0200] If the first driver's probability of canceling a target cargo order is greater than the second cancellation probability threshold but less than or equal to the first cancellation probability threshold, and the number of drivers already classified as high cancellation rate drivers and medium cancellation rate drivers has not reached the target number, then the first driver is identified as a medium cancellation rate driver.

[0201] S1066. Identify multiple first drivers other than high cancellation rate drivers and medium cancellation rate drivers as low cancellation rate drivers.

[0202] The remaining first drivers who were not classified as high-cancellation-rate drivers or medium-cancellation-rate drivers were identified as low-cancellation-rate drivers.

[0203] In this embodiment, a clear stratification standard can be formed among multiple first drivers based on the number of first drivers, a preset ratio, a basic cancellation probability, a first preset value, and a second preset value. This allows multiple first drivers to be reasonably divided into three categories: high cancellation rate drivers, medium cancellation rate drivers, and low cancellation rate drivers. The stratification results are highly stable and controllable, facilitating the implementation of subsequent stratified exposure strategies based on a clear risk distribution.

[0204] In another possible embodiment, when the electronic device divides the multiple first drivers into high cancellation rate drivers, medium cancellation rate drivers, and low cancellation rate drivers based on the cancellation probability of each first driver canceling the target cargo order, it can also be achieved in the following way.

[0205] Specifically, the first drivers are first filtered in descending order of their cancellation probabilities, and after each filtering operation, the fulfillment rate of the remaining first drivers is calculated. By recording the changes in the fulfillment rate as the number of filtered first drivers gradually increases, a curve showing the relationship between the number of filtered drivers and the fulfillment rate can be plotted.

[0206] On the curve relating the number of filtered drivers to the fulfillment rate, identify the inflection point. The inflection point indicates a sharp drop in the fulfillment rate as the number of filtered first drivers continues to increase. Therefore, the number of filtered drivers, x, can be set to the number of first drivers corresponding to the inflection point of the curve. Alternatively, the number of filtered drivers, x, can be set to 20% of the total number of multiple first drivers.

[0207] Furthermore, to prevent abnormal fluctuations in fulfillment rate statistics due to excessive filtering of drivers, constraints can be set on the magnitude of changes in fulfillment rate. For example, the relative change in fulfillment rate before and after filtering can be limited to not exceeding a second preset threshold. When the relative change in fulfillment rate before and after filtering exceeds the second preset threshold, the number of drivers classified as high cancellation rate or medium cancellation rate will no longer be increased.

[0208] In this case, the first drivers with the highest cancellation rates (ranked in the top x) among multiple first drivers are classified as high cancellation rate drivers or medium cancellation rate drivers, and the remaining first drivers who are not classified are classified as low cancellation rate drivers.

[0209] Figure 4 This is a schematic diagram of a device for predicting the probability of order cancellation, provided in an embodiment of this application. Figure 4 As shown, the order cancellation probability prediction device 400 provided in this embodiment includes a first acquisition module 401, a second acquisition module 402, a third acquisition module 403, and a processing module 404.

[0210] The first acquisition module 401 is used to acquire a first feature information set related to the target cargo order. The first feature information set includes first feature information related to the historical behavior characteristics of the target cargo owner to which the target cargo order belongs, second feature information related to the attributes of the target cargo corresponding to the target cargo order, and third feature information related to the configuration of the target cargo order.

[0211] The second acquisition module 402 is used to acquire a second set of feature information related to the first driver. The second set of feature information includes fourth feature information related to the target vehicle corresponding to the first driver, and fifth feature information related to the historical behavior features of the first driver.

[0212] The third acquisition module 403 is used to acquire a third set of feature information related to the target cargo order and the first driver. The third set of feature information includes the first cross-feature information between the first driver and the target cargo owner, the second cross-feature information between the first driver and the target cargo, and the third cross-feature information between the first driver and the target cargo order.

[0213] The processing module 404 is used to process the first feature information set, the second feature information set, and the third feature information set using the target prediction model to obtain the cancellation probability of the first driver canceling the target cargo order.

[0214] It should be understood that the corresponding processes performed by each module have been described in detail in the above method embodiments, and will not be repeated here for the sake of brevity.

[0215] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 5 As shown, the electronic device 500 provided in this embodiment includes a memory 501 and a processor 502.

[0216] The memory 501 can be a separate physical unit, connected to the processor 502 via a bus 503. Alternatively, the memory 501 and processor 502 can be integrated and implemented in hardware. The memory 501 stores program instructions, which the processor 502 calls to execute the operations performed by the electronic device in any of the above method embodiments.

[0217] Optionally, when some or all of the methods in the above embodiments are implemented by software, the electronic device 500 may also include only the processor 502. A memory 501 for storing programs is located outside the electronic device 500, and the processor 502 is connected to the memory via circuits / wires to read and execute the programs stored in the memory. The processor 502 may be a central processing unit (CPU), a network processor (NP), or a combination of a CPU and an NP. The processor 502 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof.

[0218] The memory 501 may include volatile memory, such as random-access memory (RAM); the memory may also include non-volatile memory, such as flash memory, hard disk drive (HDD) or solid-state drive (SSD); the memory may also include a combination of the above types of memory.

[0219] For example, this application provides a chip including: an interface circuit and a logic circuit. The interface circuit is used to receive signals from other chips outside the chip and transmit them to the logic circuit, or to send signals from the logic circuit to other chips outside the chip. The logic circuit is used to perform the operations performed by the electronic device in the above method embodiments.

[0220] For example, this application provides a computer-readable storage medium having computer program instructions stored thereon, which are executed by the processor of an electronic device to cause the electronic device to perform the operations performed by the electronic device in the above method embodiments.

[0221] For example, this application provides a computer program product that, when run on an electronic device, causes the electronic device to perform the operations performed by the electronic device in the above method embodiments.

[0222] The above description is merely a specific embodiment of this application, enabling those skilled in the art to understand or implement this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments described herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for predicting the probability of order cancellation, characterized in that, The method includes: Obtain a first set of feature information related to the target cargo order. The first set of feature information includes first feature information related to the historical behavior features of the target cargo owner to which the target cargo order belongs, second feature information related to the inherent attributes of the target cargo corresponding to the target cargo order, and third feature information related to the configuration of the target cargo order. Obtain a second set of feature information related to the first driver. The second set of feature information includes a fourth set of feature information related to the target vehicle corresponding to the first driver, and a fifth set of feature information related to the historical behavior features of the first driver. Obtain a third set of feature information related to the target cargo order and the first driver. The third set of feature information includes first cross-feature information between the first driver and the target cargo owner, second cross-feature information between the first driver and the target cargo, and third cross-feature information between the first driver and the target cargo order. The first feature information set, the second feature information set, and the third feature information set are processed using a target prediction model to obtain the probability that the first driver will cancel the target cargo order.

2. The method according to claim 1, characterized in that, The first feature information includes the target cargo owner's transaction and cancellation information within a preset time period, as well as the target cargo owner's credit score; wherein, the transaction information includes the target cargo owner's number of transaction days and number of transactions; the cancellation information includes the target cargo owner's number of cancellation days and number of cancellations; the preset time period is located before the time when the target cargo order is published; The second feature information includes the type information, shape information, volume information, weight information, and sensitive attribute information of the target cargo, as well as the quality score of the target cargo; The third feature information includes the transportation route and associated conditions of the target cargo.

3. The method according to claim 2, characterized in that, The fourth feature information includes the target vehicle's length, vehicle type, origin, destination, and load information; The fifth feature information includes the number of times the first driver is exposed, clicked, completed, canceled, and placed orders without fulfilling them within the preset time period; wherein, the number of exposures is the number of times the target platform shows the first cargo order to the first driver; the number of clicks is the number of times the first driver clicks to view the first cargo order; the number of completed transactions is the number of times the first driver accepts the first cargo order; the number of cancellations is the number of times the first driver accepts but does not fulfill the first cargo order; and the number of placed orders without fulfilling them is the number of times the first driver completes a transaction on the target platform but does not fulfill the first cargo order through the target platform; wherein, the first cargo order is any cargo order on the target platform.

4. The method according to claim 3, characterized in that, The first cross-feature information includes transaction information and cancellation information between the first driver and the target cargo owner within the preset time period; The second cross-feature information includes the number of times the first driver exposes the second cargo order, the number of times it is clicked, the number of transactions completed, the number of cancellations, and the number of times the first driver cancels the order within the preset time period; wherein, the cargo corresponding to the second cargo order is the target cargo. The third cross-feature information includes the number of cargo orders completed by the first driver along the transportation route within the preset time period.

5. The method according to claim 1, characterized in that, Before obtaining the first feature information related to the target source order, the method further includes: A training sample set is constructed, comprising multiple training samples. Each training sample includes a set of historical feature information and label information corresponding to the set of historical feature information. The set of historical feature information includes a fourth set of feature information related to historical freight orders, a fifth set of feature information related to the second driver corresponding to the historical freight orders, and a sixth set of feature information related to both the historical freight orders and the second driver. The label information is used to indicate whether the second driver cancels the historical freight orders. The initial prediction model is trained using the training sample set to obtain the target prediction model.

6. The method according to claim 1, characterized in that, The target prediction model is a tree model or a deep neural network model.

7. The method according to claim 1, characterized in that, After processing the first feature information set, the second feature information set, and the third feature information set using a target prediction model to obtain the cancellation probability of the first driver canceling the target freight order, the method further includes: Determine whether the target cargo meets the tiered exposure conditions, wherein the tiered exposure conditions are used to determine whether the tiered exposure is effective in reducing the probability of the first driver canceling the order for the target cargo; If the target cargo meets the tiered exposure conditions, based on the cancellation probability of each first driver canceling the target cargo order, the multiple first drivers are respectively divided into high cancellation rate drivers, medium cancellation rate drivers and low cancellation rate drivers; For drivers with high cancellation rates, the exposure of the target cargo order is delayed by a first duration; for drivers with medium cancellation rates, the exposure of the target cargo order is delayed by a second duration; and for drivers with low cancellation rates, the exposure of the target cargo order is direct; wherein the first duration is longer than the second duration.

8. The method according to claim 7, characterized in that, The method of classifying the multiple first drivers into high-cancellation-rate drivers, medium-cancellation-rate drivers, and low-cancellation-rate drivers based on the cancellation probability of each first driver among multiple first drivers for the target cargo order includes: Based on the number of the plurality of first drivers and a preset ratio, a target number is determined, wherein the target number is the total number of drivers with high cancellation rates and drivers with medium cancellation rates among the plurality of first drivers; A first cancellation probability threshold is determined based on the sum of the base cancellation probability and a first preset value; wherein, the base cancellation probability is the median of the cancellation probabilities of the plurality of first drivers when they are arranged in descending order. A second cancellation probability threshold is determined based on the sum of the basic cancellation probability and the second preset value; wherein the second preset value is less than the first preset value. If the probability of the first driver canceling the target cargo order is greater than the first cancellation probability threshold, and the number of drivers with high cancellation rates is less than the target number, the first driver is identified as the driver with high cancellation rates. If the first driver's probability of canceling the target cargo order is greater than the second cancellation probability threshold, the first driver's probability of canceling the target cargo order is less than or equal to the first cancellation probability, and the number of drivers with high cancellation rates and drivers with medium cancellation rates is less than the target number, then the first driver is identified as the driver with medium cancellation rates. The first driver, excluding the high cancellation rate driver and the medium cancellation rate driver, is identified as the low cancellation rate driver.

9. An electronic device, characterized in that, include: Memory and at least one processor; The memory stores computer-executed instructions; The at least one processor executes computer execution instructions stored in the memory, causing the at least one processor to perform the method as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions that, when executed, implement the method as described in any one of claims 1 to 8.

Images