Method and device for determining user demand information and electronic equipment

Abstract

The invention discloses a method and a device for determining user demand information and electronic equipment. The method comprises the following steps: acquiring historical transaction information of a target user, and determining transaction feature data of the target user according to the historical transaction information; inputting the transaction feature data into a user classification model to obtain a target classification result of the target user, the target classification result being used for representing that the target user belongs to a target cluster in the user classification model; and obtaining public transaction feature information of each sample user in the target cluster, and determining demand information of the target user according to the public transaction feature information. Through the method and the device, the problem of relatively low accuracy of determining the user demand information in related technologies is solved.

Description

Technical Field

[0001] This application relates to the field of financial technology, and more specifically, to a method, apparatus, and electronic device for determining user demand information. Background Technology

[0002] In order to serve customers more accurately, financial institutions need to determine the needs of users and then provide them with corresponding services in a timely manner, such as personalized product promotion and service reminders, thereby enhancing customer satisfaction.

[0003] Currently, when determining user needs based on transaction information, neural network models such as logistic regression and decision trees are commonly used. However, these models typically rely on supervised learning, requiring not only massive amounts of accurately labeled data for training but also static threshold settings for customer type classification. This approach proves inadequate when dealing with the complex and ever-changing nature of customer transaction behavior, lacking sufficient dynamic adaptability, and consequently significantly reducing the accuracy of determining user needs based on transaction information.

[0004] There is currently no effective solution to the problem of low accuracy in determining user demand information in related technologies. Summary of the Invention

[0005] The main objective of this application is to provide a method, apparatus, and electronic device for determining user demand information, so as to solve the problem of low accuracy in determining user demand information in related technologies.

[0006] To achieve the above objectives, according to one aspect of this application, a method for determining user demand information is provided. The method includes: acquiring historical transaction information of a target user and determining transaction characteristic data of the target user based on the historical transaction information; inputting the transaction characteristic data into a user classification model to obtain a target classification result for the target user, wherein the target classification result is used to characterize that the target user belongs to a target cluster in the user classification model; acquiring common transaction characteristic information of each sample user in the target cluster, and determining the target user's demand information based on the common transaction characteristic information.

[0007] Optionally, the user classification model is trained as follows: N initial values ​​are set, and each initial value is used as a K value to construct N initial classification models; each initial classification model is used to classify the sample transaction information contained in the first sample data, resulting in N predicted classification results, where the first sample data includes multiple sample transaction information, each belonging to a sample user; the sum of squared errors for each predicted classification result is calculated, resulting in N sums of squared errors; the sums of squared errors of each initial classification model are arranged in ascending order of K value to obtain a first sequence; inflection points in the first sequence are determined, and the K value in the initial classification model corresponding to the sum of squared errors indicated by the inflection point is obtained, resulting in a target K value; a classification model is constructed based on the target K value to obtain a candidate classification model, and the candidate classification model is trained using the first sample data to obtain the user classification model.

[0008] Optionally, the method further includes: if the user classification model is used for a period exceeding a preset window duration, obtaining second sample data from the newly added transaction information within the preset window duration, wherein the second sample data includes multiple samples of newly added transaction information; obtaining the updated weight vector of each transaction feature at the current moment, and obtaining the historical weight vector in the user classification model; determining the target weight vector based on the updated weight vector and the historical weight vector; updating the user classification model using the target weight vector and the second sample data to obtain the target classification model; calculating the silhouette coefficient of the target classification model, and determining whether the silhouette coefficient meets the preset coefficient requirements; if the silhouette coefficient meets the preset coefficient requirements, replacing the user classification model with the target classification model; if the silhouette coefficient does not meet the preset coefficient requirements, retaining the user classification model and deleting the target classification model.

[0009] Optionally, determining the target weight vector based on the updated weight vector and the historical weight vector includes: obtaining a preset decay coefficient and using the preset decay coefficient to determine the weights of the updated weight vector and the historical weight vector; and performing a weighted summation calculation on the updated weight vector and the historical weight vector based on the weights to obtain the target weight vector.

[0010] Optionally, calculating the silhouette coefficient of the target classification model includes: for any newly added transaction information of a target sample, calculating the average distance between the newly added transaction information of the target sample and the newly added transaction information of other samples in the target cluster to obtain a first distance value, wherein the newly added transaction information of the target sample is located in the target cluster; calculating the average distance between the newly added transaction information of the target sample and the newly added transaction information of each sample in the neighboring cluster to obtain a second distance value, wherein the neighboring cluster is the cluster closest to the target cluster; calculating the sub-profile coefficient of the newly added transaction information of the target sample based on the first distance value and the second distance value; obtaining the sub-profile coefficient of the newly added transaction information of each sample in the second sample data, and calculating the average value of the sub-profile coefficients to obtain the silhouette coefficient of the target classification model.

[0011] Optionally, determining the target user's demand information based on public transaction characteristic information includes: determining the target user's user profile based on public transaction characteristic information; determining the matching degree between the target user and various business types based on the user profile, obtaining multiple matching degrees; and determining the business product demand under the business type with the highest matching degree as the target user's demand information.

[0012] Optionally, determining the target user's transaction characteristic data based on historical transaction information includes: acquiring preset transaction indicators, wherein the preset transaction indicators include at least one of the following: time-series characteristic transaction indicators, scenario-specific transaction indicators, and amount-specific transaction indicators; determining the values ​​of each indicator under the preset transaction indicators based on historical transaction information to obtain multiple indicator values; and encoding and standardizing the multiple indicator values ​​to obtain transaction characteristic data.

[0013] To achieve the above objectives, according to another aspect of this application, an apparatus for determining user demand information is provided. The apparatus includes: a first acquisition unit, configured to acquire historical transaction information of a target user and determine transaction characteristic data of the target user based on the historical transaction information; a first classification unit, configured to input the transaction characteristic data into a user classification model to obtain a target classification result for the target user, wherein the target classification result is used to characterize that the target user belongs to a target cluster in the user classification model; and a first determination unit, configured to acquire common transaction characteristic information of each sample user in the target cluster and determine the target user's demand information based on the common transaction characteristic information.

[0014] To achieve the above objectives, according to another aspect of this application, an electronic device is provided, comprising a memory storing an executable program; and a processor for running the program, wherein the program executes the aforementioned method for determining user requirement information during runtime.

[0015] To achieve the above objectives, according to another aspect of this application, a computer program product is provided, including computer instructions, which, when executed by a processor, implement the steps of the method for determining user requirement information described above.

[0016] In this embodiment, the method involves acquiring historical transaction information of the target user and determining the target user's transaction feature data based on the historical transaction information; inputting the transaction feature data into a user classification model to obtain the target user's target classification result, where the target classification result is used to characterize that the target user belongs to the target cluster in the user classification model; acquiring the common transaction feature information of each sample user in the target cluster and determining the target user's demand information based on the common transaction feature information. By using an unsupervised classification model to determine that the target user belongs to the target cluster in the user classification model, and then determining the target user's demand information based on the common transaction feature information of each sample user in the target cluster, the method achieves the goal of accurately determining the target user's demand information, thereby improving the technical effect of determining the target user's demand information and solving the technical problem of low accuracy in determining user demand information in related technologies. Attached Figure Description

[0017] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0018] Figure 1 A hardware structure block diagram of a computer terminal for implementing a method for determining user demand information is shown.

[0019] Figure 2 This is a flowchart of a method for determining user demand information according to Embodiment 1 of this application;

[0020] Figure 3 This is a flowchart of the feature extraction operation provided in Embodiment 1 of this application;

[0021] Figure 4 This is a schematic diagram of a device for determining user demand information according to Embodiment 2 of this application;

[0022] Figure 5 This is a structural block diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0023] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0024] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0025] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0026] It should be noted that the methods, devices, and electronic equipment for determining user demand information as defined in this disclosure can be used in the fintech field, or in any field other than fintech. The application fields of the methods, devices, and electronic equipment for determining user demand information as defined in this disclosure are not limited.

[0027] It should be noted that all information, user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, and displayed data) used in this application are information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of related data all comply with the relevant laws, regulations, and standards of the relevant regions, have taken necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entry points for users to choose to authorize or refuse use. If the user chooses to refuse, the process will proceed to the expert decision-making process. For example, this system has an interface with relevant users or organizations. Before obtaining relevant information, a request to obtain the information needs to be sent to the aforementioned user or organization through the interface. After receiving consent from the aforementioned user or organization, the relevant information is obtained. Users can view the purpose of data use in real time through the authorization interface and have the right to withdraw authorization or delete data at any time. After the authorization is withdrawn, the system will terminate the relevant data processing within 24 hours.

[0028] The embodiments or examples disclosed herein are not exhaustive, but merely illustrative of some embodiments or examples, and are not intended to limit the scope of protection of this disclosure. Unless otherwise specified, each step in a particular embodiment or example can be implemented as an independent embodiment, and the steps can be arbitrarily combined. For example, a solution after removing some steps in a particular embodiment or example can also be implemented as an independent embodiment, and the order of the steps in a particular embodiment or example can be arbitrarily interchanged. Furthermore, optional methods or examples in a particular embodiment or example can be arbitrarily combined; moreover, embodiments or examples can be arbitrarily combined. For example, some or all steps of different embodiments or examples can be arbitrarily combined, and a particular embodiment or example can be arbitrarily combined with optional methods or examples of other embodiments or examples.

[0029] For ease of description, the following explains some of the nouns or terms used in the embodiments of this application:

[0030] SSE: Sum of Squared Errors, is a metric for evaluating the performance of clustering algorithms.

[0031] K-means clustering algorithm: K-means clustering algorithm is an unsupervised machine learning algorithm, mainly used for clustering analysis of data.

[0032] Example 1

[0033] According to an embodiment of this application, an embodiment of a method for determining user demand information is also provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0034] The method embodiment provided in Embodiment 1 of this application can be executed on a mobile terminal, computer terminal, or similar computing device. Figure 1 A hardware block diagram of a computer terminal for determining user requirement information is shown. Figure 1As shown, the computer terminal 10 (or mobile device) may include one or more processors 102 (shown as 102a, 102b, ..., 102n in the figure) 102 (processor 102 may include, but is not limited to, processing devices such as microprocessors or programmable logic devices), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface, a universal serial bus port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, computer terminal 10 may also include... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0035] It should be noted that the aforementioned one or more processors 102 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 10 (or mobile device). As involved in the embodiments of this application, the data processing circuits serve as a processor control mechanism (e.g., selection of a variable resistor termination path connected to an interface).

[0036] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the user requirement information determination method in this embodiment. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby realizing the aforementioned user requirement information determination method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0037] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 10. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0038] The display may be, for example, a touchscreen LCD display that allows the user to interact with the user interface of the computer terminal 10 (or mobile device).

[0039] Under the aforementioned operating environment, this application provides the following: Figure 2 The method for determining user requirements information is shown. Figure 2 This is a flowchart of a method for determining user demand information according to Embodiment 1 of this application, such as... Figure 2 As shown, the method includes:

[0040] Step S201: Obtain the target user's historical transaction information and determine the target user's transaction characteristic data based on the historical transaction information.

[0041] It should be noted that the executing entity in this embodiment can be a user demand information determination system. This system can predict the user's demand information based on the user's historical transaction information, and thus provide the user with corresponding services based on the demand information.

[0042] It should be noted that historical transaction information refers to all transaction records made by the target user within a certain time frame in the past. Transaction feature data is quantitative data extracted from historical transaction information that can describe the user's transaction behavior patterns. Through data preprocessing and feature engineering, the transaction data is converted into feature vectors for further analysis and model training.

[0043] For example, when determining the required information, the transaction feature data that requires target user information can be collected. This can be done by obtaining all credit card transaction records of the target user within a certain time period, and cleaning and preprocessing the raw transaction data through feature engineering to remove outliers and irrelevant information, and standardizing all numerical features to facilitate subsequent model processing.

[0044] Furthermore, key transaction features, such as consumption frequency, consumption amount distribution, consumption location distribution, and consumption category preference, are extracted from the preprocessed transaction data to construct a transaction feature dataset for the target users. Through data preprocessing and feature construction, the raw transaction data is transformed into structured and analyzable transaction feature data, providing high-quality input feature data for subsequent user classification models and improving the effectiveness of model training and the accuracy of prediction.

[0045] Step S202: Input the transaction feature data into the user classification model to obtain the target classification result of the target user, wherein the target classification result is used to characterize that the target user belongs to the target cluster in the user classification model.

[0046] It should be noted that a user classification model refers to a predictive model built based on unsupervised learning methods (such as the K-means clustering algorithm) to identify and group users with similar transaction characteristics. The target classification result is also the output of the user classification model.

[0047] For example, after obtaining the transaction characteristic data of the target users, this data needs to be input into a pre-trained user classification model. This model can be an improved K-means clustering algorithm, capable of identifying and forming user clusters with similar transaction characteristics. Based on the calculated weighted Euclidean distance, combined with a dynamic update mechanism and anomaly cluster handling strategy, the model automatically adjusts cluster centers and divides clusters, ultimately outputting the specific cluster to which the target user belongs, i.e., the target classification result. Through cluster analysis, the model can automatically identify the target user's affiliation in consumption behavior, providing a foundation for determining subsequent demand information.

[0048] Step S203: Obtain the public transaction feature information of each sample user in the target cluster, and determine the demand information of the target user based on the public transaction feature information.

[0049] It should be noted that public transaction characteristic information refers to the commonalities in transaction characteristic data among all users within the target cluster. Statistical analysis of these common characteristics can determine the demand information of users in this cluster. Demand information is a comprehensive description of the potential needs of target users, derived from the public transaction characteristic information of the target cluster through data mining and analysis. This can include consumption preferences, spending trends, and potential financial service needs.

[0050] For example, after obtaining the model output, the common transaction characteristics of all users within the cluster can be further analyzed based on the classification results of the cluster to which the target user belongs in the output. These characteristics include, but are not limited to, average spending amount, spending frequency, and preferences for specific spending categories. Through statistical analysis, typical consumption behaviors and potential demand patterns of users in this cluster can be identified. Then, using this common transaction characteristics and comprehensively considering the individual transaction characteristics of the target user, the specific demand information of the target user can be inferred.

[0051] For example, if the target cluster shows that users prefer online small-amount, high-frequency consumption, it can be determined that the target users have a high demand for digital payment, points redemption, or micro-credit products. Based on the above information, existing product information, and product information processed by the user, the target user's demand information can be generated. Then, by analyzing the public transaction characteristics of the user group through an unsupervised training model, the specific demand information of the target user can be determined, thereby improving the technical effect of determining the demand information of the target user.

[0052] The method for determining user demand information provided in this application involves acquiring historical transaction information of a target user and determining the target user's transaction feature data based on this information; inputting the transaction feature data into a user classification model to obtain the target user's target classification result, where the target classification result represents that the target user belongs to a target cluster in the user classification model; acquiring common transaction feature information of each sample user in the target cluster and determining the target user's demand information based on this common transaction feature information. By using an unsupervised classification model to determine that the target user belongs to a target cluster in the user classification model, and then determining the target user's demand information based on the common transaction feature information of each sample user in the target cluster, the method achieves the goal of accurately determining the target user's demand information, thereby improving the technical effect of determining the target user's demand information and solving the technical problem of low accuracy in determining user demand information in related technologies.

[0053] To improve the training accuracy of the user classification model, optionally, in the method for determining user demand information provided in this application embodiment, the user classification model is trained in the following manner: N initial values ​​are set, and each initial value is used as a K value to construct N initial classification models; each initial classification model is used to classify the sample transaction information contained in the first sample data to obtain N predicted classification results, wherein the first sample data includes multiple sample transaction information, and each sample transaction information belongs to a sample user; the sum of squared errors of each predicted classification result is calculated to obtain N sums of squared errors; the sums of squared errors of each initial classification model are arranged in ascending order of K value to obtain a first sequence; the inflection point in the first sequence is determined, and the K value in the initial classification model corresponding to the sum of squared errors indicated by the inflection point is obtained to obtain the target K value; a classification model is constructed based on the target K value to obtain a candidate classification model, and the candidate classification model is trained using the first sample data to obtain the user classification model.

[0054] It's important to note that the N initial values ​​refer to a series of different K values ​​set before building the user classification model. These values ​​are used to initialize multiple K-means classification models, with the aim of finding the optimal number of clusters by comparing the performance of different models. The first sample data is the initial dataset used to train and evaluate the user classification model. It contains a large amount of transaction information representing sample users with different consumption patterns, used for cluster analysis during model training. The sum of squared errors (SSE), also known as the intra-cluster sum of squared errors, is an important indicator of clustering effectiveness. It reflects the sum of the squared distances between all samples within a cluster and their cluster centers. The smaller the SSE value, the better the clustering effect. The inflection point is the point on the curve plotted for SSE corresponding to different K values ​​where the rate of decrease in SSE significantly slows down.

[0055] For example, in the initial model building phase, a series of different K values ​​need to be set as the initial number of clusters. For instance, with N=5, K values ​​of 3, 5, 7, 9, and 11 are selected to build five initial classification models. By comparing the performance of models with different K values, the most suitable number of clusters for the current dataset can be found.

[0056] Furthermore, using the N initial classification models constructed above, the transaction information of each sample user in the first sample data is classified, resulting in N different predicted classification results. This stage is to test the performance of different K-value models on real data and evaluate their accuracy and effectiveness in classifying user groups.

[0057] After obtaining the predicted classification results, it is necessary to calculate the sum of squared errors for each predicted classification result. The sum of squared errors is determined by measuring the sum of the squared Euclidean distances from samples within a cluster to their respective cluster centers, and it directly reflects the tightness of the clustering results. By calculating N different values ​​for the sum of squared errors, a basis is provided for the subsequent selection of the K value.

[0058] Furthermore, when selecting the optimal K value, the N sums of squared errors need to be sorted in ascending order of K value to form a first sequence. Each point in the sequence corresponds to a K value and the corresponding sum of squared errors, thereby determining the trend of model performance changes under different K values. By observing the curve of the first sequence, the inflection point where the rate of decrease of the sum of squared errors begins to slow down significantly is found, and the K value at the inflection point is determined as the target K value.

[0059] Finally, after determining the target K value, a candidate classification model can be built based on it. This model is then trained and optimized in detail using the first sample data. The training process may include initializing cluster centers, iteratively calculating the distance from samples within a cluster to the cluster center, and reallocating samples to the nearest cluster based on the distance, until the convergence condition is met or the preset number of iterations is reached. Finally, by continuously optimizing the candidate classification model until it can effectively classify users in the sample data correctly into different consumer groups, the final user classification model is obtained.

[0060] This embodiment determines the target K value by calculating the sum of squared errors, uses the target K value to determine the candidate classification model, and trains the candidate classification model to obtain the user classification model, thereby achieving the technical effect of improving the accuracy of the output results of the user classification model.

[0061] For example, to improve the timeliness and accuracy of the user classification model, it is necessary to update and verify the user classification model periodically. Optionally, in the method for determining user demand information provided in this application embodiment, the method further includes: when the user classification model is used for a longer period than a preset window, obtaining second sample data from the newly added transaction information within the preset window, wherein the second sample data includes multiple samples of newly added transaction information; obtaining the updated weight vector of each transaction feature at the current moment, and obtaining the historical weight vector in the user classification model; determining the target weight vector based on the updated weight vector and the historical weight vector; updating the user classification model using the target weight vector and the second sample data to obtain the target classification model; calculating the silhouette coefficient of the target classification model, and determining whether the silhouette coefficient meets the preset coefficient requirements; if the silhouette coefficient meets the preset coefficient requirements, replacing the user classification model with the target classification model; if the silhouette coefficient does not meet the preset coefficient requirements, retaining the user classification model and deleting the target classification model.

[0062] It should be noted that the preset window duration refers to the model update cycle, i.e., a fixed time interval. The second sample data is a dataset of newly added transaction information collected after the model has been used for the preset window duration, used to evaluate and update the model. The updated weight vector is a set of feature weights calculated based on recent transaction data, reflecting the impact of the latest consumption trends on the model. The historical weight vector is a set of feature weights previously determined in the user classification model, used to record the importance of past transaction features.

[0063] For example, when the user classification model has been running for more than a preset window period, such as 30 days, the system will automatically extract data from new transaction information within those 30 days to construct a second sample dataset. This data covers the latest consumer behavior and trends, and the model can be updated based on the second sample dataset, enabling the model to reflect the latest market changes and user needs in a timely manner, thus improving the timeliness and accuracy of predictions.

[0064] For example, when the model reaches the update time, the updated weight vectors of each transaction feature at the current time are first calculated. Transaction features include, but are not limited to, consumption amount, consumption frequency, consumption location, and consumption time. The weight vectors reflect the relative importance of these features in the new dataset. Simultaneously, the original historical weight vectors from the user classification model are also obtained for comparison and fusion with the updated weight vectors.

[0065] Furthermore, after obtaining the updated weight vector and the historical weight vector, a target weight vector can be obtained through a dynamic weighting mechanism. Specifically, the value of each element of the target weight vector is calculated by the weighted average of the corresponding elements of the updated weight vector and the historical weight vector. For example, a weight ratio of λ=0.95 means that the historical weights account for a large proportion, and the impact of new data is small, which allows the model update operation to transition smoothly and avoids drastic changes in the model caused by sudden data fluctuations.

[0066] To improve the accuracy of the target weight vector, optionally, in the method for determining user demand information provided in this application embodiment, determining the target weight vector based on the updated weight vector and the historical weight vector includes: obtaining a preset attenuation coefficient, and using the preset attenuation coefficient to determine the weights of the updated weight vector and the historical weight vector; and performing a weighted summation calculation on the updated weight vector and the historical weight vector based on the weights to obtain the target weight vector.

[0067] For example, when calculating the target weight vector, the system sets a decay coefficient λ, which reflects the relative importance of historical data and new data in the calculation. Assuming a preset decay coefficient λ = 0.95, this means that when calculating the target weight vector, the historical weight vector contributes 95%, while the updated weight vector contributes 5%. In this way, model updates can gradually adapt to market changes while maintaining historical stability, avoiding model instability caused by drastic fluctuations in new data.

[0068] Furthermore, after obtaining the attenuation coefficient, the updated weight vector and the historical weight vector are weighted and summed using a preset attenuation coefficient λ to calculate the target weight vector. The following formula is an optional formula for calculating the target weight vector.

[0069]

[0070] Where: Wt is the target weight vector, Wt-1 is the historical weight vector, Wnew is the updated weight vector, and λ is the decay coefficient [0, 1], which controls the proportion of historical weights retained.

[0071] For example, assuming the historical weight vector is [0.4, 0.3, 0.2, 0.1], representing the weights of four features: cross-border consumption, large transactions, consumption frequency, and consumption location, and the updated weight vector is [0.45, 0.35, 0.15, 0.05], the target weight vector obtained by calculating using the above formula is [0.4175, 0.3225, 0.1975, 0.0925].

[0072] This embodiment dynamically adjusts the weight vector, enabling the user classification model to balance historical consumption patterns with current consumption trends, thereby improving the model's predictive ability and adaptability.

[0073] For example, after obtaining the target weight vector and the second sample data, these two data points can be used to update the user classification model to construct the target classification model. First, the cluster centers need to be recalculated according to the target weight vector. Then, the second sample data is clustered using the dynamic K-means algorithm. Simultaneously, the silhouette coefficient of the target classification model needs to be calculated to determine the classification performance. Through this process, the target classification model can not only adapt to the latest transaction data but also more accurately reflect changes in the importance of transaction features through weight adjustments, thereby improving the model's prediction accuracy and response speed.

[0074] Furthermore, after the model update is completed, the silhouette coefficient of the target classification model is calculated to validate the target classification model. The silhouette coefficient ranges from -1 to 1, with a value closer to 1 indicating a better clustering effect. Subsequently, the calculated silhouette coefficient is compared with the preset coefficient requirements to determine whether the model's clustering effect has met the expected standard.

[0075] If the silhouette coefficient of the target classification model is higher than the preset requirement, it proves that the clustering effect has improved after the model update. In this case, the target classification model will officially replace the existing user classification model for subsequent consumer behavior prediction and analysis. Conversely, if the silhouette coefficient does not meet the requirement, it means that the model update has not brought the expected results. In this case, the original user classification model will be retained to avoid a decline in model quality. This completes the check and update operation of the user classification model, achieving the technical effect of improving the predictive accuracy of the user classification model.

[0076] This embodiment achieves the technical effect of improving the output accuracy of the user classification model by updating the user classification model.

[0077] To accurately calculate the silhouette coefficient, optionally, in the method for determining user demand information provided in this application embodiment, calculating the silhouette coefficient of the target classification model includes: for any new transaction information of a target sample, calculating the average distance from the new transaction information of the target sample to the new transaction information of other samples in the target cluster to obtain a first distance value, wherein the new transaction information of the target sample is located in the target cluster; calculating the average distance from the new transaction information of the target sample to the new transaction information of each sample in the neighboring cluster to obtain a second distance value, wherein the neighboring cluster is the cluster closest to the target cluster; calculating the sub-profile coefficient of the new transaction information of the target sample based on the first distance value and the second distance value; obtaining the sub-profile coefficient of the new transaction information of each sample in the second sample data, and calculating the average value of the sub-profile coefficients to obtain the silhouette coefficient of the target classification model.

[0078] For example, when determining the silhouette coefficient of the target classification model, it is first necessary to determine the sub-profile coefficients of the newly added transaction information of each sample and calculate the average value of the sub-profile coefficients to obtain the silhouette coefficient of the target classification model. When calculating the sub-profile coefficient of the newly added transaction information of any target sample, it is necessary to obtain the average distance from the newly added transaction information of the target sample to the newly added transaction information of other samples in the target cluster to obtain the first distance value, and calculate the average distance from the newly added transaction information of the target sample to the newly added transaction information of each sample in the neighboring cluster to obtain the second distance value. The silhouette coefficient is then calculated according to the following formula:

[0079]

[0080] Where a(i) is the first distance value, b(i) is the second distance value, and s(i) is the sub-profile coefficient.

[0081] After obtaining the sub-profile coefficients, the profile coefficients of the target classification model can be determined using the following formula:

[0082]

[0083] In the formula, N represents the number of newly added transaction information entries in the second sample data.

[0084] It should be noted that the distance between newly added transaction information of two samples can be calculated using the following distance formula:

[0085]

[0086] Where d(x, μ) is the distance between sample x and sample μ, x = (x1, x2, ..., xn) represents the feature vector of sample x, μ = (μ1, μ2, ..., μn) represents the feature vector of sample μ, and wi is the weight of the i-th feature.

[0087] This embodiment achieves the technical effect of accurately determining the clustering ability of the model by calculating the silhouette coefficient of the target classification model.

[0088] Optionally, in the method for determining user demand information provided in the embodiments of this application, determining the target user's demand information based on public transaction feature information includes: determining the target user's user profile based on public transaction feature information; determining the matching degree between the target user and each business type based on the user profile, obtaining multiple matching degrees; and determining the business product demand under the business type with the highest matching degree as the target user's demand information.

[0089] For example, when determining demand information, the first step is to extract common transaction characteristic information from all sample users within the target cluster. This information is then analyzed in depth to construct user profiles for the target users. This includes quantifying key characteristics such as transaction frequency, transaction amount, consumption preferences, and consumption time. For instance, by statistically analyzing the average consumption amount of sample users within the cluster under different consumption scenarios, the degree of preference of target users for specific scenarios can be determined. These characteristics will be integrated into a comprehensive descriptive framework to form user profiles for the target users, providing a foundation for subsequent demand matching.

[0090] Furthermore, the constructed user profiles are used to calculate the match degree between the target user and all types of services offered by the financial institution. The match degree is calculated based on the similarity between features in the user profile and the features required by different service types. For example, if the user profile shows that the target user prefers large-scale cross-border spending, then the user's match degree with service types such as high-end credit card points redemption services or overseas travel insurance is likely to be high. When calculating the match degree, various algorithms can be used, such as collaborative filtering and content-based recommendation algorithms, combined with historical data and business rules to quantify the target user's demand for different service categories.

[0091] Finally, the business type with the highest matching score is selected from the results, and the specific business product requirements under that type are identified as the target user's demand information. For example, if the matching score analysis shows that the target user has the highest matching score with wealth management services, then further analysis is conducted on their preferences for wealth management products. Combined with other information in the user profile (such as risk tolerance, investment cycle preferences, etc.), the target user's high demand information for a certain type of specific wealth management product is finally determined.

[0092] This embodiment constructs a user profile by analyzing the public transaction feature information of the target cluster, calculates and determines the matching degree between the target user and each business type, and finally determines the specific business product needs of the target user. Based on the product needs, it generates demand information, thus achieving the technical effect of accurately determining the demand information of the target user.

[0093] Optionally, in the method for determining user demand information provided in the embodiments of this application, determining the transaction characteristic data of the target user based on historical transaction information includes: obtaining preset transaction indicators, wherein the preset transaction indicators include at least one of the following: time-series characteristic transaction indicators, scenario characteristic transaction indicators, and amount characteristic transaction indicators; determining the values ​​of each indicator under the preset transaction indicators based on historical transaction information to obtain multiple indicator values; and encoding and standardizing the multiple indicator values ​​to obtain transaction characteristic data.

[0094] For example, when determining the transaction characteristic data of a target user, preset transaction indicators are first obtained, and key consumption characteristics are extracted from the target user's historical transaction information based on these indicators. For instance, time-series transaction indicators may include the user's consumption distribution over a week, i.e., the amount and frequency of consumption each day from Monday to Sunday; scenario-based transaction indicators include the user's consumption preferences in different consumption scenarios; and monetary transaction indicators involve the user's average consumption amount, fluctuations in consumption amount, etc.

[0095] Furthermore, a deeper analysis of the target user's historical transaction information is conducted to extract specific values ​​for each preset transaction indicator. For example, to determine time-series characteristic transaction indicators, it may be necessary to calculate the target user's transaction frequency and average transaction amount in different time periods to understand their cyclical consumption patterns; for scenario-based characteristic transaction indicators, it is necessary to analyze the target user's consumption distribution among various merchants; for monetary characteristic transaction indicators, statistical measures such as the mean and standard deviation of the target user's consumption amount can be calculated to determine their consumption level and risk preference.

[0096] Furthermore, after obtaining the specific indicator values ​​for each preset transaction metric, these values ​​are then encoded and standardized. Encoding involves converting categorical features (such as consumption scenarios) into numerical forms. A common method is one-hot encoding, which decomposes each categorical feature into multiple binary features, each corresponding to a scenario category, thus representing the user's consumption preferences in each scenario. Standardization involves scaling all transaction metrics to a uniform scale. A common method is Z-score standardization, which subtracts the mean from each metric value and then divides by its standard deviation. After this processing, all metric values ​​will be within the range of zero mean and unit variance, eliminating the impact of dimensional differences on the model's prediction results. Improving encoding and standardization transforms the original features into a form that the machine learning model can understand and process, while eliminating interference from dimensional differences between different features, thus improving the fairness of model training and the accuracy of prediction results.

[0097] For example, Figure 3 This is a flowchart of the feature extraction operation provided in Embodiment 1 of this application, as follows: Figure 3 As shown, historical transaction information is first collected and outliers are removed. Then, based on the historical transaction information, feature data such as mean and standard deviation are calculated according to the calculation requirements of preset transaction indicators. The calculated feature data is then standardized and encoded to obtain the final transaction feature data.

[0098] Historical transaction information may include:

[0099] Credit card transaction history (time stamp, amount, merchant category code, geographical location).

[0100] Preset trading indicators may include:

[0101] Time-series characteristic transaction indicators: weekly / monthly consumption frequency, consumption interval variance, and proportion of consumption during holidays.

[0102] Transaction metrics for specific scenarios: online / offline consumption ratio, frequency of cross-border consumption, and consumption concentration of specific merchant categories.

[0103] Transaction indicators based on monetary value: percentage of large transactions, standard deviation of average daily spending.

[0104] This embodiment achieves the technical effect of improving the accuracy of transaction feature data and the compatibility between historical transaction information and the model by processing historical transaction information.

[0105] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0106] Example 2

[0107] This application also provides a device for determining user demand information. It should be noted that this device can be used to execute the user demand information determination method provided in the above embodiments. The following describes the device for determining user demand information provided in this application.

[0108] According to an embodiment of this application, an apparatus for implementing the above-described method for determining user demand information is also provided. Figure 4 This is a schematic diagram of a device for determining user demand information according to Embodiment 2 of this application, as shown below. Figure 4 As shown, the device includes:

[0109] The first acquisition unit 41 is used to acquire the target user's historical transaction information and determine the target user's transaction characteristic data based on the historical transaction information.

[0110] The first classification unit 42 is used to input transaction feature data into the user classification model to obtain the target classification result of the target user, wherein the target classification result is used to characterize that the target user belongs to the target cluster in the user classification model.

[0111] The first determining unit 43 is used to obtain public transaction feature information of each sample user in the target cluster, and determine the demand information of the target user based on the public transaction feature information.

[0112] The user demand information determination device provided in this application embodiment acquires historical transaction information of the target user through a first acquisition unit 41 and determines the transaction feature data of the target user based on the historical transaction information; a first classification unit 42 inputs the transaction feature data into a user classification model to obtain the target classification result of the target user, wherein the target classification result is used to characterize that the target user belongs to the target cluster in the user classification model; a first determination unit 43 acquires the common transaction feature information of each sample user in the target cluster and determines the demand information of the target user based on the common transaction feature information. By using an unsupervised classification model to determine that the target user belongs to the target cluster in the user classification model, and then determining the demand information of the target user based on the common transaction feature information of each sample user in the target cluster, the device achieves the purpose of accurately determining the demand information of the target user, thereby improving the technical effect of determining the demand information of the target user and solving the technical problem of low accuracy in determining user demand information in related technologies.

[0113] Optionally, in the user demand information determination device provided in this application embodiment, the user classification model is trained by the following device: a setting unit, used to set N initial values ​​and use each initial value as a K value to construct N initial classification models; a second classification unit, used to classify the sample transaction information contained in the first sample data using each initial classification model to obtain N predicted classification results, wherein the first sample data includes multiple sample transaction information, and each sample transaction information belongs to a sample user; a first calculation unit, used to calculate the sum of squared errors of each predicted classification result to obtain N sums of squared errors; a sorting unit, used to arrange the sums of squared errors of each initial classification model in ascending order of K value to obtain a first sequence; a second determination unit, used to determine the inflection point in the first sequence and obtain the K value in the initial classification model corresponding to the sum of squared errors indicated by the inflection point to obtain the target K value; and a training unit, used to construct a classification model based on the target K value to obtain a candidate classification model, and use the first sample data to train the candidate classification model to obtain a user classification model.

[0114] Optionally, in the user demand information determination device provided in the embodiments of this application, the device further includes: a second acquisition unit, used to acquire second sample data from newly added transaction information within a preset window duration when the user classification model usage time exceeds a preset window duration, wherein the second sample data includes multiple samples of newly added transaction information; a third acquisition unit, used to acquire the updated weight vector of each transaction feature at the current moment, and acquire the historical weight vector in the user classification model; a third determination unit, used to determine the target weight vector based on the updated weight vector and the historical weight vector; an update unit, used to update the user classification model using the target weight vector and the second sample data to obtain the target classification model; a second calculation unit, used to calculate the silhouette coefficient of the target classification model, and determine whether the silhouette coefficient meets the preset coefficient requirements; a replacement unit, used to replace the user classification model with the target classification model when the silhouette coefficient meets the preset coefficient requirements; and a retention unit, used to retain the user classification model and delete the target classification model when the silhouette coefficient does not meet the preset coefficient requirements.

[0115] Optionally, in the user demand information determination device provided in the embodiments of this application, the third determination unit includes: a first acquisition module, used to acquire a preset attenuation coefficient and use the preset attenuation coefficient to determine the weights of the updated weight vector and the historical weight vector; and a first calculation module, used to perform a weighted summation calculation on the updated weight vector and the historical weight vector according to the weights to obtain the target weight vector.

[0116] Optionally, in the user demand information determination device provided in this application embodiment, the second calculation unit includes: a second calculation module, used to calculate the average distance between the newly added transaction information of the target sample and the newly added transaction information of other samples in the target cluster for any newly added transaction information of a target sample, to obtain a first distance value, wherein the newly added transaction information of the target sample is located in the target cluster; a third calculation module, used to calculate the average distance between the newly added transaction information of the target sample and the newly added transaction information of each sample in the neighboring cluster, to obtain a second distance value, wherein the neighboring cluster is the cluster closest to the target cluster; a fourth calculation module, used to calculate the sub-profile coefficient of the newly added transaction information of the target sample based on the first distance value and the second distance value; and a second acquisition module, used to acquire the sub-profile coefficient of the newly added transaction information of each sample in the second sample data, and calculate the average value of the sub-profile coefficients to obtain the profile coefficient of the target classification model.

[0117] Optionally, in the user demand information determination device provided in the embodiments of this application, the first determination unit 43 includes: a first determination module, used to determine the user profile of the target user based on public transaction feature information; a second determination module, used to determine the matching degree between the target user and each business type based on the user profile, and obtain multiple matching degrees; and a third determination module, used to determine the business product demand under the business type with the highest matching degree as the demand information of the target user.

[0118] Optionally, in the user demand information determination device provided in the embodiments of this application, the first acquisition unit 41 includes: a third acquisition module, used to acquire preset transaction indicators, wherein the preset transaction indicators include at least one of the following: time-series characteristic transaction indicators, scenario characteristic transaction indicators, and amount characteristic transaction indicators; a fourth determination module, used to determine the values ​​of each indicator under the preset transaction indicators based on historical transaction information, thereby obtaining multiple indicator values; and a processing module, used to encode and standardize the multiple indicator values ​​to obtain transaction feature data.

[0119] It should be noted that the first acquisition unit 41, the first classification unit 42, and the first determination unit 43 mentioned above correspond to steps S201 to S203 in Embodiment 1. The instances and application scenarios implemented by each of the above units and the corresponding steps are the same, but are not limited to the content disclosed in Embodiment 1. It should be noted that the above modules or units can be hardware or software components stored in memory (e.g., memory 104) and processed by one or more processors (e.g., processors 102a, 102b, ..., 102n). The above modules can also be part of a device and can run in the computer terminal 10 provided in Embodiment 1.

[0120] Example 3

[0121] Embodiments of this application may provide an electronic device. Figure 5 This is a structural block diagram of an electronic device according to an embodiment of this application. Figure 5 As shown, the electronic device may include: one or more ( Figure 5 (Only one is shown) processor 1002, memory 1004, memory controller, and peripheral interface, wherein the peripheral interface is connected to the radio frequency module, audio module and display.

[0122] The memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the methods and apparatus in the embodiments of this application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, thereby implementing the above-described methods. The memory may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include memory remotely located relative to the processor, and these remote memories can be connected to the terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0123] Those skilled in the art will understand that Figure 5 The structure shown is for illustrative purposes only. Electronic devices can also be smartphones, tablets, handheld computers, mobile internet devices (MIDs), PADs, and other terminal devices. Figure 5 This does not limit the structure of the aforementioned electronic device. For example, electronic devices may also include components that are more... Figure 5 The more or fewer components shown (such as network interfaces, display devices, etc.), or having the same Figure 5 The different configurations shown.

[0124] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0125] Example 4

[0126] Embodiments of this application also provide a storage medium. Optionally, in this embodiment, the storage medium can be used to store the program code executed by the method for determining user requirement information provided in Embodiment 1.

[0127] Optionally, in this embodiment, the storage medium may be located in any computer terminal in a group of computer terminals in a computer network, or in any mobile terminal in a group of mobile terminals.

[0128] Embodiments of this application also provide a computer program product, which, when executed on a data processing device, is adapted to perform the steps of a method for determining user demand information.

[0129] Embodiments of this application also provide a computer-readable storage medium, which includes a stored executable program, wherein the executable program controls the device where the computer-readable storage medium is located to execute the above-described method for determining user requirement information during runtime.

[0130] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0131] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0132] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0133] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0134] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0135] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0136] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method of determining user demand information, characterized by, include: Obtain the target user's historical transaction information, and determine the target user's transaction characteristic data based on the historical transaction information; The transaction feature data is input into the user classification model to obtain the target classification result of the target user, wherein the target classification result is used to characterize that the target user belongs to the target cluster in the user classification model; Obtain public transaction feature information of each sample user in the target cluster, and determine the demand information of the target user based on the public transaction feature information.

2. The method of claim 1, wherein, The user classification model is trained in the following manner: Set N initial values ​​and use each initial value as the K value to build N initial classification models; Each initial classification model is used to classify the sample transaction information contained in the first sample data to obtain N predicted classification results. The first sample data includes multiple sample transaction information, and each sample transaction information belongs to a sample user. Calculate the sum of squared errors for each predicted classification result to obtain N sums of squared errors; The sum of squared errors of each initial classification model is arranged in ascending order of K value to obtain the first sequence; Determine the inflection point in the first sequence, and obtain the K value in the initial classification model corresponding to the sum of squared errors indicated by the inflection point, to obtain the target K value; A classification model is constructed based on the target K value to obtain a candidate classification model, and the candidate classification model is trained using the first sample data to obtain the user classification model.

3. The method of claim 1, wherein, The method further includes: If the user classification model is used for a longer period than a preset window, second sample data is obtained from the new transaction information within the preset window, wherein the second sample data includes multiple samples of new transaction information. Obtain the updated weight vectors of each transaction feature at the current moment, and obtain the historical weight vectors in the user classification model; The target weight vector is determined based on the updated weight vector and the historical weight vector; The user classification model is updated using the target weight vector and the second sample data to obtain the target classification model; Calculate the silhouette coefficient of the target classification model and determine whether the silhouette coefficient meets the preset coefficient requirements; If the contour coefficients meet the preset coefficient requirements, the target classification model is used to replace the user classification model. If the contour coefficients do not meet the preset coefficient requirements, the user classification model is retained, and the target classification model is deleted.

4. The method of claim 3, wherein, Determining the target weight vector based on the updated weight vector and the historical weight vector includes: Obtain a preset decay coefficient, and use the preset decay coefficient to determine the weights of the updated weight vector and the historical weight vector; The target weight vector is obtained by weighted summation of the updated weight vector and the historical weight vector based on the weights.

5. The method of claim 3, wherein, Calculating the silhouette coefficients of the target classification model includes: For any newly added transaction information of a target sample, calculate the average distance between the newly added transaction information of the target sample and the newly added transaction information of other samples in the target cluster to obtain a first distance value, wherein the newly added transaction information of the target sample is located in the target cluster; Calculate the average distance between the newly added transaction information of the target sample and the newly added transaction information of each sample in the neighboring cluster to obtain a second distance value, wherein the neighboring cluster is the cluster that is closest to the target cluster; Calculate the sub-profile coefficient of the newly added transaction information of the target sample based on the first distance value and the second distance value; Obtain the sub-profile coefficients of the newly added transaction information of each sample in the second sample data, and calculate the average value of the sub-profile coefficients to obtain the profile coefficient of the target classification model.

6. The method of claim 1, wherein, Determining the target user's demand information based on the public transaction characteristic information includes: Determine the user profile of the target user based on the public transaction feature information; Based on the user profile, the matching degree between the target user and each business type is determined, resulting in multiple matching degrees; The business product requirements under the business type with the highest matching degree are determined as the requirement information of the target user.

7. The method of claim 1, wherein, The transaction characteristic data of the target user determined based on the historical transaction information includes: Obtain preset transaction indicators, wherein the preset transaction indicators include at least one of the following: time-series characteristic transaction indicators, scenario characteristic transaction indicators, and amount characteristic transaction indicators; Based on the historical transaction information, determine the values ​​of each indicator under the preset transaction indicator to obtain multiple indicator values; The transaction feature data is obtained by encoding and standardizing the multiple indicator values.

8. An apparatus for determining user demand information, characterized by include: The first acquisition unit is used to acquire the historical transaction information of the target user and determine the transaction characteristic data of the target user based on the historical transaction information. The first classification unit is used to input the transaction feature data into the user classification model to obtain the target classification result of the target user, wherein the target classification result is used to characterize that the target user belongs to the target cluster in the user classification model; The first determining unit is used to acquire public transaction feature information of each sample user in the target cluster, and determine the demand information of the target user based on the public transaction feature information.

9. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the steps of the method for determining user demand information as described in any one of claims 1 to 7.

10. An electronic device, comprising: include: Memory, which stores executable programs; A processor for running the program, wherein the program, when running, executes the method for determining user demand information as described in any one of claims 1 to 7.

Images