Method, device and equipment for updating prediction model and storage medium

Abstract

The application discloses a prediction model updating method and device, equipment and a storage medium, and belongs to the technical field of computers. The method comprises the following steps: obtaining historical order data of a plurality of historical orders, wherein each piece of historical order data comprises a first predicted completion time determined based on a prediction model for a corresponding historical order, an actual completion time of the historical order and delivery process information; determining a first historical order with a timeout problem in the plurality of historical orders based on the actual completion time and the first predicted completion time in each piece of historical order data; determining a timeout factor causing the first historical order to be timed out based on the delivery process information of the first historical order; and updating the prediction model based on the timeout factor of the first historical order and the historical order data of the first historical order to obtain an updated prediction model. The scheme can improve the prediction model in a targeted manner, and improves the accuracy of the prediction model.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus, device and storage medium for updating a prediction model. Background Technology

[0002] With the development of computer technology, services such as food delivery and express delivery have emerged, and the number of orders has been increasing. To help delivery personnel allocate their time more efficiently and complete orders as quickly as possible, a method is proposed that uses a predictive model to forecast completion times. Based on the predicted completion times, orders are assigned to suitable delivery personnel. Therefore, the predicted completion time largely determines whether an order is assigned to a particular delivery person and whether that person will exceed the delivery time limit, significantly impacting the experience of both delivery personnel and users. Thus, accurately predicting completion times has become a pressing issue that needs to be addressed. Summary of the Invention

[0003] This application provides a method, apparatus, device, and storage medium for updating a prediction model, which improves the accuracy of the prediction model. The technical solution is as follows:

[0004] On the one hand, a method for updating a prediction model is provided, the method comprising:

[0005] Acquire historical order data for multiple historical orders. Each historical order data item includes a first predicted completion time determined for the corresponding historical order based on a prediction model, the actual completion time of the historical order, and delivery process information.

[0006] Based on the actual completion time and the first predicted completion time in each historical order data, the first historical order with a timeout problem is identified among the multiple historical orders.

[0007] Based on the delivery process information of the first historical order, determine the timeout factors that caused the first historical order to time out;

[0008] Based on the timeout factor of the first historical order and the historical order data of the first historical order, the prediction model is updated to obtain the updated prediction model.

[0009] On the one hand, a predictive model update apparatus is provided, the apparatus comprising:

[0010] The acquisition module is used to acquire historical order data for multiple historical orders. Each historical order data includes a first predicted completion time determined for the corresponding historical order based on a prediction model, the actual completion time of the historical order, and delivery process information.

[0011] The first determining module determines the first historical order with a timeout problem among the multiple historical orders based on the actual completion time and the first predicted completion time in each historical order data.

[0012] The second determining module is used to determine the timeout factors that cause the first historical order to time out based on the delivery process information of the first historical order;

[0013] The update module is used to update the prediction model based on the timeout factor of the first historical order and the historical order data of the first historical order, so as to obtain the updated prediction model.

[0014] In one possible implementation, the update module includes:

[0015] The processing unit is used to process the delivery process information of the first historical order through the prediction model to obtain the second predicted completion time of the first historical order.

[0016] The determining unit is used to determine the first weight corresponding to the timeout factor of the first historical order based on the correspondence between the timeout factor and the weight.

[0017] The update unit is used to update the prediction model based on the first weight, the second predicted completion time and the actual completion time of the first historical order, to obtain the updated prediction model.

[0018] In one possible implementation, the actual completion time includes the actual completion sub-times corresponding to multiple nodes in the delivery process, the second predicted completion time includes the second predicted completion sub-times corresponding to multiple nodes in the delivery process, and the timeout factor of the first historical order represents the timeout node that causes the timeout problem in the delivery process;

[0019] The update unit is configured to process the actual completion time and the second predicted completion time based on the loss function of the prediction model to obtain a first loss value; process the second predicted completion sub-time and the actual completion sub-time corresponding to the timeout node in the first historical order based on the node loss function to obtain a second loss value; process the second loss value based on the first weight to obtain a third loss value; determine a fourth loss value based on the first loss value and the third loss value; and update the prediction model based on the fourth loss value to obtain the updated prediction model.

[0020] In one possible implementation, the update module is used to update the prediction model based on the timeout factor of the first historical order, the historical order data of the first historical order, and the historical order data of the second historical order, to obtain the updated prediction model, wherein the second historical order is a historical order that has not timed out among the plurality of historical orders.

[0021] In one possible implementation, the update module is used to alternately use the historical order data of the second historical order and the timeout factor and historical order data of the first historical order to update the prediction model, thereby obtaining the updated prediction model.

[0022] In one possible implementation, the update module includes:

[0023] The discard unit is used to discard the historical order data of the first historical order if the timeout factor of the first historical order is due to human factors.

[0024] The update unit is used to update the prediction model based on the timeout factors of the remaining first historical orders and historical order data to obtain the updated prediction model.

[0025] In one possible implementation, the second determining module is configured to query the delivery process information of the first historical order based on at least one of a plurality of timeout factors to obtain a query result corresponding to the at least one timeout factor; if the query result corresponding to any timeout factor indicates that the delivery process information includes information matching the timeout factor, the timeout factor is determined as the timeout factor that caused the first historical order to time out.

[0026] In one possible implementation, the second determining module includes:

[0027] The first analysis unit is used to analyze the delivery process information corresponding to the pickup process of the first historical order to obtain the first analysis result;

[0028] The second analysis unit is used to analyze the delivery process information corresponding to the delivery process of the first historical order to obtain the second analysis result;

[0029] The determining unit is used to determine the timeout factors that cause the first historical order to time out, based on the first analysis result and the second analysis result.

[0030] In one possible implementation, the actual completion time includes the actual completion sub-times corresponding to multiple nodes in the delivery process, and the first predicted completion time includes the first predicted completion sub-times corresponding to multiple nodes in the delivery process.

[0031] The first analysis unit is configured to obtain the actual completion time and the first predicted completion time of the pickup node from the delivery process information corresponding to the pickup process of the first historical order; if the actual completion time of the pickup node is not later than the corresponding first predicted completion time, the analysis of the delivery process information corresponding to the pickup process of the first historical order is stopped, and a first analysis result indicating that there is no timeout problem in the pickup process is obtained; if the actual completion time of the pickup node is later than the corresponding first predicted completion time, the analysis of the delivery process information corresponding to the pickup process of the first historical order is continued, and a first analysis result indicating the timeout factor is obtained.

[0032] In one possible implementation, the actual completion time includes the actual completion sub-times corresponding to multiple nodes in the delivery process, and the first predicted completion time includes the first predicted completion sub-times corresponding to multiple nodes in the delivery process; the first analysis result includes timeout factors that cause the first historical order to have a timeout problem.

[0033] The first analysis unit is configured to perform at least one of the following:

[0034] Query the predicted completion time and actual completion time of the previous node. If the actual completion time of the previous node is later than the predicted completion time, determine that the timeout factor is a hindrance to the previous node.

[0035] Query the actual delivery route and planned delivery route of the first historical order. If the first historical order is a direct delivery order and the actual delivery route and the planned delivery route are inconsistent, determine that the timeout factor is inconsistent delivery routes.

[0036] Query the abnormal reporting information of the first historical order. If the abnormal reporting information is not empty, determine the timeout factor based on the abnormal reporting information. If the abnormal reporting information is empty and the actual delivery route is consistent with the planned delivery route, analyze the predicted completion sub-time and actual completion sub-time corresponding to multiple nodes in the pickup process to obtain the timeout node. Determine the timeout factor as the optimistic estimate of the completion sub-time of the timeout node.

[0037] On one hand, a computer device is provided, the computer device including one or more processors and one or more memories, the one or more memories storing at least one piece of program code, the at least one piece of program code being loaded and executed by the one or more processors to perform the operations performed by the method for updating the predictive model as described in any of the above possible implementations.

[0038] On the one hand, a computer-readable storage medium is provided, which stores at least one piece of program code, which is loaded and executed by a processor to perform the operations performed by the method for updating the predictive model as described in any of the possible implementations above.

[0039] On one hand, a computer program or computer program product is provided, the computer program or computer program product comprising: computer program code, which, when executed by a computer, causes the computer to perform the operations performed by the prediction model update method as described in any of the above possible implementations.

[0040] The beneficial effects of the technical solutions provided in this application include at least the following:

[0041] The prediction model update method, apparatus, device, and storage medium provided in this application analyze historical orders processed by the prediction model. This not only identifies historical orders with timeout issues but also determines the timeout factors for those orders. By updating the prediction model using these timeout factors and historical order data, the predicted completion time determined by the prediction model becomes closer to the actual completion time of the historical orders, reducing the occurrence of order timeouts. Furthermore, by considering the specific timeout factors of historical orders, the prediction model can be specifically improved during the update process, thereby enhancing its accuracy. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a schematic diagram of an implementation environment provided in an embodiment of this application;

[0044] Figure 2 This is a flowchart of a prediction model update method provided in an embodiment of this application;

[0045] Figure 3 This is a flowchart of a prediction model update method provided in an embodiment of this application;

[0046] Figure 4 This is a schematic diagram of a delivery process provided in an embodiment of this application;

[0047] Figure 5 This is a flowchart illustrating how to determine timeout factors, provided in an embodiment of this application.

[0048] Figure 6 This is another flowchart for determining timeout factors provided in an embodiment of this application;

[0049] Figure 7 This is a schematic diagram illustrating the updating of a prediction model provided in an embodiment of this application;

[0050] Figure 8 This is a schematic diagram of a predictive model update device provided in an embodiment of this application;

[0051] Figure 9 This is a schematic diagram of the structure of an update device for another prediction model provided in an embodiment of this application;

[0052] Figure 10 This is a schematic diagram of the terminal structure provided in the embodiments of this application;

[0053] Figure 11 This is a schematic diagram of the server structure provided in an embodiment of this application. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0055] It is understood that the terms "first," "second," etc., used in this application may be used to describe various concepts herein, but unless otherwise stated, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of this application, a first weight may be referred to as a second weight, and similarly, a second weight may be referred to as a first weight.

[0056] As used in this application, the terms "at least one", "multiple", "each", and "any" are used in the following ways: at least one includes one, two, or more; multiple includes two or more; each refers to each of the corresponding multiple; and any refers to any one of the multiple. For example, multiple weights include three weights, each refers to each of the three weights, and any refers to any one of the three weights, which can be the first, the second, or the third.

[0057] It is understood that in the specific embodiments of this application, the historical order data and other user-related data involved need to obtain user permission or consent when the above embodiments of this application are applied to specific products or technologies, and the collection, use and processing of related data need to comply with the relevant laws, regulations and standards of relevant countries and regions.

[0058] The prediction model update method provided in this application is executed by a computer device. In one possible implementation, the computer device is a terminal, such as a desktop computer, tablet computer, or mobile phone. In another possible implementation, the computer device is a server. For example, the server can be a single server, a server cluster consisting of several servers, or a cloud computing service center. In yet another possible implementation, the computer device includes both a terminal and a server.

[0059] Figure 1 This is a schematic diagram of an implementation environment provided in an embodiment of this application, such as... Figure 1 As shown, the implementation environment includes a terminal 101 and a server 102. The terminal 101 and the server 102 are connected via a wireless or wired network.

[0060] Optionally, terminal 101 can be any type of terminal such as a desktop computer, tablet computer, or mobile phone. Server 102 can be a single server, a server cluster consisting of several servers, or a cloud computing service center.

[0061] A target application, provided by server 102, is installed on terminal 101. Terminal 101 can perform functions such as data transmission and message interaction through this target application. Optionally, the target application is an application in the operating system of terminal 101, or an application provided by a third party. For example, the target application is an order processing application, which has the function of processing orders. Of course, the order processing application may also have other functions, such as review functions and sharing functions.

[0062] Optionally, server 102 obtains historical order data from multiple historical orders, updates the prediction model based on the historical order data, and sends the updated prediction model to terminal 101 so that terminal 101 can make predictions for the currently pending orders based on the updated prediction model; or, server 102 receives order data for the currently pending orders uploaded by terminal 101, processes the order data based on the updated prediction model to obtain the predicted completion time of the order, allocates appropriate delivery capacity to the order based on the predicted completion time, and sends the order to the terminal logged in by the delivery capacity.

[0063] The prediction model update method provided in this application can be applied to any order processing scenario.

[0064] For example, it can be applied to food delivery scenarios.

[0065] After a user places an order on a food delivery platform, the platform uses an updated prediction model to determine the predicted completion time of the order. Based on this predicted completion time, the platform schedules the order and assigns it to a suitable rider. This makes order allocation more accurate, allowing riders to handle orders more efficiently and deliver food to users in a timely manner.

[0066] It should be noted that the embodiments in this application are only used as examples to illustrate the order processing scenario in the context of food delivery, and do not limit the order processing scenario. Optionally, the order processing scenario can also be a ride-hailing scenario, etc.

[0067] Figure 2 This is a flowchart illustrating a method for updating a prediction model provided in an embodiment of this application. This embodiment uses a computer device as an example for illustrative purposes, and includes:

[0068] 201. The computer device acquires historical order data for multiple historical orders, each historical order data including a first predicted completion time determined for the corresponding historical order based on a prediction model, the actual completion time of the historical order, and delivery process information.

[0069] In this embodiment, historical orders refer to orders processed after the prediction model has been deployed. After training, the prediction model is deployed to process generated orders, determine their predicted completion time, and schedule orders based on this time, assigning them to suitable riders to ensure timely delivery. It should be noted that this embodiment is merely an illustrative example of the role of the predicted completion time determined by the prediction model and does not constitute a limitation.

[0070] For each historical order processed by the prediction model, the computer device records not only the first predicted completion time determined by the prediction model for that historical order, but also the actual completion time and delivery process information. The delivery process information describes the delivery process of the historical order, allowing for a complete or partial reconstruction of the delivery process. Optionally, the delivery process information includes at least one of the following: planned delivery route, actual delivery route, weather information, traffic information, actual completion sub-times and the first predicted completion sub-time for multiple nodes during the delivery process, and anomaly reporting information. It should be noted that this embodiment is merely an illustrative example of delivery process information and does not limit the scope of such information. Optionally, the delivery process information is the delivery record of the historical order, including any information recorded during the delivery process.

[0071] 202. The computer equipment determines the first historical order with a timeout problem among multiple historical orders based on the actual completion time and the first predicted completion time in each historical order data.

[0072] Optionally, a timeout problem refers to the failure of delivery capacity to complete order delivery in a timely manner. For example, if the actual completion time of a historical order is later than the first predicted completion time in the historical order data, then the historical order is determined to have a timeout problem; or, if the actual completion time of a historical order is later than the first predicted completion time in the historical order data, and the difference between the actual completion time and the first predicted completion time is greater than the deviation tolerance, then the historical order is determined to have a timeout problem. The deviation tolerance can be any duration, such as 5 minutes, 10 minutes, etc. This application embodiment does not limit the deviation tolerance.

[0073] Delivery capacity can include riders, couriers, food delivery robots, and courier robots.

[0074] It should be noted that the embodiments in this application are only used as an example to illustrate how to determine the first historical order with a timeout problem based on the actual completion time and the first predicted completion time in each historical order data. In another embodiment, after determining the first predicted completion time through a prediction model and allocating the order to the delivery capacity based on the first predicted completion time, a third predicted completion time for the order is determined based on the orders held by the delivery capacity. Based on the third predicted completion time and the actual completion time of the historical orders, the first historical order with a timeout problem among multiple historical orders is determined.

[0075] For example, the computer equipment first identifies the first historical order with a timeout problem based on the actual completion time and the third predicted completion time in each historical order data. Then, based on the actual completion time and the first predicted completion time in each historical order data, it continues to filter the first historical order.

[0076] Optionally, the actual completion time includes the actual completion sub-times corresponding to multiple nodes in the delivery process, and the first predicted completion time includes the first predicted completion sub-times corresponding to multiple nodes in the delivery process. For example, the actual completion time includes the actual completion sub-time corresponding to the pickup node, and the first predicted completion time includes the first predicted completion sub-time corresponding to the pickup node. If the first predicted completion sub-time corresponding to the pickup node is later than the corresponding actual completion sub-time, or if the first predicted completion sub-time corresponding to the pickup node is later than the sum of the corresponding actual completion sub-time and the deviation tolerance, then it is determined that the historical order has a timeout problem, and further analysis is needed to determine the timeout factors.

[0077] 203. The computer equipment determines the timeout factors that caused the first historical order to time out based on the delivery process information of the first historical order.

[0078] Since historical order delivery information reflects the actual delivery process of a given order, analyzing this information can identify the factors contributing to order timeouts. These factors can be human error or overly optimistic predictions from forecasting models. For example, a delivery delay might be caused by the delivery company not following the planned route; or by incorrect merchant location; or by a timeout occurring even though the delivery process itself was normal, indicating an overly optimistic forecasting model.

[0079] 204. The computer equipment updates the prediction model based on the timeout factor of the first historical order and the historical order data of the first historical order, and obtains the updated prediction model.

[0080] Since timeout issues in historical orders may be caused by inaccurate prediction of completion time determined by the prediction model, this application embodiment, after finding the first historical order with a timeout issue, also determines the timeout factors of the first historical order. Based on the timeout factors of the first historical order and historical order data, the prediction model is improved in a targeted manner to enhance the accuracy of the prediction model.

[0081] The prediction model update method provided in this application analyzes historical orders processed by the prediction model. This not only identifies historical orders with timeout issues but also determines the timeout factors for those orders. By updating the prediction model using these timeout factors and historical order data, the predicted completion time determined by the prediction model becomes closer to the actual completion time of the historical orders, reducing the occurrence of order timeouts. Furthermore, by considering the specific timeout factors of historical orders, the prediction model can be specifically improved during the update process, thus enhancing its accuracy.

[0082] Figure 3 This is a flowchart illustrating a method for updating a prediction model provided in an embodiment of this application. This embodiment uses a computer device as an example for illustrative purposes, and includes:

[0083] 301. The computer equipment acquires historical order data for multiple historical orders. Each historical order data includes the first predicted completion time determined for the corresponding historical order based on a prediction model, the actual completion time of the historical order, and delivery process information.

[0084] This application does not limit the method of obtaining historical order data, but only uses the following two methods as examples for illustration.

[0085] In one possible implementation, a database is used to store historical order data, and the computer device retrieves historical order data for multiple historical orders from this database. In another possible implementation, the computer device generates logs during the processing of historical orders, and retrieves historical order data for multiple historical orders from the generated logs.

[0086] Optionally, the multiple historical orders can be historical orders within any time period; alternatively, the multiple historical orders can be historical orders within a target time period, such as historical orders from the previous month. It should be noted that this embodiment does not limit the time period to which the historical orders belong.

[0087] 302. The computer equipment determines the first historical order with a timeout problem among the multiple historical orders based on the actual completion time and the first predicted completion time in each historical order data.

[0088] Optionally, a timeout issue in historical orders refers to the delivery capacity failing to complete the order delivery in a timely manner. For example, if the actual completion time in the historical order data is later than the first predicted completion time, then the historical order is determined to have a timeout issue. Similarly, if the actual completion time in the historical order data is later than the first predicted completion time, and the difference between the actual completion time and the first predicted completion time is greater than the deviation tolerance, then the historical order is determined to have a timeout issue. The deviation tolerance can be any duration, such as 5 minutes, 10 minutes, etc. This application embodiment does not limit the deviation tolerance.

[0089] In one possible implementation, the computer device determines the first historical order with a timeout problem among the plurality of historical orders based on the actual completion time and the first predicted completion time in each historical order data. This includes: for any historical order, obtaining the actual completion time and the first predicted completion time of the historical order; if the actual completion time is later than the first predicted completion time, determining the historical order as the first historical order with a timeout problem; or, if the actual completion time is later than the first predicted completion time and the difference between the actual completion time and the first predicted completion time is greater than the deviation tolerance, determining the historical order as the first historical order with a timeout problem.

[0090] Optionally, a timeout issue in historical orders refers to a timeout occurring at any point in the delivery process. Examples include delivery delays at the store or pickup times. In one possible implementation, the actual completion time includes the actual completion sub-times corresponding to multiple nodes in the delivery process, and the first predicted completion time includes the first predicted completion sub-times corresponding to multiple nodes in the delivery process. The computer device determines the first historical order with a timeout issue based on the actual completion time and the first predicted completion time in each historical order data entry. This includes: for any historical order, obtaining the actual completion sub-time and the first predicted completion sub-time corresponding to each node in the delivery process; if the actual completion sub-time of any node is later than the corresponding first predicted completion sub-time, the historical order is determined to be the first historical order with a timeout issue; or, if the actual completion sub-time of any node is later than the corresponding first predicted completion sub-time, and the difference between the actual completion sub-time and the first predicted completion sub-time is greater than the deviation tolerance, the historical order is determined to be the first historical order with a timeout issue.

[0091] like Figure 4 As shown, the delivery process includes multiple nodes. In one possible implementation, the prediction model is used to determine the predicted completion sub-time corresponding to at least one of the following nodes: (1) Delivery capacity order acceptance time: the time when the order is accepted by the delivery capacity after the user places an order and the scheduling system schedules it; (2) Delivery capacity riding time: the time the delivery capacity rides during the delivery process; (3) Delivery capacity entry and exit time: the time from when the delivery capacity gets off the vehicle to when it arrives at the store, and from when it leaves the store to when it gets back on the vehicle; (4) Delivery capacity waiting time: the waiting time from when the delivery capacity arrives at the store until it picks up the food. This time is mainly affected by the merchant's food preparation time; (5) Delivery capacity entry and exit time: the time from when the delivery capacity gets off the vehicle to when it arrives at the delivery point, and from when it leaves the delivery point to when it gets back on the vehicle; (6) Delivery capacity delivery time: the waiting time after the delivery capacity arrives at the delivery point.

[0092] 303. The computer equipment obtains the actual completion time and the first predicted completion time corresponding to the pickup node from the delivery process information corresponding to the pickup process of the first historical order.

[0093] It should be noted that in this embodiment, the delivery process information can be divided into delivery process information corresponding to the pickup process and delivery process information corresponding to the delivery process, or the delivery process information can be processed as a whole. This embodiment does not limit this. In step 303, the computer device obtains the actual completion time and the first predicted completion time corresponding to the pickup node from the delivery process information corresponding to the pickup process of the first historical order. This only indicates that the actual completion time and the first predicted completion time corresponding to the pickup node belong to the delivery process information corresponding to the pickup process, and does not indicate that the computer device has divided the delivery process information into delivery process information corresponding to the pickup process and delivery process information corresponding to the delivery process. Therefore, when executing the above step 303, the computer can divide the delivery process information into delivery process information corresponding to the pickup process and delivery process information corresponding to the delivery process, and process the delivery process information corresponding to the pickup process and delivery process information corresponding to the delivery process separately; or it can process the overall delivery process information.

[0094] In this embodiment of the application, the delivery process is divided into a pickup process and a delivery process. The pickup process and the delivery process are analyzed separately to determine the timeout factors that cause the first historical order to time out.

[0095] In this embodiment, the last node of the pickup and delivery processes is analyzed first. If a timeout occurs at the last node, the delivery process information is analyzed further to identify the timeout factors causing the timeout. If no timeout occurs at the last node, the analysis stops, confirming that the pickup or delivery process does not have a timeout issue. It should be noted that this embodiment only uses steps 303 to 309 as examples to illustrate "determining the timeout factors for the first historical order based on the delivery process information of the first historical order." In another embodiment, the pickup and delivery processes can be considered interchangeably. The computer device determines the timeout factors causing the first historical order to time out based on the delivery process information of the first historical order, including: querying the delivery process information of the first historical order based on at least one of multiple timeout factors to obtain the query result corresponding to the at least one timeout factor; if the query result corresponding to any timeout factor indicates that the delivery process information includes information matching the timeout factor, then the timeout factor is determined as the timeout factor causing the first historical order to time out.

[0096] Optionally, each timeout factor corresponds to a matching rule. When the information in the delivery process information meets the matching rule of a certain timeout factor, that timeout factor is the timeout factor that caused the first historical order to time out. Specifically, the query result corresponding to that timeout factor indicates that the delivery process includes information matching that timeout factor.

[0097] For example, the timeout factor is that the merchant has closed their store. The matching rule for this timeout factor is that the reported abnormal information includes information indicating that the merchant has closed their store. Based on this timeout factor, the computer equipment queries the abnormal reporting information in the delivery process information of the first historical order to obtain the query results corresponding to the timeout factor. If the query results corresponding to the timeout factor indicate that the abnormal information includes information indicating that the merchant has closed their store, then this timeout factor is determined to be the timeout factor that caused the first historical order to time out.

[0098] It should be noted that the embodiments of this application only take steps 303 to 309 as examples to illustrate the process of "analyzing the delivery process information corresponding to the pickup process of the first historical order to obtain a first analysis result; analyzing the delivery process information corresponding to the delivery process of the first historical order to obtain a second analysis result; and determining the timeout factor that caused the first historical order to time out based on the first analysis result and the second analysis result". In another embodiment, other methods can also be used to analyze the delivery process information corresponding to the delivery process and the pickup process. For example, based on at least one timeout factor, the delivery process information corresponding to the pickup process is queried to obtain a first query result corresponding to at least one timeout factor; based on at least one timeout factor, the delivery process information corresponding to the delivery process is queried to obtain a second query result corresponding to at least one timeout factor; and based on the first query result and the second query result, the timeout factor that caused the first historical order to time out is determined. Specifically, determining the timeout factor that caused the first historical order to time out based on the first query result and the second query result includes: if any query result corresponding to any timeout factor indicates that the delivery process information includes information that matches the timeout factor, then the timeout factor is determined as the timeout factor that caused the first historical order to time out.

[0099] 304. If the actual completion time of the pickup node is not later than the corresponding first predicted completion time, the computer device stops analyzing the delivery process information corresponding to the pickup process of the first historical order and obtains a first analysis result indicating that there is no timeout problem in the pickup process.

[0100] It should be noted that the pickup node is the last node in the pickup process. If the actual completion time of the pickup node is not later than the first predicted completion time, it means that the delivery capacity has completed the pickup process on time, and there is no need to analyze the delivery process information corresponding to the pickup process to obtain the timeout factor.

[0101] Another point to note is that the embodiments in this application are merely illustrative examples of determining whether to continue analyzing the delivery process corresponding to the pickup process based on the difference between the actual completion time and the first predicted completion time. In another embodiment, a deviation tolerance is set. If the difference between the actual completion time and the predicted completion time is within the deviation tolerance range, the pickup process of the delivery capacity can also be considered to be without problems, and there is no need to analyze the delivery process information corresponding to the pickup process to obtain the timeout factor. Optionally, if the actual completion time corresponding to the pickup node is not later than the sum of the corresponding first predicted completion time and the deviation tolerance, the computer device stops analyzing the delivery process information corresponding to the pickup process of the first historical order, and obtains a first analysis result indicating that there is no timeout problem in the pickup process.

[0102] 305. If the actual completion time of the pickup node is later than the corresponding first predicted completion time, the computer device continues to analyze the delivery process information corresponding to the pickup process of the first historical order to obtain the first analysis result used to represent the timeout factor.

[0103] If the actual completion time of the pickup node is later than the first predicted completion time, it means that the delivery capacity failed to complete the pickup process on time. Therefore, it is necessary to analyze the pickup process and find out the timeout factors that caused the pickup process to exceed the timeout.

[0104] The computer equipment can use any analysis method to continue analyzing the delivery process information corresponding to the pickup process of the first historical order. This application embodiment does not limit the analysis process. This application embodiment is only used as an example to illustrate the analysis process using the following three analysis methods.

[0105] In one possible implementation, the delivery process information includes the actual completion time and the first predicted completion time for each of the multiple nodes in the delivery process, and the first analysis result includes the timeout factors that cause the first historical order to have a timeout problem; the computer equipment analyzes the delivery process information corresponding to the pickup process of the first historical order to obtain the first analysis result, which includes at least one of the following:

[0106] (1) Query the predicted completion time and actual completion time of the previous node. If the actual completion time of the previous node is later than the predicted completion time, determine that the timeout factor is a hindrance to the previous node.

[0107] Since the pickup process involves multiple nodes, a timeout at an earlier node can affect subsequent nodes. For example, if a rider spends too much time on their way to the merchant, causing them to arrive late, and if the rider arrives at the store later than the predicted pickup time, it directly results in the rider's pickup timeout.

[0108] Therefore, if a node times out, you can first check if the previous node also times out. If the previous node also times out, then the reason why this node times out is because it was affected by the previous node.

[0109] (2) Query the actual delivery route and the planned delivery route of the first historical order. If the first historical order is a direct delivery order and the actual delivery route and the planned delivery route are inconsistent, determine that the timeout factor is due to inconsistent delivery routes.

[0110] There may be multiple routes from the rider's location to the merchant's location, and the length and time of these routes may vary. Therefore, the system plans a delivery route for the rider and determines the time required for the rider to complete each node based on the planned route. If the rider does not arrive at the corresponding node within the predicted time, it can be analyzed whether the rider followed the planned delivery route. If the rider did not follow the planned delivery route, it indicates that the delay may be due to the rider not adhering to the planned delivery route.

[0111] To more accurately analyze the factors causing timeouts, this application also considers that when a rider has multiple orders, the planned delivery route may become unsuitable as the number of orders increases. Therefore, the planned delivery route and the rider's actual delivery route are only considered when the order is a direct delivery order. A direct delivery order refers to an order where the rider currently has only this one order; in this case, delivering according to the planned delivery route is the fastest method.

[0112] (3) Query the abnormal reporting information of the first historical order. If the abnormal reporting information is not empty, determine the timeout factor based on the abnormal reporting information. If the abnormal reporting information is empty and the actual delivery route is consistent with the planned delivery route, analyze the predicted completion sub-time and actual completion sub-time corresponding to multiple nodes in the pickup process to obtain the timeout node and determine that the timeout factor is the optimistic estimate of the completion sub-time of the timeout node.

[0113] If any abnormal situation occurs during the delivery process, the delivery personnel can report it. The computer system will store the reported abnormal situation as an anomaly report. For example, if the delivery personnel arrive at the merchant's address but cannot find the store, they can report "Merchant location error." Similarly, if the delivery personnel arrive at the delivery point but cannot contact the customer, they can report "Customer unreachable."

[0114] Since these anomalies waste delivery time and cause delays, the analysis of delay factors can be conducted by analyzing the reported anomalies to determine the cause. For example, if the reported anomaly includes "merchant location error," then the delay factor can be identified as "merchant location error." Similarly, if the reported anomaly includes "unable to contact customer," then the delay factor can be identified as "unable to contact customer."

[0115] If the anomaly reporting information is empty, it means that no anomalies were reported by the delivery capacity, indicating that no anomalies occurred during the delivery process. If the planned delivery route and the actual delivery route are consistent, it means that the delivery capacity was used to deliver according to the planned route. If the anomaly reporting information is empty and the actual delivery route and the planned delivery route are consistent, it means that no anomalies occurred and the timeout is unrelated to the delivery capacity. In this case, the first prediction completion time determined by the prediction model is too optimistic.

[0116] It should be noted that the embodiments of this application are only used as an example to illustrate the analysis process by considering whether the actual delivery route and the planned delivery route are consistent. In another embodiment, the consistency between the actual delivery route and the planned delivery route is only considered when the historical order is a direct delivery order. If the historical order is not a direct delivery order, then there is no need to consider whether the actual delivery route is consistent with the planned delivery route.

[0117] It should be noted that the embodiments in this application are merely examples of the above three steps to illustrate the analysis process of delivery process information, and do not limit the analysis process.

[0118] In one possible implementation, the computer device executes the above three steps sequentially. Optionally, for the retrieval process, the following steps are executed sequentially for each node in reverse order: the computer device first queries the predicted completion time and actual completion time of the previous node; if the actual completion time of the previous node is later than the predicted completion time, the timeout factor is determined to have affected the previous node. Continue querying the actual and planned delivery routes of the first historical order. If the first historical order is a direct delivery order and the actual and planned delivery routes are inconsistent, determine that the timeout factor is a delivery route inconsistency and continue analyzing the next node. If the first historical order is not a direct delivery order, or if the actual and planned delivery routes are consistent, query the abnormal reporting information of the first historical order. If the abnormal reporting information is not empty, determine the timeout factor based on the abnormal reporting information. If the abnormal reporting information is empty and the actual and planned delivery routes are consistent, analyze the predicted completion sub-times and actual completion sub-times corresponding to multiple nodes in the pickup process to obtain the timeout node and determine that the timeout factor is an optimistic estimate of the completion sub-time of the timeout node.

[0119] Analysis of the object retrieval process, such as Figure 5As shown, the computer equipment first obtains the actual pickup time and the predicted pickup time. If the actual pickup time is not later than the sum of the predicted pickup time and the deviation tolerance, the analysis is stopped. If the actual pickup time is later than the sum of the predicted pickup time and the deviation tolerance, the actual completion time and the predicted completion time of the previous node are obtained. If the actual completion time of the previous node is later than the sum of the predicted completion time and the deviation tolerance, the timeout factor is determined to be a problem with the previous node, and the analysis continues. If the actual completion time of the next node is not later than the sum of the predicted completion time and the deviation tolerance, it is determined whether the historical order is a direct delivery order and whether the actual delivery route and the planned delivery route are consistent. If the historical order is a direct delivery order and the actual delivery route and the planned delivery route are inconsistent, the timeout factor is determined to be an inconsistent delivery route. If the historical order is not a direct delivery order, or if the actual delivery route matches the planned delivery route, then query the abnormal reporting information for this first historical order to determine if the merchant information is abnormal. If the merchant information is abnormal, then based on the abnormal reporting information, determine if the timeout factor is due to the merchant being closed / not operating or an incorrect merchant location. If the merchant information is not abnormal, then obtain the actual arrival time and the predicted arrival time. If the actual arrival time is later than the sum of the predicted arrival time and the deviation tolerance, then determine if the timeout factor is an optimistic arrival time estimate. If the actual arrival time is not later than the sum of the predicted arrival time and the deviation tolerance, then determine if the merchant has clicked "prepare". If the merchant has not clicked "prepare", then proceed to determine if the actual arrival time is later than the predicted preparation time. If the merchant has clicked "prepare", then determine if the merchant... If the merchant clicks "prepare food" earlier than the actual arrival time of the delivery vehicle, then determine if the actual pickup time is later than the sum of the actual food preparation time and the deviation tolerance. If it is later, the delay is determined to be due to human error caused by the merchant or the rider. If it is not later, proceed to the step of determining if the actual arrival time of the delivery vehicle is later than the predicted food preparation time.

[0120] If the actual arrival time of the delivery vehicle is later than the predicted food preparation time, then it is further determined whether the actual arrival time is later than the sum of the predicted food preparation time and the deviation tolerance. If it is later, the delay factor is determined to be an optimistic food preparation time prediction. If it is not later, it is determined whether the delivery vehicle has reported slow food preparation. If the delivery vehicle has reported slow food preparation, the delay factor is determined to be the merchant's slow food preparation and the analysis ends. If the delivery vehicle has not reported slow food preparation, the analysis also ends. If the actual arrival time of the delivery vehicle is not later than the predicted food preparation time, then the delivery vehicle is determined to be... If the actual arrival time at the store is later than the sum of the predicted pickup time and the deviation tolerance, then the timeout factor is determined to be an optimistic prediction of the food preparation time, and it is determined whether the delivery capacity has reported slow food preparation. If the actual arrival time at the store is not later than the sum of the predicted pickup time and the deviation tolerance, then it is determined whether the delivery capacity has reported slow food preparation. If the delivery capacity has reported slow food preparation, then the timeout factor is determined to be slow food preparation by the merchant, and the analysis ends. If the delivery capacity has not reported slow food preparation, the analysis also ends.

[0121] Another point to note is that the embodiments in this application are merely illustrative examples of determining whether to continue analyzing the delivery process corresponding to the pickup process based on the difference between the actual completion time and the first predicted completion time. In another embodiment, a deviation tolerance is set. If the difference between the actual completion time and the predicted completion time is within the deviation tolerance range, the pickup process of the delivery capacity can be considered to be without problems, and there is no need to analyze the delivery process information corresponding to the pickup process to obtain the timeout factor. Optionally, if the actual completion time corresponding to the pickup node is later than the sum of the corresponding first predicted completion time and the deviation tolerance, the computer device continues to analyze the delivery process information corresponding to the pickup process of the first historical order to obtain a first analysis result representing the timeout factor.

[0122] 306. The computer equipment obtains the actual completion time and the first predicted completion time of the delivery node from the delivery process information corresponding to the delivery process of the first historical order.

[0123] Step 306 is similar to step 303, and will not be described in detail here.

[0124] 307. If the actual completion time of the delivery node is not later than the corresponding first predicted completion time, the computer device stops analyzing the delivery process information corresponding to the delivery process of the first historical order and obtains a second analysis result indicating that there is no timeout problem in the delivery process.

[0125] It should be noted that the delivery node is the last node in the delivery process. If the actual completion time of the delivery node is not later than the first predicted completion time, it means that the delivery capacity has completed the pickup process on time, and there is no need to analyze the delivery process information corresponding to the pickup process to obtain the timeout factor.

[0126] Another point to note is that the embodiments in this application are merely illustrative examples of determining whether to continue analyzing the delivery process corresponding to the delivery process based on the difference between the actual completion time and the first predicted completion time. In another embodiment, a deviation tolerance is set. If the difference between the actual completion time and the predicted completion time is within the deviation tolerance range, the delivery process of the delivery capacity can be considered to be without problems, and there is no need to analyze the delivery process information corresponding to the delivery process to obtain the timeout factor. Optionally, if the actual completion time corresponding to the delivery node is not later than the sum of the corresponding first predicted completion time and the deviation tolerance, the computer device stops analyzing the delivery process information corresponding to the delivery process of the first historical order, and obtains a first analysis result indicating that there is no timeout problem in the delivery process.

[0127] 308. If the actual completion time of the delivery node is later than the first predicted completion time, the computer device continues to analyze the delivery process information corresponding to the delivery process of the first historical order to obtain a second analysis result to represent the timeout factor.

[0128] If the actual completion time of the delivery node is later than the first predicted completion time, it means that the delivery capacity failed to complete the delivery process on time. Therefore, it is necessary to analyze the delivery process and find out the timeout factors that caused the delivery process to exceed the timeout.

[0129] The computer equipment can use any analysis method to continue analyzing the delivery process information corresponding to the delivery process of the first historical order. This application embodiment does not limit the analysis process. This application embodiment is only used as an example to illustrate the analysis process using the following three analysis methods.

[0130] In one possible implementation, the delivery process information includes the actual completion time and the first predicted completion time for each of the multiple nodes in the delivery process; the first analysis result includes the timeout factors that cause the first historical order to have a timeout problem; the computer equipment analyzes the delivery process information corresponding to the delivery process of the first historical order to obtain a second analysis result, including at least one of the following:

[0131] (1) Query the predicted completion time and actual completion time of the previous node. If the actual completion time of the previous node is later than the predicted completion time, determine that the timeout factor is a hindrance to the previous node.

[0132] This step is the same as step (1) in step 305, and will not be repeated here.

[0133] (2) Query the actual delivery route and the planned delivery route of the first historical order. If the first historical order is a direct delivery order and the actual delivery route and the planned delivery route are inconsistent, determine that the timeout factor is due to inconsistent delivery routes.

[0134] This step is the same as step (2) in step 305, and will not be repeated here.

[0135] (3) Query the abnormal reporting information of the first historical order. If the abnormal reporting information is not empty, determine the timeout factor based on the abnormal reporting information. If the abnormal reporting information is empty and the actual delivery route is consistent with the planned delivery route, analyze the predicted completion sub-time and actual completion sub-time corresponding to multiple nodes in the pickup process to obtain the timeout node and determine that the timeout factor is the optimistic estimate of the completion sub-time of the timeout node.

[0136] This step is the same as step (3) in step 305, and will not be repeated here.

[0137] It should be noted that the embodiments in this application are merely examples of the above three steps to illustrate the analysis process of delivery process information, and do not limit the analysis process.

[0138] In one possible implementation, the computer device executes the above three steps sequentially. Optionally, for the delivery process, the following steps are executed sequentially for each node in reverse order: the computer device first queries the predicted completion time and actual completion time of the previous node; if the actual completion time of the previous node is later than the predicted completion time, the timeout factor is determined to be a hindrance to the previous node. Continue querying the actual and planned delivery routes of the first historical order. If the first historical order is a direct delivery order and the actual and planned delivery routes are inconsistent, determine that the timeout factor is a delivery route inconsistency and continue analyzing the next node. If the first historical order is not a direct delivery order, or if the actual and planned delivery routes are consistent, query the abnormal reporting information of the first historical order. If the abnormal reporting information is not empty, determine the timeout factor based on the abnormal reporting information. If the abnormal reporting information is empty and the actual and planned delivery routes are consistent, analyze the predicted completion sub-times and actual completion sub-times corresponding to multiple nodes in the delivery process to obtain the timeout node. Determine that the timeout factor is an optimistic estimate of the completion sub-time of the timeout node.

[0139] Analysis of the delivery process, such as Figure 6 As shown, the computer equipment first obtains the actual delivery time and predicted delivery time of the delivery capacity. If the actual delivery time is not later than the sum of the predicted delivery time and the deviation tolerance, the analysis is terminated. If the actual delivery time is later than the sum of the predicted delivery time and the deviation tolerance, the actual completion time and predicted completion time of the previous node are obtained. If the actual completion time of the previous node is later than the sum of the predicted completion time and the deviation tolerance, the timeout factor is determined to be a problem affecting the previous node, and the analysis continues. If the actual completion time of the next node is not later than the sum of the predicted completion time and the deviation tolerance, it is determined whether the historical order is a direct delivery order and whether the actual delivery route and the planned delivery route are consistent. The steps of obtaining the actual arrival and departure times and the predicted arrival and departure times of the delivery capacity are then executed. If the historical order is a direct delivery order and the actual delivery route and the planned delivery route are inconsistent, the timeout factor is determined to be a delivery route inconsistency. If the historical order is not a direct delivery order, or the actual delivery route and the planned delivery route are consistent, the actual arrival time and the predicted arrival time of the delivery capacity are obtained. If the actual arrival time of the delivery capacity is later than the sum of the predicted arrival time and the deviation tolerance, the overtime factor is determined to be an optimistic estimate of the passenger arrival and riding time; if the actual arrival time of the delivery capacity is not later than the sum of the predicted arrival time and the deviation tolerance, the actual arrival and departure times and the predicted arrival and departure times of the delivery capacity are obtained.

[0140] If the actual arrival and departure times of the delivery capacity are later than the sum of the predicted arrival and departure times and the deviation tolerance, the delivery time forecast is considered optimistic, and the step of determining whether there are any special anomalies is performed; if the actual arrival and departure times of the delivery capacity are not later than the sum of the predicted arrival and departure times and the deviation tolerance, the step of determining whether there are any special anomalies is performed directly. Based on the anomaly reporting information, determine whether the following anomalies exist: (1) unable to contact the customer; (2) unable to click delivery in a weak network environment; (3) unable to enter the customer's location; (4) the customer noted the delivery time; (5) the customer's actual location does not match the map; if there are no special anomalies, obtain the actual delivery time and the predicted delivery time of the delivery capacity; if the actual delivery time of the delivery capacity is later than the sum of the predicted delivery time and the deviation tolerance, the delivery time forecast is considered optimistic, and the analysis ends; if the actual delivery time of the delivery capacity is not later than the sum of the predicted delivery time and the deviation tolerance, the analysis ends.

[0141] 309. Based on the first analysis results and the second analysis results, the computer equipment determines the timeout factors that caused the first historical order to time out.

[0142] In this analysis, at least one of the first or second analysis results can indicate the timeout factor that caused the first historical order to time out. If only one analysis result indicates the timeout factor that caused the first historical order to time out, and the other analysis result indicates that there is no timeout problem, then the timeout factor indicated by the analysis result is directly determined as the timeout factor that caused the first historical order to time out; if both analysis results indicate the timeout factor that caused the first historical order to time out, then both timeout factors indicated by the analysis results are used as the timeout factors that caused the first historical order to time out.

[0143] 310. If the timeout factor of the first historical order is due to human error, the computer equipment shall discard the historical order data of the first historical order.

[0144] It should be noted that some historical orders experienced timeouts due to human factors. Therefore, the failure to complete these historical orders on time was caused by human error, not by inaccurate prediction times determined by the prediction model. Consequently, using the historical order data of these orders as sample data is meaningless. In this embodiment, the historical order data of the first historical order whose timeout was due to human factors will be discarded.

[0145] For example, timeout factors such as "inconsistent delivery route for direct delivery orders", "store closed / not open for business", "incorrect merchant location", "unable to contact customer", "unable to click delivery in weak network environment", "customer's location cannot be accessed", "customer's specified delivery time", and "customer's actual location does not match the map" are all human factors.

[0146] 311. The computer equipment updates the prediction model based on the timeout factors of the remaining first historical orders and historical order data, and obtains the updated prediction model.

[0147] Since timeout issues in historical orders may be caused by inaccurate prediction of completion time determined by the prediction model, this application embodiment, after finding the first historical order with a timeout issue, also determines the timeout factors of the first historical order. Based on the timeout factors of the first historical order and historical order data, the prediction model is improved in a targeted manner to enhance the accuracy of the prediction model.

[0148] In this embodiment of the application, considering that different timeout factors have varying degrees of impact on business, different impact weights are assigned to historical order data for different timeout factors. In one possible implementation, the computer device updates the prediction model based on the remaining timeout factors and historical order data of the first historical order to obtain an updated prediction model. This includes: processing the delivery process information of the first historical order using the prediction model to obtain a second predicted completion time for the first historical order; determining a first weight corresponding to the timeout factor of the first historical order based on the correspondence between timeout factors and weights; and updating the prediction model based on the first weight, the second predicted completion time, and the actual completion time of the first historical order to obtain the updated prediction model.

[0149] Optionally, the computer device updates the prediction model based on the first weight, the second predicted completion time and the actual completion time of the first historical order, to obtain an updated prediction model, including: processing the actual completion time and the second predicted completion time based on the loss function of the prediction model to obtain a first loss value; processing the first loss value based on the first weight to obtain a fifth loss value; and updating the prediction model based on the fifth loss value to obtain the updated prediction model.

[0150] Optionally, the actual completion time includes the actual completion sub-times corresponding to multiple nodes in the delivery process, the second predicted completion time includes the second predicted completion sub-times corresponding to multiple nodes in the delivery process, and the timeout factor of the first historical order represents the timeout node that caused the timeout problem in the delivery process; the computer equipment updates the prediction model based on the first weight, the second predicted completion time and the actual completion time of the first historical order to obtain the updated prediction model, including: processing the actual completion time and the second predicted completion time based on the loss function of the prediction model to obtain a first loss value; processing the second predicted completion sub-time and the actual completion sub-time corresponding to the timeout node in the first historical order based on the node loss function to obtain a second loss value; processing the second loss value based on the first weight to obtain a third loss value; determining a fourth loss value based on the first loss value and the third loss value; and updating the prediction model based on the fourth loss value to obtain the updated prediction model.

[0151] The loss function of the prediction model can be any kind of loss function, and this application does not limit it.

[0152] The loss function for a node can be:

[0153]

[0154]

[0155] Where L is the loss value of the node, y is the predicted completion time for that node, and w is the actual completion time for that node; i This is the first loss value corresponding to the timeout factor of this node; δ i σ is the activation function; i Let M be the tolerance for deviation, and |M| represent the absolute value of M, where M is any numerical value.

[0156] It should be noted that, in this embodiment of the application, when updating the prediction model, only historical orders with timeout issues may be considered, or both historical orders with and without timeout issues may be considered simultaneously. In one possible implementation, the computer device updates the prediction model based on the timeout factors of the first historical order and the historical order data of the first historical order to obtain an updated prediction model. This includes: updating the prediction model based on the timeout factors of the first historical order, the historical order data of the first historical order, and the historical order data of the second historical order, where the second historical order is a historical order that has not timed out among multiple historical orders.

[0157] When updating the prediction model using historical order data from the first and second historical orders, the historical order data can be selected randomly or used alternately. Optionally, the computer device updates the prediction model based on the timeout factor of the first historical order, the historical order data of the first historical order, and the historical order data of the second historical order to obtain an updated prediction model. This includes: alternately using the historical order data of the second historical order, as well as the timeout factor and historical order data of the first historical order, to update the prediction model and obtain an updated prediction model.

[0158] For example, the prediction model is first updated using historical order data from a first historical order to obtain an updated prediction model; then, historical order data from a second historical order is used to update the updated prediction model again to obtain an updated prediction model; then, historical order data from a first historical order is used to update the updated prediction model again to obtain an updated prediction model, and so on.

[0159] Another point to note is that, in the embodiments of this application, the prediction model can not only determine the total time required for delivery, but also determine the predicted completion time for each node in the delivery process. Figure 7 As shown in the example, this application embodiment provides an exemplary description of the process by which the prediction model processes delivery process information to obtain the second predicted completion time.

[0160] The delivery process information may include data such as the delivery capacity, merchant, and delivery point allocated to the order. The computer equipment can obtain the following information from the delivery process information or from other information based on the delivery process information: (1) Delivery capacity characteristics, including: the number of orders delivered by the delivery capacity in the past, delivery on-time rate, average delivery time, average delivery distance, median speed, average speed, number of orders to be delivered, number of remaining orders to be picked up, number of remaining orders not delivered, and remaining delivery time. (2) Merchant characteristics, including: historical order volume, average meal preparation time, average meal pickup time of the delivery capacity, peak order volume, average meal preparation time during peak hours, meal pickup time of the delivery capacity during peak hours, merchant's menu categories, and average order acceptance time of the merchant. (3) Delivery point characteristics (the delivery point refers to the location where the delivery capacity delivers the goods to the customer), including: map coordinates of the delivery point, average delivery time, median delivery time, 80th percentile of delivery time, and 20th percentile of delivery time. The 80th percentile of delivery time refers to the 80th delivery time when sorted by magnitude. The 20th percentile of delivery time refers to the 20th delivery time when sorted by magnitude. Other information may also be included, but this embodiment does not limit this.

[0161] The prediction model extracts features from the above information to obtain order acceptance vector features, cycling vector features, entry and exit vector features, waiting for meals vector features, and delivery vector features. These vector features are then concatenated with the environmental vector features and path vector features of historical orders. After that, the concatenated vector features are processed by the MLP (Multilayer Perceptron) model to obtain the predicted completion time of each node and the predicted completion time of the entire delivery process.

[0162] The prediction model update method, apparatus, device, and storage medium provided in this application analyze historical orders processed by the prediction model. This not only identifies historical orders with timeout issues but also determines the timeout factors for those orders. By updating the prediction model using these timeout factors and historical order data, the predicted completion time determined by the prediction model becomes closer to the actual completion time of the historical orders, reducing the occurrence of order timeouts. Furthermore, by considering the specific timeout factors of historical orders, the prediction model can be specifically improved during the update process, thereby enhancing its accuracy.

[0163] In addition, the embodiments of this application also set different weights for different timeout factors. By using these weights, the prediction model can be updated more specifically, making the prediction model more targeted for optimization, thereby reducing the occurrence of timeout problems. While improving the accuracy of the prediction model, the frequency of timeout problems is reduced.

[0164] In addition, in this embodiment of the application, the first historical order with timeout problem is analyzed to obtain the timeout factor of the first historical order. The timeout factor can also be applied to other business scenarios to identify problems and make optimizations based on the timeout factor.

[0165] Figure 8 This is a schematic diagram of the structure of a prediction model update device provided in an embodiment of this application. See also... Figure 8 The device includes:

[0166] The acquisition module 801 is used to acquire historical order data of multiple historical orders. Each piece of historical order data includes a first predicted completion time determined for the corresponding historical order based on a prediction model, the actual completion time of the historical order, and delivery process information.

[0167] The first determining module 802 determines the first historical order with a timeout problem among the multiple historical orders based on the actual completion time and the first predicted completion time in each historical order data.

[0168] The second determining module 803 is used to determine the timeout factors that cause the first historical order to time out based on the delivery process information of the first historical order;

[0169] The update module 804 is used to update the prediction model based on the timeout factor of the first historical order and the historical order data of the first historical order, so as to obtain the updated prediction model.

[0170] like Figure 9 As shown, in one possible implementation, the update module 804 includes:

[0171] The processing unit 8041 is used to process the delivery process information of the first historical order through the prediction model to obtain the second predicted completion time of the first historical order.

[0172] The determining unit 8042 is used to determine the first weight corresponding to the timeout factor of the first historical order based on the correspondence between the timeout factor and the weight.

[0173] The update unit 8043 is used to update the prediction model based on the first weight, the second predicted completion time and the actual completion time of the first historical order, to obtain the updated prediction model.

[0174] In one possible implementation, the actual completion time includes the actual completion sub-times corresponding to multiple nodes in the delivery process, the second predicted completion time includes the second predicted completion sub-times corresponding to multiple nodes in the delivery process, and the timeout factor of the first historical order represents the timeout node that causes the timeout problem in the delivery process;

[0175] The update unit 8043 is configured to process the actual completion time and the second predicted completion time based on the loss function of the prediction model to obtain a first loss value; process the second predicted completion sub-time and the actual completion sub-time corresponding to the timeout node in the first historical order based on the node loss function to obtain a second loss value; process the second loss value based on the first weight to obtain a third loss value; determine a fourth loss value based on the first loss value and the third loss value; and update the prediction model based on the fourth loss value to obtain the updated prediction model.

[0176] In one possible implementation, the update module 804 is used to update the prediction model based on the timeout factor of the first historical order, the historical order data of the first historical order, and the historical order data of the second historical order, to obtain the updated prediction model, wherein the second historical order is the historical order that has not timed out among the plurality of historical orders.

[0177] In one possible implementation, the update module 804 is used to alternately use the historical order data of the second historical order and the timeout factor and historical order data of the first historical order to update the prediction model, thereby obtaining the updated prediction model.

[0178] In one possible implementation, the update module 804 includes:

[0179] The discarding unit 8044 is used to discard the historical order data of the first historical order if the timeout factor of the first historical order is due to human factors.

[0180] The update unit 8043 is used to update the prediction model based on the historical order data of the remaining first historical orders to obtain the updated prediction model.

[0181] In one possible implementation, the second determining module 803 is configured to query the delivery process information of the first historical order based on at least one of a plurality of timeout factors to obtain a query result corresponding to the at least one timeout factor; if the query result corresponding to any timeout factor indicates that the delivery process information includes information matching the timeout factor, the timeout factor is determined as the timeout factor that caused the first historical order to time out.

[0182] In one possible implementation, the second determining module 803 includes:

[0183] The first analysis unit 8031 ​​is used to analyze the delivery process information corresponding to the pickup process of the first historical order to obtain a first analysis result;

[0184] The second analysis unit 8032 is used to analyze the delivery process information corresponding to the delivery process of the first historical order to obtain a second analysis result;

[0185] The determining unit 8033 is used to determine the timeout factors that cause the first historical order to time out based on the first analysis result and the second analysis result.

[0186] In one possible implementation, the actual completion time includes the actual completion sub-times corresponding to multiple nodes in the delivery process, and the first predicted completion time includes the first predicted completion sub-times corresponding to multiple nodes in the delivery process.

[0187] The first analysis unit 8031 ​​is configured to obtain the actual completion time and the first predicted completion time of the pickup node from the delivery process information corresponding to the pickup process of the first historical order; if the actual completion time of the pickup node is not later than the corresponding first predicted completion time, the analysis of the delivery process information corresponding to the pickup process of the first historical order is stopped, and a first analysis result indicating that there is no timeout problem in the pickup process is obtained; if the actual completion time of the pickup node is later than the corresponding first predicted completion time, the analysis of the delivery process information corresponding to the pickup process of the first historical order is continued, and a first analysis result indicating the timeout factor is obtained.

[0188] In one possible implementation, the actual completion time includes the actual completion sub-times corresponding to multiple nodes in the delivery process, and the first predicted completion time includes the first predicted completion sub-times corresponding to multiple nodes in the delivery process; the first analysis result includes timeout factors that cause the first historical order to have a timeout problem.

[0189] The first analysis unit 8031 ​​is configured to perform at least one of the following:

[0190] Query the predicted completion time and actual completion time of the previous node. If the actual completion time of the previous node is later than the predicted completion time, determine that the timeout factor is a hindrance to the previous node.

[0191] Query the actual delivery route and planned delivery route of the first historical order. If the first historical order is a direct delivery order and the actual delivery route and the planned delivery route are inconsistent, determine that the timeout factor is inconsistent delivery routes.

[0192] Query the abnormal reporting information of the first historical order. If the abnormal reporting information is not empty, determine the timeout factor based on the abnormal reporting information. If the abnormal reporting information is empty and the actual delivery route is consistent with the planned delivery route, analyze the predicted completion sub-time and actual completion sub-time corresponding to multiple nodes in the pickup process to obtain the timeout node. Determine the timeout factor as the optimistic estimate of the completion sub-time of the timeout node.

[0193] It should be noted that the prediction model updating device provided in the above embodiments is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the prediction model updating device and the prediction model updating method embodiments provided in the above embodiments belong to the same concept, and their specific implementation process can be found in the method embodiments, which will not be repeated here.

[0194] In an exemplary embodiment, a computer device is provided, the computer device including one or more processors and one or more memories, the one or more memories storing at least one piece of program code, the at least one piece of program code being loaded and executed by the one or more processors to implement the prediction model update method as described in the above embodiments.

[0195] Optionally, the computer device is provided as a terminal. Figure 10This illustration shows a structural block diagram of a terminal 1000 provided in an exemplary embodiment of this application. The terminal 1000 may be a smartphone, tablet computer, MP3 player (Moving Picture Experts Group Audio Layer III), MP4 player (Moving Picture Experts Group Audio Layer IV), laptop computer, or desktop computer. The terminal 1000 may also be referred to as a user device, portable terminal, laptop terminal, desktop terminal, or other names.

[0196] The terminal 1000 includes a processor 1001 and a memory 1002.

[0197] Processor 1001 may include one or more processing cores, such as a quad-core processor, an octa-core processor, etc. Processor 1001 may be implemented using at least one hardware form selected from DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). Processor 1001 may also include a main processor and a coprocessor. The main processor, also known as a CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, processor 1001 may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the screen. In some embodiments, processor 1001 may also include an AI (Artificial Intelligence) processor, which is used to handle computational operations related to machine learning.

[0198] The memory 1002 may include one or more computer-readable storage media, which may be non-transitory. The memory 1002 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory devices. In some embodiments, the non-transitory computer-readable storage media in the memory 1002 are used to store at least one program code, which is executed by the processor 1001 to implement the prediction model update method provided in the method embodiments of this application.

[0199] In some embodiments, the terminal 1000 may also optionally include a peripheral device interface 1003 and at least one peripheral device. The processor 1001, memory 1002, and peripheral device interface 1003 can be connected via a bus or signal line. Each peripheral device can be connected to the peripheral device interface 1003 via a bus, signal line, or circuit board. Specifically, the peripheral device includes at least one of the following: a radio frequency circuit 1004, a display screen 1005, a camera 1006, an audio circuit 1007, a positioning component 1008, and a power supply 1009.

[0200] Peripheral device interface 1003 can be used to connect at least one I / O (Input / Output) related peripheral device to processor 1001 and memory 1002. In some embodiments, processor 1001, memory 1002 and peripheral device interface 1003 are integrated on the same chip or circuit board; in some other embodiments, any one or two of processor 1001, memory 1002 and peripheral device interface 1003 can be implemented on separate chips or circuit boards, which is not limited in this embodiment.

[0201] The radio frequency (RF) circuit 1004 is used to receive and transmit RF (Radio Frequency) signals, also known as electromagnetic signals. The RF circuit 1004 communicates with communication networks and other communication devices via electromagnetic signals. The RF circuit 1004 converts electrical signals into electromagnetic signals for transmission, or converts received electromagnetic signals back into electrical signals. Optionally, the RF circuit 1004 includes: an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a codec chipset, a user identity module card, etc. The RF circuit 1004 can communicate with other terminals via at least one wireless communication protocol. This wireless communication protocol includes, but is not limited to: metropolitan area networks (MANs), various generations of mobile communication networks (2G, 3G, 4G, and 5G), wireless local area networks (WLANs), and / or WiFi (Wireless Fidelity) networks. In some embodiments, the RF circuit 1004 may also include circuitry related to NFC (Near Field Communication), which is not limited in this application.

[0202] Display screen 1005 is used to display a UI (User Interface). The UI may include graphics, text, icons, videos, and any combination thereof. When display screen 1005 is a touch display screen, it also has the ability to collect touch signals on or above its surface. These touch signals can be input as control signals to processor 1001 for processing. In this case, display screen 1005 can also be used to provide virtual buttons and / or a virtual keyboard, also known as soft buttons and / or a soft keyboard. In some embodiments, there may be one display screen 1005, serving as the front panel of terminal 1000; in other embodiments, there may be at least two display screens, respectively disposed on different surfaces of terminal 1000 or in a folded design; in still other embodiments, display screen 1005 may be a flexible display screen, disposed on a curved or folded surface of terminal 1000. Furthermore, display screen 1005 may also be configured as a non-rectangular, irregular shape, i.e., a non-rectangular screen. The display screen 1005 can be made of materials such as LCD (Liquid Crystal Display) and OLED (Organic Light-Emitting Diode).

[0203] The camera assembly 1006 is used to acquire images or videos. Optionally, the camera assembly 1006 includes a front-facing camera and a rear-facing camera. The front-facing camera is disposed on the front panel of the terminal, and the rear-facing camera is disposed on the back of the terminal. In some embodiments, there are at least two rear-facing cameras, which are any one of a main camera, a depth-sensing camera, a wide-angle camera, and a telephoto camera, to achieve background blurring by fusing the main camera and the depth-sensing camera, panoramic shooting by fusing the main camera and the wide-angle camera, VR (Virtual Reality) shooting, or other fusion shooting functions. In some embodiments, the camera assembly 1006 may also include a flash. The flash can be a single-color temperature flash or a dual-color temperature flash. A dual-color temperature flash refers to a combination of a warm light flash and a cool light flash, which can be used for light compensation at different color temperatures.

[0204] The audio circuit 1007 may include a microphone and a speaker. The microphone is used to collect sound waves from the user and the environment, converting the sound waves into electrical signals that are input to the processor 1001 for processing, or input to the radio frequency circuit 1004 for voice communication. For stereo sound acquisition or noise reduction purposes, multiple microphones may be used, each positioned at a different location on the terminal 1000. The microphone may also be an array microphone or an omnidirectional microphone. The speaker is used to convert electrical signals from the processor 1001 or the radio frequency circuit 1004 into sound waves. The speaker may be a conventional diaphragm speaker or a piezoelectric ceramic speaker. When the speaker is a piezoelectric ceramic speaker, it can convert electrical signals not only into audible sound waves but also into inaudible sound waves for purposes such as distance measurement. In some embodiments, the audio circuit 1007 may also include a headphone jack.

[0205] The positioning component 1008 is used to determine the current geographical location of the terminal 1000 in order to enable navigation or LBS (Location Based Service). The positioning component 1008 can be a positioning component based on the US GPS (Global Positioning System), China's BeiDou system, Russia's Granas system, or the EU's Galileo system.

[0206] The power supply 1009 is used to power the various components in the terminal 1000. The power supply 1009 can be AC ​​power, DC power, a disposable battery, or a rechargeable battery. When the power supply 1009 includes a rechargeable battery, the rechargeable battery can support wired charging or wireless charging. The rechargeable battery can also be used to support fast charging technology.

[0207] In some embodiments, the terminal 1000 further includes one or more sensors 1010. The one or more sensors 1010 include, but are not limited to: an accelerometer 1011, a gyroscope 1012, a pressure sensor 1013, a fingerprint sensor 1014, an optical sensor 1015, and a proximity sensor 1016.

[0208] Accelerometer 1011 can detect the magnitude of acceleration along the three coordinate axes of a coordinate system established by terminal 1000. For example, accelerometer 1011 can be used to detect the components of gravitational acceleration along the three coordinate axes. Processor 1001 can control display screen 1005 to display the user interface in either a landscape or portrait view based on the gravitational acceleration signal acquired by accelerometer 1011. Accelerometer 1011 can also be used for games or for acquiring user motion data.

[0209] The gyroscope sensor 1012 can detect the orientation and rotation angle of the terminal 1000. The gyroscope sensor 1012, in conjunction with the accelerometer sensor 1011, can collect 3D motion data from the user on the terminal 1000. Based on the data collected by the gyroscope sensor 1012, the processor 1001 can perform the following functions: motion sensing (e.g., changing the UI based on the user's tilt), image stabilization during shooting, game control, and inertial navigation.

[0210] The pressure sensor 1013 can be disposed on the side bezel of the terminal 1000 and / or on the lower layer of the display screen 1005. When the pressure sensor 1013 is disposed on the side bezel of the terminal 1000, it can detect the user's grip signal on the terminal 1000, and the processor 1001 can perform left / right hand recognition or quick operation based on the grip signal collected by the pressure sensor 1013. When the pressure sensor 1013 is disposed on the lower layer of the display screen 1005, the processor 1001 can control the operable controls on the UI interface based on the user's pressure operation on the display screen 1005. The operable controls include at least one of button controls, scroll bar controls, icon controls, and menu controls.

[0211] The fingerprint sensor 1014 is used to collect a user's fingerprint. The processor 1001 identifies the user based on the fingerprint collected by the fingerprint sensor 1014, or vice versa. When the user's identity is identified as trusted, the processor 1001 authorizes the user to perform relevant sensitive operations, including unlocking the screen, viewing encrypted information, downloading software, making payments, and changing settings. The fingerprint sensor 1014 can be located on the front, back, or side of the terminal 1000. When the terminal 1000 has physical buttons or a manufacturer's logo, the fingerprint sensor 1014 can be integrated with the physical buttons or manufacturer's logo.

[0212] An optical sensor 1015 is used to collect ambient light intensity. In one embodiment, the processor 1001 can control the display brightness of the display screen 1005 based on the ambient light intensity collected by the optical sensor 1015. Specifically, when the ambient light intensity is high, the display brightness of the display screen 1005 is increased; when the ambient light intensity is low, the display brightness of the display screen 1005 is decreased. In another embodiment, the processor 1001 can also dynamically adjust the shooting parameters of the camera assembly 1006 based on the ambient light intensity collected by the optical sensor 1015.

[0213] The proximity sensor 1016, also known as a distance sensor, is installed on the front panel of the terminal 1000. The proximity sensor 1016 is used to detect the distance between the user and the front of the terminal 1000. In one embodiment, when the proximity sensor 1016 detects that the distance between the user and the front of the terminal 1000 is gradually decreasing, the processor 1001 controls the display screen 1005 to switch from a screen-on state to a screen-off state; when the proximity sensor 1016 detects that the distance between the user and the front of the terminal 1000 is gradually increasing, the processor 1001 controls the display screen 1005 to switch from a screen-off state to a screen-on state.

[0214] Those skilled in the art will understand that Figure 10 The structure shown does not constitute a limitation on terminal 1000 and may include more or fewer components than shown, or combine certain components, or use different component arrangements.

[0215] Alternatively, the computer device may be provided as a server. Figure 11 This is a schematic diagram of a server structure provided in an embodiment of this application. The server 1100 can vary considerably due to different configurations or performance. It may include one or more Central Processing Units (CPUs) 1101 and one or more memories 1102. The memory 1102 stores at least one line of program code, which is loaded and executed by the processor 1101 to implement the methods provided in the above-described method embodiments. Of course, the server may also have wired or wireless network interfaces, a keyboard, and input / output interfaces for input and output. The server may also include other components for implementing device functions, which will not be elaborated here.

[0216] The server 1100 is used to execute the steps performed by the server in the above method embodiments.

[0217] In an exemplary embodiment, a computer-readable storage medium is also provided, such as a memory including program code that can be executed by a processor in a computer device to perform the prediction model update method in the above embodiments. For example, the computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0218] In an exemplary embodiment, a computer program or computer program product is also provided, which includes computer program code that, when executed by a computer, causes the computer to implement the prediction model update method in the above embodiments.

[0219] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0220] The above description is merely an optional embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for updating a prediction model, characterized in that, The method includes: Acquire historical order data for multiple historical orders. Each historical order data item includes a first predicted completion time determined for the corresponding historical order based on a prediction model, the actual completion time of the historical order, and delivery process information. Based on the actual completion time and the first predicted completion time in each historical order data, the first historical order with a timeout problem is identified among the multiple historical orders. Based on the delivery process information of the first historical order, determine the timeout factors that caused the first historical order to time out; Based on the timeout factor of the first historical order and the historical order data of the first historical order, the prediction model is updated to obtain the updated prediction model.

2. The method according to claim 1, characterized in that, The step of updating the prediction model based on the timeout factor of the first historical order and the historical order data of the first historical order to obtain the updated prediction model includes: The delivery process information of the first historical order is processed using the prediction model to obtain the second predicted completion time of the first historical order. Based on the correspondence between timeout factors and weights, the first weight corresponding to the timeout factor of the first historical order is determined; Based on the first weight, the second predicted completion time and the actual completion time of the first historical order, the prediction model is updated to obtain the updated prediction model.

3. The method according to claim 2, characterized in that, The actual completion time includes the actual completion sub-times corresponding to multiple nodes in the delivery process, the second predicted completion time includes the second predicted completion sub-times corresponding to multiple nodes in the delivery process, and the timeout factor of the first historical order represents the timeout node that caused the timeout problem in the delivery process; the step of updating the prediction model based on the first weight, the second predicted completion time and the actual completion time of the first historical order to obtain the updated prediction model includes: Based on the loss function of the prediction model, the actual completion time and the second predicted completion time are processed to obtain a first loss value; Based on the node-based loss function, the second predicted completion sub-time and actual completion sub-time corresponding to the timeout nodes in the first historical order are processed to obtain the second loss value; Based on the first weight, the second loss value is processed to obtain the third loss value; Based on the first loss value and the third loss value, a fourth loss value is determined; Based on the fourth loss value, the prediction model is updated to obtain the updated prediction model.

4. The method according to claim 1, characterized in that, The step of updating the prediction model based on the timeout factor of the first historical order and the historical order data of the first historical order to obtain the updated prediction model includes: Based on the timeout factor of the first historical order, the historical order data of the first historical order, and the historical order data of the second historical order, the prediction model is updated to obtain the updated prediction model, wherein the second historical order is the historical order that has not timed out among the plurality of historical orders.

5. The method according to claim 4, characterized in that, The step of updating the prediction model based on the timeout factor of the first historical order, the historical order data of the first historical order, and the historical order data of the second historical order, to obtain the updated prediction model, includes: The prediction model is updated by alternately using the historical order data of the second historical order and the timeout factor and historical order data of the first historical order, resulting in the updated prediction model.

6. The method according to claim 1, characterized in that, The step of updating the prediction model based on the timeout factor of the first historical order and the historical order data of the first historical order to obtain the updated prediction model includes: If the timeout factor of the first historical order is due to human error, discard the historical order data of the first historical order; Based on the timeout factors of the remaining first historical orders and historical order data, the prediction model is updated to obtain the updated prediction model.

7. The method according to claim 1, characterized in that, The determination of timeout factors causing the first historical order to time out based on the delivery process information of the first historical order includes: Based on at least one of multiple timeout factors, the delivery process information of the first historical order is queried to obtain the query result corresponding to the at least one timeout factor; If the query result corresponding to any timeout factor indicates that the delivery process information includes information that matches the timeout factor, then the timeout factor is determined as the timeout factor that caused the first historical order to time out.

8. The method according to claim 1, characterized in that, The determination of timeout factors causing the first historical order to time out based on the delivery process information of the first historical order includes: The delivery process information corresponding to the pickup process of the first historical order is analyzed to obtain the first analysis result; The delivery process information corresponding to the delivery process of the first historical order is analyzed to obtain the second analysis result; Based on the first analysis results and the second analysis results, the timeout factors that caused the first historical order to time out are determined.

9. The method according to claim 8, characterized in that, The delivery process information includes the actual completion time and the first predicted completion time for each of the multiple nodes in the delivery process. The analysis of the delivery process information corresponding to the pickup process of the first historical order to obtain a first analysis result includes: Obtain the actual completion time and the first predicted completion time of the pickup node from the delivery process information corresponding to the pickup process of the first historical order; If the actual completion time of the pickup node is not later than the corresponding first predicted completion time, the analysis of the delivery process information corresponding to the pickup process of the first historical order is stopped, and a first analysis result is obtained to indicate that there is no timeout problem in the pickup process. If the actual completion time of the pickup node is later than the corresponding first predicted completion time, the delivery process information corresponding to the pickup process of the first historical order is further analyzed to obtain the first analysis result representing the timeout factor.

10. The method according to claim 8, characterized in that, The delivery process information includes the actual completion time and the first predicted completion time for each of the multiple nodes in the delivery process; the first analysis result includes the timeout factors that cause the first historical order to have a timeout problem. The analysis of the delivery process information corresponding to the pickup process of the first historical order to obtain a first analysis result includes at least one of the following: Query the predicted completion time and actual completion time of the previous node. If the actual completion time of the previous node is later than the predicted completion time, determine that the timeout factor is a hindrance to the previous node. Query the actual delivery route and planned delivery route of the first historical order. If the first historical order is a direct delivery order and the actual delivery route and the planned delivery route are inconsistent, determine that the timeout factor is inconsistent delivery routes. Query the abnormal reporting information of the first historical order. If the abnormal reporting information is not empty, determine the timeout factor based on the abnormal reporting information. When the abnormal reporting information is empty and the actual delivery route is consistent with the planned delivery route, the predicted completion time and actual completion time of multiple nodes in the pickup process are analyzed to obtain the timeout nodes, and the timeout factor is determined to be the optimistic prediction of the completion time of the timeout node.

11. An update device for a predictive model, characterized in that, The device includes: The acquisition module is used to acquire historical order data for multiple historical orders. Each historical order data includes a first predicted completion time determined for the corresponding historical order based on a prediction model, the actual completion time of the historical order, and delivery process information. The first determining module determines the first historical order with a timeout problem among the multiple historical orders based on the actual completion time and the first predicted completion time in each historical order data. The second determining module is used to determine the timeout factors that cause the first historical order to time out based on the delivery process information of the first historical order; The update module is used to update the prediction model based on the timeout factor of the first historical order and the historical order data of the first historical order, so as to obtain the updated prediction model.

12. A computer device, characterized in that, The computer device includes one or more processors and one or more memories, wherein at least one piece of program code is stored in the one or more memories, and the at least one piece of program code is loaded and executed by the one or more processors to perform the operations performed by the method for updating the prediction model as described in any one of claims 1 to 10.

13. A computer-readable storage medium, characterized in that, The storage medium stores at least one piece of program code, which is loaded and executed by a processor to perform the operations performed by the method for updating the prediction model as described in any one of claims 1 to 10.

14. A computer program, characterized in that, The computer program includes: computer program code, which, when executed by a computer, causes the computer to perform the operations performed by the method for updating the prediction model as described in any one of claims 1 to 10.

Images