Coupon pushing method and apparatus, and computer device and medium

Abstract

The embodiments of the present invention relate to the technical field of the Internet. Provided are a coupon pushing method and apparatus, and a computer device and a medium. The method comprises the following steps: if a push type is an active push, acquiring customer behavior data and customer consumption record data that are generated by a customer, and if the push type is a passive push, automatically matching coupons that meet conditions; classifying customer data to generate customer data corresponding to different customer types; extracting a maximum frequent keyword set from the customer data corresponding to each customer type, and generating, by means of the maximum frequent keyword set, keywords of consumption preferences and demands that correspond to each customer type; and acquiring a push channel that is set by a merchant, matching corresponding coupons for customers of each customer type on the basis of the keywords of consumption preferences and demands that correspond to each customer type, and pushing the coupons to the push channel. The solution improves the efficiency of coupon pushing.

Description

A method, apparatus, computer device, and medium for pushing coupons. Technical Field

[0001] This invention relates to the field of Internet technology, and in particular to a method, apparatus, computer equipment, and medium for pushing coupons. Background Technology

[0002] Currently, online sales platforms have rapidly entered people's lives due to their advantages such as a wide selection of products, convenient shopping experience, and 24 / 7 shopping opportunities. However, this has also brought considerable challenges to merchants, especially small and medium-sized enterprises (SMEs). For them, attracting customers, promoting products, and ensuring information security have become major issues for business development. Furthermore, there is a current market phenomenon where online sales platforms, in order to attract customers, often distribute various coupons, resulting in many unwanted coupons, low matching accuracy, and wasted merchant discount resources. Summary of the Invention

[0003] In view of this, embodiments of the present invention provide a method for pushing coupons to solve the technical problems of low matching degree and low accuracy in coupon distribution in the prior art. The method includes:

[0004] Get the push type set by the merchant. If the push type is active push, get the customer behavior data and customer consumption record data generated by the customer. If the push type is passive push, after the payment is completed, get the return data of the payment system interface, automatically match the coupons that meet the conditions, and push the coupons to the payment completion interface.

[0005] All customer behavior data and customer consumption record data are used as customer data. If the amount of customer data is less than or equal to the preset sample size threshold and the proportion of customer behavior data in customer data meets the preset ratio, the customer data is classified to generate customer data corresponding to different customer types.

[0006] Extract the most frequent keyword set from the customer data corresponding to each customer type, and generate keywords for consumption preferences and needs corresponding to each customer type using the most frequent keyword set;

[0007] Obtain the push channels set by the merchant, match corresponding coupons to customers of each customer type based on keywords related to their consumption preferences and needs, and push the coupons to the push channels.

[0008] This invention also provides a coupon distribution device to address the technical problems of low matching accuracy and low precision in coupon distribution in the prior art. The device includes:

[0009] The push type determination module is used to obtain the push type set by the merchant. If the push type is active push, it obtains customer behavior data and customer consumption record data generated by the customer. If the push type is passive push, after the payment is completed, it obtains the return data of the payment system interface, automatically matches the coupons that meet the conditions, and pushes the coupons to the payment completion interface.

[0010] The customer classification module is used to classify customer data and customer consumption records of all customers as customer data. When the amount of customer data is less than or equal to the preset sample size threshold and the proportion of customer behavior data in customer data meets the preset ratio, the customer data is classified and customer data corresponding to different customer types is generated.

[0011] The keyword extraction module is used to extract the most frequent keyword set from the customer data corresponding to each customer type, and generate keywords for consumption preferences and needs corresponding to each customer type through the most frequent keyword set;

[0012] The coupon push module is used to obtain the push channels set by merchants, match corresponding coupons to customers of each customer type according to the keywords of consumption preferences and needs, and push the coupons to the push channels.

[0013] This invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any of the above-mentioned coupon push methods to solve the technical problems of low matching degree and low accuracy of coupon distribution in the prior art.

[0014] This invention also provides a computer-readable storage medium storing a computer program that executes any of the above-described coupon distribution methods, in order to solve the technical problems of low matching degree and low accuracy in coupon distribution in the prior art.

[0015] The implementation of this invention is a legal use.

[0016] Compared with the prior art, the beneficial effects that at least one technical solution adopted in the embodiments of this specification can achieve include at least:

[0017] By providing a more efficient and precise lead generation and message delivery mechanism through both proactive and passive push methods, customers are categorized, and keywords related to consumption preferences and needs are extracted for different customer types. Coupons are then matched and issued based on these keywords, which helps improve the matching accuracy and precision of coupon issuance, thereby effectively increasing the coupon redemption conversion rate. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 is a flowchart of a coupon push method provided by an embodiment of the present invention;

[0020] Figure 2 is a structural block diagram of a computer device provided in an embodiment of the present invention;

[0021] Figure 3 is a structural block diagram of a coupon push device provided in an embodiment of the present invention. Detailed Implementation

[0022] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0023] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. This application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] In this embodiment of the invention, a method for pushing coupons is provided, as shown in Figure 1. The method includes:

[0025] Step S101: Obtain the push type set by the merchant. If the push type is active push, obtain customer behavior data and customer consumption record data generated by the customer. If the push type is passive push, after the payment is completed, obtain the return data of the payment system interface, automatically match the coupons that meet the conditions, and push the coupons to the payment completion interface.

[0026] Step S102: Take all customer behavior data and customer consumption record data as customer data. If the amount of customer data is less than or equal to the preset sample size threshold and the proportion of customer behavior data in customer data meets the preset ratio, classify the customer data and generate customer data corresponding to different customer types.

[0027] Step S103: Extract the most frequent keyword set from the customer data corresponding to each customer type, and generate keywords for consumption preferences and needs corresponding to each customer type using the most frequent keyword set;

[0028] Step S104: Obtain the push channels set by the merchant, match corresponding coupons for each customer type based on the keywords of consumption preferences and needs corresponding to each customer type, and push the coupons to the push channels.

[0029] In practice, to push different coupons to different users and increase the efficiency of coupon delivery, the following steps are taken to treat all customer behavior data and customer consumption record data as customer data. When the amount of customer data is less than or equal to a preset sample size threshold and the proportion of customer behavior data in the customer data meets a preset ratio, the customer data is classified to generate customer data corresponding to different customer types:

[0030] Customer data is preprocessed by removing outliers and duplicate data corresponding to the same customer identifier, generating a preprocessed dataset where each customer identifier is a unique identifier for each customer. If the size of the preprocessed dataset exceeds a preset sample size threshold, data is deleted until the size of the preprocessed dataset is less than or equal to the preset sample size threshold, generating an adjusted dataset. A total number of customer types is set, and a total number of customer personal data points are randomly selected from the adjusted dataset as the initial data center set. Customer behavior data and customer consumption records for the same customer are considered as the customer's personal data. The entire adjusted dataset, excluding the initial data center sets, is then iterated through. Given personal data, the distance between each customer's personal data and each initial data center point in the initial data center point set is calculated. The personal data is then assigned to the nearest initial data center point, generating multiple initial clusters. The average distance of all personal data in the initial clusters is calculated, and the location of the average distance is used as the iterative data center point. The difference distance between the initial data center point and the iterative data center point is calculated. The iterative data center point is repeatedly obtained until the difference distance is less than a preset distance threshold. The iterative data center point is then used as the final data center point, and the final cluster corresponding to each personal data is generated. All personal data in the final cluster corresponding to each final data center point are classified as customer data corresponding to the same customer type.

[0031] Specifically, user behavior data refers to action data generated by users during commercial interactions, such as online or offline purchases, browsing, and reviews. User transaction data refers to data generated after a user completes a payment, such as order amount, quantity, and type.

[0032] In practice, to reduce the amount of customer data by lowering the proportion of behavioral data in the customer data (i.e., the weight of browsing behavior is lower than that of purchasing behavior), the following steps are taken: If the amount of data in the preprocessed dataset exceeds a preset sample size threshold, data in the preprocessed dataset is deleted until the amount of data in the preprocessed dataset is less than or equal to the preset sample size threshold, thus generating an adjusted dataset:

[0033] Perform the following operations until the amount of data in the adjusted dataset is less than or equal to a preset sample size threshold: Use customer behavior data corresponding to the customer identifier as the first data item, and customer consumption record data corresponding to the customer identifier as the second data item; set the weight of the first data item corresponding to the first data item and the weight of the second data item corresponding to the second data item, where the weight of the first data item is less than the weight of the second data item; delete data from the preprocessed dataset until the proportion of customer behavior data corresponding to the same customer identifier in the preprocessed dataset and the proportion of customer consumption record data in the preprocessed dataset satisfy the first data weight and the second data weight, respectively, to generate the adjusted dataset.

[0034] Specifically, the combined weight of the first data item and the second data item is 100%. By reducing the proportion of behavioral data in customer data and increasing the proportion of purchasing behavior in customer data, the efficiency of coupon distribution and the utilization rate of coupons can be improved.

[0035] In practice, the amount of customer data can be reduced by using batch sampling. For example, the following steps can be used to reduce the amount of data in the preprocessed dataset if the amount of data in the preprocessed dataset exceeds a preset sample size threshold, until the amount of data in the preprocessed dataset is less than or equal to the preset sample size threshold, thus generating an adjusted dataset:

[0036] Perform the following operations until the amount of data in the adjusted dataset is less than or equal to the preset sample size threshold: Set the batch sample size, and select a batch of a certain number of customers' personal data from the adjusted dataset through random sampling without replay to form a sample dataset; when the amount of data in the sample dataset is less than or equal to the preset sample size threshold and the proportion of customer behavior data in the sample dataset meets the preset ratio, stop random sampling and use the sample dataset as the adjusted dataset.

[0037] Specifically, since customer behavior data and customer consumption record data are usually very large, it is necessary to reduce the data volume. Batch sampling can quickly and effectively reduce the data volume.

[0038] In practice, after determining the initial data center location, customer data is categorized, and the following steps are taken to randomly select a total number of customers' personal data as the initial data center location from the adjusted dataset:

[0039] Randomly select a customer's personal data from the adjusted dataset as the initial sample center; repeat the following steps until a total number of customer data points are selected as the initial data center points, generating an initial data center point set: calculate the distance between the initial sample center and each personal data point in the adjusted dataset other than the initial sample center; based on the distance, use a weighted probability distribution to select the personal data of the customer with the largest distance as the new data center point, save the initial sample center to the initial data center point set, and use the new data center point as the initial sample center.

[0040] In practice, to extract different consumption preferences and needs for different user types, the following steps are taken to extract the most frequent keyword set from the customer data corresponding to each customer type, and then generate keywords for the consumption preferences and needs corresponding to each customer type using the most frequent keyword set:

[0041] For each customer type, repeat the following steps until no new candidate keywords can be generated (i.e., only one set of candidate keywords remains). The generated candidate set is taken as the maximum frequent keyword set, and the candidate keyword that appears most frequently in the maximum frequent keyword set is taken as the keyword for consumption preference and demand: Count the number of times the candidate keywords appear in the customer data corresponding to each customer type, generate a candidate set by the candidate keywords and the number of times they appear, wherein the candidate keywords are the minimum consumption object and / or behavior object; Set a minimum frequency, delete candidate keywords whose number of appearances in the candidate set is lower than the minimum frequency, and combine the remaining candidate keywords in pairs according to the number of appearances corresponding to the candidate keywords to generate a new set of candidate keywords.

[0042] Specifically, by extracting consumption preferences and needs from each user category, the accuracy of extracting consumption preferences and needs can be further improved, thereby increasing the accuracy and precision of coupon push notifications and improving the efficiency of coupon usage.

[0043] In practice, to effectively push suitable coupons after matching the payment return data with the push conditions set by the merchant, the following steps are taken to obtain the return data from the payment system interface after payment is completed, automatically match coupons that meet the conditions, and push the coupons to the payment completion interface:

[0044] The system retrieves the user push conditions set by the merchant and pushes coupons to customers whose payment completion screen matches the conditions. It also retrieves customer payment data from the returned data, retrieves the payment push conditions set by the merchant, and pushes coupons to customers whose payment data meets the conditions. Furthermore, it retrieves the customer's purchase time from the returned data, retrieves the purchase time conditions set by the merchant, and pushes coupons to customers whose purchase time meets the conditions. Finally, it retrieves the customer's purchase location from the returned data, retrieves the purchase location conditions set by the merchant, and pushes coupons to customers whose purchase location meets the conditions. Finally, it retrieves the customer's corresponding historical purchase records from the database and pushes coupons to customers whose historical purchase records meet a set number of purchases.

[0045] In one practical application example, coupons are pushed out using the following method:

[0046] To help small and medium-sized merchants operate more effectively, the payment channel provides merchants with an electronic coupon solution. Customers (i.e., users who make purchases) can receive coupons after payment, and these coupons can be automatically deducted from their next purchase if the conditions for using the coupon are met. After successful payment, the system can issue appropriate coupons to customers based on their geographical location at the time of payment or the merchant. Customers can claim the merchant-configured coupons on the payment success page, click "Use Now," and then watch the relevant video on the channel to receive further discounts.

[0047] 1. Merchants create coupons: Merchants create coupons in the mall backend, setting the coupon type, acquisition method, scope of use, validity period, etc.

[0048] 2. Coupon Activity Configuration: Merchants create or participate in "Payment Rewards" activities, and apply for, modify, or cancel coupon configurations within these activities. The redemption conditions in the coupon configuration cannot all be empty.

[0049] 3. Coupon Issuance: The reward conditions for payment-reward coupons include merchant information, member information, actual payment amount, order time, product type purchased, geographical location, number of product purchases, number of coupons claimed, whether the first type of coupon issuance method applies, whether the second type of coupon issuance method applies, and the location where the coupon can be issued.

[0050] There are two types of coupon issuance methods. The first type (passive push) applies as follows: After a customer completes payment, the system automatically matches coupons for the "Payment Rewards" activity based on data returned by the payment system interface. Therefore, coupons default to the first type of coupon issuance method. When a customer meets the reward conditions for the "Payment Rewards" activity, a prompt to claim a reward coupon will pop up on the payment success page, displaying the available coupons. When the system detects that all conditions are met after payment, the system issues the coupon; if some reward conditions are empty, the corresponding reward condition is automatically considered to be met, and all reward conditions cannot be empty.

[0051] Merchant information can be configured to determine whether reward conditions are only met when purchasing the merchant's own products; membership information can be configured to require customers to reach a certain membership level to meet the conditions; actual payment amount can be configured to determine if the user's payment exceeds a preset value; order time can be set to a range, and purchases within that time period are considered to meet the condition, with time accuracy down to the minute; product category can be configured, and purchases matching the preset product category are considered to meet the condition; geolocation can be configured, and orders placed within a specified range are considered to meet the condition, while orders placed without a specified location are considered not to meet the condition; product purchase count can be configured with a greater / less than indicator and a numerical value representing the number of purchases, and the user's product purchases are stored in the backend database, meeting the condition when the user's order count is greater than or less than this value; coupon redemption count can be configured to determine whether only the same type of coupon is counted, the greater / less than indicator, and the redemption count, and the backend stores the user's total coupon redemption count and the number of redemptions for the same type of coupon, and this count is approved when the system detects that the user meets the redemption conditions when placing an order.

[0052] The second type of coupon distribution method (proactive coupon distribution) applies to coupons where the merchant's "Whether to apply the second type of coupon distribution method" setting is "Yes". The system backend segments users based on encrypted user behavior data and consumption records; then, it analyzes the consumption preferences and needs of each user group; and finally, it matches appropriate coupons to users based on these preferences and needs. The matched coupons are then sent to users through push channels such as official accounts, mini-programs, and video accounts. The coupon distribution location (official account, mini-program, video account, etc.) is set in the coupon configuration. When this setting meets the second type of coupon distribution method and is empty, it is considered that the coupon can be distributed from any distribution location. When the second type of coupon distribution method is met but this setting is not empty, the coupon can only be distributed from the set distribution location.

[0053] 4. Coupon Redemption: When a customer redeems a coupon, it is automatically added to their wallet based on the coupon type, and a redirect to a video account or mini-program is requested to help the customer quickly use the coupon. Simultaneously, coupon information is automatically saved to the merchant database and the user's coupon redemption record is updated.

[0054] 5. Video Channel or Mini Program Redirect: Customers who click to view coupon details will be redirected to the coupon usage page. When a customer clicks to use the coupon, the system automatically redirects them to the activity's video channel or the merchant's mini program using the coupon link stored in the backend, guiding the customer to follow the video channel to claim a prize or to use the coupon to purchase discounted items in the mini program.

[0055] 6. Coupon Usage: Customers can click on the link in the video account to participate in the activity and receive prizes, or purchase goods in the mini program.

[0056] 7. Data Comparison: Based on customer information and data in the backend merchant information database, the system determines whether the coupon is valid, and then automatically uses the appropriate coupon to obtain prizes or discounts when participating in activities and making payments.

[0057] 8. Push notification reminders: Coupons can be inserted into the user's wallet. Before the coupon expires, the system will send an expiration reminder notification through the push notification platform.

[0058] 9. Coupon Redemption: Coupons claimed by customers are automatically destroyed after use. When a customer claims a prize from the video account or purchases goods using a coupon, the system passes the return value from the claiming or purchasing interface to the redemption interface, then updates the database and destroys the used coupon.

[0059] The following system architecture is used to push coupons.

[0060] 1. Activity Referral Module: Coupons are distributed via the payment completion page. A JS-SDK interface redirects customers to the activity page on the video account or mini-program using the coupons. When customers use the coupons on the video account or mini-program, they can see new recommended activities or claim new coupons on the activity page, ensuring user retention and bringing new development ideas and marketing traffic to the industry.

[0061] 2. Merchant Database Module: This module uses a relational database to store information on participating merchants, discounted products, coupons, and promotional activities, facilitating data use and maintenance.

[0062] 3. Coupon Module: The "Pay for Gifts" activity can be configured by the administrator or merchant in the backend. Merchants can apply for the creation, configuration and modification of coupons in the backend. After the application is approved, the changes to the activity and coupons are synchronized to the database and take effect.

[0063] 4. Encryption Module: The internally modified SM4 encryption algorithm is used to encrypt data in the business process. Database maintainers cannot directly see sensitive customer information, but this does not affect the actual use of the coupon function, thus ensuring customer information security.

[0064] 5. Merchant Comparison Module: When customers participate in activities, algorithms such as association rule mining are used to process the data, and then the merchant information and customer information are compared with the database. If a match is found, the goods are distributed.

[0065] 6. Discount Delivery Module: This module integrates with the merchant payment system. The system searches for discount coupons based on the return values ​​from the customer's payment interface or the merchant interface, and then combines this information with data from the merchant database to accurately and effectively send discount coupons to customers.

[0066] 7. Discount Reminder Module: Sends messages to customers' apps or mobile numbers when customers claim, use, or when coupons are about to expire, enhancing user awareness of discounts. When issuing coupons, the system stops issuing coupons when the backend detects that a certain number of customers have not used them, minimizing disruption to customers.

[0067] 8. Coupon Redemption Module: It integrates with the mall's payment system and the video account's prize redemption system, using cloud processing and quick payment technology to quickly process a large number of coupon redemption requests and update coupon information and usage status in real time.

[0068] After payment, the system pushes suitable coupons to customers based on the return values ​​from the payment interface or merchant interface. Customers can claim the coupons on the payment completion page and then be redirected to the video account to participate in activities. The system ensures targeted coupon distribution through restrictions such as region and merchant. Coupons can be added to the user's wallet, and an expiration reminder notification will be received before the coupon expires. The use of coupons drives customers to the video account to learn about products. Message notifications enhance users' awareness of merchant coupon discounts.

[0069] Coupons can be distributed not only on the payment completion page and video account, but also in the merchant's own scenarios (the channels that merchants can set up to push include: merchant mini program, H5 webpage, APP, and official account article), and can be placed in payment traffic scenarios (the channels that merchants can set up to push also include: nearby coupons, payment gifts, WeChat Moments, etc.).

[0070] The payment solution's robust security and anti-fraud capabilities prevent illegal activities by black market operators, safeguarding merchants' marketing funds. Merchants are not responsible for storing user information themselves; instead, they entrust this information to the video account's agents, whose professional operations staff protect the data from hacking, allowing merchants to focus their energy on their own development.

[0071] In this embodiment, a computer device is provided, as shown in FIG2, including a memory 201, a processor 202, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any of the above-mentioned coupon push methods.

[0072] Specifically, the computer device can be a computer terminal, a server, or a similar computing device.

[0073] In this embodiment, a computer-readable storage medium is provided, which stores a computer program that performs any of the above-described coupon push methods.

[0074] Specifically, computer-readable storage media, including both permanent and non-permanent, removable and non-removable media, can store information using any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer-readable storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable storage media does not include transient media, such as modulated data signals and carrier waves.

[0075] Based on the same inventive concept, this invention also provides a coupon distribution device, as described in the following embodiments. Since the principle of the coupon distribution device in solving the problem is similar to that of the coupon distribution method, the implementation of the coupon distribution device can refer to the implementation of the coupon distribution method, and repeated details will not be elaborated further. As used below, the terms "unit" or "module" can refer to a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0076] Figure 3 is a structural block diagram of a coupon push device according to an embodiment of the present invention. As shown in Figure 3, it includes: a push type determination module 301, a customer classification module 302, a keyword extraction module 303, and a coupon push module 304. The structure will be described below.

[0077] The push type determination module 301 is used to obtain the push type set by the merchant. If the push type is active push, it obtains customer behavior data and customer consumption record data generated by the customer. If the push type is passive push, after the payment is completed, it obtains the return data of the payment system interface, automatically matches the coupons that meet the conditions, and pushes the coupons to the payment completion interface.

[0078] The customer classification module 302 is used to classify customer data and customer consumption record data of all customers as customer data. When the amount of customer data is less than or equal to the preset sample size threshold and the proportion of customer behavior data in customer data meets the preset ratio, the customer data is generated to generate customer data corresponding to different customer types.

[0079] The keyword extraction module 303 is used to extract the most frequent keyword set from the customer data corresponding to each customer type, and generate keywords for consumption preferences and needs corresponding to each customer type through the most frequent keyword set;

[0080] The coupon push module 304 is used to obtain the push channels set by the merchant, match the corresponding coupons to customers of each customer type according to the keywords of consumption preferences and needs, and push the coupons to the push channels.

[0081] In one embodiment, the customer classification module includes:

[0082] The abnormal data processing unit is used to preprocess customer data, delete abnormal data in customer data, delete identical data corresponding to the same customer identifier, and generate a preprocessed dataset, wherein the customer identifier is a unique identifier corresponding to each customer.

[0083] The data reduction unit is used to delete data in the preprocessed dataset until the data size of the preprocessed dataset is less than or equal to the preset sample size threshold if the data size of the preprocessed dataset is greater than the preset sample size threshold, and generate the adjusted dataset.

[0084] The initial data center point setting unit is used to set the total number of customer types. It randomly selects a total number of customers' personal data from the adjusted dataset as the initial data center point set, where the customer behavior data and customer consumption record data of the same customer are used as the customer's personal data.

[0085] The final cluster partitioning unit is used to traverse all personal data in the adjusted dataset except for the initial data center point, calculate the distance between each customer's personal data and each initial data center point in the initial data center point set, partition the personal data to the nearest initial data center point, generate multiple initial clusters, calculate the average distance of all personal data in the initial clusters, take the position of the average distance as the iterative data center point, calculate the difference distance between the initial data center point and the iterative data center point, repeatedly obtain the iterative data center point until the difference distance is less than a preset distance threshold, take the iterative data center point as the final data center point and generate the final cluster corresponding to each personal data;

[0086] The customer data generation unit is used to divide all personal data in the final cluster corresponding to each final data center point into customer data corresponding to the same customer type.

[0087] In one embodiment, the data volume unit is reduced to perform the following operations until the data volume of the adjusted dataset is less than or equal to a preset sample size threshold: customer behavior data corresponding to a customer identifier is taken as a first data item, and customer consumption record data corresponding to a customer identifier is taken as a second data item; a first data item weight corresponding to the first data item and a second data item weight corresponding to the second data item are set respectively, wherein the first data item weight is less than the second data item weight; data in the preprocessed dataset is deleted until the proportion of customer behavior data corresponding to the same customer identifier in the preprocessed dataset and the proportion of customer consumption record data in the preprocessed dataset satisfy the first data weight and the second data weight respectively, thereby generating the adjusted dataset.

[0088] In one embodiment, reducing the data volume unit is further configured to perform the following operations until the data volume of the adjusted dataset is less than or equal to a preset sample size threshold: setting a batch sample size, selecting a batch sample size of customers' personal data to form a sampled dataset through random sampling without replay in the adjusted dataset; stopping random sampling when the data volume of the sampled dataset is less than or equal to the preset sample size threshold and the proportion of customer behavior data in the sampled dataset meets a preset ratio, and using the sampled dataset as the adjusted dataset.

[0089] In one embodiment, an initial data center point setting unit is used to randomly select a customer data point from the adjusted dataset as an initial sample center; repeat the following steps until a total number of customer data points are selected as initial data center points to generate the initial data center point set: calculate the distance between the initial sample center and each individual data point in the adjusted dataset other than the initial sample center; select the individual data point of the customer with the largest distance as a new data center point based on the distance using a weighted probability distribution; save the initial sample center to the initial data center point set; and use the new data center point as the initial sample center.

[0090] In one embodiment, the keyword extraction module includes:

[0091] The maximum frequent keyword set generation unit is used to iterate through the following steps for each customer type until no new candidate keywords can be generated (i.e., only one set of candidate keywords remains). The generated candidate set is then used as the maximum frequent keyword set, and the candidate keyword that appears most frequently in the maximum frequent keyword set is used as the keyword for consumer preferences and needs:

[0092] The candidate set generation unit is used to count the number of times candidate keywords appear in the customer data corresponding to each customer type, and generate a candidate set by the candidate keywords and the number of times they appear, wherein the candidate keywords are minimum consumption objects and / or behavior objects;

[0093] The next-level candidate keyword generation unit is used to set a minimum frequency, delete candidate keywords that appear less than the minimum frequency in the candidate set, and combine the remaining candidate keywords in pairs according to the frequency of occurrence of the candidate keywords to generate a new set of candidate keywords.

[0094] In one embodiment, the coupon delivery module includes:

[0095] The first push unit is used to obtain the user push conditions set by the merchant and push coupons to the payment completion interface of customers who match the user push conditions.

[0096] The second push unit is used to obtain the customer's payment data from the returned data, obtain the payment push conditions set by the merchant, and push coupons to the payment completion interface of customers whose payment data meets the payment push conditions.

[0097] The third push unit is used to obtain the customer's purchase time from the returned data, obtain the purchase time conditions set by the merchant, and push coupons to the payment completion interface of customers whose purchase time meets the purchase time conditions.

[0098] The fourth push unit is used to obtain the customer's purchase location from the returned data, obtain the purchase location conditions set by the merchant, and push coupons to the payment completion interface of customers whose purchase location meets the purchase location conditions.

[0099] The fifth push unit is used to retrieve the customer's historical purchase records from the database and push coupons to the payment completion interface of customers whose historical purchase records meet the set number of times.

[0100] The embodiments of the present invention achieve the following technical effects:

[0101] By employing both proactive and reactive push methods, a more efficient and precise customer acquisition and message delivery mechanism is provided. Customers are categorized, and keywords representing their consumption preferences and needs are extracted for different customer types. Coupons are then matched and issued based on these keywords, improving the matching accuracy and effectiveness of coupon distribution, thereby significantly increasing coupon redemption conversion rates. Furthermore, a wider range of coupon distribution scenarios and a richer array of traffic channels greatly expand the user reach of merchant promotions.

[0102] Obviously, those skilled in the art should understand that the modules or steps of the above-described embodiments of the present invention can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those presented here, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the embodiments of the present invention are not limited to any particular hardware and software combination.

[0103] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, various modifications and variations can be made to the embodiments of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for pushing coupons, characterized in that, include: Get the push type set by the merchant. If the push type is active push, get the customer behavior data and customer consumption record data generated by the customer. If the push type is passive push, after the payment is completed, get the return data of the payment system interface, automatically match the coupon that meets the conditions, and push the coupon to the payment completion interface. All customer behavior data and customer consumption record data are used as customer data. When the amount of customer data is less than or equal to a preset sample size threshold and the proportion of customer behavior data in the customer data meets a preset ratio, the customer data is classified to generate customer data corresponding to different customer types. This includes: preprocessing the customer data, deleting abnormal data in the customer data, deleting identical data corresponding to the same customer identifier, and generating a preprocessed dataset. The customer identifier is a unique identifier corresponding to each customer. If the amount of data in the preprocessed dataset is greater than the preset sample size threshold, the data in the preprocessed dataset is deleted by reducing the proportion of behavioral data in the customer data until the amount of data in the preprocessed dataset is less than or equal to the preset sample size threshold, and an adjusted dataset is generated. User behavioral data includes users' online or offline purchases, browsing, and comments, and user transaction data includes order amount, quantity, and type. The process involves setting a total number of customer types, randomly selecting personal data from the adjusted dataset of the total number of customers as an initial data center set, including: randomly selecting personal data from the adjusted dataset of one customer as an initial sample center; repeating the following steps until the personal data of the total number of customers is selected as the initial data center, generating the initial data center set: calculating the distance between the initial sample center and each personal data in the adjusted dataset other than the initial sample center; selecting the personal data of the customer with the largest distance as a new data center point based on the distance using a weighted probability distribution; saving the initial sample center to the initial data center set, and using the new data center point as the initial sample center, wherein the customer behavior data and the customer consumption record data of the same customer are considered as the customer's personal data; Iterate through all the personal data in the adjusted dataset except for the initial data center point, calculate the distance between each customer's personal data and each initial data center point in the initial data center point set, divide the personal data into the nearest initial data center point, generate multiple initial clusters, calculate the average distance of all personal data in the initial clusters, take the position of the average distance as the iterative data center point, calculate the difference distance between the initial data center point and the iterative data center point, repeatedly obtain the iterative data center point until the difference distance is less than a preset distance threshold, take the iterative data center point as the final data center point and generate the final cluster corresponding to each personal data; All the personal data in the final cluster corresponding to each final data center point are divided into customer data corresponding to the same customer type; Extract the most frequent keyword set from the customer data corresponding to each customer type, and generate keywords for consumption preferences and needs corresponding to each customer type using the most frequent keyword set; Obtain the push channels set by the merchant, match corresponding coupons to customers of each customer type according to the keywords of consumption preferences and needs corresponding to each customer type, and push the coupons to the push channels; Extract the most frequent keyword set from the customer data corresponding to each customer type, and generate keywords for consumption preferences and needs corresponding to each customer type using the most frequent keyword set, including: For each of the aforementioned customer types, repeat the following steps until only one set of candidate keywords remains. The generated candidate set is then used as the maximum frequent keyword set, and the candidate keyword that appears most frequently in this maximum frequent keyword set is selected as the keyword representing consumer preferences and needs. Count the number of times candidate keywords appear in the customer data corresponding to each customer type, and generate a candidate set by the candidate keywords and the number of times they appear, wherein the candidate keywords are the minimum consumption object and / or behavior object; Set a minimum frequency, delete candidate keywords that appear less than the minimum frequency in the candidate set, and combine the remaining candidate keywords in pairs according to the frequency of occurrence of the candidate keywords to generate a new set of candidate keywords.

2. The coupon delivery method as described in claim 1, characterized in that, If the amount of data in the preprocessed dataset exceeds the preset sample size threshold, delete data from the preprocessed dataset until the amount of data in the preprocessed dataset is less than or equal to the preset sample size threshold, and generate an adjusted dataset, including: Perform the following operations until the amount of data in the adjusted dataset is less than or equal to the preset sample size threshold: The customer behavior data corresponding to the customer identifier is used as the first data item, and the customer consumption record data corresponding to the customer identifier is used as the second data item. Set the weight of the first data item corresponding to the first data item and the weight of the second data item corresponding to the second data item respectively, wherein the weight of the first data item is less than the weight of the second data item; Delete data from the preprocessed dataset until the proportion of customer behavior data corresponding to the same customer identifier in the preprocessed dataset and the proportion of customer consumption record data in the preprocessed dataset satisfy the first data weight and the second data weight, respectively, and generate the adjusted dataset.

3. The coupon delivery method as described in claim 1, characterized in that, If the amount of data in the preprocessed dataset exceeds the preset sample size threshold, delete data from the preprocessed dataset until the amount of data in the preprocessed dataset is less than or equal to the preset sample size threshold, and generate an adjusted dataset, including: Perform the following operations until the amount of data in the adjusted dataset is less than or equal to the preset sample size threshold: Set a batch sample size, and select the personal data of the specified batch sample size customers from the adjusted dataset through random sampling without replay to form a sampled dataset; If the amount of data in the sampled dataset is less than or equal to the preset sample size threshold and the proportion of customer behavior data in the sampled dataset meets the preset ratio, random sampling is stopped, and the sampled dataset is used as the adjusted dataset.

4. The method for pushing coupons as described in any one of claims 1 to 3, characterized in that, After payment is completed, the system retrieves the return data from the payment system interface, automatically matches coupons that meet the conditions, and pushes the coupons to the payment completion screen, including: Obtain the user push conditions set by the merchant, and push the coupon to the payment completion interface of the customer who matches the user push conditions; Obtain the customer's payment data from the returned data, obtain the payment push conditions set by the merchant, and push the coupon to the customer's payment completion interface when the payment data meets the payment push conditions; The purchase time of the customer is obtained from the returned data, the purchase time conditions set by the merchant are obtained, and the coupon is pushed to the payment completion interface of the customer whose purchase time meets the purchase time conditions. The purchase location of the customer is obtained from the returned data, the purchase location conditions set by the merchant are obtained, and the coupon is pushed to the payment completion interface of the customer whose purchase location meets the purchase location conditions. Retrieve the customer's historical purchase records from the database, and push the coupon to the customer's payment completion interface when the historical purchase records meet a set number of times.

5. A coupon delivery device, characterized in that, include: The push type determination module is used to obtain the push type set by the merchant. If the push type is active push, it obtains customer behavior data and customer consumption record data generated by the customer. If the push type is passive push, after the payment is completed, it obtains the return data of the payment system interface, automatically matches the coupons that meet the conditions, and pushes the coupons to the payment completion interface. The customer classification module is used to classify the customer data and the customer consumption record data of all customers as customer data. When the amount of customer data is less than or equal to a preset sample size threshold and the proportion of customer behavior data in the customer data meets a preset ratio, the module generates customer data corresponding to different customer types. The customer classification module includes: The abnormal data processing unit is used to preprocess customer data, delete abnormal data in customer data, delete identical data corresponding to the same customer identifier, and generate a preprocessed dataset, wherein the customer identifier is a unique identifier corresponding to each customer. The data reduction unit is used to delete data in the preprocessed dataset by reducing the proportion of behavioral data in the customer data if the data volume of the preprocessed dataset is greater than the preset sample size threshold, until the data volume of the preprocessed dataset is less than or equal to the preset sample size threshold, and generate an adjusted dataset. User behavioral data includes users' online or offline purchases, browsing, and comments, and user transaction data includes order amount, quantity, and type. The initial data center point setting unit is used to set the total number of customer types. It randomly selects a total number of customers' personal data from the adjusted dataset as the initial data center point set, where the customer behavior data and customer consumption record data of the same customer are used as the customer's personal data. The initial data center point setting unit is further configured to randomly select a customer data point from the adjusted dataset as the initial sample center; repeat the following steps until a total number of customer data points are selected as initial data center points to generate the initial data center point set: calculate the distance between the initial sample center and each individual data point in the adjusted dataset other than the initial sample center; select the individual data point of the customer with the largest distance as the new data center point based on the distance using a weighted probability distribution; save the initial sample center to the initial data center point set; and use the new data center point as the initial sample center. The final cluster partitioning unit is used to traverse all personal data in the adjusted dataset except for the initial data center point, calculate the distance between each customer's personal data and each initial data center point in the initial data center point set, partition the personal data to the nearest initial data center point, generate multiple initial clusters, calculate the average distance of all personal data in the initial clusters, take the position of the average distance as the iterative data center point, calculate the difference distance between the initial data center point and the iterative data center point, repeatedly obtain the iterative data center point until the difference distance is less than a preset distance threshold, take the iterative data center point as the final data center point and generate the final cluster corresponding to each personal data; The customer data generation unit is used to divide all personal data in the final cluster corresponding to each final data center point into customer data corresponding to the same customer type. The keyword extraction module is used to extract the most frequent keyword set from the customer data corresponding to each customer type, and generate keywords for consumption preferences and needs corresponding to each customer type through the most frequent keyword set; The coupon push module is used to obtain the push channels set by the merchant, match corresponding coupons for each customer type according to the keywords of consumption preferences and needs corresponding to each customer type, and push the coupons to the push channels; The keyword extraction module includes: The maximum frequent keyword set generation unit is used to iterate through the following steps for each customer type until no new candidate keywords can be generated (i.e., only one set of candidate keywords remains). The generated candidate set is then used as the maximum frequent keyword set, and the candidate keyword that appears most frequently in the maximum frequent keyword set is used as the keyword for consumer preferences and needs: The candidate set generation unit is used to count the number of times candidate keywords appear in the customer data corresponding to each customer type, and generate a candidate set by the candidate keywords and the number of times they appear, wherein the candidate keywords are minimum consumption objects and / or behavior objects; The next-level candidate keyword generation unit is used to set a minimum frequency, delete candidate keywords that appear less than the minimum frequency in the candidate set, and combine the remaining candidate keywords in pairs according to the frequency of occurrence of the candidate keywords to generate a new set of candidate keywords.

6. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the coupon push method according to any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that performs the coupon delivery method according to any one of claims 1 to 4.

Images