Express headless piece waybill number matching method and device

Abstract

The application belongs to the technical field of data analysis and processing, and particularly relates to a method and device for matching express waybill numbers of headless parcels, which obtains characteristic information of a headless parcel, polls each piece of information of a stagnant parcel in a database, performs similarity matching on the characteristic information and the information of the stagnant parcel, and obtains a matching result, which is a waybill number list containing at least one waybill number, the waybill number being obtained from the information of the stagnant parcel with a similarity greater than a threshold value, so as to realize efficient and accurate matching of the waybill numbers of the headless parcels, and solve the problem that the existing scheme for manually matching the waybill numbers of the headless parcels cannot efficiently and accurately match the waybill numbers of the headless parcels.

Description

Technical Field

[0001] This invention belongs to the field of data analysis and processing technology, specifically relating to a method and apparatus for matching waybill numbers of express mail without headers. Background Technology

[0002] Currently, the express delivery industry is developing rapidly, and the number of parcels is constantly increasing. During transportation or sorting, waybills are prone to damage or detachment, resulting in "headless" parcels. This damage or detachment leads to parcel delays or losses, affecting continued delivery and fulfillment, further impacting user experience and increasing company operating costs. The current method for matching headless parcels involves sorting staff manually recording incomplete parcel information. Based on this partially available information, customer service personnel manually search for and determine the original tracking number by visual inspection. This method is inefficient and inaccurate. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a method and apparatus for matching waybill numbers of headless express parcels, so as to solve the problem that the existing manual matching scheme for headless express parcel waybill numbers cannot efficiently and accurately match the waybill numbers of headless parcels with those of express parcels that are delayed in logistics.

[0004] According to a first aspect of the present invention, a method for matching tracking numbers of express parcels without headers is provided, comprising:

[0005] Obtain feature information of headless packages, the feature information including at least: appearance information of the headless package and information of the express item;

[0006] Each piece of information about a stalled shipment is polled in the database, and the feature information is matched with the stalled shipment information to obtain a matching result. The matching result is a list of waybill numbers containing at least one waybill number, which is obtained from the stalled shipment information with a similarity greater than a threshold.

[0007] The information on the stalled items includes at least: the waybill number for each stalled item, the appearance information of the express item corresponding to each waybill number, and the information of the express item.

[0008] Preferably, for each piece of information about a stalled component in the polling database, similarity matching is performed between the feature information and the stalled component information, including:

[0009] If the headless item is a headless item without a label, a matching model is used to perform a single similarity matching on the feature information and the stationary item information to obtain the matching result;

[0010] If the headless item is a headless item with a damaged label, different matching models are used to perform multiple similarity matchings on the feature information and the stationary item information to obtain the matching result.

[0011] Preferably, if the headless item is a headless item without a label, a matching model is used to perform a single similarity matching on the feature information and the stationary item information to obtain the matching result, including:

[0012] By using a pre-set fuzzy matching algorithm, each piece of information about a stalled item in the database is polled, and the feature information is matched with the stalled item information to obtain a first list of waybills. The first list of waybills contains at least one waybills with a similarity greater than a first threshold.

[0013] Save the first list of waybill numbers as the matching result.

[0014] Preferably, if the headless item is a headless item with a damaged label, different matching models are used to perform multiple similarity matches on the feature information and the stationary item information to obtain matching results, including:

[0015] By using a pre-set fuzzy matching algorithm, each piece of information about a stalled item in the database is polled, and the feature information is matched with the stalled item information to obtain a first list of waybills. The first list of waybills contains at least one waybills with a similarity greater than a first threshold.

[0016] By using a pre-trained fine matching model, the information of the stagnant shipment corresponding to each waybill number in the waybill number list output by the fuzzy matching algorithm is polled in turn. The feature information and the stagnant shipment information are matched for similarity to obtain a second waybill number list. The second waybill number list stores at least one waybill number with a similarity greater than a second threshold.

[0017] Save the second list of waybill numbers as the matching result.

[0018] Preferably, the pre-set fuzzy matching algorithm polls each piece of information about a stalled shipment in the database, performs similarity matching between the feature information and the stalled shipment information, and obtains a first list of waybill numbers, including:

[0019] Each piece of information about a stalled shipment is polled in the database. The appearance information and the shipment information in the feature information are matched with the appearance information and the shipment information in the stalled shipment information using regular expressions to obtain a first list of waybill numbers.

[0020] Preferably, the fine-grained matching model is a Word2Vec model. The pre-trained fine-grained matching model iteratively polls the information on stalled shipments corresponding to each waybill number in the waybill number list output by the fuzzy matching algorithm, and performs similarity matching between the feature information and the stalled shipment information to obtain a second waybill number list, including:

[0021] The system polls the information of the stalled shipment corresponding to each waybill number in the list of waybill numbers output by the fuzzy matching algorithm, and then feeds the feature information and the stalled shipment information into the trained Word2Vec model for similarity matching to obtain the second waybill number list.

[0022] Preferably, obtaining the feature information of the headless component includes:

[0023] Obtain the ID of the headless item from the headless item declaration and claim system;

[0024] The feature information of the headless component is determined based on the mapping relationship between the headless component ID and the feature information of the headless component.

[0025] Preferably, the feature information further includes: information on the damaged waybill; the information on the stalled shipment further includes: sender information and stall duration of the stalled shipment; and the method for obtaining the information on the stalled shipment includes:

[0026] If the incomplete waybill information includes sender information for items without headers, then from the idle item information in the database, match idle item information that is identical to the sender information but has an idle time exceeding a preset time.

[0027] If the incomplete waybill information does not include sender information for items without headers, then all items with a standby time exceeding a preset time will be matched from the standby item information in the database.

[0028] Preferably, the method further includes:

[0029] Based on business rules and the branch's claiming capacity, the top N matching results with high similarity are pushed; wherein, the business rules are the relationship between the number of pushes N and the branch's claiming capacity, and the branch's claiming capacity is the number of personnel at the branch that can process the pushed orders.

[0030] According to a second aspect of the present invention, a device for matching tracking numbers for express mail without headers includes:

[0031] The data acquisition module is used to acquire feature information of headless packages, the feature information including at least: appearance information of the headless package and information of the express item;

[0032] The tracking number matching module is used to poll each piece of information about a stalled shipment in the database, perform similarity matching between the feature information and the stalled shipment information, and obtain a matching result. The matching result is a list of tracking numbers containing at least one tracking number, which is obtained from the stalled shipment information with a similarity greater than a threshold. The stalled shipment information includes at least: the tracking number of each stalled shipment, the appearance information of the express shipment corresponding to each tracking number, and the express item information.

[0033] The tracking number push module is used to push the top N tracking numbers with high similarity from the matching results based on business rules and the branch's claiming capacity; wherein, the business rules are the relationship between the number of tracking numbers N pushed and the branch's claiming capacity, and the branch's claiming capacity is the number of personnel at the branch that can process the pushed tracking numbers.

[0034] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0035] By acquiring the feature information of headless packages, polling each piece of information about stalled packages in the database, and performing similarity matching between the feature information and the stalled package information, a matching result is obtained. The matching result is a list of waybill numbers containing at least one waybill number. The waybill number is obtained from the stalled package information with a similarity greater than a threshold. This achieves efficient and accurate matching of waybill numbers for headless packages, solving the problem that existing manual matching schemes for headless package waybill numbers cannot perform efficient and accurate matching.

[0036] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0037] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0038] Figure 1 This is a flowchart illustrating a method for matching waybill numbers for headless express parcels, according to an exemplary embodiment.

[0039] Figure 2 This is a flowchart illustrating a method for matching waybill numbers for headless express parcels, according to another exemplary embodiment.

[0040] Figure 3 This is a block diagram illustrating a headless express waybill number matching device according to an exemplary embodiment. Detailed Implementation

[0041] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the invention as detailed in the appended claims.

[0042] Please see Figure 1 , Figure 1 This is a flowchart illustrating a method for matching tracking numbers for headless express shipments, according to an exemplary embodiment. Figure 1 As shown, the methods for matching tracking numbers for packages without headers include:

[0043] Step S11: Obtain the feature information of the headless package, wherein the feature information includes at least: the appearance information of the headless package and the package information;

[0044] Step S12: Poll each piece of information about a stalled shipment in the database, perform similarity matching between the feature information and the stalled shipment information, and obtain a matching result. The matching result is a list of waybill numbers containing at least one waybill number, and the waybill number is obtained from the stalled shipment information with a similarity greater than a threshold.

[0045] The information on the stalled items includes at least: the waybill number for each stalled item, the appearance information of the express item corresponding to each waybill number, and the information of the express item.

[0046] It should be noted that, in practice, the express mail waybill number matching method provided in this embodiment needs to be loaded into the controller of the electronic device to run.

[0047] The electronic devices include, but are not limited to: mobile phones, tablets, desktop computers, etc.

[0048] The "feature information of the headless package" mentioned in step S11 may include: appearance information of the headless package, express item information, information on the damaged waybill, and logistics trajectory information.

[0049] In practice, the appearance information of a headless package includes: its size, volume, and weight; the item information of a headless package includes: the name and category of the item; and the information on a damaged waybill for a headless package includes: the tracking number, sender information, and recipient information.

[0050] The "stagnation information" mentioned in step S12 includes at least the following: the tracking number of each stagnant item, the appearance information of the express item corresponding to each tracking number, the express item information, the sender information, the stagnancy duration, and the logistics trajectory information.

[0051] In practice, the appearance information of a stopped item includes: the size, volume, and weight of the item without a header; the express item information of an item without a header includes: the name of the item and the category of the item.

[0052] It is understood that the technical solution provided in this embodiment obtains the feature information of headless packages, polls each piece of information about stalled packages in the database, performs similarity matching between the feature information and the stalled package information, and obtains a matching result. The matching result is a list of waybill numbers containing at least one waybill number. The waybill number is obtained from the stalled package information with a similarity greater than a threshold, thereby achieving efficient and accurate matching of waybill numbers for headless express packages. This solves the problem that existing manual matching schemes for headless express package waybill numbers cannot perform efficient and accurate matching.

[0053] Step S11, "obtaining the feature information of headless components," specifically refers to:

[0054] Obtain the ID of the headless item from the headless item declaration and claim system;

[0055] The feature information of the headless component is determined based on the mapping relationship between the headless component ID and the feature information of the headless component.

[0056] In practice, terminal devices or PCs are used to manually identify headless parcels, collect and fill in their characteristic information, such as information related to incomplete waybills (incomplete waybill information, incomplete recipient / sender information, etc.) as well as information such as category, weight, and volume. When declaring a headless parcel, its characteristic information and the corresponding transit station information must be entered into the designated headless parcel declaration and claim system. After the information is entered, each headless parcel has a unique ID in the headless parcel declaration and claim system. Subsequent use only requires knowing the headless parcel ID to retrieve the corresponding characteristic information through the mapping relationship between the headless parcel ID and its characteristic information.

[0057] The method for "obtaining the information of the stalled component" in step S12 is as follows:

[0058] If the incomplete waybill information includes sender information for items without headers, then from the idle item information in the database, match idle item information that is identical to the sender information but has an idle time exceeding a preset time.

[0059] If the incomplete waybill information does not include sender information for items without headers, then all items with a standby time exceeding a preset time will be matched from the standby item information in the database.

[0060] Understandably, the main stages of express delivery involve five transit points: receiving, sending, arriving, dispatching, and signing. We need to use data aggregation to identify waybills that have been stalled for 24 hours at any of these transit stages. Specifically, the big data platform uses the open-source Hadoop system and its big data components Hive and Spark to collect data from each transit stage and filter out information on stalled shipments that meet the specified criteria.

[0061] In practice, if the incomplete waybill information of a headless package contains sender information, when obtaining information on delayed shipments, it is necessary to filter from the big data platform to find shipments with the same sender information as the headless package's incomplete waybill information and whose delay time exceeds 24 hours. If the headless package's incomplete waybill information does not contain sender information, when obtaining information on delayed shipments, it is only necessary to filter from the big data platform to find shipments whose delay time exceeds 24 hours. Specifically, a delay time exceeding 24 hours refers to a delay time of 24 hours or more based on the last scan record during the transit process.

[0062] In practice, step S12, "polling each piece of information about a stalled component in the database and performing similarity matching between the feature information and the stalled component information," includes:

[0063] 1. If the headless item is a headless item without a label, a matching model is used to perform a single similarity matching between the feature information and the stationary item information to obtain the matching result, specifically:

[0064] By using a pre-set fuzzy matching algorithm, each piece of information about a stalled item in the database is polled, and the feature information is matched with the stalled item information to obtain a first list of waybills. The first list of waybills contains at least one waybills with a similarity greater than a first threshold.

[0065] Save the first list of waybill numbers as the matching result.

[0066] Understandably, by using a pre-set fuzzy matching algorithm, each piece of information about stalled shipments in the database is polled, and the feature information is matched with the stalled shipment information to obtain a first list of waybill numbers. The first threshold is set according to the actual situation. In specific practice, the first list of waybill numbers is obtained by using a pre-set fuzzy matching algorithm to poll each piece of information about stalled shipments in the database, and the feature information is matched with the stalled shipment information to obtain a first list of waybill numbers.

[0067] Each piece of information about a stalled shipment is polled in the database. The appearance information and the shipment information in the feature information are matched with the appearance information and the shipment information in the stalled shipment information using regular expressions to obtain a first list of waybill numbers.

[0068] In practice, before executing the "polling of every piece of information on stalled shipments in the database," the process also includes: data cleaning of the feature information. This data cleaning includes at least formatting, which involves validating the feature information and filling in missing parts with default values. Specifically, residual waybill information and graphics are often just text and images, filled out incorrectly. Standardizing these graphics and text, and replacing missing parts with default values, facilitates matching.

[0069] In practice, the obtained information on unclaimed items is matched with existing information on stalled items. Specifically, routing and flow direction matching are performed first, and then regular expressions are used to match weight, images, shapes, etc., to obtain a list of first tracking numbers with a similarity greater than a first threshold. Taking the tracking number of a damaged item as an example, the damaged item tracking number currently only contains the four digits '4567' in the middle, so fuzzy matching will match all tracking numbers containing '4567' in the middle. In addition to the damaged item tracking number, regular expression matching is also performed on the recipient, sender, recipient address, sender address, three-segment code (the three-segment code is the first three segments of numbers at the beginning of the tracking number, the first segment is the terminal center, the second segment is the network point, and the third segment is the delivery person; the three-segment code uses numbers to represent the delivery person's information at the delivery network point), item name, category, and merchant name. The overall matching accuracy rate = matching length / residual order length is greater than 80% if it is a match. The weighted calculation of each residual information is then performed to obtain the final fuzzy matching score. If the fuzzy matching score is greater than 80%, the waybill number is added to the first waybill number list.

[0070] 2. If the headless item is a headless item with a damaged label, different matching models are used to perform multiple similarity matchings on the feature information and the stationary item information to obtain the matching results, specifically:

[0071] By using a pre-set fuzzy matching algorithm, each piece of information about a stalled item in the database is polled, and the feature information is matched with the stalled item information to obtain a first list of waybills. The first list of waybills contains at least one waybills with a similarity greater than a first threshold.

[0072] By using a pre-trained fine matching model, the information of the stagnant shipment corresponding to each waybill number in the waybill number list output by the fuzzy matching algorithm is polled in turn. The feature information and the stagnant shipment information are matched for similarity to obtain a second waybill number list. The second waybill number list stores at least one waybill number with a similarity greater than a second threshold.

[0073] Save the second list of waybill numbers as the matching result.

[0074] It is understood that the refined matching model is a Word2Vec model. Through a pre-trained refined matching model, the system iteratively queries the information on stalled shipments corresponding to each waybill number in the waybill number list output by the fuzzy matching algorithm, performs similarity matching between the feature information and the stalled shipment information, and obtains a second waybill number list. The second threshold is set according to the actual situation. In specific practice, the second waybill number list is obtained by iteratively querying the information on stalled shipments corresponding to each waybill number in the waybill number list output by the fuzzy matching algorithm, and performing similarity matching between the feature information and the stalled shipment information.

[0075] The system iterates through the list of waybill numbers output by the fuzzy matching algorithm to obtain information on delayed shipments corresponding to each waybill number. The feature information and the delayed shipment information are then fed into a trained Word2Vec model for similarity matching to obtain a second list of waybill numbers. In practice, before executing "iterate through each piece of delayed shipment information in the database," the system further includes: data cleaning of the feature information. This data cleaning includes at least formatting, which involves validating the feature information and filling in missing parts with default values. Specifically, the remaining waybill information and graphics are merely text and graphics, often filled in incorrectly. Standardizing these graphics and text, and replacing missing parts with default values, facilitates matching.

[0076] In practice, since the remaining order details may not be the beginning or end of the order, word segmentation is used for arbitrary matching. Combined with information such as address and shape, the Word2Vec model is used for precise matching at the order level, resulting in a list of second order numbers with a similarity greater than a second threshold. The Word2Vec model is trained offline, typically once a day. In practice, based on fuzzy matching, Word2Vec is used with word vector matching, and finally, combined with real-time data and user historical lost order data, further score ranking is performed to generate a final accurate ranking score. Ranking is based on the score. Here, Word2Vec is mainly used for matching addresses, product names, and categories.

[0077] The method for matching tracking numbers for unmarked parcels also includes:

[0078] Based on business rules and the branch's claiming capacity, the top N matching results with high similarity are pushed; wherein, the business rules are the relationship between the number of pushes N and the branch's claiming capacity, and the branch's claiming capacity is the number of personnel at the branch that can process the pushed orders.

[0079] Understandably, N waybill numbers can be pushed based on the specific number of personnel at the branch and the corresponding business rules, where N is determined according to the actual business rules and the branch's claiming capacity.

[0080] In practice, before pushing the matching results, the waybill number information in the matching results can be synchronized in real time. Specifically, it checks one by one whether the logistics trajectory data corresponding to each waybill number in the matching results has changed. If so, the corresponding waybill number in the matching results is deleted, and the list of deleted waybill numbers is used as the matching results, sorted in descending order of matching similarity. It is understandable that setting up real-time synchronization ensures that the latest waybill information is updated promptly, reducing the probability of false matches.

[0081] In practice, N waybill numbers are pushed based on the specific number of personnel at each branch and the corresponding business rules. For example, the push is mainly based on the personnel configuration of different branches. Some branches with sufficient human resources will generally push the waybill number with the highest similarity in the matching results; some branches with insufficient human resources will push the waybill number with the highest similarity in the matching results.

[0082] Please see Figure 2 , Figure 2 This is a flowchart illustrating a method for matching tracking numbers for headless express shipments, according to another exemplary embodiment. Figure 2 As shown, the specific process for matching tracking numbers when the package has no header but a waybill includes:

[0083] Step S21: Data synchronization between the big data platform and the source system yields a data file in HDFS format. This data file includes: basic waybill information, recipient / sender information, parcel dwell time, parcel item information, and vehicle-related information, etc. The basic waybill information specifically includes: waybill number, and the appearance information and logistics trajectory information of the parcel corresponding to the waybill number.

[0084] Step S22: The big data platform cleans the customer's historical lost item information and filters out express delivery information that is the same as the sender information in the headless item information;

[0085] Step S23: The big data platform further filters the express delivery information obtained in step S22, and filters out express delivery information that has been idle for more than 24 hours. This express delivery information is idle information.

[0086] Step S24: The machine learning platform obtains information about stalled parts;

[0087] Step S25: Obtain the information of headless items with shipping labels from the operator's reported headless item information;

[0088] Step S26: Clean and structure the residual waybill data of the acquired headless item information;

[0089] Step S27: The machine learning platform obtains the information on stalled items obtained in step S24 and the information on headless items obtained in step S26, and performs fuzzy matching using regular expressions to obtain the first list of waybill numbers.

[0090] Step S28: The machine learning platform performs fine matching based on the information of the stagnant items corresponding to the waybill numbers in the first waybill number list obtained in step S27 and the information of the headless items obtained in step S26, using the trained Word2Vec model to obtain the second waybill number list.

[0091] Step S29: The machine learning platform performs real-time synchronization and online matching based on the fine matching data (second waybill number list) obtained in step S28 and the basic waybill information in the source system data to obtain the final matching result;

[0092] Step S30: Obtain matching results from the headless recommendation service;

[0093] Step S31: Push waybill numbers based on business rules and branch recognition capabilities;

[0094] Step S32: The system for reporting and claiming unclaimed packages obtains the pushed tracking number and recommends damaged packages.

[0095] Step S33: Relevant parties claim the recommended tracking number and process shipments without headers;

[0096] Step S34: Effect evaluation.

[0097] It should be noted that the technical solution provided in this embodiment needs to be loaded and run in the controller of an electronic device. The electronic device includes, but is not limited to, mobile phones, tablets, desktop computers, etc.

[0098] In practice, the effectiveness evaluation in step S34 is mainly used to improve and track the matching strategy through two indicators: data coverage and claim accuracy. The matching strategy should cover as many headless item recommendations as possible, and have as many claim recommendation numbers for each headless item as possible. Regarding data coverage, for example, if 10,000 headless item defective items are registered, and 8,000 are recommended according to the rules, then the coverage is 8,000 / 10,000 = 80%. Regarding claim accuracy, for example, if the rules recommend 8,000 items, and 6,000 of them are ultimately claimed as recommended results, then the accuracy is 6,000 / 8,000 = 75%. The final evaluation effect is 80% * 75% = 60%, and a higher evaluation effect is better, up to a maximum of 100%. Based on the evaluation effect, the matching strategy can be continuously iterated and optimized.

[0099] It is understood that the technical solution provided in this embodiment obtains the feature information of headless packages, polls each piece of information about stalled packages in the database, performs similarity matching between the feature information and the stalled package information, and obtains a matching result. The matching result is a list of waybill numbers containing at least one waybill number. The waybill number is obtained from the stalled package information with a similarity greater than a threshold, thereby achieving efficient and accurate matching of waybill numbers for headless express packages. This solves the problem that existing manual matching schemes for headless express package waybill numbers cannot perform efficient and accurate matching.

[0100] Furthermore, the technical solution provided in this embodiment can improve and track the matching strategy through effect evaluation. Based on the evaluation effect, the matching strategy can be continuously iterated and optimized, thereby further improving the accuracy of matching.

[0101] Please see Figure 3 , Figure 3 This is a block diagram illustrating a headless express waybill number matching device according to an exemplary embodiment, such as... Figure 3 As shown, the express delivery tracking number matching device 100 specifically includes:

[0102] Data acquisition module 101 is used to acquire feature information of headless packages, the feature information including at least: appearance information of headless packages and express item information;

[0103] The tracking number matching module 102 is used to poll each piece of information about a stalled shipment in the database, perform similarity matching between the feature information and the stalled shipment information, and obtain a matching result. The matching result is a list of tracking numbers containing at least one tracking number, which is obtained from the stalled shipment information with a similarity greater than a threshold. The stalled shipment information includes at least: the tracking number of each stalled shipment, the appearance information of the express shipment corresponding to each tracking number, and the express item information.

[0104] The tracking number push module 103 is used to push the top N tracking numbers with high similarity from the matching results according to business rules and the branch's claiming capacity; wherein, the business rules are the relationship between the number of tracking numbers N pushed and the branch's claiming capacity, and the branch's claiming capacity is the number of personnel at the branch that can process the pushed tracking numbers.

[0105] The specific operation methods of each module of the express mail tracking number matching device in this embodiment have been described in detail in the embodiments of the above-mentioned related methods, and will not be elaborated here.

[0106] It is understood that the technical solution provided in this embodiment obtains the feature information of headless packages, polls each piece of information about stalled packages in the database, performs similarity matching between the feature information and the stalled package information, and obtains a matching result. The matching result is a list of waybill numbers containing at least one waybill number. The waybill number is obtained from the stalled package information with a similarity greater than a threshold, thereby achieving efficient and accurate matching of waybill numbers for headless express packages. This solves the problem that existing manual matching schemes for headless express package waybill numbers cannot perform efficient and accurate matching.

[0107] Furthermore, this application provides a computer-readable storage medium storing computer instructions for causing a computer to perform the steps of any of the methods described above. The storage medium may be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.; the storage medium may also include combinations of the above types of memory.

[0108] It is understood that the same or similar parts in the above embodiments can be referred to each other, and the contents not described in detail in some embodiments can be referred to the same or similar contents in other embodiments.

[0109] It should be noted that in the description of this application, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, in the description of this application, unless otherwise stated, "multiple" or "more" means at least two.

[0110] It should be understood that when an element is referred to as "fixed to" or "set on" another element, it may be directly on the other element or may have an intervening element present at the same time; when an element is referred to as "connected to" another element, it may be directly connected to the other element or may have an intervening element present at the same time. In addition, the term "connected" as used herein may include wireless connections; the word "and / or" as used includes any unit and all combinations of one or more of the associated listed items.

[0111] Any process or method description in the flowchart or otherwise herein can be understood as: representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the function involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0112] It should be understood that various parts of this 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 memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0113] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0114] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0115] The storage media mentioned above can be read-only memory, disk, or optical disk, etc.

[0116] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0117] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A method for matching tracking numbers of express parcels without headers, characterized in that, include: Obtain feature information of headless packages, the feature information including at least: appearance information of the headless package and information of the express item; Each piece of information about a stalled shipment is polled in the database, and a similarity match is performed between the feature information and the stalled shipment information to obtain a matching result. This includes: if the headless shipment is a headless shipment without a waybill, a single similarity match is performed between the feature information and the stalled shipment information using a matching model to obtain a matching result; if the headless shipment is a headless shipment with a damaged waybill, multiple similarity matches are performed between the feature information and the stalled shipment information using different matching models to obtain matching results; the matching result is a list of waybill numbers containing at least one waybill number, which is obtained from stalled shipment information with a similarity greater than a threshold. The information on the stalled items includes at least: the waybill number of each stalled item, the appearance information of the express item corresponding to each waybill number, and the information of the express item. If the headless item is a headless item with a damaged label, different matching models are used to perform multiple similarity matches on the feature information and the stationary item information to obtain matching results, including: Using a pre-set fuzzy matching model, each piece of information about a stalled shipment in the database is polled, and the feature information is matched with the stalled shipment information to obtain a first list of waybill numbers. The first list of waybill numbers contains at least one waybill number with a similarity greater than a first threshold. Using a pre-trained fine matching model, the stalled shipment information corresponding to each waybill number in the waybill number list output by the fuzzy matching model is polled, and the feature information is matched with the stalled shipment information to obtain a second list of waybill numbers. The second list of waybill numbers contains at least one waybill number with a similarity greater than a second threshold. The second list of waybill numbers is then saved as the matching result. The fine-grained matching model is a Word2Vec model. The pre-trained fine-grained matching model iteratively polls the information on stalled shipments corresponding to each waybill number in the waybill number list output by the fuzzy matching model, and performs similarity matching between the feature information and the stalled shipment information to obtain a second waybill number list, including: The system polls the information of the stalled shipment corresponding to each waybill number in the waybill number list output by the fuzzy matching model, and then feeds the feature information and the stalled shipment information into the trained Word2Vec model for similarity matching to obtain the second waybill number list. The feature information also includes: information on the damaged waybill; the information on the stalled shipment also includes: sender information and stall duration; the method further includes: If the incomplete waybill information includes sender information for items without headers, then from the inactive item information in the database, match inactive item information that has the same sender information but has been inactive for more than a preset duration; if the incomplete waybill information does not include sender information for items without headers, then from the inactive item information in the database, match all inactive item information that has been inactive for more than a preset duration.

2. The method according to claim 1, characterized in that, If the headless item is a headless item without a shipping label, a matching model is used to perform a single similarity match between the feature information and the stationary item information to obtain the matching result, including: By using a pre-set fuzzy matching model, each piece of information about a stalled item in the database is polled, and the feature information is matched with the stalled item information to obtain a first list of waybills. The first list of waybills contains at least one waybill number with a similarity greater than a first threshold. Save the first list of waybill numbers as the matching result.

3. The method according to claim 2, characterized in that, The process involves using a pre-set fuzzy matching model to iterate through each piece of information about stalled shipments in the database, performing similarity matching between the feature information and the stalled shipment information to obtain a first list of waybill numbers, including: Each piece of information about a stalled shipment is polled in the database. The appearance information and the shipment information in the feature information are matched with the appearance information and the shipment information in the stalled shipment information using regular expressions to obtain a first list of waybill numbers.

4. The method according to claim 1, characterized in that, The acquisition of headless feature information includes: Obtain the ID of the headless item from the headless item declaration and claim system; The feature information of the headless component is determined based on the mapping relationship between the headless component ID and the feature information of the headless component.

5. The method according to any one of claims 1 to 4, characterized in that, The method further includes: Based on business rules and the branch's claiming capacity, the top N matching results with high similarity are pushed; wherein, the business rules are the relationship between the number of pushes N and the branch's claiming capacity, and the branch's claiming capacity is the number of personnel at the branch that can process the pushed orders.

6. A device for matching tracking numbers for express mail without headers, characterized in that, include: The data acquisition module is used to acquire feature information of headless packages, the feature information including at least: appearance information of the headless package and information of the express item; The tracking number matching module is used to poll each piece of information about a stalled shipment in the database, perform similarity matching between the feature information and the stalled shipment information, and obtain a matching result. This includes: if the stalled shipment is without a waybill, a single similarity matching is performed between the feature information and the stalled shipment information using one matching model to obtain a matching result; if the stalled shipment is with a damaged waybill, multiple similarity matching is performed between the feature information and the stalled shipment information using different matching models to obtain a matching result; the matching result is a list of tracking numbers containing at least one tracking number, obtained from stalled shipment information with a similarity greater than a threshold; wherein, the stalled shipment information includes at least: the tracking number of each stalled shipment, the appearance information of the corresponding express shipment, and the information of the express item; If the headless item is a headless item with a damaged label, different matching models are used to perform multiple similarity matches on the feature information and the stationary item information to obtain matching results, including: Using a pre-set fuzzy matching model, each piece of information about a stalled shipment in the database is polled, and the feature information is matched with the stalled shipment information to obtain a first list of waybill numbers. The first list of waybill numbers contains at least one waybill number with a similarity greater than a first threshold. Using a pre-trained fine matching model, the stalled shipment information corresponding to each waybill number in the waybill number list output by the fuzzy matching model is polled, and the feature information is matched with the stalled shipment information to obtain a second list of waybill numbers. The second list of waybill numbers contains at least one waybill number with a similarity greater than a second threshold. The second list of waybill numbers is then saved as the matching result. The fine-grained matching model is a Word2Vec model. The pre-trained fine-grained matching model iteratively polls the information on stalled shipments corresponding to each waybill number in the waybill number list output by the fuzzy matching model, and performs similarity matching between the feature information and the stalled shipment information to obtain a second waybill number list, including: The system polls the information of the stalled shipment corresponding to each waybill number in the waybill number list output by the fuzzy matching model, and then feeds the feature information and the stalled shipment information into the trained Word2Vec model for similarity matching to obtain the second waybill number list. The feature information also includes: incomplete waybill information; the stalled shipment information also includes: sender information and stall duration of the stalled shipment; the tracking number matching module is further configured to: if the incomplete waybill information includes sender information for shipments without headers, then match stalled shipment information from the stalled shipment information in the database that has the same sender information but a stall duration exceeding a preset duration; if the incomplete waybill information does not include sender information for shipments without headers, then match all stalled shipment information from the stalled shipment information in the database that has a stall duration exceeding a preset duration; The tracking number push module is used to push the top N tracking numbers with high similarity from the matching results based on business rules and the branch's claiming capacity; wherein, the business rules are the relationship between the number of tracking numbers N pushed and the branch's claiming capacity, and the branch's claiming capacity is the number of personnel at the branch that can process the pushed tracking numbers.

Images