Enterprise right management method and system based on SaaS
By designing data storage, classification and delivery modules on the SaaS platform and calculating the best equity configuration using the BP neural network model, the problem of insufficient equity delivery accuracy caused by the traditional SaaS platform relying on third-party data is solved, and the customer's precise classification is achieved and personalized delivery improve marketing effectiveness.
Patent Information
- Application Number
- CN202411488049.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-24
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-10-24
AI Technical Summary
Traditional SaaS platforms rely on third-party customer consumption habit data when banking corporate equity is invested, resulting in insufficient accuracy and personalization of equity is invested.
Design a SaaS-based enterprise equity management system, including data storage module, vertical classification module, equity delivery module and data security module, and classify customer identity tags by extracting customer consumption habit data in one piece and secondary categories, generate customer identity tags, and use them The pre-trained BP neural network model calculates the best equity configuration for delivery.
It realizes accurate classification and personalized equity allocation of customers, reduces calculation complexity, improves marketing effectiveness and controls costs.
Smart Images

Figure CN119006047A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data management, and in particular to a SaaS-based enterprise rights management method and system. Background Art
[0002] In recent years, thanks to the development of cloud computing and big data technology, SaaS systems have continuously made breakthroughs and innovations in the ability of banks and enterprises to invest, issue and manage equity. They can build customer preferences based on consumption data from third-party platforms, and achieve accurate equity matching and personalized services. It is an emerging trend to classify customers based on artificial intelligence and machine learning algorithms, and cooperate with various consumer platforms and institutions to provide customers with a rich variety of equity types.
[0003] At present, the Chinese invention patent with application number CN202311269650.1 discloses a digital rights SAAS system that supports multi-merchant cooperative marketing. It has basic functions including digital rights issuance, operation management, and confirmation and cancellation, and can build a business scenario that integrates the needs of the virtual and real business ecosystem. The specific technical solution of the invention is: the invention uses SAAS technology to support brand companies to pay only a small access fee to launch their own digital rights functions in a short time, thereby quickly reaching users. Using the HRC721 protocol, the issuance, confirmation and cancellation of rights are executed through smart contracts, and the entire life cycle of rights is put on the chain. Based on the blockchain, the rights data can be trusted, the cancellation can be verified, and the synthesis mechanism can be executed, and the cooperative marketing function can be innovatively realized to promote the development of the entire business ecosystem. However, the invention only focuses on the security of corporate rights, and in the rights management, it neglects the classification of customers and the personalization of rights and interests, and relies on the customer consumption habit data obtained by the third-party platform. Summary of the invention
[0004] The technical problem solved by the present invention is that traditional SaaS platforms rely on customer consumption habit data obtained from third-party platforms when allocating benefits to banking enterprises. In the absence of customer consumption habit data, they only rely on account data to classify customers, resulting in insufficient accuracy and personalization when allocating benefits.
[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: A SaaS-based enterprise equity management system, comprising: a data storage module, a vertical classification module, an equity delivery module and a data security module; The data storage module is used to store customer data, equity configuration data, historical data and system log data; The vertical classification module is used to extract the consumption habit data of customers from the data storage module, perform a primary classification process on the customer data to obtain a first customer group and a second customer group, perform a secondary classification on the first customer group to obtain the nearest customer link with the second customer group, generate an identity tag for the second customer group based on the third data set, and add an identity tag identical to the nearest customer link to each customer in the first customer group; The equity delivery module is used to obtain the best equity configuration and carry out equity delivery. The pre-trained BP neural network model is used to perform classification tasks on the input identity tags to obtain the best equity configuration, and the API platform interface is used to deliver the equity to the customer according to the best equity configuration, and record the equity feedback; The data security module is used to implement access control, transmission encryption and audit the evaluation of equity investment, hierarchical control of system access rights according to administrator rights, operator rights and operator rights, and audit equity configuration changes.
[0006] Preferably, the data storage module includes a customer data unit, an equity configuration data unit, a historical data unit and a system log unit; The customer data unit is used to store basic customer information data, account data and consumption habit data; The equity configuration data unit is used to store the equity type data, equity placement rule data and equity inventory data in the current marketing activity; The historical data unit is used to store equity distribution records, equity usage records and customer feedback records in past marketing activities; The system log unit is used to store administrator operation logs, system operation logs, and backup logs.
[0007] Preferably, the basic customer information data includes the customer's name, gender, age, place of residence, family status and contact information; The account data includes the customer's account number, card opening time, account type, deposit data, loan data, and income and expenditure data; The consumption habit data includes the customer's consumption expenditure area data, consumption platform search data, product browsing data, consumption channel data and financial expenditure data; The equity type data includes discount coupons, bank points, discount packages and financial services; The equity placement rule data includes placement objects, placement time, types of placement equity, fields of placement equity, and the number of placement equity; The equity inventory data includes uninvested equity data, invested but unused equity data and used equity data; The equity placement record includes the type of equity placed in historical marketing activities, equity placement rules and equity usage; The equity use record includes the time of equity use, the occasion of equity use and the amount of equity use; The administrator operation log includes the administrator's login data, equity creation data, equity placement data, and equity change data; The backup log includes full log copies and incremental data backups.
[0008] Preferably, the vertical classification module includes a customer grouping unit, a feature extraction unit, and an identity labeling unit; The customer grouping unit is used to extract the consumption habit data of customers from the data storage module, and classify the customers based on the consumption habit data to obtain a first customer group and a second customer group; The logic of the customer grouping unit for classifying customer data is as follows: obtaining the consumption habit data of all customers, when the number of consumptions in the customer consumption expenditure field data is less than the first threshold, the number of searches in the consumption platform search data is less than the second threshold, and the number of visits in the product visit data is less than the third threshold, the judgment result is that the consumption habit data is insufficient, otherwise the judgment result is that the consumption habit data is sufficient; The customer grouping unit groups customers whose consumption habit data are insufficient as a result of a judgment into a first customer group, and groups customers whose consumption habit data are sufficient as a result of a judgment into a second customer group.
[0009] Preferably, the customer grouping unit extracts the customer basic information data and account data corresponding to the first customer group from the data storage module to obtain a first data set; The customer grouping unit extracts the customer basic information data and account data corresponding to the second customer group from the data storage module to obtain a second data set; The customer grouping unit extracts the customer basic information data, account data and consumption habit data corresponding to the second customer group from the data storage module to obtain a third data set.
[0010] Preferably, the feature extraction unit is used to generate a first feature matrix set and a second feature matrix set, and perform secondary classification on the first customer group to obtain the nearest customer links between each customer in the first customer group and the customers in the second customer group; The feature extraction unit performs interval division and one-hot encoding on the gender, age, place of residence, family status, account type, deposit data, loan data, and income and expenditure data in the first data set to obtain a first feature vector set, performs vector concatenation on each vector set in the first feature vector set to obtain a first concatenated vector set, and performs matrix processing on the first feature vector set to obtain a first feature matrix set; The feature extraction unit performs interval division and one-hot encoding on the gender, age, place of residence, family status, account type, deposit data, loan data, and income and expenditure data in the second data set to obtain a second feature vector set, performs vector concatenation on each vector set in the second feature vector set to obtain a second concatenated vector set, and performs matrix processing on the second feature vector set to obtain a second feature matrix set; The processing logic of the feature extraction unit for secondary classification includes: respectively calculating the cosine similarity of all vectors in the first splicing vector set and each vector in the second splicing vector set, respectively calculating the mutual correlation coefficient of all matrices in the first feature matrix set and each matrix in the second feature matrix set, obtaining the feature proximity value by linear weighted calculation of the cosine similarity and the mutual correlation coefficient, sorting the feature proximity values from large to small to obtain the maximum feature proximity value, and linking each customer in the first customer group with the corresponding nearest customer in the second customer group based on the maximum feature proximity value, completing the secondary classification, and obtaining the nearest customer link, and its calculation expression is:
[0011] Among them, DIS represents the feature proximity value, represents the cosine similarity between the vector with index i in the first concatenated vector set and the vector with index j in the second concatenated vector set, represents the cosine similarity weight, represents the mutual correlation coefficient between the matrix with index i in the first characteristic matrix set and the matrix with index j in the second characteristic matrix set, Represents the weight of the mutual correlation coefficient.
[0012] Preferably, the identity tag unit is used to generate identity tags for customers, generate identity tags for the second customer group based on the customers' gender, age, family status, account type, deposit data, loan data, consumer expenditure area data, consumer platform search data, product browsing data, consumer channel data and financial expenditure data in the third data set, and add an identity tag for each customer in the first customer group that is the same as the link to each customer's nearest neighbor.
[0013] Preferably, the equity delivery module includes a personalized delivery unit and a delivery terminal unit, the personalized delivery unit is used to select an equity configuration scheme according to the customer identity tag, and to obtain the best equity configuration by performing a classification task on the input identity tag through a pre-trained BP neural network model, and the best equity configuration is a combination of the delivery object, delivery time, delivery equity type, delivery equity field and delivery equity quantity output by the BP neural network model; The delivery terminal unit is used to use the API platform interface to deliver equity to customers according to the best equity configuration and record equity feedback; API interface platforms include banking apps, payment apps, SMS, official accounts and mini-programs; Equity feedback includes customer participation rate, equity utilization rate and customer complaint rate.
[0014] Preferably, the data security module includes a data security unit and an audit unit. The data security unit is used to implement access control and transmission encryption. When transmitting customer data and equity data, encryption is performed using the SSL algorithm and the TLS algorithm. User access is managed in a hierarchical manner, and system access rights are controlled in a hierarchical manner according to administrator rights, operator rights, and operator rights. The audit unit is used to audit the evaluation equity investment. The setting and adjustment of equity investment parameters by operator authority accounts and operator authority accounts require the audit confirmation of the administrator authority account through the audit unit before they can be executed.
[0015] A SaaS-based enterprise equity management method, comprising: Step S1, extracting the customer's consumption habit data from the data storage module, including the customer's consumption expenditure field data, consumption platform search data, product browsing data, consumption channel data and financial expenditure data; Perform a classification process on the customer data, perform threshold judgment on the number of consumptions in the customer's consumption expenditure field data, the number of searches in the consumption platform search data, and the number of visits in the product visit data to obtain a consumption habit data judgment result, and divide the customer into a first customer group and a second customer group based on the consumption habit data judgment result; Step S2, performing secondary classification processing on the first customer group, extracting customer data corresponding to the first customer group to obtain a first data set, extracting customer data corresponding to the second customer group to obtain a second data set and a third data set, performing interval partitioning and one-hot encoding based on the first data set to obtain a first feature vector set, performing matrix processing to obtain a first feature matrix set, performing vector concatenation to obtain a first concatenated vector set, performing interval partitioning and one-hot encoding based on the second data set to obtain a second feature vector set, performing matrix processing to obtain a second feature matrix set, and performing vector concatenation to obtain a second concatenated vector set; Calculate the cosine similarity based on the first concatenated vector set and the second concatenated vector set, calculate the mutual correlation coefficient based on all matrices in the first feature matrix set and the second feature matrix set, perform linear weighted calculation on the cosine similarity and the mutual correlation coefficient to obtain a feature proximity value, sort based on the feature proximity value to obtain a maximum feature proximity value, classify the first customer group based on the maximum feature proximity value, and form a nearest customer link between the first customer group and the nearest customer in the corresponding second customer group; Step S3, based on the identity tags of the customers in the second customer group, encode the identity tags to obtain identity tag vectors, perform classification tasks on the input identity tag vectors through the BP neural network model to obtain the optimal equity configuration, complete equity delivery to customers through the API platform interface based on the optimal equity configuration, and record equity feedback.
[0016] The beneficial effects of the present invention are as follows: by performing primary and secondary classification on customer data to obtain the nearest neighbor link, it is beneficial to accurate classification and reduce the complexity of subsequent calculations. The machine learning model makes full use of customer data with sufficient consumption habit data to calculate the optimal equity configuration. It is beneficial to accurately provide personalized equity placement plans for different types of customers and customers with insufficient consumption habit data, effectively control costs and improve marketing effects. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Figure 1 A schematic diagram of a framework of a SaaS-based enterprise equity management system provided by an embodiment of the present invention; Figure 2 A basic flow chart of a SaaS-based enterprise equity management method provided for one embodiment of the present invention. DETAILED DESCRIPTION
[0018] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, but not all of the embodiments.
[0019] Example 1, reference Figure 1 , is an embodiment of the present invention, and provides a SaaS-based enterprise equity management method, including: a SaaS-based enterprise equity management system, including: a data storage module, a vertical classification module, an equity delivery module and a data security module; The data storage module is used to store customer data, equity configuration data, historical data and system log data; The vertical classification module is used to extract the consumption habit data of customers from the data storage module, perform a primary classification process on the customer data to obtain a first customer group and a second customer group, perform a secondary classification on the first customer group to obtain the nearest customer link with the second customer group, generate an identity tag for the second customer group based on the third data set, and add an identity tag identical to the nearest customer link to each customer in the first customer group; The equity delivery module is used to obtain the best equity configuration and carry out equity delivery. The pre-trained BP neural network model is used to perform classification tasks on the input identity tags to obtain the best equity configuration, and the API platform interface is used to deliver the equity to the customer according to the best equity configuration, and record the equity feedback; The data security module is used to implement access control, transmission encryption and audit the evaluation of equity investment, hierarchical control of system access rights according to administrator rights, operator rights and operator rights, and audit equity configuration changes.
[0020] In this embodiment, the data storage module includes a customer data unit, an equity configuration data unit, a historical data unit and a system log unit; The customer data unit is used to store basic customer information data, account data and consumption habit data; The equity configuration data unit is used to store the equity type data, equity placement rule data and equity inventory data in the current marketing activity; The historical data unit is used to store equity distribution records, equity usage records and customer feedback records in past marketing activities; The system log unit is used to store administrator operation logs, system operation logs, and backup logs.
[0021] In this embodiment, the basic customer information data includes the customer's name, gender, age, place of residence, family status and contact information; The account data includes the customer's account number, card opening time, account type, deposit data, loan data, and income and expenditure data; The consumption habit data includes the customer's consumption expenditure area data, consumption platform search data, product browsing data, consumption channel data and financial expenditure data; The equity type data includes discount coupons, bank points, discount packages and financial services; The equity placement rule data includes placement objects, placement time, types of placement equity, fields of placement equity, and the number of placement equity; The equity inventory data includes uninvested equity data, invested but unused equity data and used equity data; The equity placement record includes the type of equity placed in historical marketing activities, equity placement rules and equity usage; The equity use record includes the time of equity use, the occasion of equity use and the amount of equity use; The administrator operation log includes the administrator's login data, equity creation data, equity placement data, and equity change data; The backup log includes full log copies and incremental data backups.
[0022] In this embodiment, the vertical classification module includes a customer grouping unit, a feature extraction unit, and an identity labeling unit; The customer grouping unit is used to extract the consumption habit data of customers from the data storage module, and classify the customers based on the consumption habit data to obtain a first customer group and a second customer group; The logic of the customer grouping unit for classifying customer data is as follows: obtaining the consumption habit data of all customers, when the number of consumptions in the customer consumption expenditure field data is less than the first threshold, the number of searches in the consumption platform search data is less than the second threshold, and the number of visits in the product visit data is less than the third threshold, the judgment result is that the consumption habit data is insufficient, otherwise the judgment result is that the consumption habit data is sufficient; The customer grouping unit groups customers whose consumption habit data are insufficient as a result of a judgment into a first customer group, and groups customers whose consumption habit data are sufficient as a result of a judgment into a second customer group.
[0023] In this embodiment, the customer grouping unit extracts the basic customer information data and account data corresponding to the first customer group from the data storage module to obtain a first data set; The customer grouping unit extracts the customer basic information data and account data corresponding to the second customer group from the data storage module to obtain a second data set; The customer grouping unit extracts the customer basic information data, account data and consumption habit data corresponding to the second customer group from the data storage module to obtain a third data set.
[0024] In this embodiment, the feature extraction unit is used to generate a first feature matrix set and a second feature matrix set, and perform secondary classification on the first customer group to obtain the nearest customer links between each customer in the first customer group and the customers in the second customer group; The feature extraction unit performs interval division and one-hot encoding on the gender, age, place of residence, family status, account type, deposit data, loan data, and income and expenditure data in the first data set to obtain a first feature vector set, performs vector concatenation on each vector set in the first feature vector set to obtain a first concatenated vector set, and performs matrix processing on the first feature vector set to obtain a first feature matrix set; The feature extraction unit performs interval division and one-hot encoding on the gender, age, place of residence, family status, account type, deposit data, loan data, and income and expenditure data in the second data set to obtain a second feature vector set, performs vector concatenation on each vector set in the second feature vector set to obtain a second concatenated vector set, and performs matrix processing on the second feature vector set to obtain a second feature matrix set; The processing logic of the feature extraction unit for secondary classification includes: respectively calculating the cosine similarity of all vectors in the first splicing vector set and each vector in the second splicing vector set, respectively calculating the mutual correlation coefficient of all matrices in the first feature matrix set and each matrix in the second feature matrix set, obtaining the feature proximity value by linear weighted calculation of the cosine similarity and the mutual correlation coefficient, sorting the feature proximity values from large to small to obtain the maximum feature proximity value, and linking each customer in the first customer group with the corresponding nearest customer in the second customer group based on the maximum feature proximity value, completing the secondary classification, and obtaining the nearest customer link, and its calculation expression is:
[0025] Among them, DIS represents the feature proximity value, represents the cosine similarity between the vector with index i in the first concatenated vector set and the vector with index j in the second concatenated vector set, represents the cosine similarity weight, represents the mutual correlation coefficient between the matrix with index i in the first characteristic matrix set and the matrix with index j in the second characteristic matrix set, Represents the weight of the mutual correlation coefficient.
[0026] Among them, the customers are classified once based on insufficient consumption habit data and sufficient consumption habit data to obtain the first customer group and the second customer group. The characteristic proximity value is calculated through the basic information data and account data, and the customers closest to each customer in the first customer group are identified in the second customer group to obtain the nearest customer link. After the identity label is generated for the second customer group data with sufficient consumption habit data, the first customer group with insufficient consumption habit data is given the same identity label through the nearest customer link. This is conducive to improving the accuracy of customer classification and completing the personalized configuration of subsequent rights and interests, and reducing the computational complexity of calculating the optimal rights and interests configuration.
[0027] In this embodiment, the identity tag unit is used to generate identity tags for customers, and generates identity tags for the second customer group based on the gender, age, family status, account type, deposit data, loan data, consumer expenditure area data, consumer platform search data, product browsing data, consumer channel data and financial expenditure data of the customers in the third data set, and adds an identity tag that is the same as the link to the nearest customer of each customer to each customer in the first customer group.
[0028] Among them, customer identity tags include: Gender labels: including male and female; Age labels: adolescent, middle-aged, elderly; Family status labels: single without children, single with children, married with children, and married without children; Account type labels: savings account, credit card account; Deposit data labels and loan data labels: less than 10,000, 10,000 to 50,000, 50,000 to 200,000, 200,000 to 1 million, and more than 1 million; Consumption data tags: catering, shopping, education, medical care and tourism; Consumer platform search data tags and browsing data tags: electronic products, household goods, clothing and beauty, sports and fitness, mid- and low-end brands, and high-end brands; Consumption channel data labels: online and offline channels; Financial expenditure data labels: conservative and aggressive; In this embodiment, the equity delivery module includes a personalized delivery unit and a delivery terminal unit. The personalized delivery unit is used to select an equity configuration scheme according to the customer identity tag, and to obtain the best equity configuration by performing a classification task on the input identity tag through a pre-trained BP neural network model. The best equity configuration is a combination of the delivery object, delivery time, delivery equity type, delivery equity field, and delivery equity quantity output by the BP neural network model; The delivery terminal unit is used to use the API platform interface to deliver equity to customers according to the best equity configuration and record equity feedback; API interface platforms include banking apps, payment apps, SMS, official accounts and mini-programs; Equity feedback includes customer participation rate, equity utilization rate and customer complaint rate.
[0029] Among them, the classification task is performed on the input identity label through the pre-trained BP neural network model, and its processing logic includes: in the network input layer, the identity label is encoded by interval division and unique hot encoding to obtain the identity label vector, the identity label vector is input into the network hidden layer, and the bias term is added after weighted operation through the network node connection weight, and the nonlinear expression is added through the ReLU activation function and the result is passed to the network output layer. After the node weighted calculation and the Softmax activation function, the probability distribution of the output category is obtained. The output category corresponds to the best delivery object, the best delivery time, the best delivery equity type, the best delivery equity field and the best delivery equity quantity. The output results are combined to obtain the best equity configuration.
[0030] Among them, the BP neural network model includes in the training logic: the customer's identity label includes gender label, age label, family status label, account type label, deposit data label, loan data label, consumption field data label, platform search data label, tour data label, consumption channel data label and financial expenditure data label. According to the types of the above customer identity labels, the number of input layer nodes is determined to be 11. Based on prior experience, the number of initial hidden layers and the number of nodes in each hidden layer are set, and the number of output layer nodes is determined to be 5, corresponding to the target of investment, investment time, investment equity type, investment equity field and investment equity number. The pre-prepared data set is divided into training set, verification set and test set according to 7:2:1. The data set contains customer identity labels and the corresponding optimal equity configuration. Set the training parameters, including learning rate, number of iterations and batch size. Usually, the random initialization method is used to assign initial values to the connection weights and bias items. The customer identity labels in the training set are encoded. The output equity configuration is obtained through forward propagation calculation. Then, the difference between the output equity configuration and the actual optimal equity configuration is calculated through the cross entropy loss function. The number of hidden layers, the number of nodes in each hidden layer and the corresponding network connection weights are updated through back propagation. The accuracy, precision and recall rate of the BP neural network model are calculated through the test set and validation set to complete the training.
[0031] Among them, comprehensive consideration of the customer's gender, age, family status, account type, deposit and loan status, consumption field preferences, consumption platform search and tour preferences, consumption channel preferences and financial expenditure preferences is conducive to accurately analyzing the customer's personalized needs from a multi-dimensional perspective. In view of the high-dimensionality, complexity and diversity of current customer data, the advantages of the BP neural network model's feature extraction ability and strong nonlinear processing ability are fully utilized to accurately classify and predict. For example, for young customers who like to travel, travel-related coupons, points doubling and other benefits can be provided. For middle-aged customers who are married and have children and have high consumption in the education field, discount information of education and training institutions can be pushed. It is conducive to providing personalized rights and interests delivery plans for different types of customers, effectively controlling costs while improving marketing effectiveness.
[0032] In this embodiment, the data security module includes a data security unit and an audit unit. The data security unit is used to implement access control and transmission encryption. When transmitting customer data and equity data, encryption is performed using the SSL algorithm and the TLS algorithm. User access is managed in a hierarchical manner, and system access rights are controlled in a hierarchical manner according to administrator rights, operator rights, and operator rights. The audit unit is used to audit the evaluation equity investment. The setting and adjustment of equity investment parameters by operator authority accounts and operator authority accounts require the audit confirmation of the administrator authority account through the audit unit before they can be executed.
[0033] Example 2, reference Figure 2 , is an embodiment of the present invention, and provides a SaaS-based enterprise rights management method, including: A SaaS-based enterprise rights management method, including: Step S1, extracting the customer's consumption habit data from the data storage module, including the customer's consumption expenditure field data, consumption platform search data, product browsing data, consumption channel data and financial expenditure data; Perform a classification process on the customer data, perform threshold judgment on the number of consumptions in the customer's consumption expenditure field data, the number of searches in the consumption platform search data, and the number of visits in the product visit data to obtain a consumption habit data judgment result, and divide the customer into a first customer group and a second customer group based on the consumption habit data judgment result; Step S2, performing secondary classification processing on the first customer group, extracting customer data corresponding to the first customer group to obtain a first data set, extracting customer data corresponding to the second customer group to obtain a second data set and a third data set, performing interval partitioning and one-hot encoding based on the first data set to obtain a first feature vector set, performing matrix processing to obtain a first feature matrix set, performing vector concatenation to obtain a first concatenated vector set, performing interval partitioning and one-hot encoding based on the second data set to obtain a second feature vector set, performing matrix processing to obtain a second feature matrix set, and performing vector concatenation to obtain a second concatenated vector set; Calculate the cosine similarity based on the first concatenated vector set and the second concatenated vector set, calculate the mutual correlation coefficient based on all matrices in the first feature matrix set and the second feature matrix set, perform linear weighted calculation on the cosine similarity and the mutual correlation coefficient to obtain a feature proximity value, sort based on the feature proximity value to obtain a maximum feature proximity value, classify the first customer group based on the maximum feature proximity value, and form a nearest customer link between the first customer group and the nearest customer in the corresponding second customer group; Step S3, based on the identity tags of the customers in the second customer group, encode the identity tags to obtain identity tag vectors, perform classification tasks on the input identity tag vectors through the BP neural network model to obtain the optimal equity configuration, complete equity delivery to customers through the API platform interface based on the optimal equity configuration, and record equity feedback.
[0034] It should be understood by those skilled in the art that the embodiments of the present invention may be provided as methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program codes. Among them, the storage medium may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic memory, flash memory, magnetic disk or optical disk. These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0035] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.
Claims
1. A SaaS-based enterprise equity management system, characterized in that: include: Data storage module, vertical classification module, equity delivery module and data security module; The data storage module is used to store customer data, equity configuration data, historical data and system log data; The vertical classification module is used to extract the consumption habit data of customers from the data storage module, perform a primary classification process on the customer data to obtain a first customer group and a second customer group, perform a secondary classification on the first customer group to obtain the nearest customer link with the second customer group, generate an identity tag of the second customer group according to the third data set, and add an identity tag identical to the nearest customer link to each customer in the first customer group; The equity delivery module is used to obtain the best equity configuration and carry out equity delivery. The pre-trained BP neural network model is used to perform classification tasks on the input identity tags to obtain the best equity configuration, and the API platform interface is used to deliver the equity to the customer according to the best equity configuration, and record the equity feedback; The data security module is used to implement access control, transmission encryption and audit the evaluation of equity investment, hierarchical control of system access rights according to administrator rights, operator rights and operator rights, and audit equity configuration changes.
2. The SaaS-based enterprise equity management system according to claim 1, characterized in that: The data storage module includes a customer data unit, an equity configuration data unit, a historical data unit and a system log unit; The customer data unit is used to store basic customer information data, account data and consumption habit data; The equity configuration data unit is used to store the equity type data, equity placement rule data and equity inventory data in the current marketing activity; The historical data unit is used to store equity distribution records, equity usage records and customer feedback records in past marketing activities; The system log unit is used to store administrator operation logs, system operation logs, and backup logs.
3. The SaaS-based enterprise equity management system according to claim 2, characterized in that: The basic customer information data includes the customer's name, gender, age, place of residence, family status and contact information; The account data includes the customer's account number, card opening time, account type, deposit data, loan data, and income and expenditure data; The consumption habit data includes the customer's consumption expenditure area data, consumption platform search data, product browsing data, consumption channel data and financial expenditure data; The equity type data includes discount coupons, bank points, discount packages and financial services; The equity placement rule data includes placement objects, placement time, types of placement equity, fields of placement equity, and the number of placement equity; The equity inventory data includes uninvested equity data, invested but unused equity data and used equity data; The equity placement record includes the type of equity placed in historical marketing activities, equity placement rules and equity usage; The equity use record includes the time of equity use, the occasion of equity use and the amount of equity use; The administrator operation log includes the administrator's login data, equity creation data, equity placement data, and equity change data; The backup log includes full log copies and incremental data backups.
4. The SaaS-based enterprise equity management system according to claim 1, characterized in that: The vertical classification module includes a customer grouping unit, a feature extraction unit, and an identity labeling unit; The customer grouping unit is used to extract the consumption habit data of customers from the data storage module, and classify the customers based on the consumption habit data to obtain a first customer group and a second customer group; The logic of the customer grouping unit for classifying customer data is as follows: obtaining the consumption habit data of all customers, when the number of consumptions in the customer consumption expenditure field data is less than the first threshold, the number of searches in the consumption platform search data is less than the second threshold, and the number of visits in the product visit data is less than the third threshold, the judgment result is that the consumption habit data is insufficient, otherwise the judgment result is that the consumption habit data is sufficient; The customer grouping unit groups customers whose consumption habit data are insufficient as a result of a judgment into a first customer group, and groups customers whose consumption habit data are sufficient as a result of a judgment into a second customer group.
5. The SaaS-based enterprise equity management system according to claim 4, characterized in that: The customer grouping unit extracts the customer basic information data and account data corresponding to the first customer group from the data storage module to obtain a first data set; The customer grouping unit extracts the customer basic information data and account data corresponding to the second customer group from the data storage module to obtain a second data set; The customer grouping unit extracts the customer basic information data, account data and consumption habit data corresponding to the second customer group from the data storage module to obtain a third data set.
6. The SaaS-based enterprise equity management system according to claim 5, characterized in that: The feature extraction unit is used to generate a first feature matrix set and a second feature matrix set, and perform secondary classification on the first customer group to obtain the nearest customer links between each customer in the first customer group and the customers in the second customer group; The feature extraction unit performs interval division and one-hot encoding on the gender, age, place of residence, family status, account type, deposit data, loan data, and income and expenditure data in the first data set to obtain a first feature vector set, performs vector concatenation on each vector set in the first feature vector set to obtain a first concatenated vector set, and performs matrix processing on the first feature vector set to obtain a first feature matrix set; The feature extraction unit performs interval division and one-hot encoding on the gender, age, place of residence, family status, account type, deposit data, loan data, and income and expenditure data in the second data set to obtain a second feature vector set, performs vector concatenation on each vector set in the second feature vector set to obtain a second concatenated vector set, and performs matrix processing on the second feature vector set to obtain a second feature matrix set; The processing logic of the feature extraction unit for secondary classification includes: respectively calculating the cosine similarity of all vectors in the first splicing vector set and each vector in the second splicing vector set, respectively calculating the mutual correlation coefficient of all matrices in the first feature matrix set and each matrix in the second feature matrix set, obtaining the feature proximity value by linear weighted calculation of the cosine similarity and the mutual correlation coefficient, sorting the feature proximity values from large to small to obtain the maximum feature proximity value, and linking each customer in the first customer group with the corresponding nearest customer in the second customer group based on the maximum feature proximity value, completing the secondary classification, and obtaining the nearest customer link, and its calculation expression is: ; Among them, DIS represents the feature proximity value, represents the cosine similarity between the vector with index i in the first concatenated vector set and the vector with index j in the second concatenated vector set, represents the cosine similarity weight, represents the mutual correlation coefficient between the matrix with index i in the first characteristic matrix set and the matrix with index j in the second characteristic matrix set, Represents the weight of the mutual correlation coefficient.
7. The SaaS-based enterprise equity management system according to claim 6, characterized in that: The identity tag unit is used to generate identity tags for customers, generate identity tags for the second customer group based on the gender, age, family status, account type, deposit data, loan data, consumer expenditure field data, consumer platform search data, product browsing data, consumer channel data and financial expenditure data of the customers in the third data set, and add an identity tag that is the same as the link to the nearest neighbor of each customer to each customer in the first customer group.
8. The SaaS-based enterprise equity management system according to claim 1, characterized in that: The equity delivery module includes a personalized delivery unit and a delivery terminal unit. The personalized delivery unit is used to select an equity configuration plan according to the customer identity tag, and to obtain the best equity configuration by performing a classification task on the input identity tag through a pre-trained BP neural network model. The best equity configuration is a combination of the delivery object, delivery time, delivery equity type, delivery equity field, and delivery equity quantity output by the BP neural network model; The delivery terminal unit is used to use the API platform interface to deliver equity to customers according to the best equity configuration and record equity feedback; API interface platforms include banking apps, payment apps, SMS, official accounts and mini-programs; Equity feedback includes customer participation rate, equity utilization rate and customer complaint rate.
9. The SaaS-based enterprise equity management system according to claim 1, characterized in that: The data security module includes a data security unit and an audit unit. The data security unit is used to implement access control and transmission encryption. When transmitting customer data and equity data, the data is encrypted using the SSL algorithm and the TLS algorithm. User access is managed in a hierarchical manner, and system access rights are controlled in a hierarchical manner according to administrator rights, operator rights, and operator rights. The audit unit is used to audit the evaluation equity investment. The setting and adjustment of equity investment parameters by operator authority accounts and operator authority accounts require the audit confirmation of the administrator authority account through the audit unit before they can be executed.
10. A SaaS-based enterprise equity management method, characterized in that: include: Step S1, extracting the customer's consumption habit data from the data storage module, including the customer's consumption expenditure field data, consumption platform search data, product browsing data, consumption channel data and financial expenditure data; Perform a classification process on the customer data, perform threshold judgment on the number of consumptions in the customer's consumption expenditure field data, the number of searches in the consumption platform search data, and the number of visits in the product visit data to obtain a consumption habit data judgment result, and divide the customer into a first customer group and a second customer group based on the consumption habit data judgment result; Step S2, performing secondary classification processing on the first customer group, extracting customer data corresponding to the first customer group to obtain a first data set, extracting customer data corresponding to the second customer group to obtain a second data set and a third data set, performing interval partitioning and one-hot encoding based on the first data set to obtain a first feature vector set, performing matrix processing to obtain a first feature matrix set, performing vector concatenation to obtain a first concatenated vector set, performing interval partitioning and one-hot encoding based on the second data set to obtain a second feature vector set, performing matrix processing to obtain a second feature matrix set, and performing vector concatenation to obtain a second concatenated vector set; Calculate the cosine similarity based on the first concatenated vector set and the second concatenated vector set, calculate the mutual correlation coefficient based on all matrices in the first feature matrix set and the second feature matrix set, perform linear weighted calculation on the cosine similarity and the mutual correlation coefficient to obtain a feature proximity value, sort based on the feature proximity value to obtain a maximum feature proximity value, classify the first customer group based on the maximum feature proximity value, and form a nearest customer link between the first customer group and the nearest customer in the corresponding second customer group; Step S3, based on the identity tags of the customers in the second customer group, encode the identity tags to obtain identity tag vectors, perform classification tasks on the input identity tag vectors through the BP neural network model to obtain the optimal equity configuration, complete equity delivery to customers through the API platform interface based on the optimal equity configuration, and record equity feedback.
Citation Information
Patent Citations
Digital rights SAAS system supporting multi-merchant cooperative marketing
CN117217761A
User portrait method and device, computer readable storage medium and electronic equipment
CN109784367A
User interest mining method and device, storage medium and computer equipment
CN110413897A
User portrait-based right and interest recommendation method and device, equipment and storage medium
CN113946754A
Label completion method and label completion device
CN113987223A
