Goods source pushing method, electronic device, storage medium and program product

Abstract

This application provides a cargo source recommendation method, electronic device, storage medium, and program product, relating to the field of freight technology. The cargo source recommendation method includes: responding to target cargo source information published by a cargo owner, determining a target heavy cargo bin in a preset cargo source pool; updating the target heavy cargo bin based on the target cargo source information; determining the ranking score of the target cargo source in the target heavy cargo bin based on preset scoring rules; generating heavy cargo recommendation information based on the ranking score of the target cargo source and the ranking scores of other cargo sources in the target heavy cargo bin; and displaying the heavy cargo recommendation information to the target driver group when pushing the target cargo source to the target driver group. This application provides a technical solution that allows drivers to have as complete a understanding of heavy cargo information as possible, enabling them to quickly identify the most attractive cargo sources among these duplicate cargo sources, thereby improving the efficiency of cargo acceptance decisions.

Description

TECHNICAL FIELD

[0001] The present application relates to the freight transportation technical field, and in particular to a freight source pushing method, an electronic device, a storage medium and a program product. BACKGROUND

[0002] Heavy freight is a common problem encountered by current freight platforms. In order to find a carrier as soon as possible, a company with transportation needs will send the demand to multiple different freight agents. In order to make the freight sources they are agents for be seen by more drivers, these freight agents will publish these freight sources on the freight platform multiple times to increase the chance of transaction. If the freight platform does not process these repeated freight sources and directly displays them to the drivers, it will seriously interfere with the efficiency of the drivers' decision-making when looking for freight, increase the possibility of the drivers missing more attractive freight sources, and lead to the dissatisfaction of the drivers with the matching platform. In severe cases, the drivers will give up using the matching platform.

[0003] The existing heavy freight identification method directly makes decisions such as freight source unloading and display folding after identifying heavy freight. These display methods are not conducive to the drivers to master the complete information of heavy freight, and do not distinguish the attraction level of heavy freight, which cannot help the drivers quickly locate the better choice in heavy freight, and thus affect the efficiency of receiving freight. SUMMARY

[0004] The present application provides a freight source pushing method, an electronic device, a storage medium and a program product to provide a technical solution that can enable the drivers to understand the heavy freight information as completely as possible, so that the drivers can quickly identify the most attractive freight source in these repeated freight sources, thereby improving the efficiency of receiving freight decision-making.

[0005] In a first aspect, the present application provides a freight source pushing method, which comprises:

[0006] In response to the information of the target freight source published by the shipper end, determining a target heavy freight bucket in a preset freight source pool;

[0007] Updating the target heavy freight bucket based on the information of the target freight source;

[0008] Determining the ranking score of the target freight source in the target heavy freight bucket based on a preset scoring rule;

[0009] Generating heavy freight recommendation information according to the ranking score of the target freight source and the ranking score of other freight sources in the target heavy freight bucket, and displaying the heavy freight recommendation information to the target driver group when pushing the target freight source to the target driver group.

[0010] In an optional implementation, the determining of the target heavy freight bucket in the preset freight source pool in response to the information of the target freight source published by the shipper end comprises:

[0011] In response to the target cargo information released by the cargo owner, a target label corresponding to the target cargo is constructed based on a preset discrete value labeling rule;

[0012] Match the target heavy cargo barrel corresponding to the target label in the preset cargo pool.

[0013] In one optional implementation, in response to the target cargo information published by the cargo owner, constructing a target tag corresponding to the target cargo based on a preset discrete value tagging rule includes:

[0014] In response to the target cargo information published by the cargo owner, the origin address, destination address, loading and unloading method, classification identifier and address granularity of the target cargo are extracted from the target cargo information;

[0015] According to the address granularity, the origin address and destination address of the target cargo are converted into standardized origin address and standardized destination address;

[0016] Based on the loading and unloading method, determine the standardized loading and unloading method code corresponding to the target cargo source, and based on the classification identifier, determine the standardized classification code corresponding to the target cargo source;

[0017] According to the preset discrete value labeling rules, a target label corresponding to the target cargo source is constructed based on the standardized origin address, the standardized destination address, the standardized loading and unloading method code, and the standardized classification code.

[0018] In one optional implementation, updating the target heavy cargo container based on the target cargo source information includes:

[0019] Based on the information of the target goods source, construct a target goods source tuple;

[0020] The target heavy cargo bins are updated based on the target cargo source tuple and the preset heavy cargo bin update rules.

[0021] In one optional implementation, updating the target heavy cargo bin according to the target cargo source tuple and a preset heavy cargo bin update rule includes:

[0022] If the number of elements in the target heavy cargo bin is less than a preset capacity threshold, the system determines whether there is a heavy cargo tuple that is duplicated with the target cargo tuple based on the deduplication verification field in the target cargo tuple.

[0023] If it exists, then replace the specified element of the heavy source tuple with the specified element in the target source tuple;

[0024] If it does not exist, the target source tuple is written into the target source bucket;

[0025] The deduplication verification fields include: an identifier field representing the cargo owner, a field representing the freight, a field representing the transportation deposit, and a field representing the cargo owner's requirements for the vehicle.

[0026] The specified elements include: an identifier field representing the source of goods and a field representing the latest update time of the source of goods.

[0027] In one optional implementation, updating the target heavy cargo bin according to the target cargo source tuple and a preset heavy cargo bin update rule includes:

[0028] If the number of elements in the target heavy cargo bin is greater than or equal to the preset capacity threshold, based on the deduplication verification field in the target cargo tuple, it is determined whether there is a heavy cargo tuple that is duplicated with the target cargo tuple in the target cargo tuple.

[0029] If it exists, then replace the specified element of the heavy source tuple with the specified element in the target source tuple;

[0030] If it does not exist, a preset compression algorithm is used to compress multiple source tuples with the same deduplication verification field in the target source bucket, and the source tuple with the latest update time is retained to update the target source bucket.

[0031] If the number of elements in the updated target source bucket is less than the preset capacity threshold, the target source tuple is written into the updated target source bucket.

[0032] If the number of elements in the updated target source bucket is greater than or equal to the preset capacity threshold, a preset sampling algorithm is used to sequentially delete the delisted source tuples, the source tuples with shipping costs less than the preset shipping costs, and the source tuples with deposits higher than the preset deposits, until the number of elements in the updated target source bucket is less than the preset capacity threshold, and then the target source tuples are written into the updated target source bucket.

[0033] The deduplication verification fields include: a field representing the shipper's identifier, a field representing the freight, a field representing the transportation deposit, and a field representing the minimum vehicle length requirement.

[0034] The specified elements include: an identifier field representing the source of goods and a field representing the latest update time of the source of goods.

[0035] In one optional implementation, determining the ranking score of the target cargo source in the target heavy cargo bin based on preset scoring rules includes:

[0036] Identify a set of similar products on the shelves within the target product container;

[0037] Based on the deposit amount, freight, platform service fee, and minimum vehicle length requirement of each similar cargo tuple in the set of similar cargo sources, determine the deposit type and net freight for each of the similar cargo sources;

[0038] Based on the relationship between the deposit types of the target goods and other similar goods in the target goods container, the corresponding scoring rules are determined;

[0039] Based on the corresponding scoring rules and the relationship between the net freight cost of the target cargo and the net freight cost of each of the similar cargo sources, the ranking score of the target cargo in the target heavy cargo barrel is determined.

[0040] In one optional implementation, determining the corresponding scoring rules based on the relationship between the deposit types of the target source and other similar sources in the target source container includes:

[0041] If the deposit type for each similar source in the target source container includes non-refundable deposit, the corresponding scoring rule is determined to be a mixed deposit type scoring rule.

[0042] If all similar sources in the target source container have refundable deposits, then the corresponding scoring rule is determined to be the fully refundable deposit scoring rule.

[0043] In one optional implementation, heavy cargo recommendation information is generated based on the ranking score of the target cargo source and the ranking scores of other cargo sources in the target heavy cargo bin. When the target cargo source is pushed to the target driver group, displaying the heavy cargo recommendation information to the target driver group includes:

[0044] Based on each source tuple in the target heavy cargo bin, the sorting score of the target source, and the sorting scores and shelf status of other sources in the target heavy cargo bin, a shelf-ready source sorting list is determined; wherein, the sources in the shelf-ready source sorting list are arranged in descending order according to their sorting scores.

[0045] If the ranking score of the target cargo source is less than or equal to the ranking score of at least one similar cargo source in the cargo source ranking list, then heavy cargo recommendation information is generated based on the information of the at least one similar cargo source. While pushing the target cargo source to the target driver group, the at least one similar cargo source is also pushed to the target driver through a preset viewing entry.

[0046] If the ranking score of the target cargo is greater than the ranking score of any similar cargo in the cargo ranking list, then the target cargo is pushed to the target driver group.

[0047] If the target cargo is the only one listed in the inventory sorting list, then the target cargo is pushed to the target driver group.

[0048] In an optional implementation, before generating heavy cargo recommendation information based on the ranking score of the target cargo source and the ranking scores of other cargo sources in the target heavy cargo bin, the method further includes: determining the recommendation score of the target cargo source based on the information of the target cargo source;

[0049] Based on the recommendation score of the target source of goods, the push method of the target source of goods is determined, wherein the push method includes: immediate push and delayed push.

[0050] When pushing the target cargo source to the target driver group, displaying the heavy cargo recommendation information to the target driver group includes:

[0051] According to the target cargo source push method, when the target cargo source is pushed to the target driver group, the heavy cargo recommendation information is displayed to the target driver group.

[0052] In a second aspect, this application provides an electronic device, comprising: a memory and a processor; the memory being configured to store computer program instructions; and the processor being configured to execute the computer program instructions, such that the electronic device performs the method described in the first aspect.

[0053] Thirdly, this application provides a computer-readable storage medium including computer program instructions, wherein an electronic device executes the computer program instructions to cause the electronic device to perform the method as described in the first aspect.

[0054] Fourthly, this application provides a computer program product containing instructions that, when run on a computer, causes the computer to perform the steps of the method provided in the first aspect.

[0055] This application provides a cargo recommendation method, an electronic device, a storage medium, and a program product. The cargo recommendation method provided in this application includes: responding to information about target cargo posted by a cargo owner, determining a target heavy cargo bin in a preset cargo pool; updating the target heavy cargo bin based on the information about the target cargo; determining a ranking score for the target cargo in the target heavy cargo bin based on preset scoring rules; generating cargo recommendation information based on the ranking scores of each cargo in the target heavy cargo bin, and pushing the cargo recommendation information to a target group of drivers; wherein the cargo recommendation information includes the target cargo.

[0056] As can be seen, the embodiments of this application aggregate target cargo sources and cargo sources that overlap with the target cargo sources by targeting heavy cargo bins. At the same time, the heavy cargo is stratified by attraction based on the ranking score. Drivers do not need to manually filter and can quickly locate the better choice among the heavy cargo, which greatly reduces the time cost of finding cargo.

[0057] Furthermore, unlike the delisting / folding method, this embodiment aggregates heavy goods and retains them in the target heavy goods bin. High-value goods (including target goods) are displayed to drivers through cargo recommendation information. In addition, this embodiment sorts and classifies the attractiveness level of each heavy goods, so that highly attractive heavy goods are displayed in the recommendation information. This satisfies the exposure needs of cargo owners / agents and helps drivers reach high-quality cargo first, thus improving matching efficiency and transaction probability in both directions. It solves the problem of information redundancy and ensures that drivers can grasp the information of heavy goods, reducing the decline in platform trust caused by information blocking. Attached Figure Description

[0058] Figure 1 A flowchart illustrating the steps of a product delivery method provided in this application embodiment;

[0059] Figure 2 A flowchart illustrating the specific steps of a product delivery method provided in this application embodiment;

[0060] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0061] Heavy freight is a common problem for freight platforms. Companies with transportation needs will send their requests to multiple freight agents in order to find carriers as quickly as possible. In order to increase the chances of closing deals, these freight agents will post these freight listings multiple times on the freight platform. If the freight platform does not process these duplicate listings and directly displays them to drivers, it will seriously interfere with the efficiency of drivers' decision-making when looking for freight, increase the possibility that drivers will miss more attractive freight, and lead to dissatisfaction with the matching platform. In severe cases, drivers will abandon the matching platform altogether.

[0062] Existing heavy cargo identification methods make decisions such as removing the cargo from the shelves or folding the display after identifying heavy cargo. These display methods are not conducive to drivers grasping complete information about heavy cargo, and they do not distinguish the attractiveness levels of heavy cargo, which cannot help drivers quickly locate the better choice among heavy cargo, thus affecting the efficiency of cargo receiving.

[0063] Based on this, the technical concept of this application embodiment is: to provide a way to allow drivers to see as complete repeated cargo information as possible, understand the composition of selectable cargo sources on the current cargo-finding route, and make selection decisions by comparing representative cargo sources from heavy cargo with non-heavy cargo in the same list, and quickly identify the most attractive cargo source among these repeated cargo sources, thereby improving the efficiency of drivers' cargo acceptance decisions.

[0064] The technical solutions shown in this application will now be described in detail through specific embodiments. It should be noted that the following embodiments may exist independently or in combination with each other; for identical or similar content, the description will not be repeated in different embodiments.

[0065] It is worth noting that in the embodiments of this application, the terms "heavy goods" and "similar goods" refer to the same type of entity. Generally, "heavy goods" is used in the context of backend calculation, while "similar goods" is used in the context of product frontend display.

[0066] Reference Figure 1 In a first aspect, embodiments of this application provide a method for pushing goods to a system, the method comprising:

[0067] S100, in response to the target cargo information published by the cargo owner, determines the target heavy cargo barrel in the preset cargo pool.

[0068] In this embodiment of the application, after the cargo owner completes the publication of the target cargo information in the cargo source push system, the cargo source push system receives and verifies the legality of the target cargo information, and then determines the target heavy cargo barrel in the preset cargo source pool.

[0069] Next, the core discrete fields of the target cargo information (originating address, destination address, loading / unloading method, secondary category target identification, and address granularity) are extracted and standardized according to preset rules, such as converting the address to an administrative division ID and the loading / unloading method / secondary category to a unified code. Based on the standardized discrete fields, a unique target label is formed. The format of the target label can be: Origin ID corresponding to the address granularity - Destination ID - Loading / unloading method code - Secondary category code.

[0070] Optionally, a preset supply pool is used to store existing supplies in the supply push system. The storage structure can be an in-memory key-value storage structure, where Key = discrete tag, and Value = set of similar supplies, i.e., heavy goods bins. The system queries the Key corresponding to the target tag:

[0071] If the key is matched, the corresponding value (a set of similar goods) is the target heavy cargo container;

[0072] Key not hit: Create a new empty heavy cargo container as the target heavy cargo container, and store the target tag-new heavy cargo container mapping relationship in the cargo source pool.

[0073] Based on the above description, by using discrete tags and KV storage binning logic, duplicate cargo sources with the same transportation needs are concentrated into the same heavy cargo bin, which fundamentally avoids the scattered display of duplicate cargo sources and reduces information redundancy.

[0074] S200, based on the information of the target cargo source, update the target heavy cargo barrel.

[0075] Optionally, the information of the target cargo source is first encapsulated into standardized target cargo source tuples, and then written into the target heavy cargo bin. It should be understood that the cargo sources stored in the target heavy cargo bin are all stored in tuple form.

[0076] In some examples, the target cargo is written directly into the target heavy cargo bin;

[0077] In other instances, the target cargo is deduplicated before being written into the target heavy cargo bin.

[0078] S300, based on preset scoring rules, determine the ranking score of the target cargo source in the target heavy cargo barrel.

[0079] It should be understood that the ranking of target cargo within the target cargo bucket is used to quantify the attractiveness of the target cargo to drivers in closing deals.

[0080] The preset scoring rules can be a divide-and-conquer approach. By using this approach, fuzzy decision-making factors such as freight and deposits are transformed into ranking scores, allowing drivers to quickly assess the quality of cargo and shorten decision-making time.

[0081] S400: Based on the ranking score of the target cargo source and the ranking scores of other cargo sources in the target heavy cargo bin, generate heavy cargo recommendation information, and display the heavy cargo recommendation information to the target driver group when pushing the target cargo source to the target driver group.

[0082] In this embodiment, based on the attributes of the target cargo (e.g., minimum vehicle length requirement, transportation route, secondary category), driver registration information (vehicle type, frequently used routes, historical order categories) is matched to filter out a target driver group that has the qualifications to accept orders and whose needs match. This target driver group can be a single driver or multiple drivers; this embodiment does not impose any special limitations on this.

[0083] The recommended cargo information includes similar cargo sources identified in the target heavy cargo bin, as well as the target cargo source itself.

[0084] Based on the above description, the embodiments of this application aggregate target cargo sources and cargo sources that overlap with the target cargo sources by targeting heavy cargo bins. At the same time, the heavy cargo is stratified by attractiveness based on the ranking score. Drivers do not need to manually filter and can quickly locate the better choice among the heavy cargo, which greatly reduces the time cost of finding cargo.

[0085] Furthermore, unlike the delisting / folding method, this application embodiment aggregates heavy goods and retains them in the target heavy goods bin. The full amount of heavy goods (including the target cargo) is displayed to drivers through cargo source recommendation information. In addition, this application embodiment also sorts and classifies the attractiveness level of each heavy goods, so that highly attractive heavy goods are displayed in the recommendation information. This satisfies the exposure needs of cargo owners / agents and helps drivers to reach high-quality cargo sources first, thus improving matching efficiency and transaction probability in both directions. It solves the problem of information redundancy and ensures that drivers can grasp the information of heavy goods, reducing the decline in platform trust caused by information blocking.

[0086] In one optional implementation, determining the target heavy cargo container in a preset cargo pool in response to the target cargo source information published by the cargo owner includes:

[0087] First, in response to the target cargo information released by the cargo owner, a target label corresponding to the target cargo is constructed based on a preset discrete value labeling rule.

[0088] In this embodiment of the application, the target label is a unified code for similar cargo sources, which is used to classify duplicate cargo sources with the same transportation needs into the same target heavy cargo container through standardized discrete field combinations.

[0089] Duplicate cargo sources with the same transportation needs (such as the same route, the same loading and unloading method, and the same cargo type) will generate the same target label, while cargo sources with different needs will generate different labels, thereby realizing a classification method that puts similar heavy cargo into the same bin and distinguishes different cargo sources.

[0090] The preset discrete value labeling rules are standardized operations for constructing target labels. They are used to standardize the extraction, encoding, and concatenation logic of discrete fields, ensuring that similar products of the same type generate unique and consistent target labels.

[0091] Then, the target heavy cargo barrel corresponding to the target label is matched in the preset cargo pool.

[0092] Optionally, the preset cargo pool uses a KV database to store cargo information. The Key serves as a unique ID for a set of similar cargo sources, ensuring that similar cargo sources of the same type are accurately assigned to the same heavy cargo bin, and that different types of cargo sources are strictly distinguished.

[0093] In this embodiment of the application, the target label is used as the unique query key, and the exact matching mode of the KV database is adopted. Only when there is a key in the database that is exactly the same as the target label, mismatch of goods source is avoided.

[0094] The system sends a Get(Key=target tag) request to the in-memory key-value database. Based on the database's response, it handles two scenarios: a match and a no-match, ensuring the effective identification of the target heavy cargo bin.

[0095] If a match is found, the KV database returns a non-empty Value, which is the heavy cargo bin (a set of standardized tuples storing similar goods) bound to the target tag. This heavy cargo bin is then used as the target heavy cargo bin.

[0096] If no match is found, the KV database returns an empty result, meaning that the current preset source pool does not contain the key corresponding to the target tag, or the key once existed but the corresponding heavy cargo bin has no valid sources and has been cleaned up. Then, an empty heavy cargo bin is initialized, with a default maximum capacity of 50 source tuples, sorted by update timestamp, and basic rules such as 30-day expiration and cleanup. A key-value pair (Key=target tag, Value=new heavy cargo bin) is added to the KV database to complete the mapping association between the tag and the bin. This newly created empty bin is then used as the target heavy cargo bin.

[0097] Based on the above description, the target heavy cargo bin defines a unique data range of similar cargo sources for subsequent stages. Heavy cargo updates only perform deduplication / capacity control on cargo sources within the bin, and the scoring and ranking only compare the attractiveness of cargo sources within the bin, which can avoid logical confusion caused by cross-category operations.

[0098] Optionally, in response to the target cargo information published by the cargo owner, constructing the target tag corresponding to the target cargo based on a preset discrete value tagging rule may include:

[0099] The first step is to respond to the target cargo information published by the cargo owner and extract the origin address, destination address, loading and unloading method, classification identifier, and address granularity of the target cargo from the target cargo information.

[0100] Among them, the origin address of the target cargo is the starting location of the cargo filled in by the cargo owner, the destination address is the ending location of the cargo filled in by the cargo owner, the loading and unloading method is the cargo loading and unloading mode selected by the cargo owner, the category identifier is the category description of the cargo by the cargo owner, and the address granularity is the address matching accuracy preset by the system, which is selected by the cargo owner or defaulted by the system according to the business scenario, such as city level, district / county level, and township level.

[0101] The second step is to convert the origin and destination addresses of the target cargo into standardized origin and destination addresses according to the address granularity.

[0102] In this embodiment of the application, in order to eliminate the ambiguity of the address text description, the origin and destination addresses of the target cargo are converted into standardized administrative division IDs according to the selected address granularity, so as to obtain standardized origin addresses and standardized destination addresses.

[0103] Specifically, first remove irrelevant information (such as house number, street name, and park name) from the origin and destination addresses of the target goods, and only retain the administrative district information that matches the address granularity:

[0104] Then, the administrative region information is matched with the preset administrative division standard library to generate standardized origin address and standardized destination address corresponding to the address granularity. The format is provincial administrative region + municipal administrative region (municipal granularity) or provincial administrative region + municipal administrative region + district / county administrative region (district / county granularity).

[0105] Example: If the granularity is at the city level and the origin address is Hangzhou, then the corresponding standardized origin address is Province A, City B;

[0106] If the granularity is at the district / county level, and the origin address is District C, then the standardized origin address is District C, City B, Province A.

[0107] The third step is to determine the standardized loading and unloading method code corresponding to the target cargo source based on the loading and unloading method, and to determine the standardized classification code corresponding to the target cargo source based on the classification identifier.

[0108] In this embodiment of the application, for the two non-address fields of loading and unloading method and classification identifier, the text description is converted into discrete code by a preset encoding rule to avoid label mismatch caused by different expressions of the same meaning.

[0109] Specifically, the preset coding rules for loading and unloading methods to standardized coding can be as follows: door-to-door is coded as 1, station-to-station is coded as 2, door-to-station is coded as 3, and station-to-door is coded as 4.

[0110] The preset coding rule for the classification label to the standardized classification code can use a 6-digit code, with the first 2 digits for the first-level category, the middle 2 digits for the second-level category, and the last 2 digits for the third-level category, to cover the mainstream freight categories, and reserve space for extension.

[0111] If the cargo owner fills in a multi-level category (such as general goods-home appliances-refrigerator), the code will be matched directly; if it is a single-level description (such as fresh produce), the third-level category will be matched by default.

[0112] The fourth step is to construct a target label corresponding to the target cargo source based on the standardized origin address, the standardized destination address, the standardized loading and unloading method code, and the standardized classification code, according to the preset discrete value labeling rules.

[0113] In this step, according to the preset discrete value labeling rules, the standardized origin address and standardized destination address generated in the second step are combined with the standardized loading and unloading method code and standardized classification code generated in the third step in a fixed format to form the target label.

[0114] For example, the target label can be formatted in the order of standardized origin address, standardized destination address, standardized loading / unloading method code, and standardized classification code.

[0115] It is worth noting that the target tags are in string format, and the length varies slightly depending on the address granularity, but the structure is uniform, which facilitates memory key-value storage and querying.

[0116] Based on the above description, this application embodiment, through cleaning and standard library matching, converts all addresses, regardless of whether the consignor fills in City A or the area code of City A, into the same standardized address, thus improving the accuracy of geographical matching. Furthermore, this application embodiment strictly distinguishes between city-level and district / county-level granularity to avoid addresses of different granularities being misclassified as similar cargo sources; combined with the standardization of loading / unloading methods and classification coding, the accuracy of identifying similar cargo sources is improved, ensuring that heavy goods with the same transportation needs are grouped into the same target label.

[0117] As a specific example of the above content, since the severity of heavy cargo varies greatly in different regions, the amount of heavy cargo in a single shipment can reach more than 400 shipments in a 24-hour period of a certain time and space. Therefore, this example takes into account the handling of such extreme cases to avoid serious degradation of the response time of front-end requests due to the huge quantity.

[0118] This example maintains a Key Valuation pair table in the system's in-memory KV database. It uses pre-constructed tag buckets to store the tuple information of the goods to enable fast querying between a given goods and its similar goods, as detailed below;

[0119] Target labels are pre-constructed using discrete-valued variables: these labels serve as the IDs of similar product sets and are designed using a structured approach, as detailed below:

[0120] The tag structure is as follows: When using city-level address judgment: departure city ID - destination city ID - loading / unloading method code - secondary category code; When using district / county-level address judgment: departure district ID - destination district ID - loading / unloading method code - secondary category code; When querying duplicate cargo sources, construct target tags based on the information of the given cargo source, use the target tags to query all similar cargo sources that meet the conditions and their corresponding target cargo bins, and then, based on the weight of the target cargo source, filter the cargo sources in the target weight bins with a weight difference within 20% as the results.

[0121] In one optional implementation, updating the target heavy cargo container based on the target cargo source information includes:

[0122] First, based on the information of the target source of goods, a target source of goods tuple is constructed.

[0123] The information regarding the target source of goods has already been described above and will not be repeated here.

[0124] Optionally, the target cargo tuple includes: a unique identifier for the target cargo, a unique identifier for the cargo owner / agent who posted the target cargo, the latest update timestamp of the target cargo (automatically refreshed when posted / modified), the original freight amount filled in by the cargo owner when posting, the order deposit amount set by the cargo owner, the minimum length requirement of the vehicle required for the target cargo, and the current shelf status of the target cargo.

[0125] Then, the target heavy cargo bins are updated according to the target cargo source tuple and the preset heavy cargo bin update rules.

[0126] In this embodiment, the preset heavy cargo bin update rule is used to remove redundant data from the cargo sources in the target heavy cargo bin in order to control the bin capacity and retain valid cargo sources. The update operation can be performed in the order of deduplication, capacity control, and state synchronization based on the comparison between the target cargo source tuple and the existing data in the target cargo source bin.

[0127] As an example, updating the target heavy cargo bin based on the target cargo source tuple and the preset heavy cargo bin update rules includes:

[0128] If the number of elements in the target heavy cargo bin is less than a preset capacity threshold, the system determines whether there is a heavy cargo tuple that is duplicated with the target cargo tuple based on the deduplication verification field in the target cargo tuple.

[0129] If it exists, then replace the specified element of the heavy source tuple with the specified element in the target source tuple;

[0130] If it does not exist, the target source tuple is written into the target source bucket;

[0131] The deduplication verification fields include: an identifier field representing the cargo owner, a field representing the freight, a field representing the transportation deposit, and a field representing the cargo owner's requirements for the vehicle.

[0132] The specified elements include: an identifier field representing the source of goods and a field representing the latest update time of the source of goods.

[0133] It should be understood that the identifier field representing the shipper is used to ensure that the target cargo and the heavy cargo tuples belong to the same entity that posted the cargo. The freight field is used to ensure that the freight amounts for the target cargo and the heavy cargo tuples are exactly the same. The freight deposit field is used to ensure that the deposit amounts for the target cargo and the heavy cargo tuples are exactly the same. The vehicle requirements field representing the shipper is used to ensure that the vehicle length requirements for the target cargo and the heavy cargo tuples are exactly the same.

[0134] When there are duplicate source tuples, replacing the specified element of the duplicate source tuple with the specified element in the target source tuple includes: replacing the identifier field representing the source in the duplicate source tuple with the identifier field representing the source in the target source tuple, and replacing the field representing the latest update time of the source in the duplicate source tuple with the field representing the latest update time of the source in the target source tuple.

[0135] Specifically, replacing the identifier field representing the source in the duplicate source tuple with the identifier field representing the source in the target source tuple ensures that the same source uses the newly published / modified source ID. Replacing the field representing the latest update time of the source in the duplicate source tuple with the field representing the latest update time of the source in the target source tuple ensures that the target source bucket stores the latest version of the duplicate source.

[0136] When there are no duplicate cargo tuples, insert the target cargo tuple completely into the target heavy cargo bin. After insertion, sort the tuples in descending order by update timestamp to facilitate the priority display of the latest cargo and improve the efficiency of driver viewing.

[0137] It is worth noting that after the target source tuple is written into the target source bucket, the number of source tuples in the target source bucket must still be less than or equal to the preset capacity threshold.

[0138] For example: The target source bucket originally contains 30 tuples (<50 threshold). The target source tuples are unique. After direct insertion, the number of elements in the bucket becomes 31, and the tuple is sorted by timestamp to the corresponding position (the newest one is ranked first).

[0139] Based on the above description, in this example, multiple deduplication verification fields are used to jointly verify that only goods with the same exact demand published by the same subject are judged as duplicate goods, resulting in high deduplication accuracy. This avoids duplicate storage of the same goods, reduces redundant data in the heavy goods bin, and improves storage efficiency and subsequent data processing speed.

[0140] Furthermore, in this example, there is no need to replace all tuples; only the specified elements are updated. This ensures that the source identification and timeliness are up-to-date while maintaining the consistency of other fields, such as shipping costs and deposits, which do not need to be rewritten if they remain unchanged.

[0141] Finally, in this example, writing / replacing cargo sources is only performed when the number of tuples in the bin is less than a preset capacity threshold, ensuring that the heavy cargo bin is always in a lightweight state (the maximum number of tuples does not exceed the threshold). Subsequent rating, sorting, and recommendation displays only need to process a limited number of tuples, significantly reducing the computational load and meeting the real-time display requirements of the driver's end.

[0142] As another example, updating the target heavy cargo bin based on the target cargo source tuple and the preset heavy cargo bin update rules includes:

[0143] If the number of elements in the target heavy cargo bin is greater than or equal to the preset capacity threshold, based on the deduplication verification field in the target cargo tuple, it is determined whether there is a heavy cargo tuple that is duplicated with the target cargo tuple in the target cargo tuple.

[0144] If it exists, then replace the specified element of the heavy source tuple with the specified element in the target source tuple;

[0145] If it does not exist, a preset compression algorithm is used to compress multiple source tuples with the same deduplication verification field in the target source bucket, and the source tuple with the latest update time is retained to update the target source bucket.

[0146] If the number of elements in the updated target source bucket is less than the preset capacity threshold, the target source tuple is written into the updated target source bucket.

[0147] If the number of elements in the updated target source bucket is greater than or equal to the preset capacity threshold, a preset sampling algorithm is used to sequentially delete the delisted source tuples, the source tuples with shipping costs less than the preset shipping costs, and the source tuples with deposits higher than the preset deposits, until the number of elements in the updated target source bucket is less than the preset capacity threshold, and then the target source tuples are written into the updated target source bucket.

[0148] The deduplication verification fields include: a field representing the shipper's identifier, a field representing the freight, a field representing the transportation deposit, and a field representing the minimum vehicle length requirement.

[0149] The specified elements include: an identifier field representing the source of goods and a field representing the latest update time of the source of goods.

[0150] In this example, all tuples in the target heavy cargo bin are traversed. If the deduplication check field of a certain tuple is exactly the same as the deduplication check field of the target cargo source tuple, it is determined to be a heavy cargo source tuple; if all tuples have at least one field difference, they are determined to be a non-duplicate cargo source.

[0151] If it exists, replace the identifier field representing the source of goods in the heavy source tuple with the identifier field representing the source of goods in the target source tuple, and replace the field representing the latest update time of the source of goods in the heavy source tuple with the field representing the latest update time of the source of goods in the target source tuple, without changing other valid fields of the heavy source tuple (such as freight, deposit, vehicle length requirements) to avoid information loss.

[0152] If not, the preset compression algorithm is used to group all tuples in the bucket according to the deduplication verification field. The four deduplication fields of the tuples in the same group are completely identical. Only the tuple with the latest update timestamp is retained in each group, and other old tuples in the group are removed to eliminate the historical duplicate sources of goods published by the same cargo owner (such as old versions that have not been deleted after multiple modifications) and reduce the number of elements in the bucket without affecting the data quality.

[0153] If the number of elements in the updated target source bucket is less than the preset capacity threshold, the target source tuple is written into the updated target source bucket.

[0154] If the number of elements in the updated target source bucket is greater than or equal to the preset capacity threshold, a preset sampling algorithm is used to sequentially delete the delisted source tuples, the source tuples with shipping costs less than the preset shipping costs, and the source tuples with deposits higher than the preset deposits, until the number of elements in the updated target source bucket is less than the preset capacity threshold, and then the target source tuples are written into the updated target source bucket.

[0155] Based on the above description, this example uses a preset compression algorithm and a preset sampling algorithm to control the capacity of heavy cargo barrels within a threshold (e.g., 50 barrels), so that subsequent processes only need to process finite element sets.

[0156] This example first identifies duplicate listings using a deduplication validation field, then removes older data from the same group using a compression algorithm. This reduces the proportion of redundant data within the listing bucket. Furthermore, compression retains tuples with the most recent update time, and sampling deletion prioritizes high-value, valid listings, improving both the efficiency and freshness of the data within the listing bucket. Finally, this example deletes low-value listings in the order of being on the shelf, low shipping costs, and high deposits, while prioritizing the retention of highly attractive listings that are already on the shelf, have high shipping costs, and low deposits, thus improving the matching accuracy of listing recommendations.

[0157] In a specific example, the data structure of the similar product bucket can be as follows: the similar product bucket is constructed as a set, and the data structure of the elements in the set is an n-tuple. The structure and special value conventions of the tuple content are as follows:

[0158] (cargo_id (shipper ID), shipper_id (shipper ID), cargo_time (shipper ID), shipper fills in the freight, deposit amount, minimum required vehicle length, and on-shelf status).

[0159] If the goods are priceless, the freight cost should be recorded as 0 by the consignor. If the price is per ton or cubic meter, it should be uniformly processed as a trip price. If the deposit is refundable, the deposit amount should be recorded as 0.

[0160] Preset compression and sampling algorithms: When the number of heavy goods is large, in order to cope with the risk of query timeout caused by extreme data skew, preset compression and sampling methods are adopted to limit the number of tuples in the bucket to below the target value while retaining representative heavy goods and the most attractive heavy goods, thereby eliminating the negative impact of extreme cases.

[0161] Preset compression algorithm: For tuples with the same shipper_id, only the latest record with different combinations of (shipper_id, freight rate filled in by the shipper, deposit amount, minimum required vehicle length) is retained (the new and old records are determined by the create_time value).

[0162] Preset sampling algorithm: A maximum of 50 tuples are retained in the similar product bin. When the number of tuples exceeds 50, products are deleted in the following order:

[0163] Update the inventory status, delete all delisted inventory, and stop if there are no more than 50 remaining elements;

[0164] Next, group by minimum vehicle length, and delete the bottom 40% of non-zero freight owners sorted in descending order. Stop if there are no more than 50 remaining elements.

[0165] Finally, group by minimum vehicle length and delete the top 40% of tuples sorted in descending order of deposit amount from the sources of goods with a freight charge of 0 (do not delete if the top 40% is less than 1). Stop when the remaining elements in the bucket do not exceed 50.

[0166] If the number of elements in the bucket is still greater than 50 after one round of deletion, repeat the above steps until the number of elements in the bucket does not exceed 50.

[0167] Target cargo addition update: When the number of elements in the bucket exceeds 50, if the (shipper_id, freight, deposit amount, minimum required vehicle length) of the target cargo is duplicated in the corresponding heavy cargo bucket, the cargo_id and cargo_time in the tuple are updated with the latest cargo_time according to the preset compression algorithm.

[0168] Add elements directly if the number of elements in the bucket does not exceed 50.

[0169] Existing product source modifications: If a product source is modified, the similar product source set will be updated according to the following rules:

[0170] The modifiable fields for the source of goods have no effect on the combined key (the name of the goods, address, and loading / unloading method cannot be modified), so the list of similar sources of goods itself is unaffected;

[0171] When the cargo type is changed from carpooling to full truckload, the tagging calculation of similar cargo needs to be added;

[0172] If the deposit amount, whether the deposit is refundable, or the minimum required vehicle length are changed, the heavy cargo calculation needs to be updated.

[0173] Tag lifecycle management and expired source deletion: If a new element is added to the value set of a structured federated key, the TTL of that key is renewed, extending it to 24 hours. If no new sources are updated for that key within 24 hours, the key is automatically dropped upon expiration.

[0174] In one optional implementation, determining the ranking score of the target cargo source in the target heavy cargo bin based on preset scoring rules includes:

[0175] First, identify a set of similar products on the shelf in the target product bin.

[0176] In this embodiment of the application, all tuples in the target heavy cargo bin are traversed, only tuples with the status word "on shelf" are retained, and tuples with the status of "off shelf", "sold", or "invalid" are automatically removed to obtain a set of similar cargo sources for the shelf.

[0177] Therefore, the resulting set of similar product sources includes the target product source tuple and other similar product source tuples that are on the shelves, and all tuples satisfy the similar product source attribute under the same target label.

[0178] Second, based on the deposit amount, freight, platform service fee, and minimum vehicle length requirement of each similar cargo tuple in the similar cargo set, determine the deposit type and net freight for each of the similar cargo sources.

[0179] In this embodiment of the application, if the cargo owner sets the deposit to be refundable, then when scoring heavy cargo, the non-refundable deposit amount is recorded as 0; otherwise, it is the deposit amount entered.

[0180] In this embodiment of the application, the net freight cost = freight cost filled in by the consignor - non-refundable deposit amount; in the case of a refundable deposit, the non-refundable deposit amount is recorded as 0.0 yuan.

[0181] For special scenarios: If the cargo owner fills in freight = 0.00 yuan (for goods without a price), the net freight will be uniformly recorded as 0.00; if it is a ton-volume quoted price (such as 300 yuan / ton), the ton-volume to trip price will be converted according to the platform's preset rules, such as combining the weight, volume and common loading capacity of the goods to the trip price, and then the platform service fee will be subtracted to obtain the net freight.

[0182] Third, based on the relationship between the deposit types of the target source of goods and other similar sources of goods in the target source of goods container, the corresponding scoring rules are determined.

[0183] In this embodiment of the application, the corresponding scoring rules are matched according to the deposit type distribution (refundable / non-refundable combination) of all tuples in the similar source set to ensure that the scoring logic fits the actual order-taking scenario.

[0184] Optionally, determining the corresponding scoring rules based on the relationship between the deposit types of the target source and other similar sources in the target source container includes:

[0185] First, if the deposit type for each similar source in the target source barrel includes non-refundable deposit, the corresponding scoring rule is determined to be a mixed deposit type scoring rule.

[0186] Then, if the deposit type for each similar source in the target source bin is refundable, the corresponding scoring rule is determined to be the fully refundable deposit scoring rule.

[0187] In this embodiment of the application, based on the aforementioned standardized deposit type field (deposit type field = 0 indicates that the deposit is refundable, otherwise it is non-refundable), deposit type statistics are performed on all tuples in the set of similar goods sources on the target heavy goods barrel: the number of goods sources with non-refundable deposits in the set is counted, denoted as N1, and the number of goods sources with refundable deposits is denoted as N2.

[0188] Example: The set of similar products on the shelf contains 30 similar product tuples (including the target product). Among the 29 tuples excluding the target product, 15 are non-refundable deposits (N1=15) and 14 are refundable deposits (N2=14), so there are two types of deposits. If all 29 tuples are refundable deposits (N1=0, N2=29), then there is only one type of deposit.

[0189] If N1≥1 and N2≥1 (both non-refundable and refundable deposits exist simultaneously), the corresponding scoring rule is a mixed deposit type scoring rule. In this case, the driver needs to make a decision between high-yield (potentially accompanied by high risk) and low-risk (potentially lower yield) cargo sources, and needs to balance the yield and risk weights.

[0190] Fourth, based on the corresponding scoring rules and the relationship between the net freight cost of the target cargo and the net freight cost of each of the similar cargo sources, the ranking score of the target cargo in the target heavy cargo barrel is determined.

[0191] In this embodiment of the application, based on the scoring rules of the third step of matching, and combined with the comparison of the net freight cost of the target cargo source with other cargo sources in the set, the ranking score of the target cargo source is calculated. The ranking score of the target cargo source ranges from [0,1], and the higher the score, the stronger the attractiveness.

[0192] For the mixed deposit type scoring rules, drivers need to balance net freight and deposit type to avoid overestimating high-risk cargo due to high returns, or underestimating low-risk cargo due to slightly lower returns.

[0193] Extract the net freight cost of all goods in the set of similar goods on the shelf, and determine the maximum value Max_F;

[0194] Set risk weight coefficients: if the deposit is refundable, the weight = 1.0; if the deposit is non-refundable, the weight = 0.

[0195] The ranking score is calculated as follows: Ranking Score = (Net Freight Cost of Target Cargo / Max_F) × Risk Weight Coefficient of Target Cargo.

[0196] If N1=0 and N2≥1 (all similar goods are refundable deposits), then the corresponding scoring rule is the fully refundable deposit scoring rule. In this case, drivers have no difference in order acceptance risk, and drivers only focus on actual income (net freight). The scoring logic can focus on the income dimension.

[0197] For the fully refundable deposit rating rules, when there is no difference in risk among drivers, the core evaluation criterion is only the actual income. The higher the net freight earned, the higher the ranking score.

[0198] Extract the net freight cost of all goods in the set of similar goods on the shelf, and determine the maximum value Max_F;

[0199] Ranking score calculation formula: Ranking score = Net freight cost of target cargo / Max_F;

[0200] Example: In the set of similar products on the shelf, Max_F = 3600 yuan, the net shipping cost of the target product is 3200 yuan, and the deposit is refundable. The ranking score is 3200 / 3600 ≈ 0.8889.

[0201] Based on the above description, the embodiments of this application divide the deposit scoring rules according to the distribution of deposit types. The mixed deposit type scoring rules balance the benefits and risks, and the fully refundable deposit scoring rules focus on benefits. The scoring logic is highly consistent with the driver's decision-making logic.

[0202] In scenarios corresponding to the hybrid deposit scoring rule, the risk of non-refundable deposits is quantified by the risk weight coefficient, which rationally reduces the ranking score of high-risk cargo sources, allowing drivers to intuitively perceive the risk differences. Before accepting an order, drivers can indirectly judge the balance between cargo source risk and income through the ranking score, which reduces the dispute rate and improves order acceptance satisfaction.

[0203] In a specific example, based on the Bayesian prior probability concept, we examine the historical transaction patterns of heavy cargo to determine the ranking score of the target cargo source within the target heavy cargo bins. The specific scheme is shown in Table 1 below:

[0204] Table 1

[0205]

[0206] Where S represents similar goods, Son represents similar goods in stock, D represents deposit, Di represents deposit for the i-th similar goods in stock, Dnew represents deposit for the target goods, DSon represents deposit for similar goods in stock, NF represents net shipping cost, NF_i represents net shipping cost for the i-th similar goods in stock, NFnew represents net shipping cost for the target goods, NFSon represents net shipping cost for the in-stock goods, M represents ranking score, Mnew represents ranking score for the target goods, and Mi represents ranking score for the i-th similar goods in stock.

[0207] In one optional implementation, heavy cargo recommendation information is generated based on the ranking score of the target cargo source and the ranking scores of other cargo sources in the target heavy cargo bin. When the target cargo source is pushed to the target driver group, displaying the heavy cargo recommendation information to the target driver group includes:

[0208] Based on each source tuple in the target heavy cargo bin, the sorting score of the target source, and the sorting scores and shelf status of other sources in the target heavy cargo bin, a shelf-ready source sorting list is determined; wherein, the sources in the shelf-ready source sorting list are arranged in descending order according to their sorting scores.

[0209] If the ranking score of the target cargo source is less than or equal to the ranking score of at least one similar cargo source in the cargo source ranking list, then heavy cargo recommendation information is generated based on the information of the at least one similar cargo source. While pushing the target cargo source to the target driver group, the at least one similar cargo source is also pushed to the target driver through a preset viewing entry.

[0210] If the ranking score of the target cargo is greater than the ranking score of any similar cargo in the cargo ranking list, then the target cargo is pushed to the target driver group.

[0211] If the target cargo is the only one listed in the inventory sorting list, then the target cargo is pushed to the target driver group.

[0212] In this embodiment, an ordered list of goods on the shelf is first constructed based on all the cargo tuples within the target heavy cargo bin, combined with the sorting score and shelf status of each cargo, providing a foundation for subsequent recommendations. The list of goods on the shelf includes the target cargo tuple and other similar cargo tuples on the shelf. All elements are similar heavy cargo that drivers can actually accept, and the sorting result directly reflects the attractiveness level of the cargo to drivers.

[0213] The list of available products is sorted in descending order by sorting score. The higher the sorting score, the higher the display priority. If the sorting scores are the same, they can be sorted in descending order by update timestamp to ensure that the newest products are given priority.

[0214] Next, the system iterates through the list of available goods. If there is at least one similar goods with a higher ranking score than the target goods, the system displays the target goods to the driver while recommending similar goods with a better ranking to improve order acceptance satisfaction.

[0215] Specifically, all similar cargo sources with a higher ranking score than the target cargo source are extracted from the list of available cargo sources. If the number is ≥2, the entry point for recommending similar cargo sources uses the top 2 ranked cargo source. After entering the list through the entry point, the complete list of available cargo sources can be viewed, with 10 cargo sources per page and a maximum of 5 pages. If there is only 1 similar cargo source, that cargo source is extracted. The information for recommending similar cargo sources includes the origin / destination, net freight cost, deposit type, minimum required vehicle length, attractiveness level corresponding to the ranking score, and is labeled with tags such as "currently best". The recommended information is displayed through two preset entry points: the recommendation bar at the bottom of the cargo source details page on the driver's side and a dedicated entry point for similar cargo sources. This does not interfere with the viewing of the target cargo source and allows drivers to quickly switch to view similar cargo sources.

[0216] The detailed description of the similar product list is as follows: list title, similar product recommendations;

[0217] The recall and sorting strategy retrieves all similar products corresponding to the original product based on the similar product tagging logic.

[0218] sort by:

[0219] a) If similar products exist in stock, remove products that have been taken down from the similar products. The remaining similar products in stock are sorted by M_i from high to low and displayed in order.

[0220] (b) If similar products exist, but they have all been removed from the shelves (either sold or removed voluntarily), only the products currently on the shelves will be displayed and pinned to the top.

[0221] The first shipment will have its card specially rendered.

[0222] The list page can be closed in the upper right corner. Clicking the close button will close the current list of similar products and leave you on the product details page.

[0223] Optionally, when pushing target cargo to the target driver group, heavy cargo recommendation information is simultaneously loaded at the above-mentioned preset entry point. Drivers can click on the entry point to view all recommended better similar cargo.

[0224] If the target cargo has the highest ranking score in the list of available cargo, no additional heavy cargo recommendations will be generated. Since the target cargo is already the most attractive option among similar heavy cargo, there is no need to recommend other similar cargo, thus avoiding redundant information that could interfere with the driver's decision-making.

[0225] Optionally, when pushing target cargo to the target driver group, highlight their current best tag (generated based on ranking score) to enhance drivers' perception of the attractiveness of the target cargo and promote quick order acceptance.

[0226] If the target cargo tuple exists only in the cargo sorting list (with no other similar heavy cargo on the shelves), then there is no comparison object, and the target cargo is directly pushed to ensure that the cargo owner's needs are properly reached by the driver.

[0227] Optionally, when pushing target cargo to the target driver group, mark it with the tag "No similar cargo available" to avoid confusion for drivers due to lack of comparison information, while increasing the attention of the target cargo.

[0228] It is worth noting that the selection of the target driver group can be determined based on the core attributes of the target cargo (minimum required vehicle length, transportation route, secondary cargo category) and matched with the driver's registration information (vehicle type, frequently traveled routes, and historical order categories). Only drivers whose vehicle type is greater than or equal to the minimum required vehicle length, whose frequently traveled routes include the target route, and whose historical order categories match the cargo category are selected, ensuring that the cargo pushed is consistent with the driver's ability to accept the cargo and their demand preferences.

[0229] Based on the above description, this application embodiment only recommends similar cargo sources with better ranking (at most Top 2) to avoid information overload. Drivers do not need to compare one by one to obtain the best option of the same type, and the cargo search decision time is shortened.

[0230] Furthermore, the embodiments of this application actively recommend better cargo sources when they exist, and do not redundantly push information when there are no better cargo sources. This ensures that drivers obtain complete information on similar cargo sources while avoiding interference from invalid information.

[0231] In one example, if there are similar products available for the target product and the ranking score of these similar products is higher than that of this product, then the bottom of the details page will display a recommendation of similar products, showing the product with the highest ranking score among the similar products available for sale.

[0232] If similar products already exist and the target product is the one with the highest ranking among similar products, similar product recommendations will not be displayed at the bottom of the product details page.

[0233] Similar products exist but have all been removed from the platform; therefore, similar product recommendations are not displayed at the bottom of the product details page.

[0234] In this example, a new entry point for viewing similar products has been added to the product details page. The entry point is located in the middle of the list of product owners and recommended products, and the module title is "Current Best | Compare Similar Products".

[0235] In this example, a) if there are >= 2 similar items in stock, the top 2 items in the similar items in stock (excluding the items in the same stock with details) will be displayed from left to right. If there is only 1 similar item in stock, only 1 item will be displayed.

[0236] b) If all similar products have been taken down, the top two similar products will be displayed based on their shipping time, from most recent to oldest. If there is only one similar product, only that one product will be displayed.

[0237] This scenario adds a switch. When the switch is on, similar products are displayed; when the switch is off, similar products are not displayed.

[0238] When the switch is on, blacklist configuration is supported. For blacklisted users, the module for similar goods will not be exposed in scenario b).

[0239] For users who can see the module recommending similar products, clicking anywhere in the module will take them to the list of recommended similar products.

[0240] In an optional implementation, before generating heavy cargo recommendation information based on the ranking score of the target cargo source and the ranking scores of other cargo sources in the target heavy cargo bin, the method further includes: determining the recommendation score of the target cargo source based on the information of the target cargo source;

[0241] Based on the recommendation score of the target source of goods, the push method of the target source of goods is determined, wherein the push method includes: immediate push and delayed push.

[0242] In this embodiment, the recommendation score for the target cargo is a comprehensive value score calculated based on the multi-dimensional attributes of the target cargo, taking into account factors such as driver willingness to accept the cargo, cargo owner credit, and cargo timeliness, and is used to determine the priority of push notifications.

[0243] In one example, the recommendation score = driver attractiveness score + shipper credit rating score + cargo availability score + freight payment method score + anomaly penalty score; where, the driver attractiveness score can be the ranking score of the target cargo. The shipper credit rating score can be the shipper's registration duration, historical transaction rate, and driver rating. The cargo availability score can be the interval between the loading time filled in by the shipper and the current time (the shorter the interval, the stronger the timeliness). The freight payment method score can be the payment method chosen by the shipper, and the anomaly penalty score can be the shipper's historical order cancellation rate.

[0244] Then, two recommendation score thresholds can be preset. Based on the range in which the recommendation score of the target product falls, the corresponding push method can be determined to ensure that high-quality products are reached more efficiently.

[0245] If it is a high-priority push, it will be pushed immediately. That is, after the cargo is published / updated, the push process will be triggered immediately, and the target driver will be reached through push notifications, pinning the cargo search list, etc.

[0246] If it is a medium priority push, the push will be delayed. After the product is published / updated, it will be pushed after a preset time delay (configurable) to avoid competing for traffic with high-value products.

[0247] If it is a low-priority push, it will only be passively displayed. After the cargo is published / updated, it will not be actively pushed (no push notification). It will only be displayed in the driver's cargo search list in natural order and will not occupy the active push resources.

[0248] Optionally, when pushing the target cargo source to the target driver group, displaying the heavy cargo recommendation information to the target driver group includes:

[0249] According to the target cargo source push method, when the target cargo source is pushed to the target driver group, the heavy cargo recommendation information is displayed to the target driver group.

[0250] In this embodiment of the application, for the target cargo source to be pushed immediately, after generating heavy cargo recommendation information, the target cargo source is pushed through the driver's push notification and displayed at the top of the cargo search list; and the heavy cargo recommendation information is simultaneously displayed on the push notification jump page and the target cargo source details page.

[0251] For target goods with delayed push notifications, after generating heavy goods recommendation information, the target goods will be pushed through the search list in a regular sorting manner and displayed on the category filtering results page after a preset delay period, without triggering push notifications. The heavy goods recommendation information will be displayed synchronously on the target goods details page without a dedicated pinned label.

[0252] For target cargo sources that are only passively displayed, after generating heavy cargo recommendation information, the target cargo sources are only displayed in natural sorting (sorting is in descending order) when drivers filter by route and category. There is no active push action, and the heavy cargo recommendation information is only displayed on the target cargo source details page. There is no bottom recommendation bar, and only a dedicated entry for similar cargo sources is retained.

[0253] As a concrete example, refer to Figure 2After a shipper publishes a target cargo source, the system extracts the origin and destination address information. Based on the address, loading and unloading methods, and other information, it finds the corresponding target heavy cargo bin in the cargo source pool, adds the new cargo source, and updates the bin's data. After updating the heavy cargo bin, an identifier (such as a tag) is generated for the cargo source, and it enters the subsequent scoring stage. Then, based on dimensions such as the cargo source's net freight cost, deposit type, and shipper credit, a comprehensive recommendation score is calculated. The system checks if the recommendation score equals 1: if yes, the cargo source is displayed normally and proceeds to the listing process; otherwise, it checks if a trigger switch is triggered: if the trigger switch is triggered, listing is delayed according to the recommendation score (low-priority cargo sources are displayed at off-peak times with corresponding delayed push notifications); if the trigger switch is not triggered, the target cargo source is not displayed to filter out low-value cargo sources.

[0254] When a driver initiates a freight search (such as searching for routes or product categories), the driver's freight search criteria are converted into discrete value tags (keys). The corresponding heavy cargo bins are found in the freight pool, the freight list within the heavy cargo bins is obtained, and the available and valid freight bins are filtered. A ranking score is calculated for similar freight bins, and they are sorted by attractiveness. The ranked similar freight bins are then recommended to the driver.

[0255] Finally, check if the product listings have been flagged as unlisted. If so, remove unlisted products from the list: remove unlisted, sold, or expired products from the heavy stock / recommended list. If not, continue with the list and retain the current list (only displaying valid products).

[0256] Based on the above description, the duplicate cargo identification strategy provided in this application is robust. If the cargo owner intentionally modifies the duplicate cargo information or uses different accounts to ship goods, as long as the origin and destination cities remain unchanged, modifications to other information dimensions will not affect the determination of heavy cargo. A dedicated list is used to organize and display representative heavy cargo, providing drivers with the most complete information on heavy cargo and helping them make more pragmatic decisions. An algorithm is introduced to evaluate the attractiveness of duplicate cargo to drivers, helping them quickly find the most profitable option among a set of duplicate cargo. For situations with a large amount of heavy cargo, a specially designed compression and sampling algorithm is used to retain representative heavy cargo and the driver's optimal choice, effectively reducing the driver's decision-making burden while avoiding degradation of front-end response speed in extreme cases.

[0257] Furthermore, existing technologies directly control the display, and the effectiveness depends on the accuracy of heavy cargo identification. Low accuracy in heavy cargo identification can have negative effects. However, the embodiments of this application organize the heavy cargo in an orderly manner and display it to the driver as completely as possible, giving the driver the decision-making power. This places higher demands on the recall capability of heavy cargo identification and relaxes the restrictions on identification accuracy.

[0258] In determining the scoring formula for ranking the attractiveness of heavy goods to drivers, this application adopts the idea of ​​Bayesian prior probability to select key influencing factors and uses the divide-and-conquer approach to design the scoring formula.

[0259] Furthermore, this application's embodiments draw on the Hamming distance algorithm to design a method for storing and querying heavy cargo information. By using bucketing and secondary filtering, the time complexity of heavy cargo query calculation is reduced from O(n) to O(1).

[0260] Finally, this application embodiment designs a set of sampling and compression algorithms in combination with business objectives to solve the extreme data skew problem encountered in heavy goods storage and query, and avoid significant degradation of front-end request response time affecting user experience.

[0261] Secondly, this application also provides an electronic device. Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 3 As shown, the electronic device may include: transceiver 121, processor 122, and memory 123.

[0262] Processor 122 executes computer execution instructions stored in memory, causing processor 122 to perform the scheme in the above method embodiments. Processor 122 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0263] The memory 123 is connected to the processor 122 via the system bus and completes communication between them. The memory 123 is used to store computer program instructions.

[0264] Transceiver 121 can be used to obtain the task to be run and its configuration information.

[0265] The system bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The system bus can be divided into address bus, data bus, control bus, etc. For ease of representation, only one thick line is used in the diagram, but this does not indicate that there is only one bus or one type of bus. Transceivers are used to enable communication between database access devices and other computers (e.g., clients, read-write libraries, and read-only libraries). Memory may include random access memory (RAM) and may also include non-volatile memory.

[0266] This application also provides a chip for executing instructions, which is used to execute the technical solutions of the methods described in the above embodiments.

[0267] Thirdly, embodiments of this application also provide a computer-readable storage medium storing computer instructions that, when executed on a computer, cause the computer to perform the technical solutions of the methods described in the above embodiments.

[0268] Fourthly, embodiments of this application also provide a computer program product, which includes a computer program stored in a computer-readable storage medium. At least one processor can read the computer program from the computer-readable storage medium, and when the at least one processor executes the computer program, it can implement the technical solutions of the methods in the above embodiments.

[0269] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms.

[0270] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to implement the solution of this embodiment according to actual needs.

[0271] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The unit composed of the above modules can be implemented in hardware or in the form of hardware plus software functional units.

[0272] The integrated modules described above, implemented as software functional modules, can be stored in a computer-readable storage medium. These software functional modules, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods of the various embodiments of this application.

[0273] It should be understood that the aforementioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly manifested as execution by a hardware processor, or execution by a combination of hardware and software modules within the processor.

[0274] The memory may include high-speed RAM, and may also include non-volatile storage (NVM), such as at least one disk storage device, and may also be a USB flash drive, external hard drive, read-only memory, disk or optical disc, etc.

[0275] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0276] The aforementioned storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0277] An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Alternatively, the storage medium can be an integral part of the processor. The processor and storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and storage medium can exist as discrete components in an electronic control unit or main control device.

[0278] The above description is merely an embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for pushing product sources, characterized in that: The method includes: In response to the target cargo information released by the cargo owner, the target heavy cargo barrel is determined from the preset cargo pool; The target heavy cargo barrels are updated based on the information of the target cargo source; Based on preset scoring rules, the ranking score of the target cargo source in the target heavy cargo barrel is determined; Based on the ranking score of the target cargo source and the ranking scores of other cargo sources in the target heavy cargo bin, heavy cargo recommendation information is generated, and when the target cargo source is pushed to the target driver group, the heavy cargo recommendation information is displayed to the target driver group. The step of updating the target heavy cargo barrel based on the information of the target cargo source includes: Based on the information of the target goods source, construct a target goods source tuple; The target heavy cargo barrel is updated based on the target cargo source tuple and the preset heavy cargo barrel update rules. The step of updating the target heavy cargo bins according to the target cargo source tuple and the preset heavy cargo bin update rules includes: If the number of elements in the target heavy cargo bin is less than a preset capacity threshold, the system determines whether there is a heavy cargo tuple that is duplicated with the target cargo tuple based on the deduplication verification field in the target cargo tuple. If it exists, then replace the specified element of the heavy source tuple with the specified element in the target source tuple; If it does not exist, the target source tuple is written into the target source bucket; The deduplication verification fields include: an identifier field representing the cargo owner, a field representing the freight, a field representing the transportation deposit, and a field representing the cargo owner's requirements for the vehicle. The specified elements include: an identifier field representing the source of goods and a field representing the latest update time of the source of goods.

2. The method according to claim 1, characterized in that, The process of responding to the target cargo source information released by the cargo owner, and determining the target heavy cargo container in the preset cargo source pool, includes: In response to the target cargo information released by the cargo owner, a target label corresponding to the target cargo is constructed based on a preset discrete value labeling rule; A target heavy cargo container matching the target label is identified in the preset cargo pool.

3. The method according to claim 2, characterized in that, In response to the target cargo information published by the cargo owner, the target tags corresponding to the target cargo are constructed based on preset discrete value tagging rules, including: In response to the target cargo information published by the cargo owner, the origin address, destination address, loading and unloading method, classification identifier and address granularity of the target cargo are extracted from the target cargo information; According to the address granularity, the origin address and destination address of the target cargo are converted into standardized origin address and standardized destination address; Based on the loading and unloading method, determine the standardized loading and unloading method code corresponding to the target cargo source, and based on the classification identifier, determine the standardized classification code corresponding to the target cargo source; According to the preset discrete value labeling rules, a target label corresponding to the target cargo source is constructed based on the standardized origin address, the standardized destination address, the standardized loading and unloading method code, and the standardized classification code.

4. The method according to claim 1, characterized in that, The step of updating the target heavy cargo bins according to the target cargo source tuple and the preset heavy cargo bin update rules includes: If the number of elements in the target heavy cargo bin is greater than or equal to the preset capacity threshold, based on the deduplication verification field in the target cargo tuple, it is determined whether there is a heavy cargo tuple that is duplicated with the target cargo tuple in the target cargo tuple. If it exists, then replace the specified element of the heavy source tuple with the specified element in the target source tuple; If it does not exist, a preset compression algorithm is used to compress multiple source tuples with the same deduplication verification field in the target source bucket, and the source tuple with the latest update time is retained to update the target source bucket. If the number of elements in the updated target source bucket is less than the preset capacity threshold, the target source tuple is written into the updated target source bucket. If the number of elements in the updated target source bucket is greater than or equal to the preset capacity threshold, a preset sampling algorithm is used to sequentially delete the delisted source tuples, the source tuples with shipping costs less than the preset shipping costs, and the source tuples with deposits higher than the preset deposits, until the number of elements in the updated target source bucket is less than the preset capacity threshold, and then the target source tuples are written into the updated target source bucket. The deduplication verification fields include: a field representing the shipper's identifier, a field representing the freight, a field representing the transportation deposit, and a field representing the minimum vehicle length requirement. The specified elements include: an identifier field representing the source of goods and a field representing the latest update time of the source of goods.

5. The method according to claim 1, characterized in that: Based on preset scoring rules, the ranking score of the target cargo source in the target heavy cargo barrel is determined as follows: Identify a set of similar products on the shelves within the target product container; Based on the deposit amount, freight, platform service fee, and minimum vehicle length requirement of each similar cargo tuple in the set of similar cargo sources, determine the deposit type and net freight for each of the similar cargo sources; Based on the relationship between the deposit types of the target goods and other similar goods in the target goods container, the corresponding scoring rules are determined; Based on the corresponding scoring rules and the relationship between the net freight cost of the target cargo and the net freight cost of each of the similar cargo sources, the ranking score of the target cargo in the target heavy cargo barrel is determined.

6. The method according to claim 5, characterized in that, The scoring rules, based on the relationship between the deposit types of the target goods and other similar goods in the target goods container, include: If the deposit type for each similar source in the target source container includes non-refundable deposit, the corresponding scoring rule is determined to be a mixed deposit type scoring rule. If all similar sources in the target source container have refundable deposits, then the corresponding scoring rule is determined to be the fully refundable deposit scoring rule.

7. The method according to any one of claims 1-4, characterized in that: Based on the ranking score of the target cargo source and the ranking scores of other cargo sources in the target heavy cargo bin, heavy cargo recommendation information is generated, and when the target cargo source is pushed to the target driver group, the heavy cargo recommendation information is displayed to the target driver group, including: Based on each source tuple in the target heavy cargo bin, the sorting score of the target source, and the sorting scores and shelf status of other sources in the target heavy cargo bin, a shelf-ready source sorting list is determined; wherein, the sources in the shelf-ready source sorting list are arranged in descending order according to their sorting scores. If the ranking score of the target cargo source is less than or equal to the ranking score of at least one similar cargo source in the cargo source ranking list, then heavy cargo recommendation information is generated based on the information of the at least one similar cargo source. While pushing the target cargo source to the target driver group, the at least one similar cargo source is also pushed to the target driver through a preset viewing entry. If the ranking score of the target cargo is greater than the ranking score of any similar cargo in the cargo ranking list, then the target cargo is pushed to the target driver group. If the target cargo is the only one listed in the inventory sorting list, then the target cargo is pushed to the target driver group.

8. The method according to any one of claims 1-4, characterized in that, Before generating heavy cargo recommendation information based on the ranking score of the target cargo source and the ranking scores of other cargo sources in the target heavy cargo bin, the method further includes: Based on the information of the target source of goods, determine the recommendation score of the target source of goods; Based on the recommendation score of the target product source, the push method of the target product source is determined, wherein the push method includes: immediate push and delayed push; When pushing the target cargo source to the target driver group, displaying the heavy cargo recommendation information to the target driver group includes: According to the target cargo source push method, when the target cargo source is pushed to the target driver group, the heavy cargo recommendation information is displayed to the target driver group.

9. An electronic device, characterized in that, include: A memory and a processor; the memory is configured to store computer program instructions; the processor is configured to execute the computer program instructions, causing the electronic device to perform the method as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, include: Computer program instructions; The electronic device executes the computer program instructions, causing the electronic device to perform the method as described in any one of claims 1 to 8.

11. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method described in any one of claims 1-8.

Images