Express false delivery identification method, device and equipment and storage medium

Abstract

The invention relates to the field of logistics, and discloses an express false delivery identification method, device and equipment and a storage medium, and the method is used for automatically identifying false delivery behaviors and improving supervision efficiency. The method comprises the following steps: marking and classifying collected express order data, and dividing the express order data into real delivery order data and false delivery order data; extracting time features and text features in the real delivery order data, training an LSTM (Long Short Term Memory) network based on the time features and the text features, and constructing a merchant normal delivery behavior model; collecting to-be-identified express order data in real time, comparing the to-be-identified express order data with a preset standard and a merchant normal delivery behavior model, and judging whether the to-be-identified express order data is false delivery order data; and if it is determined that the to-be-identified express order data is the false delivery order data, sending an early warning to an e-commerce platform, and filtering the express number of the false delivery order data.

Description

Technical Field

[0001] The present invention relates to the field of logistics technology, and in particular to a method, device, equipment and storage medium for identifying false delivery of express deliveries. Background Art

[0002] With the booming development of the e-commerce industry, the volume of express delivery business has increased explosively. However, some unscrupulous merchants engage in false delivery behavior for personal gain. False delivery violates consumers' right to know and right to choose, and damages consumers' legitimate rights and interests. Consumers may not receive the goods in time due to false delivery, or the goods received do not match the order, which not only affects consumers' shopping experience, but may also cause economic losses, leading to problems such as extremely poor shopping experience and increased complaints. It also disrupts the normal e-commerce market order and affects the fair competition environment. False delivery behavior can be classified into five common situations: using duplicate or invalid logistics waybills, no update of logistics information for a long time, abnormal stagnation of logistics track, inconsistent receiving addresses, and consumers not actually receiving the goods. At present, the means for identifying false delivery in the express delivery logistics system are relatively limited, mostly relying on manual sampling inspection or simple timeout judgment of logistics nodes, and it is difficult to comprehensively, efficiently and accurately detect false delivery behavior.

[0003] Therefore, the existing technology still needs to be improved and developed. Summary of the Invention

[0004] The present invention provides a method, device, equipment and storage medium for identifying false delivery of express deliveries, aiming to solve the problem that the means for identifying false delivery in the existing express delivery logistics system are relatively limited, mostly relying on manual sampling inspection or simple timeout judgment of logistics nodes, and it is difficult to comprehensively, efficiently and accurately detect false delivery behavior.

[0005] In a first aspect of the present invention, a method for identifying false delivery of express deliveries is provided. The method for identifying false delivery of express deliveries includes: marking and classifying the collected express order data, and dividing the express order data into real delivery order data and false delivery order data; extracting time features and text features from the real delivery order data, training the LSTM (Long Short-Term Memory) network based on the time features and text features, and constructing a normal delivery behavior model of merchants; collecting real-time express order data to be identified, comparing the express order data to be identified with a preset standard and the normal delivery behavior model of merchants, and determining whether the express order data to be identified is false delivery order data; if it is determined that the express order data to be identified is false delivery order data, an early warning is sent to the e-commerce platform, and the logistics waybill numbers of the false delivery order data are filtered.

[0006] Optionally, in the first implementation manner of the first aspect of the present invention, the marking and classification of the collected express order data, and the division of the express order data into real delivery order data and false delivery order data include: collecting express order data using an API open platform; cleaning the collected express order data to remove invalid or incomplete records; using a marking tool to mark the cleaned express order data and verifying the address consistency, and dividing the express order data into real delivery order data and false delivery order data.

[0007] Optionally, in the second implementation manner of the first aspect of the present invention, the extraction of the time features and text features in the real delivery order data, and the training of the LSTM long short-term memory network based on the time features and text features to construct a merchant's normal delivery behavior model include: extracting the time features corresponding to the order information in the real delivery order data; using natural language technology to analyze the text data in the real delivery order data and extract text features; taking the time features, text features and the marked real delivery order data as inputs and importing them into the LSTM long short-term memory network for training and constructing a merchant's normal delivery model.

[0008] Optionally, in the third implementation manner of the first aspect of the present invention, the real-time collection of the express order data to be identified, and the comparison of the express order data to be identified with a preset standard and a merchant's normal delivery behavior model to determine whether the express order data to be identified is false delivery order data include: setting a preset standard, where the preset standard includes that there is a delivery notice but no delivery within 48 hours after the order payment, and the logistics track has not been updated within 48 hours and exceeds the normal delivery delay range of the same route; real-time collecting the express order data to be identified, comparing the express order data to be identified with the preset standard to determine whether the express order data to be identified is false delivery order data; if it is determined that there is no behavior of having a delivery notice but no delivery within 48 hours after the order payment, and the logistics track has not been updated within 48 hours and exceeds the normal delivery delay range of the same route, then comparing the express order data to be identified with the merchant's normal delivery behavior model; if it is determined that there is any behavior of having a delivery notice but no delivery within 48 hours after the order payment, and the logistics track has not been updated within 48 hours and exceeds the normal delivery delay range of the same route, then determining that the express order data to be identified is false delivery order data.

[0009] Optionally, in the fourth implementation manner of the first aspect of the present invention, if it is determined that there is no behavior that the goods are not shipped within 48 hours after the order payment but there is a shipping notice, or the logistics track is not updated within 48 hours and exceeds the normal delivery delay range of the same route, then the express order data to be identified is compared with the merchant's normal shipping behavior model, including: comparing the express order data to be identified with the merchant's normal shipping behavior model, and judging the deviation degree between the express order data to be identified and the merchant's normal shipping behavior model; if it is determined that the deviation degree is less than the preset threshold, it is determined that the express order data to be identified is real order data; if it is determined that the deviation degree is greater than the preset threshold, it is determined that the express order data to be identified is false shipping order data.

[0010] Optionally, in the fifth implementation manner of the first aspect of the present invention, if it is determined that the express order data to be identified is false shipping order data, a warning is sent to the e-commerce platform, and the express order number of the false shipping order data is filtered, including: after determining that the express order data to be identified is false shipping order data, sending a warning to the e-commerce platform and sending an instant message notification to the platform customer service team; transmitting the order details corresponding to the false shipping order data to the supervision background of the e-commerce platform, and the order details include the order number, merchant information, express order number, and suspicious feature description; identifying the express company to which the express order number belongs, and determining and filtering the express order number of the false shipping order data.

[0011] Optionally, in the sixth implementation manner of the first aspect of the present invention, identifying the express company to which the express order number belongs, and determining and filtering the express order number of the false shipping order data includes: using the logistics query interface to automatically identify the express company to which the express order number belongs; based on the order number verification rules and blacklist library of the express company, judging whether the express order number is a false express order number; if it is determined that the express order number is a false express order number, the determined false express order number is filtered.

[0012] Optionally, a false shipping identification device for express delivery is provided in the second aspect of the present invention, including: a first collection module, configured to mark and classify the collected express order data, and divide the express order data into real shipping order data and false shipping order data; an extraction module, configured to extract the time features and text features in the real shipping order data, and train the LSTM long short-term memory network based on the time features and text features to construct a merchant's normal shipping behavior model; a second collection module, configured to collect the express order data to be identified in real time, compare the express order data to be identified with the preset standard and the merchant's normal shipping behavior model, and judge whether the express order data to be identified is false shipping order data; a sending module, configured to send a warning to the e-commerce platform and filter the express order number of the false shipping order data if it is determined that the express order data to be identified is false shipping order data.

[0013] Optionally, in the first implementation manner of the second aspect of the present invention, the first collection module includes: a first collection unit, configured to collect express order data by using an API open platform; a cleaning unit, configured to clean the collected express order data to remove invalid or incomplete records; a partitioning unit, configured to mark the cleaned express order data by using a marking tool, check the address consistency, and partition the express order data into real delivery order data and false delivery order data.

[0014] Optionally, in the second implementation manner of the second aspect of the present invention, the extraction module includes: an extraction unit, configured to extract time features corresponding to order information from the real delivery order data; an analysis unit, configured to analyze text data in the real delivery order data by using natural language technology to extract text features; a construction unit, configured to import the time features, text features, and marked real delivery order data as inputs into an LSTM (Long Short-Term Memory) network to train and construct a merchant normal delivery model.

[0015] Optionally, in the third implementation manner of the second aspect of the present invention, the second collection module includes: a setting unit, configured to set a preset standard, where the preset standard includes that there is a delivery notice but no delivery within 48 hours after the order is paid, and the logistics track is not updated within 48 hours and exceeds the normal delivery delay range of the same route; a second collection unit, configured to collect the express order data to be identified in real time, compare the express order data to be identified with the preset standard, and determine whether the express order data to be identified is false delivery order data; a first determination unit, configured to, if it is determined that there is no such behavior as no delivery within 48 hours after the order is paid but there is a delivery notice, and the logistics track is not updated within 48 hours and exceeds the normal delivery delay range of the same route, compare the express order data to be identified with the merchant normal delivery behavior model; a second determination unit, configured to, if it is determined that there is such behavior as no delivery within 48 hours after the order is paid but there is a delivery notice, and the logistics track is not updated within 48 hours and exceeds the normal delivery delay range of the same route, determine that the express order data to be identified is false delivery order data.

[0016] Optionally, in the fourth implementation manner of the second aspect of the present invention, the first determination unit includes: a first judgment subunit, configured to compare the express order data to be identified with the merchant normal delivery behavior model to judge the deviation degree between the express order data to be identified and the merchant normal delivery behavior model; a first determination subunit, configured to, if it is determined that the deviation degree is less than a preset threshold, determine that the express order data to be identified is real order data; a second determination subunit, configured to, if it is determined that the deviation degree is greater than the preset threshold, determine that the express order data to be identified is false delivery order data.

[0017] Optionally, in the fifth implementation manner of the second aspect of the present invention, the sending module includes: a sending unit, configured to send a warning to an e-commerce platform and send an instant message notification to the platform customer service team after determining that the express order data to be recognized is false delivery order data; a transmission unit, configured to transmit the order details corresponding to the false delivery order data to the supervision background of the e-commerce platform, where the order details include an order number, merchant information, express waybill number, and a description of suspicious features; an identification unit, configured to identify the express company to which the express waybill number belongs, and determine and filter the express waybill numbers of the false delivery order data.

[0018] Optionally, in the sixth implementation manner of the second aspect of the present invention, the identification unit includes: an identification subunit, configured to automatically identify the express company to which the express waybill number belongs by using a logistics query interface; a second judgment subunit, configured to judge whether the express waybill number is a false express waybill number based on the waybill number verification rule of the express company and a blacklist library; a filtering subunit, configured to filter the determined false express waybill number if it is determined that the express waybill number is a false express waybill number.

[0019] Optionally, a third aspect of the present invention provides an express false delivery identification device, including a memory and at least one processor, where computer-readable instructions are stored in the memory, and the memory and the at least one processor are interconnected by a line; the at least one processor calls the computer-readable instructions in the memory to cause the express false delivery identification device to execute each step of the express false delivery identification method as described above.

[0020] Optionally, a fourth aspect of the present invention provides a computer-readable storage medium, where computer-readable instructions are stored in the computer-readable storage medium, and when it runs on a computer, it causes the computer to execute each step of the express false delivery identification method as described above.

[0021] In the technical solution provided by the present invention, by marking and classifying the collected express order data, the collected express order data can be divided into real delivery order data and false delivery order data, so that a normal delivery model of merchants can be constructed using the real delivery order data. Secondly, based on the time features and text features in the extracted real delivery order data, the LSTM long short-term memory network is trained, and a normal delivery behavior model highly adapted to the behaviors of each merchant can be constructed. When comparing the express order data to be identified with the constructed normal delivery behavior model of merchants, the misjudgment of false delivery order data can be reduced, the accuracy of judging false delivery order data can be improved, and the cost and time of manual review can be reduced. After determining the false express order data, a warning is sent to the e-commerce platform, enabling the e-commerce platform to take regulatory measures such as restricting the merchant's fund withdrawal, taking off the shelves of illegal goods, and reducing the store's search weight, enhancing the control ability of express false delivery behaviors, protecting the legitimate rights and interests of all parties, avoiding consumers suffering economic losses and bad shopping experiences due to false delivery, and promoting the healthy and sustainable development of the e-commerce industry. In addition, after sending a warning to the e-commerce platform, the express waybill numbers of false delivery order data are filtered using the waybill number query system of the express company, optimizing the logistics data environment and ensuring the authenticity and reliability of the express delivery chain. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 The first flowchart of the express false delivery identification method provided by the embodiment of the present invention;

[0023] Figure 2 The second flowchart of the express false delivery identification method provided by the embodiment of the present invention;

[0024] Figure 3 The third flowchart of the express false delivery identification method provided by the embodiment of the present invention;

[0025] Figure 4 The fourth flowchart of the express false delivery identification method provided by the embodiment of the present invention;

[0026] Figure 5 The fifth flowchart of the express false delivery identification method provided by the embodiment of the present invention;

[0027] Figure 6 The sixth flowchart of the express false delivery identification method provided by the embodiment of the present invention;

[0028] Figure 7 The seventh flowchart of the express false delivery identification method provided by the embodiment of the present invention;

[0029] Figure 8 The structural schematic diagram of the express false delivery identification device provided by the embodiment of the present invention;

[0030] Figure 9 This is a schematic structural diagram of the express false delivery identification device provided by the embodiments of the present invention. Detailed implementation manners

[0031] The embodiments of the present invention provide an express false delivery identification method, device, equipment and storage medium. The method is used to automatically identify false delivery behaviors and improve supervision efficiency. The method includes: marking and classifying the collected express order data, and dividing the express order data into real delivery order data and false delivery order data; extracting time features and text features from the real delivery order data, training the LSTM long short-term memory network based on the time features and text features, and constructing a merchant's normal delivery behavior model; collecting the express order data to be identified in real time, comparing the express order data to be identified with preset standards and the merchant's normal delivery behavior model, and judging whether the express order data to be identified is false delivery order data; if it is determined that the express order data to be identified is false delivery order data, an early warning is sent to the e-commerce platform, and the express order numbers of the false delivery order data are filtered.

[0032] The terms "first", "second", "third", "fourth", etc. (if any) in the description and claims of the present invention and the above drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" or "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0033] For ease of understanding, the specific process of the embodiments of the present invention will be described below. Please refer to Figure 1 , the first embodiment of an express false delivery identification method in the embodiments of the present invention includes:

[0034] S101. Mark and classify the collected express order data, and divide the express order data into real delivery order data and false delivery order data.

[0035] It can be understood that the execution subject of the present invention can be an express false delivery identification device, or a terminal or a server. Specifically, no limitation is made here. The embodiments of the present invention will be described by taking the server as the execution subject as an example.

[0036] Specifically, in this embodiment, express order data is collected from various channels such as the database of the e-commerce platform and the express delivery system. These express order data should include, but are not limited to, order information (including product name, quantity specification, order placement time, payment time, merchant promised delivery time, order remarks, etc.), logistics information (including timestamps, locations, operator identifiers, etc. of each node feedback from the express company), user feedback (including evaluations and complaints collected from consumers on the e-commerce platform regarding the order delivery experience, such as "haven't shipped for a long time after placing the order" and "logistics has not been updated"), etc. By collecting the above express order data again, it can intuitively reflect the merchant's shipping situation and the needs and problems of users during the shopping and receiving process. After collecting the express order data, the express order data is marked. For example, for logistics information, if there is no pickup record within 24 hours after the shipping notice and there is still no effective logistics update in the subsequent 48 hours, it is marked and classified as false shipping order data. For example, for order information, if it is found that a large number of orders are quickly marked as shipped after payment by the same merchant within a short period (such as within 1 hour), but there is no query result for the corresponding logistics order number, it is also marked and classified as false shipping order data.

[0037] S102. Extract the time features and text features from the real shipping order data, and train the LSTM long short-term memory network based on the time features and text features to construct a merchant's normal shipping behavior model.

[0038] Specifically, since the real shipping order data can reflect the merchant's normal shipping situation, in this embodiment, it is necessary to construct a merchant's normal shipping behavior model based on the real shipping order data. To make the constructed merchant's normal shipping behavior model conform to the merchant's normal shipping rhythm, when training the LSTM long short-term memory network in this embodiment, by extracting the time features in the real shipping order data, such as the shipping time and the receiving time, etc., it helps to analyze the merchant's shipping efficiency and habits in different time periods (such as weekdays and weekends, different time periods of a day), so that the LSTM long short-term memory network can learn these rules. Since time factors such as different seasons, months, and weeks will affect the merchant's shipping behavior. For example, during e-commerce promotion seasons (such as "Double 11" and "618"), the merchant's order volume will increase significantly, and the shipping time may be extended; during some special festivals, the logistics distribution may also be affected. By extracting time features such as the residence duration of the package at each transfer station and the deviation value between the expected delivery time and the actual receipt time, the model can consider the shipping situation under various time scenarios, so as to accurately judge the merchant's normal shipping behavior under different time conditions and enhance the generalization ability of the merchant's normal shipping behavior model.

[0039] Due to the differences in the supply chain, inventory management, shipping characteristics, and user requirements among different categories of goods, in order to improve the generality and adaptability of the normal shipping behavior model for merchants, this embodiment also extracts text features from real shipping order data (including product descriptions, special requirements in order remarks, etc.), and enables the LSTM (Long Short-Term Memory) network to learn the information contained in these text features, so as to more accurately understand the shipping behavior patterns of merchants for different products and improve the generality and adaptability of the normal shipping behavior model for merchants.

[0040] Regarding the time features in real express order data, such as the order placement time, shipping time, timestamps of each logistics node, etc., they constitute a time series. In this series, there are long-term dependencies between data at different time points. For example, a merchant's current shipping behavior may be affected by factors such as the order volume and logistics status in a previous period. The LSTM network is specifically designed with memory units and gating mechanisms, which can effectively preserve and transmit long-term information. Its forget gate, input gate, and output gate can dynamically determine when to forget old information, when to update new information, and when to output information according to the input data, thus well capturing the long-term dependencies in the time series data of express orders and more accurately learning the changing rules of a merchant's normal shipping behavior over time.

[0041] Moreover, there may be significant differences in the order processing processes and shipping times among different merchants, resulting in different lengths of the time series in real express order data. Some merchants may process orders quickly, with short shipping time intervals and relatively short sequence lengths; while some merchants may have long shipping time intervals and long sequence lengths due to complex business and large order volumes. The LSTM network has good adaptability to the sequence length. It can process input sequences of any length without truncating or padding the input data to a fixed length, and can better retain the information of the original data, thus more accurately learning the shipping behavior patterns of different merchants. During the training process of this embodiment, through feeding a large amount of order data of different merchants, different time periods, and different categories of products, the trained model is enabled to learn the complex pattern features of a merchant's normal shipping, such as enabling the model to master the rule that when the order volume of a certain clothing merchant surges during the promotion season, the shipping time will be slightly delayed but still remain within a certain time range, or the rule that a certain digital product merchant has a faster shipping speed on weekdays than on weekends, etc., so as to construct a normal shipping behavior model that highly adapts to the behavior characteristics of each merchant.

[0042] S103. Collect real-time the express order data to be identified, compare the express order data to be identified with the preset standard and the normal shipping behavior model of the merchant, and determine whether the express order data to be identified is false shipping order data.

[0043] In this embodiment, the express order data to be recognized is collected in real time, and then the time features and text features in the express order data to be recognized are extracted. Then, the time features and text features extracted from the express order data to be recognized are compared with the preset standards and the normal shipping behavior model of the merchant to determine whether the express order data to be recognized is false shipping order data. For example, for the text features composed of keywords such as "customized", "pre-sale", and "in stock", "customized" products usually have a long production cycle. If the shipping time of such products is too fast in the comparison of the express order data to be recognized with the normal shipping behavior model of the merchant, it is determined that the express order data to be recognized is false shipping order data; for "pre-sale" products, it needs to be specifically compared and determined according to the promised shipping period of the merchant (the promised shipping time). For text features such as special requirements in the order remarks, such as "urgently needed, hope to ship as soon as possible", if the special requirements in the order remarks are extracted from the subsequent collected express order data to be recognized, and the logistics is sluggish (for example, the logistics track has not been updated for a long time), then the express order data to be recognized is determined to be false shipping order data.

[0044] S104. If it is determined that the express order data to be recognized is false shipping order data, an early warning is sent to the e-commerce platform, and the express order numbers of the false shipping order data are filtered.

[0045] In this embodiment, after it is determined that the express order data to be recognized is false shipping order data, an early warning is sent to the e-commerce platform, which is convenient for the e-commerce platform operators to quickly locate the merchant, enabling the e-commerce platform to take regulatory measures such as restricting the merchant's fund withdrawal, taking off the shelves of the illegal products, and reducing the store search weight to crack down on the false shipping behavior. By filtering the express order numbers of the false shipping order data, it prevents the express order numbers of the false shipping order data from being used and causing confusion in the subsequent logistics process, further optimizing the logistics data environment and ensuring the authenticity and reliability of the express delivery chain.

[0046] This embodiment provides a method for identifying false delivery of express deliveries. By marking and classifying the collected express order data, the collected express order data can be divided into real delivery order data and false delivery order data, enabling the construction of a normal delivery model for merchants using the real delivery order data. Secondly, based on the time features and text features in the extracted real delivery order data, the LSTM long short-term memory network is trained to construct a normal delivery behavior model that highly adapts to the behaviors of each merchant. When comparing the express order data to be identified with the constructed normal delivery behavior model of the merchant, it can reduce the misjudgment of false delivery order data, improve the accuracy of judging false delivery order data, and reduce the cost and time of manual review. After determining the false express order data, a warning is sent to the e-commerce platform, enabling the e-commerce platform to take regulatory measures such as restricting the merchant's fund withdrawal, taking off the shelves of illegal goods, and reducing the store's search weight, enhancing the control ability of false express delivery behaviors, protecting the legitimate rights and interests of all parties, preventing consumers from suffering economic losses and bad shopping experiences due to false deliveries, and promoting the healthy and sustainable development of the e-commerce industry. In addition, after sending a warning to the e-commerce platform, the express order numbers of false delivery order data are filtered using the express company's order number query system, optimizing the logistics data environment and ensuring the authenticity and reliability of the express delivery chain.

[0047] Please refer to Figure 2 , the second embodiment of a method for identifying false delivery of express deliveries in the embodiments of the present invention includes:

[0048] S201. Use the API open platform to collect express order data;

[0049] S202. Clean the collected express order data to remove invalid or incomplete records;

[0050] S203. Use a marking tool to mark the cleaned express order data, check the address consistency, and divide the express order data into real delivery order data and false delivery order data.

[0051] Specifically, in this embodiment, express order data is collected from multiple channels such as the database of the e-commerce platform and the express system by using the API open platform. For example, order information including product name, quantity, specification, order time, payment time, and merchant promised delivery time is collected through the e-commerce platform API. User feedback is collected through the user behavior and feedback API. The complete logistics track of the order can be automatically tracked through the logistics express API, and the logistics status can be real-time throughout the process. Key information including delivery speed, transportation time, and receipt status can be obtained, and then express order data including logistics information is collected based on the above key information.

[0052] After collecting the express order data, clean the collected express order data to ensure the reliability of the quality of the collected express order data. When training and building a model for normal merchant shipments, it can reduce noise interference and improve the accuracy and reliability of the model. For example, mark and correct the logistics time records whose time format does not conform to the standard of "YYYY-MM-DD HH:MM:SS"; eliminate the error data with non-numeric characters in the order amount field. Secondly, remove invalid or incomplete records through logical judgment, such as logistics information lacking the key node time (the pickup time is empty), records lacking the key attributes of the goods in the order information, etc., to ensure the reliability of the quality of the collected express order data.

[0053] Moreover, mark the express order data after cleaning through a marking tool, and compare the order receiving address with the logistics delivery end address through address parsing technology to check the address consistency in the logistics information, so as to divide the real order data, enabling the real order data to be used as a learning sample for building a model for normal merchant shipments.

[0054] As an example, the marking tool can be a database management system. For example, using database management systems such as MySQL and Oracle, write SQL statements to create a database table containing fields such as order information, logistics information, user feedback, and marking results. Through SQL query statements and in combination with the association of multiple database tables, filter out the real shipment order data and false shipment order data that meet the conditions and mark them.

[0055] As an example, when checking the address consistency in the logistics information, if there are situations where the province, city, or even district and county do not match, mark the express order data containing such logistics information and classify it as false shipment order data.

[0056] Please refer to Figure 3 , the third embodiment of a method for identifying false express shipments in the embodiments of the present invention includes:

[0057] S301. Extract the time features corresponding to the order information in the real shipment order data;

[0058] S302. Use natural language technology to analyze the text data in the real shipment order data and extract text features;

[0059] S303. Import the time features, text features, and the marked real shipment order data as inputs into the LSTM long short-term memory network to train and build a model for normal merchant shipments.

[0060] Specifically, in this embodiment, first, from the real delivery order data, based on order information, such as order placement time, delivery time, logistics pickup time, receipt time, etc., the data is preliminarily screened and sorted to ensure the integrity and accuracy of the data, removing abnormal or incorrect data records. Then, the diverse time representation forms that may exist in different data sources are converted into a standard time format for subsequent processing. Next, time features are extracted from the screened and sorted data. To improve the training effect of the model, the extracted time features can be normalized and mapped to a specific interval (such as [0, 1]) to eliminate the dimensional differences between different features.

[0061] To reduce resource consumption during model training and improve model accuracy, for the extraction of text features, in this embodiment, it is first necessary to preprocess the text data to remove noise information in the text, such as special characters, HTML tags, punctuation marks, etc., and unify the case of the text. Then, a word segmentation tool in natural language technology (such as jieba for Chinese and NLTK for English) is used to split the text into individual words or phrases, removing stop words that contribute little to the semantic expression of the text (such as "de", "shi", "zai", etc.) to reduce data redundancy. Then, text features can be extracted from the preprocessed text data through methods such as word frequency statistics and topic modeling. For example, by counting the frequency of each word in the text and determining text features based on frequently occurring words, because frequently occurring words may represent important information about the order, such as product name, key attributes, and other product descriptions. To facilitate the import of text features into the LSTM (Long Short-Term Memory) network, word embedding technology (such as Word2Vec, GloVe) can be used to convert the extracted text features into numerical vectors. By extracting time features and text features from real delivery order data, it is possible to comprehensively consider time and text information, fully characterize the merchant's delivery situation, and utilize the powerful time series processing ability of the LSTM network to effectively capture the dynamic change rules in the data, improving the model's ability to accurately judge and predict the merchant's delivery status and assisting in the identification of false delivery behaviors.

[0062] Please refer to Figure 4 , the fourth embodiment of a method for identifying false delivery of express in an embodiment of the present invention includes:

[0063] S401. Set a preset standard, where the preset standard includes that there is a delivery notice but no delivery within 48 hours after the order is paid, and the logistics track has not been updated within 48 hours and exceeds the normal delivery delay range of the same route;

[0064] S402. Real-time collect the express order data to be identified, compare the express order data to be identified with the preset standard, and determine whether the express order data to be identified is false delivery order data;

[0065] S403. If it is determined that there is no behavior that the goods are not shipped within 48 hours after the order payment but there is a shipping notice, or the logistics track is not updated within 48 hours and exceeds the normal delivery delay range of the same route, the express order data to be identified will be compared with the normal shipping behavior model of the merchant.

[0066] Specifically, in this embodiment, first, a preset standard is set based on the general industry norms and reasonable expectations of consumers. Then, the express order data to be identified is collected in real time using the API platform. Next, the express order data to be identified is compared with the preset standard. Since the preset standard is set based on the general industry norms and reasonable expectations of consumers, fully considering the general shipping situation in the industry, when comparing the express order data to be identified with the preset standard, it is possible to quickly screen out the data of orders that are obviously false shipments and exclude most of the express order data of normal shipments, avoiding comparing all the express order data to be identified with the model, reducing the unnecessary computational workload of the merchant's normal shipping model, and enabling the computational resources of the model to be concentrated on the express order data to be identified that is suspected of false shipments, optimizing the resource allocation and enabling the system to operate efficiently. When it is determined that the express order data to be identified does not have any behavior of the preset standard, the express order data to be identified is deeply compared with the normal shipping model that highly adapts to the behavior characteristics of each merchant, taking into account both the general industry situation and the individual differences of the merchants, thereby improving the accuracy of identifying false shipment order data.

[0067] Please refer to Figure 5 , the fifth embodiment of a method for identifying false shipments of express deliveries in the embodiments of the present invention includes:

[0068] S501. Compare the express order data to be identified with the normal shipping behavior model of the merchant, and judge the deviation degree between the express order data to be identified and the normal shipping behavior model of the merchant;

[0069] S502. If it is determined that the deviation degree is less than the preset threshold, it is determined that the express order data to be identified is real order data;

[0070] S503. If it is determined that the deviation degree is greater than the preset threshold, it is determined that the express order data to be identified is false shipment order data.

[0071] Specifically, in this embodiment, a preset threshold can be used to compare with the deviation degree. Considering the differences in the difficulty of logistics distribution in different regions, the standard can be appropriately relaxed. For example, for orders in remote areas, the preset threshold can be set larger. If it is determined that the deviation degree is less than the preset threshold, it is directly determined that the express order data to be identified is real order data. For example, during major e-commerce promotions, the order volume of merchants will increase significantly. Take a clothing merchant as an example. Usually, the order volume of this merchant is small and it can complete the delivery within 24 hours, and the logistics track update is relatively stable. However, during the promotion period, the residence time of the package at the transfer station in the logistics track also increases slightly, which is 1-2 hours more than usual, but it does not exceed the normal delivery delay range of the same route. At this time, the deviation degree is 1-2 hours, and when the deviation degree is less than the preset threshold (such as 3 hours), it is determined that the express order data to be identified is still real order data.

[0072] If it is determined that the deviation degree is greater than the preset threshold, it is determined that the express order data to be identified is false delivery order data. For example, if the delivery duration in the same type of area is more than twice the normal delivery duration, it is determined that the express order data including this delivery duration is false delivery order data. Therefore, by comparing the express order data to be identified with the normal delivery behavior model of the merchant, considering the changes in the delivery rhythm during the promotion period and the local delays caused by the logistics peak period and other situations, accurate and intelligent identification of false delivery order data can be achieved.

[0073] Please refer to Figure 6 , the sixth embodiment of a method for identifying false delivery of express in the embodiment of the present invention includes:

[0074] S601. After determining that the express order data to be identified is false delivery order data, send a warning to the e-commerce platform and send an instant message notification to the platform customer service team;

[0075] S602. Transmit the order details corresponding to the false delivery order data to the supervision background of the e-commerce platform, and the order details include the order number, merchant information, express order number, and suspicious feature description;

[0076] S603. Identify the express company to which the express order number belongs, and determine and filter the express order numbers of false delivery order data.

[0077] Specifically, in this embodiment, after determining that the express order data to be recognized is false delivery order data, an early warning is sent to the e-commerce platform, and the order details associated with the false delivery order data are transmitted to the regulatory background of the e-commerce platform in real time in the form of API data push, which facilitates the platform operators to quickly locate the problematic merchants and take regulatory measures, enabling the e-commerce platform to crack down on false delivery behaviors, safeguard the legitimate rights and interests of all parties, and avoid economic losses and bad shopping experiences of consumers. By sending instant message notifications to the platform customer service team, the platform customer service team can prepare in advance to handle consumer consultations and complaints, improving the quality of consumer services. By using the waybill query system of the express company to determine and filter the waybills of false delivery order data, a healthy and sustainable e-commerce environment is created.

[0078] Please refer to Figure 7 , the seventh embodiment of a method for identifying false delivery of express in the embodiments of the present invention includes:

[0079] S701. Use the logistics query interface to automatically identify the express company to which the waybill belongs;

[0080] S702. Based on the waybill verification rules and blacklist library of the express company, determine whether the waybill is a false waybill;

[0081] S703. If it is determined that the waybill is a false waybill, filter out the determined false waybill.

[0082] In this embodiment, by using the logistics query interface, real-time interaction with the waybill query systems of major express companies is realized. After an early warning occurs, the express company to which the waybill belongs is automatically identified. By combining the waybill verification rules and blacklist library of the express company, the authenticity of the waybill is quickly identified. For example, for waybills that do not conform to the waybill coding rules of the express company, have no records in the express company system, or are marked as abnormal multiple times in a short period, they are directly determined as false waybills, and the false waybills are filtered, thereby constructing an efficient interception mechanism, effectively blocking the circulation of false logistics information, reducing the operation risks of merchants, and at the same time improving the credibility of logistics data and the industry standardization.

[0083] The method for identifying false delivery of express in the embodiments of the present invention has been described above. Next, the device in the embodiments of the invention will be described. Please refer to Figure 8 , the implementation manner of the device for identifying false delivery of express in the embodiments of the present invention includes:

[0084] The first collection module 81 is used to mark and classify the collected express order data, and divide the express order data into real delivery order data and false delivery order data;

[0085] An extraction module 82, configured to extract time features and text features from the real delivery order data, train an LSTM long short-term memory network based on the time features and text features, and construct a normal delivery behavior model for merchants;

[0086] A second collection module 83, configured to collect the express order data to be identified in real time, compare the express order data to be identified with a preset standard and the normal delivery behavior model of merchants, and determine whether the express order data to be identified is false delivery order data;

[0087] A sending module 84, configured to, if it is determined that the express order data to be identified is false delivery order data, send a warning to the e-commerce platform and filter the express order numbers of the false delivery order data.

[0088] In this embodiment, the first collection module 81 includes: a first collection unit 811, configured to collect express order data by using an API open platform; a cleaning unit 812, configured to clean the collected express order data and remove invalid or incomplete records; a partitioning unit 813, configured to mark the cleaned express order data by using a marking tool, check the address consistency, and partition the express order data into real delivery order data and false delivery order data.

[0089] In this embodiment, the extraction module 82 includes: an extraction unit 821, configured to extract time features corresponding to order information from the real delivery order data; an analysis unit 822, configured to analyze the text data in the real delivery order data by using natural language technology and extract text features; a construction unit 823, configured to import the time features, text features, and marked real delivery order data as inputs into an LSTM long short-term memory network, and train and construct a normal delivery model for merchants.

[0090] In this embodiment, the second collection module 83 includes: a setting unit 831 for setting a preset standard, where the preset standard includes that there is a shipping notice but the goods are not shipped within 48 hours after the order payment, and the logistics track is not updated within 48 hours and exceeds the normal delivery delay range of the same route; a second collection unit 832 for collecting data of express orders to be identified in real time, comparing the data of express orders to be identified with the preset standard, and determining whether the data of express orders to be identified is false shipping order data; a first determination unit 833 for, if it is determined that there is no such behavior as there is a shipping notice but the goods are not shipped within 48 hours after the order payment, and the logistics track is not updated within 48 hours and exceeds the normal delivery delay range of the same route, comparing the data of express orders to be identified with the normal shipping behavior model of the merchant; and a second determination unit 834 for, if it is determined that there is any such behavior as there is a shipping notice but the goods are not shipped within 48 hours after the order payment, and the logistics track is not updated within 48 hours and exceeds the normal delivery delay range of the same route, determining that the data of express orders to be identified is false shipping order data.

[0091] In this embodiment, the first determination unit 833 includes: a first judgment subunit 8331 for comparing the data of express orders to be identified with the normal shipping behavior model of the merchant and judging the deviation degree between the data of express orders to be identified and the normal shipping behavior model of the merchant; a first determination subunit 8332 for, if it is determined that the deviation degree is less than a preset threshold, determining that the data of express orders to be identified is real order data; and a second determination subunit 8333 for, if it is determined that the deviation degree is greater than the preset threshold, determining that the data of express orders to be identified is false shipping order data.

[0092] In this embodiment, the sending module 84 includes: a sending unit 841 for, after determining that the data of express orders to be identified is false shipping order data, sending a warning to the e-commerce platform and sending an instant message notification to the platform customer service team; a transmission unit 842 for transmitting the order details corresponding to the false shipping order data to the supervision background of the e-commerce platform, where the order details include the order number, merchant information, express order number, and suspicious feature description; and an identification unit 843 for identifying the express company to which the express order number belongs and determining and filtering the express order numbers of false shipping order data.

[0093] In this embodiment, the identification unit 843 includes: an identification subunit 8431 for automatically identifying the express company to which the express order number belongs by using a logistics query interface; a second judgment subunit 8432 for judging whether the express order number is a false express order number based on the single-number verification rule and blacklist library of the express company; and a filtering subunit 8433 for, if it is determined that the express order number is a false express order number, filtering the determined false express order number.

[0094] In this embodiment, by marking and classifying the collected express order data, the collected express order data can be divided into real delivery order data and false delivery order data, so that a normal delivery model of merchants can be constructed using the real delivery order data. Secondly, based on the time features and text features in the extracted real delivery order data, the LSTM long short-term memory network is trained to construct a normal delivery behavior model that highly adapts to the behaviors of each merchant. When comparing the express order data to be identified with the constructed normal delivery behavior model of merchants, the misjudgment of false delivery order data can be reduced, the accuracy of judging false delivery order data can be improved, and the cost and time of manual review can be reduced. After determining the false express order data, a warning is sent to the e-commerce platform, enabling the e-commerce platform to take regulatory measures such as restricting the merchant's fund withdrawal, taking off the shelves of illegal goods, and reducing the store's search weight, improving the control ability of false express delivery behaviors, protecting the legitimate rights and interests of all parties, avoiding economic losses and bad shopping experiences suffered by consumers due to false delivery, and promoting the healthy and sustainable development of the e-commerce industry. In addition, after sending a warning to the e-commerce platform, the express waybill numbers of false delivery order data are filtered using the waybill number query system of the express company, optimizing the logistics data environment and ensuring the authenticity and reliability of the express delivery chain.

[0095] Figure 8 The structure of the express false delivery identification device shown does not limit the express false delivery identification device and can implement the steps of the express false delivery identification method provided by the above method embodiments.

[0096] Above Figure 8 The express false delivery identification device in the embodiments of the present invention is described in detail from the perspective of modular functional entities. Next, the express false delivery identification device in the embodiments of the present invention is described in detail from the perspective of hardware processing.

[0097] Figure 9 FIG. is a schematic structural diagram of an express false delivery identification device provided by an embodiment of the present invention. The device 90 may vary greatly due to different configurations or performances and may include one or more processors (central processing units, CPUs) 910 (for example, one or more processors) and a memory 920, and one or more storage media 930 (for example, one or more mass storage devices) storing application programs 933 or data 932. Among them, the memory 920 and the storage media 930 may be transient storage or persistent storage. The program stored in the storage media 930 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the device 90. Further, the processor 910 may be set to communicate with the storage media 930 and execute a series of instruction operations in the storage media on the device 90.

[0098] The device 90 may further include one or more power supplies 940, one or more wired or wireless network interfaces 950, one or more input / output interfaces 960, and / or one or more operating systems 931, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, and so on.

[0099] The embodiment of the present invention also provides a computer-readable storage medium. The computer-readable storage medium can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are run on a computer, the computer is caused to execute the steps of the method for identifying false delivery of express items.

[0100] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, or units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0101] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0102] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for identifying false express delivery, characterized in that: The express delivery false delivery identification method comprises: Marking and classifying the collected express order data, and dividing the express order data into real delivery order data and false delivery order data; Extracting time features and text features from the real shipping order data, training an LSTM long short-term memory network based on the time features and text features, and building a normal shipping behavior model for merchants; Collect the express order data to be identified in real time, compare the express order data to be identified with preset standards and the merchant's normal shipping behavior model, and determine whether the express order data to be identified is false shipping order data; If it is determined that the express order data to be identified is false delivery order data, an early warning is issued to the e-commerce platform, and the express order numbers of the false delivery order data are filtered.

2. The method for identifying false express delivery according to claim 1, characterized in that: The collected express order data is marked and classified, and the express order data is divided into real delivery order data and false delivery order data, including: Use the API open platform to collect express order data; Clean the collected express order data and remove invalid or incomplete records; The cleaned express order data is marked using a marking tool, and the address consistency is verified to divide the express order data into real delivery order data and false delivery order data.

3. The method for identifying false express delivery according to claim 1, characterized in that: The extracting time features and text features from the real shipping order data, training the LSTM long short-term memory network based on the time features and text features, and building a normal shipping behavior model for merchants includes: Extracting time features corresponding to order information from the real shipping order data; Using natural language technology, analyzing text data in the real shipping order data to extract text features; The time features and text features as well as the marked real delivery order data are imported as input into the LSTM long short-term memory network to train and build a normal delivery model for merchants.

4. The method for identifying false express delivery according to claim 1, characterized in that: The real-time collection of the express order data to be identified, comparing the express order data to be identified with a preset standard and a normal shipping behavior model of a merchant, and determining whether the express order data to be identified is false shipping order data, includes: Set preset standards, including that the order has not been shipped 48 hours after payment but there is a shipping notice, the logistics track has not been updated within 48 hours and exceeds the normal delivery delay range of the same route; Collecting the express order data to be identified in real time, comparing the express order data to be identified with a preset standard, and determining whether the express order data to be identified is false delivery order data; If it is determined that there is no delivery notification within 48 hours after order payment, or the logistics track has not been updated within 48 hours and exceeds the normal delivery delay range of the same route, the express order data to be identified will be compared with the merchant's normal delivery behavior model; If it is determined that the goods have not been shipped within 48 hours after the order payment but there is a shipping notification, or the logistics track has not been updated within 48 hours and exceeds the normal delivery delay range of the same route, then the express order data to be identified is determined to be false shipping order data.

5. The method for identifying false express delivery according to claim 4, characterized in that: If it is determined that there is no such behavior as the delivery notification is not issued within 48 hours after the order is paid, or the logistics track is not updated within 48 hours and exceeds the normal delivery delay range of the same route, the express order data to be identified is compared with the merchant's normal delivery behavior model, including: Compare the express order data to be identified with the merchant's normal shipping behavior model to determine the degree of deviation between the express order data to be identified and the merchant's normal shipping behavior model; If it is determined that the degree of deviation is less than the preset threshold, the express order data to be identified is determined to be real order data; If it is determined that the degree of deviation is greater than a preset threshold, the express order data to be identified is determined to be false delivery order data.

6. The method for identifying false express delivery according to claim 1, characterized in that: If it is determined that the express order data to be identified is false delivery order data, a warning is issued to the e-commerce platform, and the express order number of the false delivery order data is filtered, including: After determining that the express order data to be identified is false delivery order data, issuing an early warning to the e-commerce platform and sending an instant message notification to the platform customer service team; Transmit the order details corresponding to the fake delivery order data to the supervision backend of the e-commerce platform, wherein the order details include the order number, merchant information, express delivery number, and suspicious feature description; Identify the courier company to which the express tracking number belongs, and determine and filter the express tracking numbers of false delivery order data.

7. The method for identifying false express delivery according to claim 6, characterized in that: The method of identifying the express company to which the express order number belongs and determining and filtering the express order number of false delivery order data includes: Use the logistics query interface to automatically identify the courier company to which the express order number belongs; Based on the express company's tracking number verification rules and blacklist library, determine whether the express tracking number is a fake one; If the courier tracking number is determined to be a false courier tracking number, the false courier tracking number will be filtered out.

8. A device for identifying false express delivery, characterized in that: include: A first collection module is used to mark and classify the collected express order data, and divide the express order data into real delivery order data and false delivery order data; An extraction module is used to extract time features and text features from the real shipping order data, train an LSTM long short-term memory network based on the time features and text features, and build a normal shipping behavior model for merchants; The second collection module is used to collect the express order data to be identified in real time, compare the express order data to be identified with the preset standard and the merchant's normal shipping behavior model, and determine whether the express order data to be identified is false shipping order data; The sending module is used to send an early warning to the e-commerce platform if it is determined that the express order data to be identified is false delivery order data, and filter the express delivery order numbers of the false delivery order data.

9. A device for identifying false express delivery, characterized in that: comprising a memory and at least one processor, wherein the memory has computer-readable instructions stored therein; The at least one processor calls the computer-readable instructions in the memory to execute the various steps of the express false shipment identification method as described in any one of claims 1-7.

10. A computer-readable storage medium having computer-readable instructions stored thereon, characterized in that: When the computer-readable instructions are executed by the processor, the various steps of the method for identifying false express delivery as described in any one of claims 1-7 are implemented.

Images