Bank information pushing method based on big data model

Abstract

The invention discloses a bank information pushing method based on a big data model, and the method comprises the steps: obtaining daily consumption history details of a target user under the condition of user authorization, carrying out the keyword extraction to obtain daily consumption merchant information of the user, and constructing a consumption merchant feature recognition model of the target user; acquiring current address information of the target user through a mobile terminal carried by the target user regularly, and performing identification according to the current address information of the target user and a target user consumption merchant feature identification model to obtain preferential merchants interested by the target user; and pushing to a mobile phone bank APP of a mobile terminal according to preferential merchants interested by the target user, and generating preferential tasks from preferential information of each preferential merchant and storing the preferential tasks in a server. According to the invention, personalized recommendation of the bank preferential information can be realized.

Description

Technical Field

[0001] The present invention relates to bank information push, and in particular to a bank information push method based on a big data model. Background Art

[0002] At present, the push of bank information is generally done by the bank system regularly pushing some preferential information to the user's device in the form of SMS or APP notification. However, this kind of push has information confusion, and users need to filter out the notifications they are interested in from a lot of information, which brings a bad experience to users; at the same time, the push of messages is a unified push, which cannot meet the personalized needs of users. At the same time, due to the emergence of many SMS frauds, many people block or choose not to read SMS messages from unknown sources, resulting in the inability to push preferential information to users in need. Summary of the invention

[0003] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a bank information push method based on a big data model, which can solve the problem that the bank information push in the prior art is confusing and cannot meet the personalized needs of users.

[0004] The purpose of the present invention is achieved by the following technical solutions:

[0005] The bank information push method based on the big data model includes:

[0006] Model building steps: With the user's authorization, obtain the target user's historical daily consumption details, extract keywords from the target user's historical daily consumption details to obtain the target user's daily consumption merchant information, and then combine the neural network model to build a target user's consumption merchant feature recognition model;

[0007] Screening step: regularly obtaining the target user's address information through the mobile terminal carried by the target user, and obtaining all preferential merchants corresponding to the target user's authorized bank type within a preset range according to the target user's address information, and then matching and identifying each obtained preferential merchant with the target user's consumer merchant feature recognition model to screen out a number of candidate preferential merchants for the target user;

[0008] Recommended steps: obtain the merchant discount information of each candidate discount merchant according to several candidate discount merchants of the target user, generate discount notifications according to the merchant discount information of each candidate discount merchant and push them to the mobile banking APP of the mobile terminal carried by the target user in sequence, and store the merchant discount information of each candidate discount merchant in the server.

[0009] Furthermore, the model building step also includes: building a data set based on the collected daily consumption merchant information of the target user, and dividing the data set into a training set and a verification set, then training the pre-built neural network model based on the training set, and then verifying the trained model based on the verification set, and when the verification passes, deriving the target user's consumption merchant feature recognition model based on the verified model; the target user's consumption merchant feature recognition model includes the association relationship between the target user and the merchant features of his consumption.

[0010] Furthermore, the model building step also includes: obtaining the target user's configuration information to obtain the target user's working days and rest days, and dividing the data set into a working day data set and a rest day data set; and then performing model training and verification based on the working day data set, the rest day data set and a pre-built neural network model, thereby obtaining a target user's working day consumer merchant feature recognition model and a target user's rest day consumer merchant feature recognition model.

[0011] Furthermore, while obtaining the address information of the target user in the screening step, it also includes: obtaining the date of the day and selecting a corresponding consumer merchant feature recognition model according to the date of the day to match and identify so as to screen out a number of candidate preferential merchants for the target user; wherein, if the date of the day is a weekday, each preferential merchant is matched and identified with the target user's weekday consumer merchant feature recognition model; if the date of the day is a holiday, each preferential merchant is matched with the target user's holiday consumer merchant feature recognition model.

[0012] Furthermore, the recommendation step further includes: when it is detected that the mobile banking APP on the mobile terminal is opened, the latest promotion notification is displayed on the main page of the mobile banking APP in a pop-up window.

[0013] Furthermore, the recommendation step further includes: when the corresponding discount notification is read or viewed, obtaining the merchant discount information of the corresponding discount notification from the server by obtaining the click operation of the target user, and displaying the merchant discount information of the corresponding discount notification to the target user; the merchant discount information of the corresponding discount notification includes the name of the discount merchant, the address of the discount merchant, the type of discount merchant, the picture of the discount merchant, the discount amount, the discount time, the discount payment method and a brief description of the discount participation.

[0014] Furthermore, the recommendation step also includes: when the corresponding discount notification is pushed, timing is performed according to the push time of the corresponding discount notification, and when the timing time exceeds the first preset push time and the corresponding discount notification has not been read or viewed, generating a text message notification according to the text message notification template for the corresponding discount notification and pushing the text message notification to the text message end of the mobile terminal bound to the target user.

[0015] Furthermore, the recommendation step further includes: when the discount of the corresponding discount notification expires or the push time of the corresponding discount notification exceeds the second preset push time, deleting the corresponding discount notification from the mobile banking APP of the mobile terminal.

[0016] Furthermore, the recommendation step also includes: calculating the priority of each candidate preferential merchant of the target user according to a preset priority calculation rule, and pushing the preferential notification of each candidate preferential merchant to the mobile banking APP of the mobile terminal in sequence according to the order of priority.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] The present invention uses a machine learning model to perform data analysis, learning and other studies on the historical consumption details of the target user to obtain the characteristic data of the merchants consumed by the user, and then screens and matches the preferential merchants corresponding to the authorized bank for the user according to the user's current address information data and pushes them to the user's mobile banking APP in a timely manner, so that the preferential information of merchants with preferential offers and users are interested in will be notified to the user in a timely manner for viewing and consumption, etc., to prevent users from missing out on the preferential offers, and to achieve personalized preferential recommendations. At the same time, it can also relatively increase the user activity of the bank, ensure the user stickiness to the bank, and improve the user's recognition and experience of the bank. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flow chart of the bank information push method based on the big data model provided by the present invention. DETAILED DESCRIPTION

[0020] The present invention is further described below in conjunction with the accompanying drawings and specific implementation methods. It should be noted that, under the premise of no conflict, the various embodiments or technical features described below can be arbitrarily combined to form a new embodiment.

[0021] Based on the defects of existing bank information push, the present invention provides a bank information push method based on big data model, such as Figure 1 As shown, including:

[0022] Step S1: Obtain the target user's daily consumption history details with the target user's authorization, extract keywords from the target user's daily consumption history details to obtain the user's daily consumption merchant information, and build a target user's consumption merchant feature recognition model based on the user's daily consumption merchant information.

[0023] Specifically, this embodiment implements message push in the mobile banking APP installed by the user. Therefore, the target users here refer to users who have opened mobile banking of the corresponding bank. These users need to install the corresponding mobile banking APP in the mobile terminals they carry.

[0024] In addition, in order to ensure privacy, in this embodiment, the acquisition of the target user's daily consumption history details is completed with the user's authorization, such as WeChat, Alipay, and various bank cards to provide specific consumption details. After the user authorizes, the target user's daily consumption history details can be obtained and keywords can be extracted to obtain the user's daily consumption merchant information. The consumption merchant information here includes merchant type, merchant name, merchant address, consumption time, payment fee, payment method, whether there is a payment discount, discount method, discount amount and other information.

[0025] By conducting data analysis and mining on the above-mentioned users' daily consumption merchant information to construct a user's daily consumption merchant feature recognition model, the merchant information of interest to the user should be screened according to the consumption merchant feature recognition model to achieve personalized discount recommendations for the user.

[0026] More specifically, by collecting the user's daily consumption history details and performing keyword extraction, the user's daily consumption merchant information is obtained and a data set is constructed, and the data set is divided into a training set and a validation set.

[0027] Then, through the pre-constructed neural network model, the neural network model will be trained according to the above training set to obtain the trained neural network model, and then the trained neural network model will be verified according to the verification set to determine whether it meets the requirements. If so, the trained neural network model will be defined as the consumer merchant feature recognition model of the target user; if not, it will return to retrain until the trained consumer merchant feature recognition model of the target user is obtained.

[0028] The present invention utilizes machine learning models, such as neural network models, to achieve data learning and mining, and then derives the characteristics of merchants consumed by users every day, so as to recommend preferential merchants of interest to users. While achieving preferential merchant push, it can also achieve user personalized push, guide users to further consume, and solve the problems in the prior art that preferential consumption push is chaotic and not accurate enough, and cannot meet user needs, resulting in poor user experience of the bank, poor user stickiness, and reduced user activity.

[0029] Step S2: periodically obtain the target user's current address information through the mobile terminal carried by the target user, and obtain all preferential merchants corresponding to the target user's authorized bank type within a preset range based on the target user's current address information, and then match each preferential merchant with the target user's consumer merchant feature recognition model to identify the preferential merchants that the target user is interested in.

[0030] Specifically, this embodiment uses a mobile terminal carried by the user to obtain the current address information of the target user, and then uses the address information as the center to define the corresponding area according to preset rules to filter the merchant information for the target user. More preferably, when filtering the merchant information here, it can also be filtered according to pre-set keywords, such as the merchant type is food, entertainment, furniture, etc. At the same time, the merchant type that the user is interested in can also be estimated based on the user's current address information. For example, if the user is in an office building or a shopping mall, the types of interest are mostly food, entertainment, etc.; on the contrary, when the user is in a furniture mall, the types of interest are mostly furniture, and of course, food, etc. The specific selection can be made according to actual interests. In addition, the merchants with bank discounts recommended to users by this application, that is, when filtering users, it is also necessary to filter merchants with bank discounts. For some merchants that do not have bank discounts, they can be ignored.

[0031] In addition, the target user's authorized bank type includes not only the type of the current bank corresponding to the mobile banking APP, but also the bank cards of other banks under the target user's name can be bound. In this way, the mobile banking APP can lead to the bank type and bank card type of the bank card under the target user's name. Among them, the bank card type includes debit card and credit card.

[0032] Furthermore, the preset range in this embodiment refers to a certain reasonable range, such as the takeaway delivery range, or the area that can be reached within 20 minutes by car. Normally, for a user, the consumption of a merchant is through takeaway or directly in the store. Then when the recommended user is far away, it is very likely that the user will not choose to run a long distance to choose consumption for some discounts. Therefore, a consumption range can be set according to actual experience to increase the user's consumption possibility. The range here can also be based on the user's actual daily consumption needs and the distance between the user's home address, workplace, etc. and the consumer merchant to dig out a reasonable consumption area.

[0033] After all preferential merchants corresponding to the target user's authorized bank type are obtained within the preset range, each preferential merchant is matched with the target user's consumer merchant feature recognition model to obtain the preferential merchants that the target user is interested in. Among them, the target user's consumer merchant feature recognition model is mined based on the user's daily consumption details. Therefore, the selected preferential merchants are matched with the pre-trained model to obtain the target merchants that the user is interested in.

[0034] By mining the characteristics of the merchants where users consume, we can recommend preferential merchants that users are interested in, guide users to consume, and provide users with a better experience.

[0035] Step S3: Obtain the preferential information of each preferential merchant according to the preferential merchants that the target user is interested in and generate a preferential notification, then push the preferential notification to the mobile banking APP of the mobile terminal at a regular time, and generate a preferential task for the preferential information of each preferential merchant and store it in the server.

[0036] Specifically, in order to avoid that all the data of the preferential information is pushed to the mobile banking APP, which may cause freezes on the user's mobile phone and affect the user experience, this embodiment, when generating preferential notifications, makes corresponding preferential notifications based on the preferential information, such as making preferential posters in the form of pictures, and directly pushes the preferential posters to the mobile banking APP of the mobile terminal. This method can avoid storing all the data in the mobile terminal, causing device freezes, etc. In addition, this poster method is more concise, more attractive to users, and improves user interest. The preferential poster is pushed to the APP end. When the user views the poster and is interested, he will click on the button such as "Participate Now" on the preferential poster. In this way, the preferential data related to the preferential treatment is called from the server, so that the user can further understand the details of the preferential treatment and determine whether to participate in the preferential treatment.

[0037] That is, when the discount notification is read, if the target user is interested, he or she can click the operation button on the poster to jump to the server to obtain the discount merchant information corresponding to the discount notification, including the name of the discount merchant, the address of the discount merchant, the type of discount merchant, the picture of the discount merchant, the discount amount, the discount time, the discount payment method, a brief description of the discount participation and other information, so that the user can check whether the discount meets their own psychological needs, whether to participate in the discount, and how to participate in the discount.

[0038] More preferably, the present embodiment further obtains the user's working day and rest day according to the user's configuration information, and then divides the data set consisting of the user's daily consumption details into a working day data set and a rest day data set, and trains the constructed neural network model according to the working day data set and the rest day data set to obtain the target user's working day consumption merchant feature recognition model and rest day consumption merchant feature recognition model. Specifically, normally, for some users, especially some office workers, their working day consumption is mostly based on the consumption of daily work meals. On the contrary, rest days may be more inclined to other types of consumption, such as dinners, and family-related consumption, etc. For example, for some entertainment consumption, large-amount consumption, etc., more may be completed on rest days, that is, the consumption habits on working days and rest days will be different. Therefore, the present embodiment also divides the user's consumption into working days and rest days, so as to recommend more accurate consumption to users.

[0039] Furthermore, step S2 also includes, while obtaining the current address information of the target user, obtaining the current system date, and selecting the corresponding consumer merchant feature recognition model according to the system date of the day to obtain the preferential merchants that the target user is interested in. Specifically, if the current system date is a weekday, each preferential merchant is matched with the target user's weekday consumer merchant feature recognition model; if the system date of the day is a holiday, each preferential merchant is matched with the target user's holiday consumer merchant feature recognition model.

[0040] In addition, in order not to disturb the user's work and rest, this embodiment also sets a timer when obtaining the user's address information and system date, such as setting it between 9 am and 10 am every day, and setting the do not disturb mode at other times. In addition, in order to push more accurately, this embodiment also needs to time the user's address information when obtaining the user's address information, that is, to determine whether the user's address information has changed significantly within a period of time. If so, the user is in a mobile state, and there is no need to push discounts to him, so as to avoid frequent pushes, which will cause system confusion and affect the user's experience.

[0041] More preferably, after the preferential notice is pushed to the user, if it is detected that the mobile banking APP is opened, the latest preferential notice pushed is also displayed on the main page of the mobile banking APP in the form of a pop-up window. The pop-up window can guide the user to view the preferential notice.

[0042] Similarly, when a discount notification is pushed, the push time of the discount notification starts counting. When the timing time exceeds the preset time or the discount notification is not read, the corresponding discount notification will also generate a text message notification according to the text message notification template, and then the corresponding text message notification will be pushed to the text message end of the target user's mobile terminal. In actual use, some users may turn off the notification permission of the mobile banking APP, which will cause the target user to miss the discount notification. Therefore, when the discount notification is pushed, if it is not read or viewed within a certain period of time, in order to avoid the user missing the discount notification, it is necessary to generate a text message notification according to the text message notification template and push it to the text message end of the target user's mobile terminal to remind the user to check the discount notification. That is, the user can be reminded again to check the discount notification through text messages to avoid missing the discount.

[0043] More preferably, when the discount of the discount notification expires or the push time of the discount notification exceeds the preset time, the corresponding discount notification is deleted from the mobile banking APP. By regularly deleting the discount notification, the normal operation of the mobile banking APP can be guaranteed, and the excessive cache occupation can be avoided, which affects the user's use. When deleting the discount notification, it can be determined whether to delete it based on whether the discount of the discount notification expires and the push time.

[0044] When the discount corresponding to the discount notification is used, the time when the target user uses the discount, the discount merchant, the discount merchant address, the discount payment method, the payment fee details, etc. will also be recorded, and the discount usage details record of the target user will be generated and stored in the system, and keywords will be set for the discount usage details record of the target user. In step S3, when multiple discount merchants that the target user is interested in are obtained, the priority of each discount merchant is calculated according to the priority calculation rule, and then the discount information of multiple discount merchants is pushed to the mobile banking APP of the mobile terminal in the order of priority. When calculating the priority of the discount merchant, the priority of the discount merchant is realized by weighted average, specifically by setting multiple key factors, each key factor has a preset weight ratio, and multiple weighted averages of multiple key factors are calculated to calculate the priority of the discount merchant. At the same time, the engineer also pre-constructs the score of each key factor, such as setting multiple scores in sequence according to the size of the discount, etc., and setting according to the actual experience value, which is not specifically introduced in this embodiment.

[0045] Preferably, this embodiment also supports user-specified push, that is, the target user actively opens the mobile banking APP and searches for relevant discounts on the discount page. By pushing the page with search keywords to the target user, the target user can enter data information related to the discount, including address, merchant type, time, bank type, maximum consumption amount, etc. After obtaining the relevant information entered by the user, the relevant discount information can be searched according to the above information. At the same time, the found discount merchants are matched with the merchant feature recognition model constructed in the system, and several discount merchants are screened out and displayed to the user in sequence on the search page in the form of a list, so as to realize personalized recommendation for the user.

[0046] The above-mentioned embodiments are only preferred embodiments of the present invention and cannot be used to limit the scope of protection of the present invention. Any non-substantial changes and substitutions made by technicians in this field on the basis of the present invention shall fall within the scope of protection required by the present invention.

Claims

1. A bank information push method based on a big data model, characterized in that: include: Model building steps: With the user's authorization, obtain the target user's historical daily consumption details, extract keywords from the target user's historical daily consumption details to obtain the target user's daily consumption merchant information, and then combine the neural network model to build a target user's consumption merchant feature recognition model; Screening step: regularly obtaining the target user's address information through the mobile terminal carried by the target user, and obtaining all preferential merchants corresponding to the target user's authorized bank type within a preset range according to the target user's address information, and then matching and identifying each obtained preferential merchant with the target user's consumer merchant feature recognition model to screen out a number of candidate preferential merchants for the target user; Recommended steps: obtain the merchant discount information of each candidate discount merchant according to several candidate discount merchants of the target user, generate discount notifications according to the merchant discount information of each candidate discount merchant and push them to the mobile banking APP of the mobile terminal carried by the target user in sequence, and store the merchant discount information of each candidate discount merchant in the server.

2. The bank information push method based on the big data model according to claim 1 is characterized in that: The model building step also includes: building a data set based on the collected daily consumer merchant information of the target user, and dividing the data set into a training set and a verification set, then training a pre-built neural network model based on the training set, and then verifying the trained model based on the verification set, and when the verification is passed, deriving a consumer merchant feature recognition model for the target user based on the verified model; the consumer merchant feature recognition model for the target user includes an association relationship between the target user and the merchant features of the target user's consumption.

3. The bank information push method based on the big data model according to claim 2 is characterized in that: The model building step also includes: obtaining the target user's configuration information to obtain the target user's working days and rest days, and dividing the data set into a working day data set and a rest day data set; and then respectively performing model training and verification based on the working day data set, the rest day data set and a pre-built neural network model, thereby obtaining a target user's working day consumer merchant feature recognition model and a target user's rest day consumer merchant feature recognition model.

4. The bank information push method based on the big data model according to claim 3 is characterized in that: While obtaining the address information of the target user in the screening step, the step also includes: obtaining the date of the day and selecting a corresponding consumer merchant feature recognition model according to the date of the day to match and identify so as to screen out a number of candidate preferential merchants for the target user; wherein, if the date of the day is a weekday, each preferential merchant is matched and identified with the target user's weekday consumer merchant feature recognition model; if the date of the day is a holiday, each preferential merchant is matched with the target user's holiday consumer merchant feature recognition model.

5. The bank information push method based on big data model according to claim 1 is characterized in that: The recommendation step also includes: when it is detected that the mobile banking APP on the mobile terminal is opened, the latest promotion notification is displayed on the main page of the mobile banking APP in a pop-up window.

6. The bank information push method based on the big data model according to claim 5 is characterized in that: The recommendation step also includes: when the corresponding discount notification is read or viewed, obtaining the merchant discount information of the corresponding discount notification from the server by obtaining the click operation of the target user, and displaying the merchant discount information of the corresponding discount notification to the target user; the merchant discount information of the corresponding discount notification includes the name of the discount merchant, the address of the discount merchant, the type of discount merchant, the picture of the discount merchant, the discount amount, the discount time, the discount payment method and a brief description of the discount participation.

7. The bank information push method based on big data model according to claim 1 is characterized in that: The recommendation step also includes: when the corresponding discount notification is pushed, timing is performed according to the push time of the corresponding discount notification, and when the timing time exceeds the first preset push time and the corresponding discount notification has not been read or viewed, generating a text message notification according to the text message notification template for the corresponding discount notification and pushing the text message notification to the text message end of the mobile terminal bound to the target user.

8. The bank information push method based on big data model according to claim 1 is characterized in that: The recommendation step also includes: when the discount of the corresponding discount notification expires or the push time of the corresponding discount notification exceeds the second preset push time, deleting the corresponding discount notification from the mobile banking APP of the mobile terminal.

9. The bank information push method based on big data model according to claim 1, characterized in that: The recommendation step also includes: calculating the priority of each candidate preferential merchant of the target user according to a preset priority calculation rule, and pushing the preferential notification of each candidate preferential merchant to the mobile banking APP of the mobile terminal in sequence according to the order of priority.

Images