Intelligent financial product cross-recommendation and optimization system
Patent Information
- Application Number
- TW115201107
- Authority / Receiving Office
- TW · TW
- Patent Type
- Utility models
- Current Assignee / Owner
- Filing Date
- 2026-02-03
- Publication Date
- 2026-07-11
- Estimated Expiration
- 2036-02-02
Smart Images
Figure IMG-2_DRAW_115201107-A0305-14-0001-1 
Figure IMG-2_DRAW_115201107-A0305-14-0002-2 
Figure IMG-2_DRAW_115201107-A0305-14-0003-3
Abstract
Description
Intelligent financial product cross-recommendation and optimization system Technical Field
[0001] This invention relates to a cross-recommendation and optimization system for financial products, specifically an intelligent cross-recommendation and optimization system for financial products that combines user identity information, user suitability information, user operation trajectory data, and multi-dimensional characteristic information of financial products, and uses recommendation algorithms and dynamic optimization mechanisms to automatically select and iteratively update the recommended content of financial products. Prior Technology
[0002] In recent years, digital financial services have become increasingly widespread, with a significant increase in users browsing, comparing, calculating, and purchasing financial products through bank websites or mobile banking. However, traditional financial product recommendation mechanisms are mostly based on static rules, typically relying solely on basic user attributes (such as age, risk tolerance, and existing relationship levels) for categorized recommendations. This fails to integrate dynamic data such as actual browsing behavior, interaction depth, and calculation context, resulting in a discrepancy between recommended results and users' current needs. Furthermore, existing recommendation technologies largely focus on the popularity or historical click-through rates of individual products, failing to calculate cross-application correlations, complementary features, and multi-dimensional product characteristics between different financial products, and making it difficult to construct recommendation combinations with cross-marketing effects.
[0003] Furthermore, previous technologies typically lacked comprehensive integration of user action data, failing to adjust recommendation logic based on user behavior patterns on financial product pages, such as repeated clicks, dwell time, trial calculations, comparison list behavior, and access to in-depth information pages. This static recommendation method, lacking a real-time response mechanism, prevents product ranking from adjusting to changes in user behavior, resulting in recommendations that do not fully reflect users' true preferences.
[0004] Furthermore, previous technologies lacked the ability to continuously adjust algorithms and dynamically optimize recommendation portfolios. Existing recommendation processes often remained unchanged for a long time after the initial generation of recommended content, lacking a mechanism for iterative updates based on user feedback data (such as click-through rates, application conversion rates, deep interaction behavior, and dwell time). As a result, long-term recommendation results gradually became distorted, even including products with consistently poor performance. In addition, past recommendation methods failed to establish performance evaluation indicators for each financial product in the recommendation portfolio, nor could they automatically replace or add products with better cross-marketing indicators when performance was poor, leaving the overall recommendation portfolio lacking self-adjustment and optimization capabilities.
[0005] In summary, previous technologies have long suffered from insufficient recommendation accuracy, inability to integrate multi-source data, lack of cross-product correlation analysis, lack of dynamic update mechanism for recommendation combination levels, and failure to automatically adjust recommended content based on feedback from all users. Therefore, it is necessary to propose improved technical means to solve this problem. Summary of the Invention
[0006] In view of the shortcomings of previous technologies, such as insufficient recommendation accuracy, inability to integrate multi-source data, lack of cross-product correlation analysis, lack of dynamic update mechanism for recommendation combination levels, and failure to automatically adjust recommendation content based on feedback from all users, this work discloses an intelligent financial product cross-recommendation and optimization system, in which:
[0007] First, this invention discloses an intelligent financial product cross-recommendation and optimization system, which includes: a user device and a bank server. The bank server further includes: a financial product recommendation database, a user operation database, a financial product recommendation module, a page generation module, a feedback collection module, and a recommendation combination optimization module.
[0008] The user device logs into the bank's website using the user's identity information. When financial product information is triggered on the bank's website, the financial product information is provided to the bank's server, and the financial product page is retrieved and displayed from the bank's server.
[0009] The bank server obtains financial product information from the user's device and provides financial product pages to the user's device. The bank server's financial product recommendation database stores multiple pieces of financial product information and their corresponding characteristics. The bank server's user operation database stores user identity information, user compatibility information, operation trajectory data, and optimized financial product recommendation combinations. The bank server's financial product recommendation module is connected to both the financial product recommendation database and the user operation database. When the optimized financial product recommendation combination corresponding to the user's identity information is empty, a recommendation algorithm is used to select a preset number of financial product pieces from the financial product recommendation database based on the user compatibility information, operation trajectory data, and the financial product characteristics corresponding to the financial product information to create a financial product recommendation combination. The bank server's page generation module is connected to the financial product recommendation module. When the optimized financial product recommendation combination corresponding to the user's identity information is not empty, the optimized financial product recommendation combination is integrated into the financial product page corresponding to the financial product information. When the optimized financial product recommendation combination corresponding to the user's identity information is empty, the optimized financial product recommendation combination is integrated into the page corresponding to the financial product information. The system includes: a financial product page corresponding to product information; a feedback collection module on the bank server that monitors and collects user feedback on recommended or optimized financial product combinations on the financial product page; and a recommendation optimization module on the bank server connected to the feedback collection module and the financial product recommendation database. Based on all user feedback on recommended or optimized financial product combinations, the system identifies and removes financial product information that needs to be replaced. It also eliminates negative correlations between unreplaced financial product information in recommended or optimized combinations and other financial product information in non-financial product combinations. Furthermore, it calculates cross-marketing metrics between unreplaced financial product information in recommended or optimized combinations and other financial product information in non-financial product combinations. The system selects the financial product information corresponding to the best cross-marketing metric to add to the recommended financial product combination to establish an optimized financial product combination, or selects the financial product information corresponding to the best cross-marketing metric to add to the optimized financial product combination to update the optimized financial product combination.
[0010] The system disclosed in this work, as described above, when any financial product information is triggered on a bank's website, uses a recommendation algorithm to select a preset number of financial product information from a financial product recommendation database, establishes an initial financial product recommendation combination, continuously monitors the behavioral feedback on the financial product recommendation combination, identifies and removes financial product information that needs to be replaced in the financial product recommendation combination, and then calculates the cross-marketing index of the remaining financial product information with other financial product information, thereby automatically replacing the recommended content to form an optimal financial product recommendation combination. This allows the recommendation combination to be automatically updated and iterated, and is then integrated with the financial product page and presented to the user's device in real time.
[0011] Through the aforementioned technical means, this invention can dynamically improve the accuracy and timeliness of financial product recommendations, and continuously adjust and optimize the recommended content based on user behavior, thereby enhancing the cross-marketing effect of financial products and improving the overall application conversion rate. Simple Explanation of the Diagram
[0012] Figure 1 shows the system block diagram of the intelligent financial product cross-recommendation and optimization system of this creation. Figure 2A illustrates a bank website that utilizes the intelligent cross-recommendation and optimization of financial products in this invention. Figure 2B illustrates the page layout of the intelligent financial product cross-recommendation and optimized integrated financial product recommendation combination in this creation. Figure 2C illustrates the page layout of the intelligent financial product cross-recommendation and optimization combination financial product page. Figures 3A and 3B illustrate the flowcharts for the intelligent cross-recommendation and optimization of financial products in this invention. Figure 4 illustrates the computer system architecture for intelligent cross-recommendation and optimization of financial products in this invention. Implementation
[0013] The following will describe in detail the implementation of this invention with reference to the drawings and embodiments, so that you can fully understand how this invention uses technical means to solve technical problems and achieve technical effects and implement it accordingly.
[0014] The following section will first explain the intelligent financial product cross-recommendation and optimization system disclosed in this work, and please refer to "Figure 1", which is a system block diagram of the intelligent financial product cross-recommendation and optimization system of this work.
[0015] First, this invention discloses an intelligent financial product cross-recommendation and optimization system, which includes: a user device 10 and a bank server 20. The bank server 20 further includes: a financial product recommendation database 21, a user operation database 22, a financial product recommendation module 23, a page generation module 24, a feedback collection module 25, and a recommendation combination optimization module 26.
[0016] User device 10 logs into the bank website using user identity information, such as a combination of bank account number and password, financial certificates, etc. This is only an example and does not limit the application scope of this invention. When financial product information 31 is triggered on the bank website 30, user device 10 provides financial product information 31 to the bank server 20. For a diagram of the bank website 30 and financial product information 31, please refer to "Figure 2A". "Figure 2A" is a schematic diagram of the bank website for intelligent cross-recommendation and optimization of financial products in this invention.
[0017] A financial product recommendation database 21 is pre-established in the bank server 20, storing information on multiple financial products and their corresponding characteristics. The aforementioned financial product information includes, but is not limited to, basic product attributes, fee information, market characteristic indicators, return and compensation features, suitable user classification suggestions, and product restrictions or special qualifications. The aforementioned basic product attributes include, for example, funds, foreign currency fixed deposits, insurance, structured products, etc. (product type); RR1, RR2, RR3, etc. (product risk level); short-term, medium-term, long-term, flexible period, etc. (product duration). The aforementioned fee information includes, for example, commission rates (e.g., 1% for subscription, 0% for redemption). 5%), management fees, trust fees, etc.; the aforementioned product market characteristic indicators include: market region (e.g., US equity funds, global balanced funds, emerging markets, etc.), investment target type (e.g., stocks, bonds, ETFs, foreign exchange, etc.); the aforementioned product return and return characteristics include: historical annualized rate of return (e.g., 1 year, 3 years, 5 years), return volatility indicators; the aforementioned suitable user classification suggestions include: suitable for conservative users, suitable for short-term capital use, suitable for those who pursue high volatility products, etc.; the aforementioned product restrictions or special qualifications include: whether new accounts are restricted, whether specific account levels are restricted, whether APP subscriptions are restricted, etc. These are only examples and are not intended to limit the application scope of this work.
[0018] A user operation database 22 is pre-established in the bank server 20, storing user identity information, user adaptation information, operation trajectory data, and optimized recommendation combinations of financial products. The aforementioned user adaptation information includes, but is not limited to: user basic attribute information, financial risk tolerance classification, user historical financial product preferences, user interaction and dwell behavior characteristics, and credit behavior and transaction record summaries, etc. The aforementioned user basic attribute information includes, for example: age range (e.g., 20 to 30 years old), whether it is a new or old customer, income range (e.g., 30,000 to 60,000 yuan), occupation type (e.g., salaried worker, self-employed), etc.; the aforementioned financial risk tolerance classification includes, for example: Conservative, stable, balanced, growth-oriented, or aggressive risk profiles can be automatically estimated from a user's previously completed risk attribute questionnaire or transaction preference algorithm. The aforementioned user's historical financial product preferences include, for example, browsing history showing a preference for investment-linked insurance, funds, and foreign currency deposits. The aforementioned user interaction and dwell time characteristics include the user's dwell time on a particular type of product, frequent use of banking auxiliary tools (e.g., product comparison, calculators, etc.). The aforementioned summary of credit behavior and transaction records includes credit card usage (activity level, payment patterns), loan records, average fund transfer levels, etc. These are merely illustrative examples and do not limit the application scope of this work.
[0019] The aforementioned operation trajectory data is automatically generated by the user's device 10 while browsing the bank's website 30. The operation trajectory data includes, but is not limited to, page browsing sequence, click behavior data, dwell time, interaction depth data, swipe trajectory, and device and usage environment information. The aforementioned page browsing sequence is, for example: homepage → fund zone → a fund product page → calculation page → back to fund list → foreign currency zone; the aforementioned click behavior data is, for example: records of clicking on product descriptions, risk disclosure documents, return trend charts, comparison buttons, FAQs, and immediate subscription buttons; the aforementioned dwell time is, for example: 25 seconds spent in the "product return calculation area," indicating high attention to the relevant functions; the aforementioned interaction depth data includes whether advanced explanations are expanded, whether product PDF manuals are downloaded, and whether risk assessment functions are used; the aforementioned swipe trajectory is, for example: quickly swiping over content, repeatedly pulling back to view a certain section, etc., to infer the information area that the user values most; the aforementioned device and usage environment information includes the type of device used (mobile / desktop), browser version, whether the operation is within an APP, etc., which are only examples and do not limit the application scope of this creation.
[0020] The aforementioned optimized financial product recommendation portfolio is the result of continuously running this invention and optimizing the cross-marketing among the financial products. Examples of such financial product recommendations include: Financial Product Information A, Financial Product Information B, and Financial Product Information C. Optimized financial product recommendation portfolios after running this invention may include: Financial Product Information A, Financial Product Information B, and Financial Product Information D; Financial Product Information A, Financial Product Information E, and Financial Product Information C; Financial Product Information A, Financial Product Information D, and Financial Product Information F, etc. These are merely illustrative examples and do not limit the application scope of this invention.
[0021] When financial product information is triggered on the bank's website 30, the user device 10 provides financial product information to the bank server 20. The bank server 20 can then obtain the financial product information from the user device 10. The financial product recommendation module 23 is connected to the financial product recommendation database 21 and the user operation database 22. When the financial product recommendation module 23 confirms that the optimal recommendation combination of financial products corresponding to the user's identity information is empty, the financial product recommendation module 23 uses a recommendation algorithm to select a preset number of financial product information from the financial product characteristic database 22 based on the user adaptation information and operation trajectory data corresponding to the user's identity information and the financial product characteristics corresponding to the financial product information to establish a financial product recommendation combination.
[0022] The financial product recommendation module 23 of the bank server 20 obtains user-adaptive information corresponding to the user's identity information from the user operation database 22, such as: the user's risk tolerance level (conservative, moderate, aggressive), investment preferences (preference for fixed deposits, bonds, funds, structured products), investment period preferences (short-term, medium-term, long-term), liquidity needs, and regulatory compliance results, etc. It also obtains operation trajectory data corresponding to the user's identity information from the user operation database 22, such as: financial product pages viewed or clicked by the user, dwell time, records of products added to the comparison list, previously calculated product plans, historical application records and application results, etc. The financial product recommendation module 23 also obtains the characteristics of the financial products corresponding to the financial product information from the financial product recommendation module 23, such as: the risk level, expected rate of return, investment period, currency, minimum subscription threshold, whether it has principal protection, default, redemption conditions, target customer attributes, and other multi-dimensional characteristics of each financial product, etc.
[0023] Specifically, the financial product recommendation module 23, based on the aforementioned user adaptation information, operation trajectory data, and financial product characteristics, uses a pre-built recommendation algorithm for calculation. The recommendation algorithm, for example, combines rule-based filtering with a machine learning model: based on user adaptation information related to user identity, it filters out suitable financial product information, and then, using a trained machine learning model, calculates the matching score of the filtered financial products to the user identity information. For instance, the recommendation algorithm can input behavioral characteristics from operation trajectory data such as "repeatedly browsing foreign currency fixed deposit pages," "recently checking foreign exchange rates multiple times," and "not yet holding foreign currency products," along with financial product characteristics such as "foreign currency fixed deposits," "foreign currency bond funds," and "multi-currency portfolio products," into the machine learning model to estimate the matching score of the user identity information to the filtered financial product information.
[0024] The financial product recommendation module 23 sorts the selected financial product information according to the matching score and selects a preset number (e.g., 2, 3, 5, etc.) of financial product information to create a financial product recommendation portfolio.
[0025] Next, when the optimized recommendation combination of financial products corresponding to the user's identity information is empty, the page generation module 24 of the bank server 20 integrates the financial product recommendation combination 41 into the financial product page 40 corresponding to the financial product information. For a schematic diagram of the financial product page 40 including the financial product recommendation combination 41, please refer to "Figure 2B". "Figure 2B" is a schematic diagram of the integrated financial product recommendation combination financial product page of the intelligent cross-recommendation and optimization of financial products in this invention. Then, the bank server 20 provides the financial product page 40 to the user device 10, and the user device 10 obtains the financial product page 40 from the bank server 20 and displays it.
[0026] Meanwhile, the feedback collection module 25 of the bank server 20 monitors and collects user identity information regarding behavioral feedback on the financial product recommendation combination 41 on the financial product page 40. The aforementioned behavioral feedback includes, but is not limited to, the click-through rate of the financial product information in the financial product recommendation combination 41, the conversion rate of the financial product information in the financial product recommendation combination 41, and the time or behavior of subsequently staying on the financial product information in the financial product recommendation combination 41, etc. This is only an example and is not intended to limit the application scope of this work.
[0027] Specifically, when a domestic bond fund in the recommended portfolio 41 of financial products on the financial product page 40 is clicked (by touch or mouse), the feedback collection module 25 records the click event to calculate the click-through rate of the domestic bond fund in the recommended portfolio 41. The feedback collection module 25 also records the dwell time on the domestic bond fund in the recommended portfolio 41; differences in dwell time are considered as the degree of user interest in the recommended portfolio 41. The feedback collection module 25 monitors whether the user's information enters the detailed pages of the domestic bond fund in the recommended portfolio 41 (e.g., expanding the product risk level, enabling the return trend chart, using calculation tools, etc.) to confirm whether the user's information involves in-depth interaction with the domestic bond fund in the recommended portfolio 41. The data collection module 25 monitors whether the user has added the recommended financial product portfolio 41 (China's domestic bond fund) to their favorites list, added it to their comparison list, downloaded its prospectus, or initiated its online subscription process, etc., to calculate the conversion rate of the recommended financial product portfolio 41 (China's domestic bond fund). The feedback module 25 records the time sequence of these behavioral responses, i.e., it records each event according to timestamps. For example, if the user clicks on the recommended financial product portfolio 41 (China's domestic bond fund) within 5 seconds, or enters its calculation page 20 seconds later, etc., this is merely an example and does not limit the application scope of this invention.
[0028] Next, the recommendation portfolio optimization module 26 of the bank server 20 is connected to the financial product recommendation database 21. Based on the behavioral feedback of all users' identity information on the financial product recommendation portfolio, it identifies and removes financial product information that needs to be replaced in the financial product recommendation portfolio. Then, it calculates the cross-marketing index between the financial product information that has not been replaced in the financial product recommendation portfolio and the other financial product information in the non-financial product recommendation portfolio. It selects the financial product information corresponding to the best cross-marketing index and adds it to the financial product recommendation portfolio to establish the optimized financial product recommendation portfolio.
[0029] The recommended portfolio optimization module 26 calculates the performance indicators for each individual financial product in each recommended portfolio, such as: click-through rate (calculated from the number of clicks and applications), application approval rate, holding rate, subsequent subscription rate, user cancellation rate, etc., and combines the above performance indicators to form a weighted performance score.
[0030] When the recommendation portfolio optimization module 26 determines that the weighted performance score of a specific financial product information in a certain financial product recommendation portfolio is lower than a preset threshold, or that the financial product information continues to cause the overall financial product recommendation portfolio application conversion rate to be significantly lower than the average of the same group, the financial product information is marked as "financial product information that needs to be replaced" and temporarily removed from the financial product recommendation portfolio to form a recommendation portfolio to be optimized, in which only the financial product information that has not yet been marked for replacement is retained.
[0031] Next, the recommendation portfolio optimization module 26 calculates cross-marketing metrics for the financial product information in the recommendation portfolio that has not been replaced, and for the remaining financial product information in the financial product recommendation database 21 that is not included in the financial product recommendation portfolio. The recommendation portfolio optimization module 26 uses the historical application behavior of all users to determine whether there is a higher probability of applying for financial product information B when a user's identity information simultaneously holds financial product information A, and calculates the cross-application rate of financial product information A and financial product information B accordingly. At the same time, it can also consider the trend of financial product information A and financial product information B. Whether the insurance attributes, investment period, currency, and target customer groups are complementary or synergistic (e.g., fixed deposit products paired with medium-risk fund products, protection insurance paired with investment insurance, etc.), and combining the above cross-application rates and product complementarity indicators into a cross-marketing indicator score. Specifically, if the recommended portfolio optimization module 26 finds that "users holding foreign currency fixed deposit products are significantly more likely to apply for foreign currency bond funds later than average" or "users applying for term life insurance are more likely to apply for medical insurance later," then the recommended portfolio optimization module 26 will set a higher cross-marketing indicator score for these product combinations.
[0032] After completing the cross-marketing metric calculation, the recommended portfolio optimization module 26 sorts all candidate "other financial product information in non-financial product recommended portfolios" according to their cross-marketing metric scores. It can also check regulatory compliance conditions, risk level matching strategies, and product duplication restrictions (e.g., avoiding too many highly similar or functionally overlapping products in the same recommended portfolio). The recommended portfolio optimization module 26 selects at least one piece of financial product information with the highest cross-marketing metric and that meets the above restrictions from the sorting results and adds it to the recommended portfolio to be optimized, so that the gaps created by the removal can be filled by financial product information with better cross-marketing potential.
[0033] Specifically, if the recommended portfolio of financial products includes "NTD time deposits", "domestic bond funds" and "USD structured products", and the "USD structured products" have a low long-term conversion rate and are marked as needing to be replaced, then the recommended portfolio optimization module 26 can calculate based on historical data that the cross-marketing index of "NTD time deposits" and "USD high-rated bond funds" is higher than that of other candidate products. Therefore, "USD high-rated bond funds" are selected to replace "USD structured products" to form a new optimized recommended portfolio of financial products.
[0034] It is worth noting that before calculating cross-marketing metrics, the recommended portfolio optimization module 26 further excludes negative correlations between unreplaced financial product information in the financial product recommended portfolio or the optimized financial product recommended portfolio and other financial product information in the non-financial product recommended portfolio. This is to avoid simultaneously allocating financial product information with negative relationships to the same financial product recommended portfolio or the optimized financial product recommended portfolio. That is, if holding or applying for the first financial product information leads to a significant decrease in the application conversion rate, click-through rate, or deep interaction behavior of the second financial product information (the decrease exceeds the preset negative correlation judgment threshold), it is determined that there is a negative correlation between the first financial product information and the second financial product information.
[0035] Finally, the recommended combination optimization module 26 writes the optimized financial product recommendation combination 42, after replacing and adding new financial product information, back to the financial product recommendation database 21, and updates the user group tags and version information corresponding to the financial product recommendation combination 41. This allows the subsequent financial product recommendation module 23 to directly read the latest optimized financial product recommendation combination 42 when the corresponding user logs in with their identity information. For a schematic diagram of the financial product page 40 containing the optimized financial product recommendation combination 42, please refer to "Figure 2C". "Figure 2C" is a schematic diagram of the integrated optimized financial product recommendation combination financial product page of the intelligent cross-recommendation and optimization of financial products in this creation.
[0036] When user device 10 logs into the bank's website again using user identity information, and financial product information is triggered on the bank's website 30, user device 10 provides financial product information to the bank's server 20. The bank's server 20 can then obtain the financial product information from user device 10. The financial product recommendation module 23 confirms that the optimal recommendation combination of financial products corresponding to the user identity information is not empty, and integrates the optimal recommendation combination of financial products 42 into the financial product page 40 corresponding to the financial product information.
[0037] Meanwhile, the feedback collection module 25 monitors and collects user identity information regarding behavioral feedback on the optimized financial product recommendation combination 42 on the financial product page 40. Based on all user identity information regarding behavioral feedback on the optimized financial product recommendation combination 42, the recommendation combination optimization module 26 identifies and removes financial product information that needs to be replaced in the optimized financial product recommendation combination 42. It then calculates the cross-marketing index between the financial product information that has not been replaced in the optimized financial product recommendation combination 42 and the remaining financial product information in the non-optimized financial product recommendation combination 42. The financial product information corresponding to the best cross-marketing index is selected and added to the optimized financial product recommendation combination 42 to update the optimized financial product recommendation combination 42. The update process of the optimized financial product recommendation combination 42 is described above in the process of establishing the optimized financial product recommendation combination 42, and will not be repeated here. This is to achieve adaptive iterative updates of the optimized financial product recommendation combination 42.
[0038] Next, the operation process of this creation will be explained below. Please also refer to "Figure 3A" and "Figure 3B". "Figure 3A" and "Figure 3B" show the flowchart of the intelligent cross-recommendation and optimization of financial products in this creation.
[0039] First, the user device logs into the bank's website using the user's identity information. When financial product information is triggered on the bank's website, the user device provides the financial product information to the bank's server (step 501). Next, the bank's server pre-establishes a financial product recommendation database storing multiple pieces of financial product information and their corresponding financial product characteristics (step 502). Next, the bank's server pre-establishes a user operation database storing user identity information, user adaptation information, operation trajectory data, and optimized financial product recommendation combinations (step 503). Next, when the optimized financial product recommendation combination corresponding to the user's identity information is empty, the bank's server uses a recommendation algorithm to select a preset number of financial product information from the financial product recommendation database based on the user adaptation information and operation trajectory data corresponding to the user's identity information, as well as the financial product characteristics corresponding to the financial product information, to create a financial product recommendation combination (step 504). Next, when the optimized financial product recommendation combination corresponding to the user's identity information is empty, the bank's server integrates the financial product recommendation combination into the financial product page corresponding to the financial product information (step 505). Next, when the user's identity information... When the corresponding optimized financial product recommendation portfolio is not empty, the bank server integrates the optimized financial product recommendation portfolio into the financial product page corresponding to the financial product information (step 506); then, the bank server provides the financial product page to the user's device and displays it (step 507); next, the bank server monitors and collects user identity information's behavioral feedback on the financial product recommendation portfolio or optimized financial product recommendation portfolio on the financial product page, confirms and removes the financial product information that needs to be replaced in the financial product recommendation portfolio or optimized financial product recommendation portfolio (step 508); next, the bank server calculates the cross-marketing index between the financial product information that has not been replaced in the financial product recommendation portfolio or optimized financial product recommendation portfolio and the other financial product information in the non-financial product recommendation portfolio (step 509); next, the bank server selects the financial product information corresponding to the best cross-marketing index and adds it to the financial product recommendation portfolio to establish the optimized financial product recommendation portfolio (step 510); and the bank server selects the financial product information corresponding to the best cross-marketing index and adds it to the optimized financial product recommendation portfolio to update the optimized financial product recommendation portfolio (step 511).
[0040] Please refer to Figure 4, which illustrates the computer system architecture for the intelligent financial product cross-recommendation and optimization of this invention. It should be noted that the computer system 600 of the electronic device shown in Figure 4 is merely an example and should not impose any limitations on the functionality or scope of use of the embodiments of this invention.
[0041] As shown in Figure 4, the computer system 600 includes a Central Processing Unit (CPU) 601, which can perform various appropriate actions and processes, such as executing the methods described in the above embodiments, based on programs stored in Read-Only Memory (ROM) 602 or programs loaded from Storage Unit 608 into Random Access Memory (RAM) 603. The RAM 603 also stores various programs and data required for system operation. The CPU 601, ROM 602, and RAM 603 are interconnected via bus 604. An Input / Output (I / O) interface 605 is also connected to bus 604.
[0042] The following components are connected to the I / O interface 605: an input section 606 including a keyboard, mouse, etc.; an output section 607 including a cathode ray tube (CRT), liquid crystal display (LCD), and speakers, etc.; a storage section 608 including a hard drive, etc.; and a communication section 609 including a network interface card such as a LAN (Local Area Network) card and a modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as needed. Removable media 611, such as magnetic disks, optical discs, magneto-optical disks, semiconductor memory, etc., are installed on the drive 610 as needed so that computer programs read from them can be installed into the storage section 608 as needed.
[0043] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including computer code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 609, and / or installed from removable media 611. When the computer program is executed by central processing unit (CPU) 601, it performs various functions defined in the system of the present invention.
[0044] It should be noted that the computer-readable media shown in this embodiment can be a computer-readable signal media, a computer-readable storage media, or any combination thereof. Computer-readable storage media can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), electronically erasable programmable read-only memory (EPROM), flash memory, optical fiber, compact disc read-only memory (CD-ROM), optical memory devices, magnetic memory devices, or any suitable combination thereof. In this invention, computer-readable signal media can include data signals propagated in a baseband frequency or as part of a carrier wave, carrying computer-readable computer programs. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit programs for use by or in connection with an instruction execution system, apparatus, or device. Computer programs contained on a computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.
[0045] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. Each block in the flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, or they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram or flowchart, and combinations of blocks in the block diagram or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.
[0046] The units described in this embodiment can be implemented in software or hardware, and can also be located in a processor. The names of these units do not necessarily limit the specific unit itself. Therefore, the technical solution according to this embodiment can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, portable hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, touch terminal, or network device, etc.) to execute the method according to this embodiment.
[0047] In summary, when any financial product information is triggered on a bank's website, a recommendation algorithm selects a preset number of financial product information from the financial product recommendation database to establish an initial financial product recommendation combination. The algorithm continuously monitors behavioral feedback on the financial product recommendation combination, identifies and removes financial product information that needs to be replaced, and calculates the cross-marketing metrics of the remaining financial product information with other financial product information. This automatically replaces the recommended content, forming an optimized financial product recommendation combination. The recommendation combination is automatically updated and iterated, and after integration with the financial product page, it is presented to the user's device in real time.
[0048] This technology can solve the problems of insufficient recommendation accuracy, inability to integrate multi-source data, lack of cross-product correlation analysis, lack of dynamic update mechanism for recommendation combination levels, and failure to automatically adjust recommendation content based on the behavioral feedback of all users. It can then achieve the technical efficacy of dynamically improving the accuracy and timeliness of financial product recommendations, and continuously adjusting and optimizing recommendation content based on user behavior, thereby enhancing the cross-marketing effect of financial products and improving the overall application conversion rate.
[0049] While the embodiments disclosed in this invention are as described above, the content is not intended to directly limit the scope of patent protection for this invention. Anyone skilled in the art to which this invention pertains may make minor modifications in form and detail without departing from the spirit and scope disclosed herein. The scope of patent protection for this invention shall still be determined by the appended claims.
[0050] 10: User Device
[0051] 20: Bank Server
[0052] 21: Financial Product Recommendation Database
[0053] 22: User Operation Database
[0054] 23: Financial Product Recommendation Module
[0055] 24: Page Generation Module
[0056] 25: Feedback Collection Module
[0057] 26: Recommended optimal module combination
[0058] 30: Bank websites
[0059] 31: Financial Products Information
[0060] 40: Financial Products Page
[0061] 41: Recommended portfolio of financial products
[0062] 42: Optimized Recommendation Portfolio for Financial Products
[0063] 501-511: Steps
[0064] 601: CPU
[0065] 602:ROM
[0066] 603: RAM
[0067] 604: Busbar
[0068] 605:I / O interface
[0069] 606: Input section
[0070] 607: Output Section
[0071] 608: Storage Section
[0072] 609: Communication Section
[0073] 610: Driver
[0074] 611: Uninstallable media
Claims
1. An intelligent financial product cross-recommendation and optimization system, comprising: a user device that logs into a bank's website using user identity information; when financial product information is triggered on the bank's website, providing the financial product information to a bank server, and retrieving and displaying a financial product page from the bank server; the bank server retrieving the financial product information from the user device, and providing the financial product page to the user device; the bank server further comprising: a financial product database storing multiple pieces of financial product information and corresponding financial product characteristics; and a user operation database storing the user identity information, user adaptation information, operation trajectory data, and an optimized combination of financial products. A financial product recommendation module, connected to a financial product recommendation database and the user operation database, when the optimal recommendation combination of financial products corresponding to the user's identity information is empty, uses a recommendation algorithm to select a preset number of financial product information from the financial product recommendation database based on the user adaptation information corresponding to the user's identity information, the operation trajectory data, and the characteristics of the financial product information corresponding to the financial product information to establish a financial product recommendation combination; a page generation module, connected to the financial product recommendation module, integrates the optimal recommendation combination of financial products corresponding to the user's identity information into the financial product page corresponding to the financial product information when the optimal recommendation combination of financial products corresponding to the user's identity information is not empty, and integrates the financial product recommendation combination into the financial product page corresponding to the financial product information when the optimal recommendation combination of financial products corresponding to the user's identity information is empty; a feedback collection module monitors and collects the user's behavioral feedback on the financial product recommendation combination or the optimal recommendation combination of financial products on the financial product page based on the user's identity information.A recommendation portfolio optimization module, connected to the feedback collection module and the financial product recommendation database, is used to identify and remove financial product information that needs to be replaced in the recommended financial product portfolio or the optimized financial product portfolio based on the behavioral feedback of all users' identity information. This includes identifying and removing negative correlations between unreplaced financial product information in the recommended financial product portfolio or the optimized financial product portfolio and other financial product information outside the recommended financial product portfolio; calculating a cross-marketing index between unreplaced financial product information in the recommended financial product portfolio or the optimized financial product portfolio and other financial product information outside the recommended financial product portfolio; and selecting the financial product information corresponding to the best cross-marketing index to add to the recommended financial product portfolio to establish the optimized financial product portfolio, or selecting the financial product information corresponding to the best cross-marketing index to add to the optimized financial product portfolio to update the optimized financial product portfolio.
2. The intelligent financial product cross-recommendation and optimization system as described in claim 1, wherein the feedback collection module monitors and collects the user's identity information and provides behavioral feedback on the financial product recommendation combination on the financial product page, including the click-through rate of the financial product information in the financial product recommendation combination, the conversion rate of the financial product information in the financial product recommendation combination, and the time or behavior of subsequently staying on the financial product information in the financial product recommendation combination.
3. The intelligent financial product cross-recommendation and optimization system as described in claim 1, wherein the recommendation combination optimization module identifies and removes the financial product information that needs to be replaced in the financial product recommendation combination or the financial product optimized recommendation combination by calculating the performance index of each individual financial product information in the financial product recommendation combination, and forming a weighted performance score by combining the performance indexes. When the weighted performance score of the financial product information in the financial product recommendation combination or the financial product optimized recommendation combination is lower than a preset threshold, the financial product information in the financial product recommendation combination or the financial product optimized recommendation combination is the financial product information that needs to be replaced.
4. The intelligent financial product cross-recommendation and optimization system as described in claim 1, wherein the recommendation combination optimization module confirms and removes the financial product information that needs to be replaced in the financial product recommendation combination or the financial product optimized recommendation combination when the application conversion rate of the financial product information in the financial product recommendation combination or the financial product optimized recommendation combination is significantly lower than the average of the same group, then the financial product information in the financial product recommendation combination or the financial product optimized recommendation combination is the financial product information that needs to be replaced.