Abnormal order detection method and system

Abstract

The embodiment of the invention provides an abnormal order detection method and system, and relates to the technical field of artificial intelligence, and the abnormal order detection method comprises the steps: obtaining the first logistics information of a first order; the first logistics information comprises first logistics track information and first order address information; calling an artificial intelligence model, and obtaining a cue word of the artificial intelligence model; the prompt word is used for indicating an anomaly detection strategy corresponding to the first order; inputting the prompt word and the first logistics information into an artificial intelligence model, and performing text analysis on the prompt word and the first logistics information through the artificial intelligence model to obtain an anomaly detection result of the first order; the anomaly detection result comprises whether the first order is an abnormal order. According to the method and the device, the effect of intelligently analyzing the abnormal order by using the artificial intelligence model is realized, the efficiency and the accuracy of abnormal order detection are improved, and the efficiency and the investment rate of logistics circulation are improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method and system for detecting abnormal orders. Background Art

[0002] In the actual operational flow of internet service platforms, some orders may display a logistics status of "received," but the user has not actually received the corresponding item. For these abnormal orders, manual verification of the order's logistics trajectory can reveal that the order has not actually reached the recipient. However, manual verification is not only labor-intensive and inefficient, but also prone to oversight of abnormal order detection. Therefore, a method for automatically detecting abnormal orders is needed. Summary of the Invention

[0003] The embodiments of the present application provide a method and system for detecting abnormal orders to alleviate or solve one or more technical problems existing in the prior art.

[0004] In a first aspect, an embodiment of the present application provides a method for detecting abnormal orders, comprising:

[0005] Obtaining first logistics information of a first order; the first logistics information includes first logistics track information and first order address information;

[0006] Invoking an artificial intelligence model and obtaining a prompt word of the artificial intelligence model; the prompt word is used to indicate an anomaly detection strategy corresponding to the first order;

[0007] The prompt word and the first logistics information are input into the artificial intelligence model, and the prompt word and the first logistics information are subjected to text analysis by the artificial intelligence model to obtain an anomaly detection result of the first order; the anomaly detection result includes: whether the first order is an abnormal order.

[0008] Optionally, before obtaining the logistics information of the first order, the method further includes:

[0009] Obtaining second logistics information of historical abnormal orders; the second logistics information includes second logistics track information and second order address information;

[0010] Analyzing abnormality-related information of the historical abnormal orders based on the second logistics information; the abnormality-related information includes abnormality categories and / or abnormality reasons;

[0011] The prompt word is generated according to the abnormality association information.

[0012] Optionally, generating the prompt word according to the abnormality association information includes:

[0013] Determining the anomaly detection strategy based on the anomaly association information; the anomaly detection strategy is a strategy for detecting an order corresponding to logistics trajectory information including first content as an abnormal order; the first content including at least one of the following: a keyword used to characterize a rejection of receipt, trajectory information returned to the order's starting address, and abnormal address information that does not match the order's target address;

[0014] The prompt word is generated according to the anomaly detection strategy.

[0015] Optionally, the abnormality detection result further includes the abnormality category and / or abnormality cause;

[0016] The performing text analysis on the prompt word and the first logistics information by the artificial intelligence model to obtain an anomaly detection result of the first order includes:

[0017] In the case where it is determined that the first order is the abnormal order, determining the abnormality category and / or abnormality cause of the first order based on the abnormality detection strategy;

[0018] Among them, the abnormality category includes at least one of the following: refusal to sign, reverse logistics, and address abnormality; the abnormality reason includes at least one of the following: user refusal to sign, the first logistics track information does not match the first order address information, and the actual signing address does not match the order target address.

[0019] Optionally, the method further includes:

[0020] determining an accuracy rate of the artificial intelligence model based on anomaly detection results of the artificial intelligence model on the plurality of first orders;

[0021] When the accuracy rate is lower than or equal to a preset threshold, obtaining a third order with an abnormality detection result error;

[0022] The prompt word is optimized according to the abnormal association information of the third order.

[0023] Optionally, before obtaining the first logistics information of the first order, the method further includes:

[0024] Obtaining third-party logistics information of the candidate order; the third-party logistics information includes third-party logistics track information and third-party order address information;

[0025] According to the third logistics information, the candidate order that meets the first condition is determined to be the first order; the first condition includes: the order has been signed for, and the order logistics track does not match the order target address.

[0026] Optionally, obtaining the first logistics information of the first order includes:

[0027] Generate a request for obtaining the first logistics information; the request includes order identification information of the first order;

[0028] Sending the acquisition request to the logistics service end; the logistics service end is used to obtain the first logistics information based on the acquisition request, and send the first logistics information to the application service end;

[0029] Receive the first logistics information sent by the logistics service end.

[0030] Optionally, after performing text analysis on the prompt word and the first logistics information by the artificial intelligence model to obtain an anomaly detection result of the first order, the method further includes:

[0031] If it is determined that the first order is the abnormal order, generating a return and refund request for the first order; the return and refund request includes order identification information of the first order;

[0032] The return and refund request is sent to the logistics service end, so that the logistics service end generates a return and refund instruction for the first order according to the return and refund request, and sends the return and refund instruction to the transportation terminal allocated for the first order; the return and refund instruction includes the order identification information and the first logistics information; the return and refund instruction is used to instruct the transportation terminal to transport the order items of the first order from the actual receipt address to the order starting address according to the first logistics information.

[0033] In a second aspect, an embodiment of the present application provides a method for detecting abnormal orders, including:

[0034] Receiving a request from the application server for obtaining first logistics information of a first order; the obtaining request includes order identification information of the first order; the first logistics information includes first logistics track information and first order address information;

[0035] According to the order identification information, the first logistics information of the first order is obtained, and the first logistics information is sent to the application server, so that the application server calls the artificial intelligence model based on the first logistics information to obtain an anomaly detection result of the first order, and the anomaly detection result includes: whether the first order is an abnormal order.

[0036] Optionally, the method further includes:

[0037] receiving a return and refund request for the first order sent by the application server, wherein the return and refund request includes order identification information of the first order;

[0038] generating a return and refund instruction for the first order based on the return and refund request, wherein the return and refund instruction includes the order identification information and the first logistics information;

[0039] A transport terminal is allocated for the first order, and the return and refund instruction is sent to the transport terminal; the return and refund instruction is used to instruct the transport terminal to transport the order items of the first order from the actual receipt address to the order starting address according to the first logistics information.

[0040] In a third aspect, an embodiment of the present application provides an abnormal order detection device, comprising:

[0041] A first acquisition module is used to acquire first logistics information of a first order; the first logistics information includes first logistics track information and first order address information;

[0042] A calling module, configured to call an artificial intelligence model and obtain a prompt word of the artificial intelligence model; the prompt word is used to indicate an anomaly detection strategy corresponding to the first order;

[0043] A detection module is used to input the prompt word and the first logistics information into the artificial intelligence model, perform text analysis on the prompt word and the first logistics information through the artificial intelligence model, and obtain an abnormality detection result of the first order; the abnormality detection result includes: whether the first order is an abnormal order.

[0044] In a fourth aspect, an embodiment of the present application provides an abnormal order detection device, comprising:

[0045] a receiving module, configured to receive a request from an application server for obtaining first logistics information of a first order; the obtaining request includes order identification information of the first order; the first logistics information includes first logistics track information and first order address information;

[0046] The second acquisition module is used to obtain the first logistics information of the first order according to the order identification information, and send the first logistics information to the application server, so that the application server calls the artificial intelligence model based on the first logistics information to obtain an anomaly detection result of the first order, and the anomaly detection result includes: whether the first order is an abnormal order.

[0047] In a fifth aspect, an embodiment of the present application provides an abnormal order detection system, including:

[0048] AI server, used to store AI models;

[0049] A logistics service end, configured to store first logistics information of a first order; the first logistics information includes first logistics track information and first order address information;

[0050] An application server is used to obtain the first logistics information from the logistics server; call the artificial intelligence model from the artificial intelligence server and obtain a prompt word of the artificial intelligence model; the prompt word is used to indicate the anomaly detection strategy corresponding to the first order; input the prompt word and the first logistics information into the artificial intelligence model, and perform text analysis on the prompt word and the first logistics information through the artificial intelligence model to obtain an anomaly detection result of the first order; the anomaly detection result includes: whether the first order is an abnormal order.

[0051] Optionally, the application server includes:

[0052] A task management platform, configured to create an artificial intelligence detection task for the first order and send the artificial intelligence detection task to an application service platform and an artificial intelligence open platform;

[0053] The application service platform is configured to obtain the first logistics information from the logistics service end in response to the artificial intelligence detection task; and send the first logistics information to the artificial intelligence open platform;

[0054] The artificial intelligence open platform is used to call the artificial intelligence model from the artificial intelligence server in response to the artificial intelligence detection task and obtain the prompt word; input the prompt word and the first logistics information into the artificial intelligence model to obtain an anomaly detection result of the first order.

[0055] In a sixth aspect, an embodiment of the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory, wherein the processor implements any method of the embodiment of the present application when executing the computer program.

[0056] In a seventh aspect, an embodiment of the present application provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the method of any one of the embodiments of the present application is implemented.

[0057] In an eighth aspect, an embodiment of the present application provides a computer program product, including a computer program, which implements any method of the embodiments of the present application when executed by a processor.

[0058] According to the technical solution provided by the embodiment of the present application, by obtaining the first logistics information of the first order, the first logistics information includes the first logistics track information and the first order address information, and calling the artificial intelligence model to obtain the prompt word of the artificial intelligence model, the prompt word is used to indicate the abnormality detection strategy corresponding to the first order; then the prompt word and the first logistics information are input into the artificial intelligence model, and the artificial intelligence model performs text analysis on the prompt word and the first logistics information to obtain the abnormality detection result of the first order, and the abnormality detection result includes whether the first order is an abnormal order. It can be seen that based on the powerful text analysis and logical reasoning capabilities of the artificial intelligence model, by providing the artificial intelligence model with prompt words for indicating the abnormality detection strategy, the effect of intelligently analyzing abnormal orders using the artificial intelligence model can be achieved, so that the detection of abnormal orders no longer depends on manpower, avoiding the detection omission problem that is easy to occur when manually detecting orders, and improving the efficiency and accuracy of abnormal order detection. In addition, based on the high efficiency of artificial intelligence detection of abnormal orders, abnormal situations in the logistics track can be perceived in advance, thereby improving the efficiency and delivery rate of logistics flow and reducing the losses of ordering users.

[0059] The above description is only an overview of the technical solution of this application. In order to more clearly understand the technical means of this application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of this application more obvious and easy to understand, the specific implementation methods of this application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In the accompanying drawings, unless otherwise specified, the same reference numerals throughout the multiple drawings represent the same or similar components or elements. These drawings are not necessarily drawn to scale. It should be understood that these drawings only depict some embodiments according to the present application and should not be regarded as limiting the scope of the present application.

[0061] Figure 1 The following diagram shows an application scenario of the abnormal order detection method provided by an embodiment of the present application;

[0062] Figure 2 A flowchart of the abnormal order detection method provided by an embodiment of the present application is shown;

[0063] Figure 3 A flowchart of the abnormal order detection method provided by an embodiment of the present application is shown;

[0064] Figure 4 A block diagram of an abnormal order detection system provided by an embodiment of the present application is shown;

[0065] Figure 5 The following is a logical architecture diagram of the abnormal order detection system provided by an embodiment of the present application;

[0066] Figure 6 A system interaction diagram of the abnormal order detection system provided by an embodiment of the present application is shown;

[0067] Figure 7 A timing diagram of the abnormal order detection method provided in an embodiment of the present application is shown;

[0068] Figure 8 A block diagram of an abnormal order detection device provided in an embodiment of the present application is shown;

[0069] Figure 9 A block diagram of an abnormal order detection device provided in an embodiment of the present application is shown;

[0070] Figure 10 A block diagram of an electronic device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0071] Hereinafter, only certain exemplary embodiments are briefly described. As will be appreciated by those skilled in the art, the described embodiments may be modified in various ways without departing from the spirit or scope of the present application. Therefore, the drawings and description are to be regarded as illustrative in nature and not restrictive.

[0072] To facilitate understanding of the technical solutions of the embodiments of the present application, the following describes the related technologies of the embodiments of the present application. The following related technologies can be combined with the technical solutions of the embodiments of the present application as optional solutions, and all of them fall within the scope of protection of the embodiments of the present application.

[0073] The following terms will be used in the following text:

[0074] Delivery Rate: This refers to the ratio of the number of packages successfully delivered to recipients to the total number of packages delivered within a specific timeframe. The calculation formula is: Delivery Rate = Number of Packages Successfully Delivered / Total Number of Packages Delivered × 100%.

[0075] Artificial intelligence model: refers to a deep learning model with a large number of parameters in the field of artificial intelligence, which is usually trained through a large-scale data set, aiming to learn complex patterns and structures in the data, so as to achieve efficient processing and prediction of the data. The embodiment of the present application does not limit the model category of the artificial intelligence model. As long as it is a model with natural language processing and text analysis capabilities, it can be used to implement the abnormal order detection method of the embodiment of the present application. For example, an artificial intelligence model built using an open source engineering platform (such as model qwen2.5-72b or qwen2-72b, etc.) is used to implement the abnormal order detection method of the embodiment of the present application. The present application can directly use the open source artificial intelligence model to detect abnormal orders, or it can pre-fine-tune the parameters of the open source artificial intelligence model, and then use the fine-tuned artificial intelligence model to detect abnormal orders.

[0076] Prompts: These are textual inputs used to guide the generation or response of an AI model when interacting with it. Prompts can be questions, statements, commands, or other forms of text. Their purpose is to guide the AI ​​model to generate appropriate output, thereby helping the model understand the user's intent and generate more accurate and contextually relevant results.

[0077] The embodiments of this application are intended to provide a method and system for detecting abnormal orders, which utilize artificial intelligence models to intelligently analyze abnormal orders. This eliminates the need for manual labor to detect abnormal orders, avoids oversights that can easily occur when manually detecting orders, and improves the efficiency and accuracy of abnormal order detection. Furthermore, based on the high efficiency of artificial intelligence in detecting abnormal orders, abnormalities in logistics trajectories can be detected in advance, thereby improving the efficiency and delivery rate of logistics flows and reducing losses to ordering users.

[0078] Figure 1 The following diagram shows an application scenario of the abnormal order detection method provided by the embodiment of the present application. Figure 1As shown, the application scenario of the abnormal order detection method includes an artificial intelligence server, a logistics server, and an application server. The artificial intelligence server is used to store and manage artificial intelligence models. The logistics server is used to store order logistics information, which includes at least logistics trajectory information and order address information. The order address information includes at least the order origin and destination addresses. The application server is used to obtain the logistics information of the order to be detected from the logistics server, invoke the artificial intelligence model from the artificial intelligence server, and obtain the artificial intelligence model's prompt word, which indicates the anomaly detection strategy corresponding to the order to be detected. The prompt word and logistics information are then input into the artificial intelligence model. The artificial intelligence model performs text analysis on the prompt word and logistics information to obtain an anomaly detection result for the order to be detected. The anomaly detection result includes whether the order to be detected is an abnormal order. If the order to be detected is determined to be an abnormal order, the anomaly detection result may also include information such as the abnormality category and cause of the abnormal order. Through the interaction between the artificial intelligence server, the logistics server, and the application server, the artificial intelligence model is used to intelligently analyze abnormal orders. This eliminates the reliance on human labor for abnormal order detection, avoids the detection omissions that are common in manual order inspection, and improves the efficiency and accuracy of abnormal order detection. Furthermore, the high efficiency of detecting abnormal orders based on artificial intelligence enables abnormal situations in logistics tracks to be perceived in advance, thereby improving the efficiency and delivery rate of logistics flow and reducing the losses of ordering users.

[0079] It should be noted that the above-mentioned application scenarios or application examples provided in the embodiments of the present application are for ease of understanding, and the embodiments of the present application do not specifically limit the application of the technical solution. In addition, the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data need to comply with the relevant laws, regulations and standards of the relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0080] The following describes in detail the technical solution of this application and how it solves the aforementioned technical problems using specific embodiments. The several specific embodiments listed can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments. The following describes the embodiments of this application in detail with reference to the accompanying drawings.

[0081] Figure 2 The flowchart of the abnormal order detection method provided by the embodiment of the present application is shown as follows: Figure 2 As shown, this method can be applied to Figure 1The application server in the illustrated scenario includes the following steps S201, S202, and S203.

[0082] Step S201, obtaining first logistics information of a first order, where the first logistics information includes first logistics track information and first order address information.

[0083] Among them, the first logistics track information is used to represent the logistics track of the first order, and the first order address information includes the order start address and order target address of the first order.

[0084] Optionally, the first order includes all or part of the orders within a specified time period (eg, the last month).

[0085] Optionally, to narrow the scope of order anomaly detection, before executing step S201, a first order may be screened from multiple candidate orders and used as the order for anomaly detection using the artificial intelligence model. The candidate orders may be all or part of the orders within a specified time period (e.g., the last two months).

[0086] When selecting the first order from multiple candidate orders, the third logistics information of the candidate orders is first obtained. Based on the third logistics information, the candidate order that meets the first condition is determined to be the first order. The third logistics information includes the third logistics track information and third order address information of the candidate order. The third logistics track information is used to represent the logistics track of the candidate order, and the third order address information includes the order origin address and order destination address of the candidate order.

[0087] The first condition includes: the order has been signed for, and the order logistics track does not match the order target address. For example, in the third logistics information of the candidate order, it is shown that the candidate order has been signed for, but the signing address is not the order target address of the candidate order. In this case, the candidate order is considered to be the first order that needs to be detected for anomalies. For another example, in the third logistics information of the candidate order, it is shown that the candidate order has been signed for, but the logistics track has backtracked, such as being returned to the order starting address or the address of the previous logistics node after signing. In this case, the candidate order is considered to be the first order that needs to be detected for anomalies. It can be seen that although orders that are shown to have been signed for are usually considered normal orders, in fact, such orders often have various anomalies, such as incorrect signing addresses, backtracking of logistics tracks, etc. Therefore, by using an artificial intelligence model to perform anomaly detection on the first order that meets the first condition, abnormal orders that are easily overlooked can be accurately detected, thereby improving the order delivery rate.

[0088] Optionally, when obtaining the first logistics information, a request for obtaining the first logistics information is first generated, the request including the order identification information of the first order; the request for obtaining the first logistics information is then sent to the logistics service end, which is configured to obtain the first logistics information based on the request and send the first logistics information to the application service end; and the first logistics information sent by the logistics service end is then received. The order identification information may include at least one of the following: an order number, a unique order identification code, an order logistics number (such as a courier number), etc.

[0089] Step S202: call the artificial intelligence model and obtain the prompt word of the artificial intelligence model, where the prompt word is used to indicate the anomaly detection strategy corresponding to the first order.

[0090] Among them, the anomaly detection strategy can be determined based on the anomaly category and / or anomaly cause of historical abnormal orders, so as to generate prompt words according to the anomaly detection strategy. The generation method of the prompt words will be described in detail in the following embodiments.

[0091] The exception categories may include at least one of the following: refusal to sign, reverse logistics, and address exception. Examples of abnormal orders in the refusal to sign category are as follows: the user refuses to sign, the user returns the order, etc. Examples of abnormal orders in the reverse logistics category are as follows: the order is returned to the order starting address, the order is returned to the address of the previous logistics node, etc. Examples of abnormal orders in the address exception category are as follows: the actual signing address and the order target address do not match, the logistics track represented by the logistics track information does not match the order address information, etc. The cause of the exception may include at least one of the following: the user refuses to sign, the logistics track information does not match the order address information, the actual signing address does not match the order target address, etc.

[0092] In step S203, the prompt word and the first logistics information are input into the artificial intelligence model, and the artificial intelligence model performs text analysis on the prompt word and the first logistics information to obtain an anomaly detection result of the first order, where the anomaly detection result includes whether the first order is an abnormal order.

[0093] When inputting the prompt word and the first logistics information into the artificial intelligence model, any of the following input methods may be used:

[0094] Method 1: Input the prompt word and the first logistics information into the artificial intelligence model separately. For example, input the prompt word and the first logistics information sequentially through the same model input channel, or input the prompt word and the first logistics information in parallel through two different model input channels.

[0095] Method 2: Concatenate the prompt word and the first logistics information and input them into the artificial intelligence model. For example, concatenate the prompt word and the first logistics information in the order of "prompt word + first logistics information" or "first logistics information + prompt word", and input the concatenated information into the artificial intelligence model.

[0096] Optionally, when the anomaly detection result shows that the first order is an abnormal order, the anomaly detection result also includes the abnormal reason and / or abnormal category of the first order.

[0097] According to the technical solution provided by the embodiment of the present application, by obtaining the first logistics information of the first order, the first logistics information includes the first logistics track information and the first order address information, and calling the artificial intelligence model to obtain the prompt word of the artificial intelligence model, the prompt word is used to indicate the abnormality detection strategy corresponding to the first order; then the prompt word and the first logistics information are input into the artificial intelligence model, and the artificial intelligence model performs text analysis on the prompt word and the first logistics information to obtain the abnormality detection result of the first order, and the abnormality detection result includes whether the first order is an abnormal order. It can be seen that based on the powerful text analysis and logical reasoning capabilities of the artificial intelligence model, by providing the artificial intelligence model with prompt words for indicating the abnormality detection strategy, the effect of intelligently analyzing abnormal orders using the artificial intelligence model can be achieved, so that the detection of abnormal orders no longer depends on manpower, avoiding the detection omission problem that is easy to occur when manually detecting orders, and improving the efficiency and accuracy of abnormal order detection. In addition, based on the high efficiency of artificial intelligence detection of abnormal orders, abnormal situations in the logistics track can be perceived in advance, thereby improving the efficiency and delivery rate of logistics flow and reducing the losses of ordering users.

[0098] In some embodiments, before obtaining the logistics information of the first order, a prompt word is pre-generated and stored. The method for generating the prompt word may include the following steps A1, A2, and A3:

[0099] Step A1: Obtain the second logistics information of the historical abnormal order, where the second logistics information includes the second logistics track information and the second order address information.

[0100] The historical abnormal order may include historical orders that meet the second condition: the order has been signed for and the order's logistics track does not match the order's destination address. The second logistics track information is used to characterize the logistics track of the historical abnormal order, and the second order address information includes the order's origin address and destination address.

[0101] Step A2: Analyze the abnormality association information of the historical abnormal orders based on the second logistics information. The abnormality association information includes the abnormality category and / or the abnormality cause.

[0102] Exception categories may include at least one of the following: refused delivery, reverse logistics, and address anomaly. Exception reasons may include at least one of the following: user refused delivery, logistics tracking information does not match order address information, the actual delivery address does not match the order destination address, etc.

[0103] Step A3: Generate prompt words based on the abnormality-related information of historical abnormal orders.

[0104] Optionally, step A3 may be executed as the following steps A31 and A32:

[0105] Step A31: Determine an anomaly detection strategy based on the anomaly association information.

[0106] Among them, the anomaly detection strategy refers to a strategy for detecting orders corresponding to logistics trajectory information including the first content as abnormal orders. The first content may include at least one of the following: keywords used to characterize refusal to sign, trajectory information returned to the order starting address, and abnormal address information that does not match the order target address. Keywords used to characterize refusal to sign include "reject", "return", "withdrawal", etc. Abnormal address information that does not match the order target address may include: address information that is different from the order target address, and address information whose distance from the order target address exceeds a preset distance threshold.

[0107] Step A32: Generate prompt words according to the anomaly detection strategy.

[0108] The following are some abnormal related information of historical abnormal orders:

[0109] Historical abnormal order 1: The abnormal type is the rejection of receipt, and the abnormal reason is the user's rejection of receipt.

[0110] Historical abnormal order 2: The abnormal type is reverse logistics, and the abnormal reason is that the logistics track information contains track information of the return order starting address.

[0111] Historical abnormal order 3: The abnormal type is address abnormality, and the abnormal reason is that the actual receipt address is different from the order target address.

[0112] It should be noted that the above-mentioned exception causes and exception categories are only exemplary descriptions. In actual applications, the exception causes and exception categories can be expanded according to actual scenarios. For example, the exception category can also include the signatory exception class, and the corresponding exception cause can include the actual signatory being different from the order signatory corresponding to the order recipient information; and so on.

[0113] Based on the abnormal correlation information of historical abnormal orders, the following prompt words can be generated:

[0114] As a customer service representative at a courier company, you need to perform anomaly detection on orders. Your goals are as follows:

[0115] (1) Detect the existence of rejection, return, and cancellation scenarios in the logistics trajectory information, and include the trajectory of returning to the order starting address.

[0116] (2) Detect whether the logistics track information contains track information returning to the order's starting address, or whether there is an abnormal signature that may result in the actual user not receiving the ordered items.

[0117] (3) Detect that the actual receipt address in the logistics track information is different from the order target address.

[0118] (4) If at least one of the above objectives (1) to (3) exists and there is no successful re-delivery or successful modification of the order target address, it is determined to be an abnormal order and 1 is returned; otherwise, it is determined to be a normal order and 0 is returned.

[0119] (5) For orders that include a re-delivery scenario, if it is detected that the order has not been successfully signed for at the order destination address again, it is determined to be an abnormal order and returns 1; otherwise, it is determined to be a normal order and returns 0.

[0120] In this embodiment, the anomaly detection strategy includes the following strategies: determining an order that meets at least one of the above objectives (1) to (3) as an abnormal order; and, for an order that includes a re-delivery scenario, if the order is not successfully signed for again at the order target address, it is determined to be an abnormal order.

[0121] In this embodiment, the abnormality detection strategy is determined based on the abnormality category and / or abnormality cause of historical abnormal orders, and prompt words are generated according to the abnormality detection strategy. When the artificial intelligence model is used to detect abnormal orders, based on the abnormality detection strategy indicated by the prompt word, abnormal orders that meet the categories of refusal to sign, reverse logistics, address abnormality, etc. can be accurately detected. This not only improves the efficiency and accuracy of abnormal order detection, but also enables abnormal situations in the logistics trajectory to be perceived in advance, thereby improving the efficiency and delivery rate of logistics flow and reducing the losses of ordering users.

[0122] In some embodiments, the prompt word can be optimized in at least one of the following situations: a preset prompt word optimization frequency is met, and the accuracy of the artificial intelligence model is lower than or equal to a preset threshold.

[0123] For the first case, when the preset prompt word optimization frequency is met, the prompt words are optimized, which can optimize the prompt words at a certain frequency, thereby helping to improve the artificial intelligence model's ability to understand order-related texts (such as order address information, logistics track information, order consignee information, etc.), and improve the accuracy of abnormal order detection. For example, if the preset prompt word optimization frequency is 2 months, the prompt words are optimized once every 2 months. Optionally, the prompt word optimization method can be: screening historical orders that are detected incorrectly by the artificial intelligence model, and analyzing the abnormal related information of this part of the historical orders, thereby optimizing the prompt words according to the abnormal related information of this part of the historical orders. The abnormal related information includes the abnormal category and / or the abnormal reason. Artificial intelligence model detection error refers to the situation where a normal order is detected as an abnormal order by the artificial intelligence model, or an abnormal order is detected as a normal order.

[0124] For the second case, when the accuracy of the artificial intelligence model is lower than or equal to the preset threshold, the prompt words are optimized. This enables the artificial intelligence model to improve the accuracy of the artificial intelligence model by optimizing the prompt words in a timely manner when the model performance does not meet the requirements, thereby ensuring the accuracy of abnormal order detection. Optionally, first, the accuracy of the artificial intelligence model is determined based on the abnormality detection results of multiple first orders by the artificial intelligence model; secondly, when the accuracy is lower than or equal to the preset threshold, a third order with an incorrect abnormality detection result is obtained; and thirdly, the prompt words are optimized based on the abnormality-related information of the third order. The abnormality-related information includes the abnormality category and / or the abnormality cause. Among them, the abnormality detection strategy can be optimized first based on the abnormality-related information of the third order, and then the prompt words can be optimized based on the optimized abnormality detection strategy.

[0125] The following are examples of optimized prompt words:

[0126] As a customer service representative at a courier company, you need to perform anomaly detection on orders. Your goals are as follows:

[0127] (1) Detect the existence of rejection, return, and cancellation scenarios in the logistics trajectory information, and include the trajectory information of returning to the order starting address.

[0128] (2) Detect whether the logistics track information contains track information returning to the order's starting address, or whether there is an abnormal signature that may result in the actual user not receiving the ordered items.

[0129] (3) Detect that the actual receipt address in the logistics track information is different from the order target address.

[0130] (4) If at least one of the above objectives (1) to (3) exists and there is no successful re-delivery or successful modification of the order target address, it is determined to be an abnormal order and 1 is returned; otherwise, it is determined to be a normal order and 0 is returned.

[0131] (5) For orders that include a re-delivery scenario, if it is detected that the order has not been successfully signed for at the order target address again, it is determined to be an abnormal order and returns 1; otherwise, it is determined to be a normal order and returns 0.

[0132] (6) If it is detected as an abnormal order, please determine whether it meets the following scenarios: abnormal information of the signatory, refusal of the signatory, re-delivery by the courier, placement in the express cabinet, signing for at the front desk, signing for by family members or colleagues, etc. For any of the above scenarios, if the logistics track information does not include the track information of returning to the starting address of the order, it is determined to be a normal order and 0 is returned; if the logistics track information includes the track information of returning to the starting address of the order, it is determined to be an abnormal order and 1 is returned.

[0133] Please note that the following special scenarios may exist:

[0134] Scenario A: If the order is detected as an abnormal order, but the order is ultimately signed for at the order target address, or at an address whose distance from the order target address does not exceed the preset distance threshold, it is determined to be a normal order and returns 0; if the order is signed for at an address whose distance from the order target address exceeds the preset distance threshold, it is determined to be an abnormal order and returns 1.

[0135] Scenario B: If the order start address and the order target address are in the same province but not the same city, if the order is signed for at the order start address, it is determined to be an abnormal order and 1 is returned; otherwise, it is determined to be a normal order and 0 is returned.

[0136] Scenario C: The courier is rerouted after the order is rejected by the recipient. For example, a courier is sent from City 1 to City 2, but is rejected and ultimately received in City 3. If City 1 and City 2 are in different provinces, but City 2 and City 3 are in the same province, then this order can be considered normal and the result is 0.

[0137] In this embodiment, the optimized anomaly detection strategy may include the following strategies: determining orders that meet at least one of the above objectives (1) to (3) as abnormal orders; for orders that include a re-delivery scenario, if the order is not successfully signed for again at the order target address, it is determined to be an abnormal order; and, for cases detected as abnormal orders, further detecting whether the following scenarios are met: abnormal information of the signatory, refusal of the signatory, re-delivery by the courier, placement in the courier cabinet, signing for at the front desk, signing for by family members and colleagues, etc. For orders that meet any of the above scenarios, if the logistics track information contains track information returning to the order starting address, it is determined to be an abnormal order.

[0138] It should be noted that the prompt words and anomaly detection strategies listed above are only exemplary. In actual applications, the prompt words and anomaly detection strategies can be customized, adjusted, and optimized according to actual scenarios.

[0139] As can be seen from the above example, the optimized prompt words have been expanded to include target (6) and three special scenarios, namely, scenarios A, B, and C, compared to the pre-optimized prompt words. The added content of the optimized prompt words is the optimized anomaly detection strategy based on the abnormal correlation information of the third order. During the use of the artificial intelligence model, as the number of abnormal order detections increases, by continuously optimizing the prompt words, the artificial intelligence model can continuously improve the accuracy of abnormal order detection based on higher-quality prompt words.

[0140] In some embodiments, the anomaly detection result further includes an anomaly category and / or an anomaly cause. When the artificial intelligence model performs text analysis on the prompt word and the first logistics information to obtain an anomaly detection result for the first order (i.e., executing step S203), the following steps may be executed: if the first order is determined to be an abnormal order, determine the anomaly category and / or anomaly cause for the first order based on the anomaly detection strategy.

[0141] Among them, the exception categories include at least one of the following: refusal to sign, reverse logistics, and address exception; the exception reasons include at least one of the following: user refusal to sign, the first logistics track information does not match the first order address information, and the actual signing address does not match the order target address.

[0142] Optionally, in the predetermined anomaly detection strategies, each strategy corresponds to one or more anomaly categories and anomaly causes. When determining the anomaly category and / or anomaly cause of the first order based on the anomaly detection strategies, the strategy that was hit when the first order was detected as an abnormal order may be first determined, and then the corresponding anomaly category and anomaly cause may be determined based on the hit strategy.

[0143] For example, when the first order is detected as an abnormal order, the strategy hit is "If it is detected that the actual receipt address in the logistics track information is different from the order target address, then the corresponding order is determined to be an abnormal order". It can be determined that the abnormality category of the order is the address abnormality category, and the cause of the abnormality is that the actual receipt address in the logistics track information is different from the order target address.

[0144] In this embodiment, when it is determined that the first order is an abnormal order, by further determining the abnormal category and / or abnormal cause of the first order, not only can the abnormal order be detected through the artificial intelligence model, but the abnormal category and / or abnormal cause can also be analyzed for the abnormal order, providing accurate data basis for subsequent processing of abnormal orders.

[0145] In some embodiments, if the first order is determined to be an abnormal order, a return and refund request for the first order may be generated and sent to the logistics service end. The logistics service end then generates a return and refund instruction for the first order based on the return and refund request and sends the return and refund instruction to the transport terminal assigned to the first order.

[0146] Among them, the return and refund request includes the order identification information of the first order, and the return and refund instruction includes the order identification information and the first logistics information of the first order. The return and refund instruction is used to instruct the transportation terminal to perform the following actions: according to the first logistics information, the order items of the first order are transported from the actual receipt address to the order start address. The order identification information may include at least one of the following: order number, order unique identification code, order logistics number (such as express number), etc. The first logistics information includes the first logistics track information and the first order address information. The first logistics track information is used to characterize the logistics track of the first order. The first order address information includes the order start address and order target address of the first order.

[0147] In this embodiment, by actively generating a return and refund request for the abnormal order and sending the return and refund request to the logistics service end, the logistics service end can be triggered to actively allocate a transportation terminal for the first order and instruct the transportation terminal to complete the return and refund event of the first order. In this way, the return and refund event of the abnormal order is automatically executed without the user's perception, which simplifies the user's subsequent processing operations on the abnormal order and improves the user's ordering experience.

[0148] Figure 3 A flowchart of an abnormal order detection method provided by another embodiment of the present application is shown as follows: Figure 3 As shown, this method can be applied to Figure 1 The logistics service end in the illustrated scenario includes the following steps S301 and S302.

[0149] Step S301: Receive a request from an application server for obtaining first logistics information of a first order, where the request includes order identification information of the first order.

[0150] The first logistics information includes first logistics track information and first order address information. The first logistics track information is used to represent the logistics track of the first order, and the first order address information includes the order origin address and order destination address of the first order. The order identification information may include at least one of the following: an order number, a unique order identification code, an order logistics number (such as a courier number), etc.

[0151] Step S302: Obtain the first logistics information of the first order according to the order identification information, and send the first logistics information to the application server, so that the application server calls the artificial intelligence model based on the first logistics information to obtain the anomaly detection result of the first order. The anomaly detection result includes: whether the first order is an abnormal order.

[0152] According to the technical solution provided in the embodiment of the present application, by receiving a request from the application server for obtaining the first logistics information of the first order, obtaining the first logistics information of the first order, and sending the first logistics information to the application server, the application server can call the artificial intelligence model based on the first logistics information to detect abnormal orders, thereby realizing the effect of using the artificial intelligence model to intelligently analyze abnormal orders, so that the detection of abnormal orders no longer relies on manpower, avoiding the problem of detection omissions that are easy to occur when manually detecting orders, and improving the efficiency and accuracy of abnormal order detection. In addition, the high efficiency of abnormal order detection based on artificial intelligence allows abnormal situations in the logistics trajectory to be perceived in advance, thereby improving the efficiency and delivery rate of logistics flow and reducing the losses of ordering users.

[0153] In some embodiments, the following steps B1, B2, and B3 may also be performed:

[0154] Step B1: Receive a return and refund request for a first order sent by an application server, where the return and refund request includes order identification information of the first order.

[0155] The order identification information may include at least one of the following: order number, order unique identification code, order logistics number (such as express delivery number), etc. The return and refund request is automatically generated by the application server when it is detected that the first order is an abnormal order.

[0156] Step B2: Based on the return and refund request, generate a return and refund instruction for the first order, where the return and refund instruction includes the order identification information and the first logistics information of the first order.

[0157] Step B3: Allocate a transport terminal for the first order and send a return and refund instruction to the transport terminal.

[0158] Among them, the return and refund instruction is used to instruct the transportation terminal to transport the order items of the first order from the actual receipt address to the order starting address according to the first logistics information.

[0159] Optionally, when the logistics service end allocates a transport terminal for the first order, it may determine the transport terminal serving the first order based on at least one of the following information: the order address information of the first order, the order item information of the first order, the transportation status information of the transport terminal, the transportation space information of the transport terminal, the location information of the transport terminal, etc. The order address information of the first order includes the order start address and the order destination address of the first order. The order item information includes information such as the item size and item type of the order item. The transportation status information of the transport terminal includes information such as whether the item is in transit and whether it can be transported normally. The transportation space information is used to represent the currently remaining space of the transport terminal and can be used to measure whether the order items can be carried.

[0160] For example, if the distance between the actual delivery address of the first order and the current location of the transport terminal is less than or equal to the preset distance threshold, and the transportation space of the transport terminal is sufficient to carry the order items of the first order, the transport terminal can be allocated to the first order.

[0161] If the return trajectory of the first order substantially matches the transport trajectory of the transport terminal, for example, if there is at least 80% overlap, and the transport terminal has sufficient space to carry the items for the first order, the transport terminal can be assigned to the first order. Return trajectory information includes the trajectory from the actual receipt address back to the order's originating address.

[0162] In this embodiment, after receiving the return and refund request for the first order sent by the application server, a return and refund instruction for the first order is generated, and the return and refund instruction is sent to the transportation terminal to instruct the transportation terminal to transport the order items of the first order from the actual receipt address to the order starting address, thereby automatically executing the return and refund event of the abnormal order without the user's perception, simplifying the user's subsequent processing operations on the abnormal order and improving the user's ordering experience.

[0163] Figure 4 A block diagram of an abnormal order detection system provided by an embodiment of the present application is shown in FIG. Figure 4 As shown, the abnormal order detection system includes an artificial intelligence server 41, a logistics server 42, and an application server 43, wherein:

[0164] The artificial intelligence server 41 is used to store artificial intelligence models.

[0165] The logistics service terminal 42 is used to store the first logistics information of the first order; the first logistics information includes the first logistics track information and the first order address information.

[0166] The application server 43 is used to obtain the first logistics information from the logistics server 42; call the artificial intelligence model from the artificial intelligence server 41 and obtain the prompt word of the artificial intelligence model; the prompt word is used to indicate the anomaly detection strategy corresponding to the first order; the prompt word and the first logistics information are input into the artificial intelligence model, and the prompt word and the first logistics information are subjected to text analysis by the artificial intelligence model to obtain the anomaly detection result of the first order; the anomaly detection result includes: whether the first order is an abnormal order.

[0167] In this embodiment, through the interaction between the AI ​​server, logistics server, and application server, an AI model is used to intelligently analyze abnormal orders. This eliminates the need for manual labor to detect abnormal orders, avoids the oversights that often occur with manual order checking, and improves the efficiency and accuracy of abnormal order detection. Furthermore, the high efficiency of AI-based abnormal order detection enables the early detection of anomalies in logistics trajectories, thereby improving the efficiency and delivery rate of logistics flows and reducing losses for ordering users.

[0168] In some embodiments, the application server 43 includes a task management platform, an artificial intelligence open platform, and an application service platform, wherein:

[0169] The task management platform is used to create the AI ​​inspection task for the first order and send the AI ​​inspection task to the application service platform and the AI ​​open platform;

[0170] The application service platform is configured to obtain first logistics information from the logistics service end in response to the artificial intelligence detection task; and send the first logistics information to the artificial intelligence open platform;

[0171] The artificial intelligence open platform is used to respond to the artificial intelligence detection task, call the artificial intelligence model from the artificial intelligence server, and obtain the prompt word; input the prompt word and the first logistics information into the artificial intelligence model to obtain the anomaly detection result of the first order.

[0172] Figure 5 The following is a logical architecture diagram of the abnormal order detection system provided by the embodiment of the present application. Figure 5 As shown, the overall logical architecture of the abnormal order detection system may include: basic service layer, data layer and application layer. Among them:

[0173] The basic service layer is used to provide at least one of the following: a logistics information query platform, a database, an artificial intelligence model basic platform, an RPC (Remote Procedure Call) basic service, and an artificial intelligence open platform. The logistics information query platform is used to query and obtain the logistics information of an order, and the logistics information query platform is configured on the logistics server. The database is configured on the application server and is used to store the data involved when the application server executes the abnormal order detection method, including the logistics information obtained from the logistics server, pre-generated prompt words, the abnormal detection results of the first order, etc. The RPC basic service is used to provide the application server with the ability to remotely call the artificial intelligence model, so that the application server can call the artificial intelligence model from the artificial intelligence server. The artificial intelligence model basic platform is configured on the artificial intelligence server and is used to store and manage artificial intelligence models. The artificial intelligence open platform is configured on the application server and is used to provide the ability to establish a connection between the application server and the artificial intelligence server, so that the artificial intelligence model managed by the artificial intelligence server can be applied to the application server. This application does not limit the type of database, which can be any database with data storage capabilities, such as a MySQL database.

[0174] The data layer is used to provide services related to order data, logistics information and other data (hereinafter referred to as data-related services). Data-related services may include at least one of the following: obtaining logistics information of the first order; if the first order is determined to be an abnormal order, executing a return and refund event for the first order.

[0175] The application layer is equipped with a task management platform, which is used to create and execute AI inspection tasks for the first order. When executing the AI ​​inspection task, the application layer uses the logistics information acquisition capabilities provided by the data layer to obtain the first logistics information for the first order from the logistics information query platform. Utilizing the AI ​​open platform and RPC basic services of the basic service layer, the application layer invokes the AI ​​model from the AI ​​model foundation platform, retrieves the prompt word from the database, and inputs the prompt word and the first logistics information into the AI ​​model. The AI ​​model then performs text analysis on the prompt word and the first logistics information to obtain an anomaly detection result for the first order.

[0176] Optionally, the application layer can schedule orders regularly and create corresponding AI inspection tasks for these scheduled orders. For example, if the preset scheduling period is one month, the application layer will schedule orders generated within the past month every month and create AI inspection tasks for these scheduled orders.

[0177] Optionally, the application layer can schedule orders within a specified time period and create corresponding AI inspection tasks for these scheduled orders. For example, if the specified time period is set to the last two months, the application layer will schedule orders generated within the last two months and create AI inspection tasks for these scheduled orders.

[0178] based on Figure 5 The system logical architecture shown in the figure is Figure 6 The system interaction diagram of the abnormal order detection system of the embodiment of the present application is shown. Figure 6 As shown, the abnormal order detection system includes a task management platform, an application service platform, an artificial intelligence open platform, a logistics service end, and an artificial intelligence service end. The task management platform, the application service platform, and the artificial intelligence open platform are configured on the application service end. During the system interaction process, the task management platform is used to create an artificial intelligence detection task for the first order. Based on the artificial intelligence detection task, it triggers the application service platform to pull order data, including the first logistics information of the first order. Triggered by the artificial intelligence detection task, the logistics service end queries the first logistics information. The application service platform sends the acquired first logistics information to the artificial intelligence development platform. The task management platform initiates the artificial intelligence detection task to the artificial intelligence open platform and configures prompt words for the artificial intelligence open platform so that the artificial intelligence development platform, based on the first logistics information and prompt words, calls the artificial intelligence model from the artificial intelligence server end to detect anomalies in the first order. The artificial intelligence server end returns the anomaly detection results of the first order to the artificial intelligence open platform.

[0179] Optionally, the artificial intelligence open platform can be used to store and manage artificial intelligence detection tasks created by the task management platform. When the number of artificial intelligence detection tasks is large, the artificial intelligence open platform can execute each artificial intelligence detection task in sequence according to the order in which the artificial intelligence detection tasks are created.

[0180] Optionally, the artificial intelligence open platform is also used to store prompt words of the artificial intelligence model, so that each time an artificial intelligence detection task is executed, it is only necessary to obtain the logistics information of the order corresponding to the artificial intelligence detection task without repeatedly obtaining the prompt words.

[0181] Figure 7 The timing diagram of the abnormal order detection method provided by the embodiment of the present application is shown as follows: Figure 7 As shown, the abnormal order detection method includes the following steps S7.1 to S7.7:

[0182] Step S7.1: The application server starts the artificial intelligence detection task for the first order.

[0183] Optionally, the task management platform configured by the application server creates an artificial intelligence detection task for the first order, and triggers the application service platform and the artificial intelligence development platform to interact and execute the artificial intelligence detection task.

[0184] In step S7.2, the application server sends a request for obtaining the first logistics information of the first order to the logistics server based on the artificial intelligence detection task.

[0185] The first logistics information includes first logistics track information and first order address information. The first logistics track information is used to represent the logistics track of the first order, and the first order address information includes the order start address and order destination address of the first order.

[0186] In step S7.3, the logistics server queries the first logistics information of the first order based on the acquisition request, and sends the first logistics information to the application server.

[0187] The acquisition request includes the order identification information of the first order, and the order identification information may include at least one of the following: order number, order unique identification code, order logistics number (such as express delivery number), etc.

[0188] In step S7.4, the application server calls the artificial intelligence model from the artificial intelligence server to perform anomaly detection on the first order through the artificial intelligence model.

[0189] The application server invokes the AI ​​model from the AI ​​server and inputs the prompt word and the first logistics information into the AI ​​model. The AI ​​model then performs text analysis on the prompt word and the first logistics information to determine an anomaly detection result for the first order. The prompt word is pre-generated and stored on the application server. The generation of the prompt word and the anomaly detection method for the first order have been described in detail in the above embodiments and will not be repeated here.

[0190] In step S7.5, the artificial intelligence server returns the anomaly detection result to the application server.

[0191] Optionally, after obtaining the anomaly detection result of the first order, the application server may store the anomaly detection result in a local database, for example, by associating the order identification information of the first order and the anomaly detection result and writing them into a local artificial intelligence analysis log.

[0192] Step S7.6: When the anomaly detection result indicates that the first order is an abnormal order, the application server generates a return and refund request for the first order and sends the return and refund request to the logistics server.

[0193] Step S7.7: The logistics service end executes the return and refund event of the first order according to the return and refund request of the first order.

[0194] Among them, the way in which the logistics service end executes the return and refund event may include the following steps: generating a return and refund instruction for the first order, and allocating a transportation terminal for the first order; sending the return and refund instruction to the transportation terminal allocated for the first order to instruct the transportation terminal to transport the order items of the first order from the actual receipt address to the order starting address according to the first logistics information; the return and refund instruction includes order identification information and the first logistics information.

[0195] It can be seen that according to the technical solution of the embodiment of the present application, it is possible to use the artificial intelligence model's powerful text analysis and logical reasoning capabilities to achieve the effect of intelligently analyzing abnormal orders using the artificial intelligence model, thereby making the detection of abnormal orders no longer dependent on manpower, avoiding the detection omission problem that is easy to occur when manually detecting orders, and improving the efficiency and accuracy of abnormal order detection. In addition, the high efficiency of abnormal order detection based on artificial intelligence enables abnormal situations in the logistics trajectory to be perceived in advance, thereby improving the efficiency and delivery rate of logistics flow and reducing the losses of ordering users. In addition, when an abnormal order is detected by the artificial intelligence model, a return and refund request for the first order is generated and sent to the logistics service end, so that the logistics service end allocates a transport terminal for the first order and instructs the transport terminal to transport the ordered items of the first order from the actual receipt address to the order starting address. This achieves the proactive execution of return and refund events for abnormal orders without the ordering user's awareness, providing users with a more complete and convenient intelligent service and improving the ordering experience.

[0196] Corresponding to the application scenario and method of the method provided in the embodiment of the present application, the embodiment of the present application also provides an abnormal order detection device, such as Figure 8 As shown, the device includes:

[0197] A first acquisition module 81 is configured to acquire first logistics information of a first order; the first logistics information includes first logistics track information and first order address information;

[0198] A calling module 82 is configured to call an artificial intelligence model and obtain a prompt word of the artificial intelligence model; the prompt word is used to indicate an anomaly detection strategy corresponding to the first order;

[0199] The detection module 83 is used to input the prompt word and the first logistics information into the artificial intelligence model, perform text analysis on the prompt word and the first logistics information through the artificial intelligence model, and obtain an abnormality detection result of the first order; the abnormality detection result includes: whether the first order is an abnormal order.

[0200] In some embodiments, the apparatus further comprises:

[0201] A third acquisition module is configured to acquire second logistics information of a historical abnormal order before acquiring the logistics information of the first order; the second logistics information includes second logistics track information and second order address information;

[0202] An analysis module, configured to analyze abnormality-related information of the historical abnormal orders based on the second logistics information; the abnormality-related information includes abnormality categories and / or abnormality causes;

[0203] The first generating module is used to generate the prompt word according to the abnormality association information.

[0204] In some embodiments, when generating the prompt word according to the abnormality association information, the first generating module performs the following steps:

[0205] Determining the anomaly detection strategy based on the anomaly association information; the anomaly detection strategy is a strategy for detecting an order corresponding to logistics trajectory information including first content as an abnormal order; the first content including at least one of the following: a keyword used to characterize a rejection of receipt, trajectory information returned to the order's starting address, and abnormal address information that does not match the order's target address;

[0206] The prompt word is generated according to the anomaly detection strategy.

[0207] In some embodiments, the abnormality detection result further includes an abnormality category and / or abnormality cause;

[0208] When the detection module 83 performs text analysis on the prompt word and the first logistics information using the artificial intelligence model to obtain an abnormality detection result for the first order, the detection module 83 performs the following steps:

[0209] In the case where it is determined that the first order is the abnormal order, determining the abnormality category and / or abnormality cause of the first order based on the abnormality detection strategy;

[0210] Among them, the abnormality category includes at least one of the following: refusal to sign, reverse logistics, and address abnormality; the abnormality reason includes at least one of the following: user refusal to sign, the first logistics track information does not match the first order address information, and the actual signing address does not match the order target address.

[0211] In some embodiments, the apparatus further comprises:

[0212] a first determination module, configured to determine an accuracy rate of the artificial intelligence model based on anomaly detection results of the artificial intelligence model on the plurality of first orders;

[0213] A fourth acquisition module is configured to acquire, when the accuracy rate is lower than or equal to a preset threshold, a third order with an abnormality detection result error;

[0214] An optimization module is used to optimize the prompt word according to the abnormality association information of the third order.

[0215] In some embodiments, the apparatus further comprises:

[0216] A fifth acquisition module is configured to acquire third logistics information of the candidate order before acquiring the first logistics information of the first order; the third logistics information includes third logistics track information and third order address information;

[0217] The second determination module is used to determine, based on the third logistics information, that a candidate order that meets a first condition is the first order; the first condition includes: the order has been signed for, and the order logistics track does not match the order target address.

[0218] In some embodiments, when acquiring the first logistics information of the first order, the first acquiring module 81 performs the following steps:

[0219] Generate a request for obtaining the first logistics information; the request includes order identification information of the first order;

[0220] Sending the acquisition request to the logistics service end; the logistics service end is used to obtain the first logistics information based on the acquisition request, and send the first logistics information to the application service end;

[0221] Receive the first logistics information sent by the logistics service end.

[0222] In some embodiments, the apparatus further comprises:

[0223] a second generating module configured to perform text analysis on the prompt word and the first logistics information using the artificial intelligence model to obtain an abnormality detection result for the first order, and then, if the first order is determined to be the abnormal order, generate a return and refund request for the first order; the return and refund request including order identification information of the first order;

[0224] A return and refund module is used to send the return and refund request to the logistics service end, so that the logistics service end generates a return and refund instruction for the first order according to the return and refund request, and sends the return and refund instruction to the transportation terminal allocated for the first order; the return and refund instruction includes the order identification information and the first logistics information; the return and refund instruction is used to instruct the transportation terminal to transport the order items of the first order from the actual receipt address to the order starting address according to the first logistics information.

[0225] According to the device provided in the embodiment of the present application, by obtaining the first logistics information of the first order, the first logistics information includes the first logistics track information and the first order address information, and calling the artificial intelligence model to obtain the prompt word of the artificial intelligence model, the prompt word is used to indicate the abnormality detection strategy corresponding to the first order; then the prompt word and the first logistics information are input into the artificial intelligence model, and the artificial intelligence model performs text analysis on the prompt word and the first logistics information to obtain the abnormality detection result of the first order, the abnormality detection result includes whether the first order is an abnormal order. It can be seen that based on the powerful text analysis and logical reasoning capabilities of the artificial intelligence model, by providing the artificial intelligence model with prompt words for indicating the abnormality detection strategy, the effect of intelligently analyzing abnormal orders using the artificial intelligence model can be achieved, so that the detection of abnormal orders no longer relies on manpower, avoiding the detection omission problem that is easy to occur when manually detecting orders, and improving the efficiency and accuracy of abnormal order detection. In addition, the high efficiency of abnormal order detection based on artificial intelligence allows abnormal situations in the logistics track to be perceived in advance, thereby improving the efficiency and delivery rate of logistics flow and reducing the losses of ordering users.

[0226] Corresponding to the application scenario and method of the method provided in the embodiment of the present application, the embodiment of the present application also provides another abnormal order detection device, such as Figure 9 As shown, the device includes:

[0227] Receiving module 91, configured to receive a request from an application server for obtaining first logistics information of a first order; the request includes order identification information of the first order; the first logistics information includes first logistics track information and first order address information;

[0228] The second acquisition module 92 is used to obtain the first logistics information of the first order according to the order identification information, and send the first logistics information to the application server, so that the application server calls the artificial intelligence model based on the first logistics information to obtain the anomaly detection result of the first order, and the anomaly detection result includes: whether the first order is an abnormal order.

[0229] In some embodiments, the apparatus further comprises:

[0230] A second receiving module is configured to receive a return and refund request for the first order sent by the application server, where the return and refund request includes order identification information of the first order;

[0231] a third generating module, configured to generate a return and refund instruction for the first order based on the return and refund request, wherein the return and refund instruction includes the order identification information and the first logistics information;

[0232] An allocation module is used to allocate a transport terminal for the first order and send the return and refund instruction to the transport terminal; the return and refund instruction is used to instruct the transport terminal to transport the order items of the first order from the actual receipt address to the order starting address based on the first logistics information.

[0233] According to the device provided in the embodiment of the present application, by receiving the application server's request to obtain the first logistics information of the first order, the first logistics information of the first order is obtained, and the first logistics information is sent to the application server, so that the application server can call the artificial intelligence model based on the first logistics information to detect abnormal orders, thereby realizing the effect of using the artificial intelligence model to intelligently analyze abnormal orders, so that the detection of abnormal orders no longer relies on manpower, avoiding the problem of detection omissions that are easy to occur when manually detecting orders, and improving the efficiency and accuracy of abnormal order detection. In addition, the high efficiency of abnormal order detection based on artificial intelligence allows abnormal situations in the logistics trajectory to be perceived in advance, thereby improving the efficiency and delivery rate of logistics flow and reducing the losses of ordering users.

[0234] The functions of each module in each device in the embodiments of the present application can be found in the corresponding description in the above method, and have corresponding beneficial effects, which will not be repeated here.

[0235] Figure 10 FIG. 1 is a block diagram of an electronic device for implementing an embodiment of the present application. Figure 10 As shown, the electronic device includes: a memory 1001 and a processor 1002. The memory 1001 stores a computer program that can be executed on the processor 1002. When the processor 1002 executes the computer program, the method of the above embodiment is implemented. The number of memory 1001 and processor 1002 can be one or more. In a specific implementation, the electronic device may also include a communication interface 1003 for communicating with external devices and exchanging data.

[0236] In a specific implementation, if the memory 1001, the processor 1002, and the communication interface 1003 are implemented independently, the memory 1001, the processor 1002, and the communication interface 1003 can be connected to each other via a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 10 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0237] Optionally, in a specific implementation, if the memory 1001, the processor 1002 and the communication interface 1003 are integrated on a chip, the memory 1001, the processor 1002 and the communication interface 1003 can communicate with each other through an internal interface.

[0238] An embodiment of the present application provides a computer-readable storage medium storing a computer program, which implements the method provided in the embodiment of the present application when the program is executed by a processor.

[0239] An embodiment of the present application provides a computer program product, including a computer program, which implements the method provided in the embodiment of the present application when executed by a processor.

[0240] An embodiment of the present application also provides a chip, which includes a processor for calling and executing instructions stored in the memory from the memory, so that a communication device equipped with the chip executes the method provided in the embodiment of the present application.

[0241] An embodiment of the present application also provides a chip, including: an input interface, an output interface, a processor and a memory. The input interface, the output interface, the processor and the memory are connected through an internal connection path. The processor is used to execute the code in the memory. When the code is executed, the processor is used to execute the method provided in the embodiment of the application.

[0242] It should be understood that the processor may be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. It is worth noting that the processor may be a processor that supports the Advanced RISC Machine (ARM) architecture.

[0243] Furthermore, optionally, the above-mentioned memory may include a read-only memory and a random access memory. The memory may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memory. Among them, the non-volatile memory may include a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may include a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available. For example, static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM) and direct memory bus random access memory (DR RAM).

[0244] In the above embodiments, all or part of the embodiments may be implemented using software, hardware, firmware, or any combination thereof. When implemented using software, all or part of the embodiments may be implemented in the form of a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions according to the present application are generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another.

[0245] In the description of this specification, the reference terms "one embodiment," "some embodiments," "example," "specific example," or "some examples" mean that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. Moreover, the specific features, structures, materials, or characteristics described may be combined in any appropriate manner in any one or more embodiments or examples. In addition, those skilled in the art may combine and combine different embodiments or examples described in this specification, as well as features of different embodiments or examples, unless they are mutually inconsistent.

[0246] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. Throughout the description of this application, "plurality" means two or more, unless otherwise specifically defined.

[0247] Any process or method described in the flowchart or otherwise described herein can be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing the steps of a specific logical function or process. The scope of the preferred embodiments of the present application includes other implementations in which the functions may be performed in a different order than shown or discussed, including performing the functions substantially simultaneously or in reverse order depending on the functions involved.

[0248] The logic and / or steps described in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, apparatus or device (such as a computer-based system, a system including a processor or other system that can fetch instructions from an instruction execution system, apparatus or device and execute instructions), or used in combination with such instruction execution systems, apparatuses or devices.

[0249] It should be understood that various parts of the present application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. All or part of the steps of the above embodiment method can be completed by instructing the relevant hardware through a program, which can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0250] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing module, or each unit may exist physically separately, or two or more units may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or in the form of software functional modules. If the aforementioned integrated modules are implemented in the form of software functional modules and sold or used as independent products, they may also be stored in a computer-readable storage medium. The storage medium may be a read-only memory, a magnetic disk, or an optical disk, etc.

[0251] The above is merely an exemplary embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any person skilled in the art can easily conceive of various modifications or substitutions within the technical scope described in this application, and such modifications or substitutions should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A method for detecting abnormal orders, characterized in that: include: Get the first logistics information of the first order; The first logistics information includes first logistics track information and first order address information; Invoking an artificial intelligence model and obtaining a prompt word of the artificial intelligence model; the prompt word is used to indicate an anomaly detection strategy corresponding to the first order; Inputting the prompt word and the first logistics information into the artificial intelligence model, and performing text analysis on the prompt word and the first logistics information by the artificial intelligence model to obtain an anomaly detection result of the first order; The abnormality detection result includes: whether the first order is an abnormal order.

2. The method according to claim 1, characterized in that Before obtaining the logistics information of the first order, the process further includes: Obtaining second logistics information of historical abnormal orders; the second logistics information includes second logistics track information and second order address information; Analyzing abnormality-related information of the historical abnormal orders based on the second logistics information; the abnormality-related information includes abnormality categories and / or abnormality reasons; The prompt word is generated according to the abnormality association information.

3. The method according to claim 2, characterized in that Generating the prompt word according to the abnormality association information includes: Determining the anomaly detection strategy based on the anomaly association information; the anomaly detection strategy is a strategy for detecting an order corresponding to logistics trajectory information including first content as an abnormal order; the first content including at least one of the following: a keyword used to characterize a rejection of receipt, trajectory information returned to the order's starting address, and abnormal address information that does not match the order's target address; The prompt word is generated according to the anomaly detection strategy.

4. The method according to claim 3, characterized in that The abnormality detection result also includes the abnormality category and / or abnormality cause; The performing text analysis on the prompt word and the first logistics information by the artificial intelligence model to obtain an anomaly detection result of the first order includes: In the case where it is determined that the first order is the abnormal order, determining the abnormality category and / or abnormality cause of the first order based on the abnormality detection strategy; Among them, the abnormality category includes at least one of the following: refusal to sign, reverse logistics, and address abnormality; the abnormality reason includes at least one of the following: user refusal to sign, the first logistics track information does not match the first order address information, and the actual signing address does not match the order target address.

5. The method according to claim 2 or 3, characterized in that Also includes: determining an accuracy rate of the artificial intelligence model based on anomaly detection results of the artificial intelligence model on the plurality of first orders; When the accuracy rate is lower than or equal to a preset threshold, obtaining a third order with an abnormality detection result error; The prompt word is optimized according to the abnormal association information of the third order.

6. The method according to claim 1, characterized in that Before obtaining the first logistics information of the first order, the method further includes: Obtaining third-party logistics information of the candidate order; the third-party logistics information includes third-party logistics track information and third-party order address information; According to the third logistics information, the candidate order that meets the first condition is determined to be the first order; the first condition includes: the order has been signed for, and the order logistics track does not match the order target address.

7. The method according to claim 6, characterized in that The obtaining of the first logistics information of the first order includes: Generate a request for obtaining the first logistics information; the request includes order identification information of the first order; Sending the acquisition request to the logistics service end; the logistics service end is used to obtain the first logistics information based on the acquisition request, and send the first logistics information to the application service end; Receive the first logistics information sent by the logistics service end.

8. The method according to claim 1, characterized in that After performing text analysis on the prompt word and the first logistics information by the artificial intelligence model to obtain an abnormality detection result of the first order, the method further includes: If it is determined that the first order is the abnormal order, generating a return and refund request for the first order; the return and refund request includes order identification information of the first order; The return and refund request is sent to the logistics service end, so that the logistics service end generates a return and refund instruction for the first order according to the return and refund request, and sends the return and refund instruction to the transportation terminal allocated for the first order; the return and refund instruction includes the order identification information and the first logistics information; the return and refund instruction is used to instruct the transportation terminal to transport the order items of the first order from the actual receipt address to the order starting address according to the first logistics information.

9. A method for detecting abnormal orders, characterized in that: include: Receive a request from the application server for obtaining first logistics information of the first order; The acquisition request includes order identification information of the first order; The first logistics information includes first logistics track information and first order address information; According to the order identification information, the first logistics information of the first order is obtained, and the first logistics information is sent to the application server, so that the application server calls the artificial intelligence model based on the first logistics information to obtain an anomaly detection result of the first order, and the anomaly detection result includes: whether the first order is an abnormal order.

10. An abnormal order detection system, characterized in that: include: AI server, used to store AI models; The logistics service end is used to store the first logistics information of the first order; The first logistics information includes first logistics track information and first order address information; An application server is configured to obtain the first logistics information from the logistics server; invoke the artificial intelligence model from the artificial intelligence server and obtain a prompt word from the artificial intelligence model; the prompt word is used to indicate an anomaly detection strategy corresponding to the first order; input the prompt word and the first logistics information into the artificial intelligence model, and perform text analysis on the prompt word and the first logistics information by the artificial intelligence model to obtain an anomaly detection result for the first order; The abnormality detection result includes: whether the first order is an abnormal order.

Images