Teaching resource recommendation method and system for financial science and technology practical training platform

Abstract

The invention provides a teaching resource recommendation method and system for a financial science and technology practical training platform, and relates to the technical field of financial teaching resource management. The method comprises the following steps: obtaining training record data of a user about a plurality of financial practical training items and group financial strategy data corresponding to the financial practical training items; extracting use characteristics of the user corresponding to different financial practical training items, calculating contribution parameters about different financial signals in the financial practical training items and dynamic preference parameters and strategy matching parameters of each financial signal under the signal use characteristics, and generating a signal characteristic resource recommendation strategy of the user; constructing a preference evolution sequence of the user about each financial signal, performing mode evolution detection on the user to determine a current mode evolution path of the user, and generating a mode evolution resource recommendation strategy of the user; and determining a target teaching resource recommendation strategy of the user and recommending teaching resources to the user. According to the invention, personalized teaching resource recommendation to the user is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of financial teaching resource management, and in particular to a teaching resource recommendation method and system for a financial technology training platform. Background Art

[0002] As a crucial tool for cultivating financial trading talent, FinTech training platforms offer a variety of trading strategy simulations and practical training environments. These platforms enable users to formulate strategies using various trading factors in simulated markets and verify their effectiveness through backtesting analysis. These platforms are widely used in financial education, vocational training, and other fields. To enhance learning outcomes, training platforms often provide personalized educational resource recommendations based on users' trading performance, helping them supplement their financial knowledge and adapt to trading needs in diverse market environments.

[0003] Because different training programs rely on different trading factors, for example, certain financial signals may be crucial in trend trading strategies but less influential in arbitrage trading or market microstructure analysis. Therefore, fully considering the degree of compatibility between financial signals across different training programs is crucial for accurately assessing users' use of different signals, thereby avoiding distorting the accuracy of recommended content. Furthermore, users' trading patterns are not static but may evolve with accumulated trading experience, changes in the market environment, and progress in learning. For example, a user may initially rely on momentum factors for trend trading but gradually shift to high-frequency trading strategies as their understanding of market microstructure deepens. In such a dynamic process, recommendation strategies generated solely through static mining of trading data may fail to accurately capture changes in user behavior, resulting in recommendations that are out of sync with actual user needs. Summary of the Invention

[0004] To solve the above technical problems, the present invention proposes a teaching resource recommendation method and system for a financial technology training platform. By analyzing the static trading performance of users in different training projects and considering the dynamic evolution of users' trading patterns over time, the teaching resource recommendation strategy for different users is comprehensively determined. Based on personalized strategies that meet the actual learning needs of different users, personalized financial teaching resources are provided to users on different platforms.

[0005] A first aspect of the present invention provides a teaching resource recommendation method for a fintech training platform, comprising:

[0006] Obtaining the user's training record data corresponding to multiple financial training projects, as well as the group financial strategy data corresponding to each financial training project;

[0007] Extract the user's usage characteristics of each financial signal in different financial training projects from the training record data, and calculate the contribution parameters of different financial signals in the financial training projects based on the group financial strategy data;

[0008] Based on the contribution parameters, the user's usage characteristics of multiple financial signals in the financial training project are analyzed, the dynamic preference parameters and strategy matching parameters of each financial signal under the signal usage characteristics are calculated, and the user's signal characteristic resource recommendation strategy is generated based on multiple strategy matching parameters;

[0009] Construct a user's preference evolution sequence for each financial signal, perform pattern evolution detection on the user based on multiple preference evolution sequences, determine the user's current pattern evolution path, and generate a pattern evolution resource recommendation strategy for the user based on the target pattern associated with the pattern evolution path;

[0010] The signal characteristic resource recommendation strategy and the pattern evolution resource recommendation strategy are combined to determine the user's target teaching resource recommendation strategy, and teaching resources are recommended to the user based on the target teaching resource recommendation strategy.

[0011] Preferably, the usage characteristics of each financial signal of the user in different financial training projects are extracted from the training record data, and the contribution parameters of different financial signals in the financial training projects are calculated based on the group financial strategy data, including:

[0012] Determine multiple financial transaction operations of the user in the financial training project based on the training record data, determine multiple decision-making financial signals for each financial transaction operation, perform usage frequency attenuation statistics on the multiple financial transaction operations, and generate usage characteristics of each financial signal of the user in the financial training project;

[0013] Multiple reference individuals are determined based on the group financial strategy data. The group financial strategy data includes the transaction record data of each training member in the financial training project. Based on the transaction record data of multiple reference individuals, transaction coverage improvement analysis and profit contribution analysis are performed on multiple financial signals respectively, and the coverage improvement index and profit contribution index of each financial signal are calculated. The contribution parameters of each financial signal are determined based on the coverage improvement index and profit contribution index.

[0014] Preferably, based on the contribution parameters, a difference analysis is performed on the usage characteristics of multiple financial signals in the financial training project by the user, and the dynamic preference parameters and strategy matching parameters of each financial signal under the signal usage characteristics are calculated, including:

[0015] Based on the transaction record data of multiple reference individuals in the financial training project, the reference usage parameters of each financial signal are determined. The user's usage characteristics are detected for deviation based on the reference usage parameters to obtain the user's behavioral deviation parameters for each financial signal. The preference suppression strategy for each financial signal is determined based on the contribution parameters of multiple financial signals. Based on the preference suppression strategy, the preference suppression optimization is performed on the user's multiple behavioral deviation parameters to calculate the user's dynamic preference parameters and strategy matching parameters for each financial signal.

[0016] Preferably, performing pattern evolution detection on the user according to multiple preference evolution sequences to determine the user's current pattern evolution path includes:

[0017] Perform time period stability tests on multiple preference evolution sequences of users, calculate the global stability parameters of users with respect to multiple financial signals in each time window, and determine the candidate steady-state periods of users based on the multiple global stability parameters;

[0018] Based on the candidate steady-state time period, multiple preference cumulative offset vectors of the user are extracted. According to multiple financial transaction behavior patterns determined based on the platform's historical training record data, the reference initial behavior pattern of the user in the candidate steady-state time period is determined, and signal offset energy analysis is performed on the user's multiple preference cumulative offset vectors. The offset energy convergence parameters of the user for each financial signal in each financial transaction behavior pattern are calculated. According to the offset energy convergence parameters, multiple candidate convergence signals of the user in each financial transaction behavior pattern are determined. According to the multiple candidate convergence signals of the user in each financial transaction behavior pattern, the reference transfer behavior pattern to which the user currently belongs is determined. According to the reference initial behavior pattern and the reference transfer behavior pattern, the user's current pattern evolution path is generated.

[0019] Preferably, determining the reference transfer behavior pattern to which the user currently belongs based on multiple candidate convergence signals of the user in each financial transaction behavior pattern includes:

[0020] Based on multiple candidate convergence signals of the user in each financial transaction behavior mode, the real-time preference vector of the user in each financial transaction behavior mode is determined, and the pattern local preference vector corresponding to the user in different financial transaction behavior modes is generated based on the real-time preference vector. Based on the real-time preference vector and the pattern local preference vector of the user in the financial transaction behavior mode, the user is matched with multiple financial transaction behavior patterns to determine the reference transfer behavior pattern to which the user currently belongs.

[0021] Preferably, a financial signal whose offset energy convergence parameter is greater than a preset offset threshold is recorded as a candidate convergence signal.

[0022] A second aspect of the present invention provides a teaching resource recommendation system for a fintech training platform, wherein the system is used to implement the above-mentioned teaching resource recommendation method for a fintech training platform, including:

[0023] The financial training data collection module is used to obtain the user's training record data corresponding to multiple financial training projects, as well as the group financial strategy data corresponding to each financial training project;

[0024] The contribution analysis module is used to extract the usage characteristics of each financial signal in different financial training projects from the training record data, and calculate the contribution parameters of different financial signals in the financial training projects based on the group financial strategy data;

[0025] The signal usage characteristics analysis module is used to perform differential analysis on the usage characteristics of multiple financial signals in financial training projects based on contribution parameters, calculate the dynamic preference parameters and strategy matching parameters of each financial signal under the signal usage characteristics, and generate a signal characteristic resource recommendation strategy for the user based on multiple strategy matching parameters;

[0026] The pattern evolution analysis module is used to construct a user's preference evolution sequence for each financial signal, perform pattern evolution detection on the user based on multiple preference evolution sequences, determine the user's current pattern evolution path, and generate a pattern evolution resource recommendation strategy for the user based on the target pattern associated with the pattern evolution path;

[0027] The teaching resource recommendation management module is used to determine the user's target teaching resource recommendation strategy by combining the signal characteristic resource recommendation strategy and the pattern evolution resource recommendation strategy, and recommend teaching resources to the user based on the target teaching resource recommendation strategy.

[0028] The present invention has the following beneficial effects:

[0029] The present invention conducts in-depth analysis of users' training record data in multiple financial training projects, fully considers the use requirements of financial signals in different financial scenarios, accurately identifies users' preferred financial signals, combines users' recent trading behaviors in different training projects, and compares them with the financial signal usage characteristics of high-yield groups, identifies users' shortcomings and optimization space in signal usage, and then generates personalized recommendation strategies based on usage characteristics. At the same time, it further explores the evolution trend of users' trading patterns, analyzes the potential changes in their trading styles, determines which trading pattern users may be migrating to, and generates recommendation strategies based on pattern transitions based on the signal usage requirements or biases under the target pattern. Finally, it constructs a personalized teaching resource recommendation plan that meets the actual learning needs of different users, enabling users to more effectively optimize trading strategies and improve the learning effect of financial training. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 The present invention is a flowchart illustrating a method for recommending teaching resources for a financial technology training platform according to an embodiment of the present invention.

[0031] Figure 2 This is a structural diagram of a teaching resource recommendation system for a financial technology training platform according to an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0032] In order to enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0033] The embodiment of the present invention provides a teaching resource recommendation method for a financial technology training platform. Figure 1 , the method includes the following contents.

[0034] Step S1: Obtain user training record data corresponding to multiple financial training projects, as well as group financial strategy data corresponding to each financial training project.

[0035] Specifically, on the FinTech training platform, users will participate in financial training projects offered by multiple platforms. These projects will simulate different market environments and trading scenarios, such as trend trading, arbitrage trading, and market microstructure trading. In order to accurately analyze user trading behavior, we first collect the user's training record data in each financial training project and obtain the group financial strategy data for the project as a reference. Among them, the user training record data includes at least the user's trading strategy usage log, such as buying, selling, and position adjustments, the financial signals used by the user in the trading strategy, the trading returns corresponding to the trading operations, and trading execution characteristics such as trading frequency and average holding time. The trading record data of multiple trainees participating in the project at the same time serves as the project's group financial strategy data, which contains information such as the financial strategies, trading operations, and corresponding profit data actually adopted by different members during the training process.

[0036] Step S2: extracting the user's usage characteristics of each financial signal in different financial training projects from the training record data, and calculating the contribution parameters of different financial signals in the financial training projects based on the group financial strategy data.

[0037] Specifically, by analyzing users' usage of various financial signals in different financial training projects, we extract the usage characteristics of different financial signals in financial transaction operations, and combine them with group financial strategy data to analyze the impact of different financial signals on overall returns. We also explore the strategies of high-return groups in using financial signals to calculate the contribution parameters of financial signals, which are used to measure the actual contribution value of financial signals to returns in this project.

[0038] As an optional implementation plan, the following contents are included for extracting user usage features of different financial signals and calculating contribution parameters of different financial signals in financial training projects:

[0039] For the extraction of usage features, the user's multiple financial transaction operations in the financial training project are determined based on the training record data, and multiple decision-making financial signals for each financial transaction operation are determined. The frequency attenuation statistics of the multiple financial transaction operations are performed to generate the user's usage features for each financial signal in the financial training project.

[0040] Specifically, financial trading operations include common operations such as buying, selling, short selling, and closing positions. Different financial trading operations will use different financial signals as decisions. For example, a buy operation is performed when the momentum factor is greater than 0.02 and the price-earnings ratio is lower than a certain threshold. Then, the user's multiple valid financial trading operations are analyzed, and features such as the frequency of use of different financial signals are extracted as usage features of financial signals. In this process, the effectiveness of multiple financial trading operations can be analyzed first. For example, only operations that actually affect positions, such as buy / sell / position adjustment, are considered valid operations, and there are clear factor annotations in the transaction records, such as which factor signal is used as a decision to execute the operation. At the same time, considering that a single trading operation may be based on multiple financial signals as a decision, in this case, if the user frequently uses a fixed factor combination, a simple statistical calculation of the frequency of use of different financial signals will result in an inflated frequency of each signal. In this embodiment, the dimensional expansion effect under the joint action of multiple signals is considered. When the number of combined signals increases, the marginal contribution of a single signal is controlled to decrease. Specifically, The effective frequency of each financial signal in a single financial transaction represents the information gain brought by multiple signals, with n representing the number of decision-making financial signals used in the transaction. This method calculates the transaction frequency of each financial signal in a single transaction, ultimately yielding the overall frequency parameter for each financial signal in the user's project and revealing the user's usage characteristics for each financial signal in the financial training project.

[0041] For the calculation of contribution parameters, multiple reference individuals are determined based on the group financial strategy data, and transaction coverage improvement analysis and return contribution analysis are performed on multiple financial signals based on the transaction record data of multiple reference individuals. The coverage improvement index and return contribution index of each financial signal are calculated, and the contribution parameters of each financial signal are determined based on the coverage improvement index and return contribution index.

[0042] Specifically, the actual returns of each participant in the project were determined based on group financial strategy data. After ranking these participants based on returns, several participants with returns above a preset percentile were selected as reference individuals. Then, based on the actual trading records of these reference individuals, a coverage lift analysis and return contribution analysis were performed on multiple financial signals. The coverage lift index and return contribution index for each financial signal were calculated, and the contribution parameters of each financial signal were determined based on these indexes.

[0043] In this process, the usage characteristics of each reference individual for each financial signal are first extracted, and the overall usage frequency parameter mean of the financial signal for the training members other than the reference individual in the financial training project, as well as the reference usage frequency parameter mean corresponding to multiple reference individuals, are calculated. The ratio of the reference usage frequency parameter mean to the overall usage frequency parameter mean is used as the coverage improvement index of the financial signal, which is used to describe the usage of different financial signals in high-income groups and the coverage improvement between ordinary users. If the usage frequency of the financial signal in high-income users is significantly higher than that in the ordinary group, that is, the coverage improvement index is larger, it means that the financial signal has strong market adaptability in the current training scenario. Otherwise, it means that the usage of the signal in the high-income group is not prominent and may have limited contribution to the income. The profit data brought by multiple financial signals are analyzed. Specifically, for multiple reference individuals, the profit growth data created by each reference individual in a specific period after different financial transaction operations are counted. Based on the profit growth data, the single transaction contribution data corresponding to one or more financial signals involved are determined. The cumulative profit of each financial signal under multiple financial transaction operations is counted separately to obtain the individual cumulative profit data corresponding to different financial signals for each reference individual. Finally, the individual cumulative profit data of multiple reference individuals on the same financial signal are summarized to obtain the cumulative profit brought by each financial signal in the representative group. The deviation between the cumulative profit of each financial signal and the average profit of multiple financial signals can be used to represent the contribution of the financial signal to the overall profit, and the profit contribution index of each financial signal is obtained. Finally, combined with the coverage improvement index, the contribution of different financial signals to the overall profit is further verified. The profit contribution index of the financial signal is weighted and corrected through the coverage improvement index, and the contribution parameters of different financial signals are calculated to better adapt to the market adaptability and stability of the financial signal.

[0044] Step S3: Based on the contribution parameters, a difference analysis is performed on the usage characteristics of multiple financial signals used by users in the financial training project, the dynamic preference parameters and strategy matching parameters of each financial signal under the signal usage characteristics are calculated, and a signal characteristic resource recommendation strategy for the user is generated based on the multiple strategy matching parameters.

[0045] Specifically, the contribution parameter quantifies the contribution, or importance, of a financial signal to the overall returns of a project. In this context, the contribution parameter is combined with an analysis of user usage patterns for different financial signals. Across different financial scenarios, high-contribution financial signals may be used more frequently than standard signals, but this does not accurately represent users' actual preferences for these signals. Therefore, based on the financial signal contribution parameter, we analyze users' trading signal usage in financial training projects and calculate two key parameters: the dynamic preference parameter and the strategy matching parameter. The dynamic preference parameter represents users' actual preferences for financial signals across different training projects. This process takes into account the impact of varying financial signal importance across different training scenarios to accurately capture users' true preferences for different financial signals. The strategy matching parameter measures how closely a user's use of financial signals matches that of a high-return group. This parameter indicates which financial signals require further training to achieve higher returns in the training project.

[0046] As an optional implementation scheme, the calculation of the dynamic preference parameters and strategy matching parameters of each financial signal under the signal usage feature specifically includes:

[0047] The reference usage parameters of each financial signal are determined based on the transaction record data of multiple reference individuals in the financial training project. The deviation of the user's usage characteristics is detected based on the reference usage parameters to obtain the user's behavioral deviation parameters for each financial signal.

[0048] Specifically, the reference usage parameters for financial signals are the usage characteristics corresponding to reference individuals, representing the actual usage of financial signals by different members of the high-yield group. These are then compared with the usage characteristics of individual users to perform deviation detection, analyze the differences, and calculate the deviation parameters for each user's behavior with respect to each financial signal. For example, if the reference usage parameter for a financial signal, such as the price-to-earnings ratio, for the high-yield group is 31% in terms of overall usage, while users only use that signal at a rate of 16%, the ratio of this 15% difference to the overall usage rate of 31% can effectively represent the difference between the user's usage of that financial signal and that of the representative group.

[0049] Based on the contribution parameters of multiple financial signals, the preference suppression strategy of each financial signal is determined. Based on the preference suppression strategy, the preference suppression optimization of multiple behavioral deviation parameters of the user is performed, and the user's dynamic preference parameters and strategy matching parameters for each financial signal are calculated.

[0050] Specifically, the preference suppression strategy is used to control the difference between the usage characteristics of financial signals and the user's actual preference for the signal. Under the influence of financial scenarios, in order to improve the final training benefits, the user's usage rate of some key financial signals will increase, but it does not mean that the user has a high preference for the financial signal. Therefore, according to the contribution parameters of each financial signal, the user's behavior deviation parameters are optimized for preference suppression. Specifically, the behavior deviation parameters corresponding to financial signals with high contribution parameters need to be suppressed to a certain extent, while the behavior deviation parameters corresponding to financial signals with low contribution parameters need to be appropriately amplified to reduce the behavioral impact brought by the financial scenario. The dynamic preference parameters of the financial signal calculated in this way can better remove the individual behavioral impact brought by the training scenario and more accurately represent the user's true preference for financial signals. At the same time, the dynamic preference parameters are used to reversely analyze the differences in users' usage of different financial signals. The user's behavioral deviation parameters and dynamic preference parameters are combined to calculate the user's strategy matching parameters for different financial signals. The behavioral deviation parameters quantify the behavioral differences between the user's financial strategy and the representative group. Signals with large differences may require supplementary relevant usage knowledge, while the dynamic preference parameters represent the user's actual usage preference for financial signals. Therefore, after determining the user's control over the use of different financial signals, we can further combine the user's real preferences and use the strategy matching parameters to characterize the overall matching degree between the usage teaching resources corresponding to different financial signals and the user. The signal characteristic resource recommendation strategy determined by the strategy matching parameters can provide users with the necessary teaching resources to improve their financial knowledge, while meeting the user's current preferences, so that users can supplement their knowledge gaps with personalized learning based on their own preferences.

[0051] Step S4: Construct a user preference evolution sequence for each financial signal, perform pattern evolution detection on the user based on multiple preference evolution sequences, determine the user's current pattern evolution path, and generate a pattern evolution resource recommendation strategy for the user based on the target pattern associated with the pattern evolution path.

[0052] Specifically, the preference evolution sequence is constructed based on the aforementioned dynamic preference parameters and time-series relationships, which can characterize how a user's preferences for financial signals change over time. Considering that a user's trading patterns will change with factors such as market experience and strategy optimization, multiple preference evolution sequences are established for the user and time-series evolution analysis is performed to determine whether the user's trading style is changing. Upon detecting a shift in the user's trading pattern, the user's current pattern evolution path is extracted, indicating the current evolution of the user's trading pattern, including historical trading patterns and the new trading pattern currently being transformed. Based on the actual situation of different financial signals in the new trading pattern and the information currently available to the user, a resource recommendation strategy for the user's pattern evolution is determined, recommending teaching resources corresponding to the knowledge that may need to be supplemented under the new trading model.

[0053] As an optional implementation scheme, a user's pattern evolution is detected based on multiple preference evolution sequences to determine the user's current pattern evolution path, specifically including:

[0054] A time period stability test is performed on multiple preference evolution sequences of the user, the global stability parameters of the user with respect to multiple financial signals in each time window are calculated, and the candidate steady-state time period of the user is determined based on the multiple global stability parameters.

[0055] Specifically, as trading experience accumulates, the market environment changes, or learning content is adjusted, users' trading patterns may shift from trend trading strategies to arbitrage trading strategies, or from fundamental investment strategies to quantitative trading strategies. The purpose of period stability testing is to identify periods in a user's history when preferences are relatively stable, in order to better determine one or more trading patterns corresponding to the user's historical behavior. In this process, multiple preference evolution sequences are analyzed using a sliding window detection method to extract the fluctuation characteristics of the user's preference for financial signals within different windows, such as the standard deviation of each financial signal. Based on the fluctuation characteristics of multiple financial signals, an overall quantitative parameter, namely a global stability parameter, is calculated to measure the fluctuation of preferences for multiple financial signals within the time window. For example, the global stability parameter is obtained by directly summing up multiple fluctuation characteristics. Based on the global stability parameter, one or more candidate steady-state periods are determined for the user, taking into account the fact that the user's historical performance may involve multiple trading pattern transitions.

[0056] Based on the candidate steady-state time period, multiple preference cumulative offset vectors of users are extracted. According to multiple financial transaction behavior patterns determined based on the platform's historical training record data, signal offset energy analysis is performed on the multiple preference cumulative offset vectors of users, and the offset energy convergence parameters of the user's financial signal under each financial transaction behavior pattern are calculated.

[0057] Specifically, considering that a user may be involved in multiple candidate steady-state periods, the user's transaction pattern in the first candidate steady-state period with a temporal position close to the current one is used as the starting point of the user's pattern evolution path. For multiple preference evolution sequences, the corresponding local transition sequence is extracted from each preference evolution sequence based on the candidate steady-state period as the starting point. The offset starting point of each financial signal is determined based on the candidate steady-state period as the starting point, and a preference cumulative offset vector is generated for each local transition sequence. Each element in the preference cumulative offset vector is the deviation value between the dynamic preference parameter at the corresponding time step and the corresponding offset starting point.

[0058] Among them, for multiple financial transaction behavior patterns determined based on the platform's historical training record data, the historical training record data includes historical training data of multiple members in the financial technology training platform. Data from the period when the members' preferences are stable can be extracted, and the signal preference vector of each member regarding the degree of preference for different financial signals can be extracted. Then, cluster analysis is performed on multiple users to identify multiple financial transaction behavior patterns. For example, the DBSCN clustering algorithm is used to perform pattern recognition on multiple members in the platform, obtain multiple financial transaction behavior patterns, and extract the group preference vector of each pattern.

[0059] Based on the group preference vector of the financial transaction behavior pattern, a signal offset energy analysis is performed on the user's multiple preference cumulative offset vectors. In this process, the user's initial preference vector in the candidate steady-state period as the starting point is first determined. Specifically, it is composed of dynamic preference parameters corresponding to each financial signal in the candidate steady-state period. Then, it is matched with multiple financial transaction behavior patterns to determine the financial transaction behavior pattern to which the user belongs in the candidate steady-state period as the starting point, and it is recorded as the user's reference initial behavior pattern. The total offset energy of the financial signal in the preference cumulative offset vector is further calculated, that is, the overall difference between the initial offset value and the corresponding mean of multiple deviation values ​​in the preference cumulative offset vector, and the baseline offset energy of the total offset energy of the financial signal is determined. Specifically, it is calculated according to the group preference vectors corresponding to the user's initial financial transaction behavior pattern and the current financial transaction behavior pattern, that is, the difference between the element values ​​of the financial signal in the two vectors. The ratio of the total offset energy to the baseline offset energy is used as the offset energy convergence parameter corresponding to the financial signal, and the offset direction of the offset energy convergence parameter is determined according to the element value corresponding to the financial signal in the group preference vector. The value close to the element value corresponding to the financial signal in the group preference vector is recorded as a positive value, otherwise it is recorded as a negative value, thereby obtaining the offset energy convergence parameter corresponding to each financial signal of the user under any financial transaction behavior pattern.

[0060] Based on the offset energy convergence parameter, multiple candidate convergence signals of the user in each financial transaction behavior mode are determined. The aforementioned offset energy convergence parameter quantifies the offset amplitude and offset direction of the preference degree. The candidate convergence signals are specifically multiple financial signals with positive offsets and offset amplitudes greater than a preset offset threshold. Based on the multiple candidate convergence signals of the user in each financial transaction behavior mode, the reference transfer behavior mode to which the user currently belongs is determined. Based on the reference initial behavior mode and the reference transfer behavior mode, the user's current mode evolution path is generated. The reference transfer behavior mode is recorded as the target mode associated with the mode evolution path, which represents the user's current mode conversion target.

[0061] Specifically, a real-time preference vector for each user's financial transaction behavior pattern is generated from multiple preference evolution sequences. This vector contains only the real-time preference values ​​corresponding to multiple candidate convergence signals in the financial transaction behavior pattern, i.e., the dynamic preference parameters corresponding to the current moment in the preference evolution sequence. Similarly, a local preference vector corresponding to the real-time preference vector is extracted from the group preference vector of the financial transaction behavior pattern. Then, based on the real-time preference vector and the local preference vector, the user is matched with the financial transaction behavior pattern. Ultimately, the financial transaction behavior pattern with the highest matching degree is selected as the user's current reference transition behavior pattern, thereby obtaining the user's pattern evolution path, which includes the reference initial behavior pattern and the reference transition behavior pattern. This process takes into account the uncertainty inherent in the dynamic evolution of the user's pattern. Specifically, based on the user's changes under different financial signals, multiple key financial signals are identified, and targeted matches are achieved with multiple financial transaction behavior patterns. This results in the user's current pattern evolution path, allowing for precise analysis of the user's conversion goals based on the overall degree of deviation from different signals.

[0062] Step S5: Determine the user's target teaching resource recommendation strategy by combining the signal characteristic resource recommendation strategy and the pattern evolution resource recommendation strategy, and recommend teaching resources to the user based on the target teaching resource recommendation strategy.

[0063] Specifically, the signal-characteristic resource recommendation strategy can be generated based on a user's recent performance in multiple financial training projects, analyzing the user's overall deviation from different financial signals. For example, if the overall performance of a financial signal in a recent financial training project is significantly different from that of a representative group, the corresponding teaching resource can be prioritized. The pattern-evolution resource recommendation strategy, on the other hand, can analyze which financial signals the user needs to further improve their mastery of if they want to transition to the reference transfer behavior model based on the group preference vector of the user's current reference transfer behavior model, thereby determining which teaching resources should be recommended first. For the integration of signal characteristic resource recommendation strategy and pattern evolution resource recommendation strategy, the corresponding thresholds can be reasonably set based on the fluctuations of the user's recent dynamic preference parameters for different financial signals. If the user's dynamic preference parameters for multiple financial signals have fluctuated greatly recently, it means that the user currently has a large demand for pattern change. The weight corresponding to the pattern evolution resource recommendation strategy can be increased. In a relatively stable situation, the signal characteristic resource recommendation strategy can be dominant, and educational resources can be recommended based on the user's recent lack of experience in using financial signals in different financial training projects. In this way, the target teaching resource recommendation strategy corresponding to the user is generated, and then personalized financial learning-related teaching resources are recommended to the user based on the teaching resources related to different financial signals.

[0064] The present invention also provides a teaching resource recommendation system for a financial technology training platform. Figure 2 , the system comprises:

[0065] The financial training data collection module is used to obtain the user's training record data corresponding to multiple financial training projects, as well as the group financial strategy data corresponding to each financial training project;

[0066] The contribution analysis module is used to extract the usage characteristics of each financial signal in different financial training projects from the training record data, and calculate the contribution parameters of different financial signals in the financial training projects based on the group financial strategy data;

[0067] The signal usage characteristics analysis module is used to perform differential analysis on the usage characteristics of multiple financial signals in financial training projects based on contribution parameters, calculate the dynamic preference parameters and strategy matching parameters of each financial signal under the signal usage characteristics, and generate a signal characteristic resource recommendation strategy for the user based on multiple strategy matching parameters;

[0068] The pattern evolution analysis module is used to construct a user's preference evolution sequence for each financial signal, perform pattern evolution detection on the user based on multiple preference evolution sequences, determine the user's current pattern evolution path, and generate a pattern evolution resource recommendation strategy for the user based on the target pattern associated with the pattern evolution path;

[0069] The teaching resource recommendation management module is used to determine the user's target teaching resource recommendation strategy by combining the signal characteristic resource recommendation strategy and the pattern evolution resource recommendation strategy, and recommend teaching resources to the user based on the target teaching resource recommendation strategy.

[0070] The foregoing description is merely a detailed description of the present invention, which is intended to enable those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art. Portions not described in detail in this specification are well known to those skilled in the art.

Claims

1. A teaching resource recommendation method for a financial technology training platform, characterized in that: include: Obtaining the user's training record data corresponding to multiple financial training projects, as well as the group financial strategy data corresponding to each financial training project; Extract the user's usage characteristics of each financial signal in different financial training projects from the training record data, and calculate the contribution parameters of different financial signals in the financial training projects based on the group financial strategy data; Based on the contribution parameters, the user's usage characteristics of multiple financial signals in the financial training project are analyzed, the dynamic preference parameters and strategy matching parameters of each financial signal under the signal usage characteristics are calculated, and the user's signal characteristic resource recommendation strategy is generated based on multiple strategy matching parameters; Construct a user's preference evolution sequence for each financial signal, perform pattern evolution detection on the user based on multiple preference evolution sequences, determine the user's current pattern evolution path, and generate a pattern evolution resource recommendation strategy for the user based on the target pattern associated with the pattern evolution path; The signal characteristic resource recommendation strategy and the pattern evolution resource recommendation strategy are combined to determine the user's target teaching resource recommendation strategy, and teaching resources are recommended to the user based on the target teaching resource recommendation strategy.

2. A teaching resource recommendation method for a financial technology training platform according to claim 1, characterized in that: Extract the user usage characteristics of each financial signal in different financial training projects from the training record data, and calculate the contribution parameters of different financial signals in the financial training projects based on the group financial strategy data, including: Determine multiple financial transaction operations of the user in the financial training project based on the training record data, determine multiple decision-making financial signals for each financial transaction operation, perform usage frequency attenuation statistics on the multiple financial transaction operations, and generate usage characteristics of each financial signal of the user in the financial training project; Multiple reference individuals are determined based on the group financial strategy data. The group financial strategy data includes the transaction record data of each training member in the financial training project. Based on the transaction record data of multiple reference individuals, transaction coverage improvement analysis and profit contribution analysis are performed on multiple financial signals respectively, and the coverage improvement index and profit contribution index of each financial signal are calculated. The contribution parameters of each financial signal are determined based on the coverage improvement index and profit contribution index.

3. A teaching resource recommendation method for a financial technology training platform according to claim 2, characterized in that: Based on the contribution parameters, the user's usage characteristics of multiple financial signals in the financial training project are analyzed, and the dynamic preference parameters and strategy matching parameters of each financial signal under the signal usage characteristics are calculated, including: Based on the transaction record data of multiple reference individuals in the financial training project, the reference usage parameters of each financial signal are determined. The user's usage characteristics are detected for deviation based on the reference usage parameters to obtain the user's behavioral deviation parameters for each financial signal. The preference suppression strategy for each financial signal is determined based on the contribution parameters of multiple financial signals. Based on the preference suppression strategy, the preference suppression optimization is performed on the user's multiple behavioral deviation parameters to calculate the user's dynamic preference parameters and strategy matching parameters for each financial signal.

4. A teaching resource recommendation method for a financial technology training platform according to claim 1, characterized in that: Detect user pattern evolution based on multiple preference evolution sequences to determine the user's current pattern evolution path, including: Perform time period stability tests on multiple preference evolution sequences of users, calculate the global stability parameters of users with respect to multiple financial signals in each time window, and determine the candidate steady-state periods of users based on the multiple global stability parameters; Based on the candidate steady-state time period, multiple preference cumulative offset vectors of the user are extracted. According to multiple financial transaction behavior patterns determined based on the platform's historical training record data, the reference initial behavior pattern of the user in the candidate steady-state time period is determined, and signal offset energy analysis is performed on the user's multiple preference cumulative offset vectors. The offset energy convergence parameters of the user for each financial signal in each financial transaction behavior pattern are calculated. According to the offset energy convergence parameters, multiple candidate convergence signals of the user in each financial transaction behavior pattern are determined. According to the multiple candidate convergence signals of the user in each financial transaction behavior pattern, the reference transfer behavior pattern to which the user currently belongs is determined. According to the reference initial behavior pattern and the reference transfer behavior pattern, the user's current pattern evolution path is generated.

5. A teaching resource recommendation method for a financial technology training platform according to claim 4, characterized in that: Based on multiple candidate convergence signals of the user in each financial transaction behavior mode, the reference transfer behavior mode to which the user currently belongs is determined, including: Based on multiple candidate convergence signals of the user in each financial transaction behavior mode, the real-time preference vector of the user in each financial transaction behavior mode is determined, and the pattern local preference vector corresponding to the user in different financial transaction behavior modes is generated based on the real-time preference vector. Based on the real-time preference vector and the pattern local preference vector of the user in the financial transaction behavior mode, the user is matched with multiple financial transaction behavior patterns to determine the reference transfer behavior pattern to which the user currently belongs.

6. A teaching resource recommendation method for a financial technology training platform according to claim 4, characterized in that: Financial signals whose offset energy convergence parameters are greater than a preset offset threshold are recorded as candidate convergence signals.

7. A teaching resource recommendation system for a financial technology training platform, characterized in that: The system is used to implement the teaching resource recommendation method for a fintech training platform as described in any one of claims 1 to 6, comprising: The financial training data collection module is used to obtain the user's training record data corresponding to multiple financial training projects, as well as the group financial strategy data corresponding to each financial training project; The contribution analysis module is used to extract the usage characteristics of each financial signal in different financial training projects from the training record data, and calculate the contribution parameters of different financial signals in the financial training projects based on the group financial strategy data; The signal usage characteristics analysis module is used to perform differential analysis on the usage characteristics of multiple financial signals in financial training projects based on contribution parameters, calculate the dynamic preference parameters and strategy matching parameters of each financial signal under the signal usage characteristics, and generate a signal characteristic resource recommendation strategy for the user based on multiple strategy matching parameters; The pattern evolution analysis module is used to construct a user's preference evolution sequence for each financial signal, perform pattern evolution detection on the user based on multiple preference evolution sequences, determine the user's current pattern evolution path, and generate a pattern evolution resource recommendation strategy for the user based on the target pattern associated with the pattern evolution path; The teaching resource recommendation management module is used to determine the user's target teaching resource recommendation strategy by combining the signal characteristic resource recommendation strategy and the pattern evolution resource recommendation strategy, and recommend teaching resources to the user based on the target teaching resource recommendation strategy.

Images