Potential customer identification method, device and equipment, medium and program product

Abstract

The invention provides a potential customer identification method and device, equipment, a medium and a program product, relates to the technical field of artificial intelligence, and can solve the problem that the accuracy of potential customer identification is poor. The specific technical scheme is as follows: in response to a first input for at least one group session in a session list interface, obtaining session data of each group session in the at least one group session; extracting a session feature vector related to the target service in the session data of each group session through a potential level determination model, and determining potential levels of all contacts in at least one group session based on the session feature vector and a potential level determination rule; and sending different messages related to the target business to the contacts of different potential levels. The accuracy of identifying potential customers is improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a potential customer identification method, apparatus, device, medium, and program product. Background Technology

[0002] With the rapid development of information technology and industry, such as in financial services, the types and number of target businesses have increased dramatically, making the screening of potential customers for these target businesses increasingly important. Screening potential customers for target businesses refers to assessing and selecting customers based on their qualifications, creditworthiness, and risk tolerance to determine their suitability for providing relevant services for the target business.

[0003] Currently, the screening of potential customers for the target business mainly relies on customer information reports. This depends on the professional skills of human staff, who manually categorize customers into potential and non-potential customers based on the information in these reports, thereby selecting qualified customers for the target business.

[0004] However, when classifying potential and non-potential customers for a target business, manual processing of a large amount of customer information is required. This is influenced by subjective factors and makes it difficult to accurately classify potential and non-potential customers for the target business, resulting in poor accuracy in identifying potential customers. Summary of the Invention

[0005] This application provides a potential customer identification method, apparatus, device, medium, and program product that can accurately determine the potential level of a contact person and improve the accuracy of identifying potential customers.

[0006] To achieve the above objectives, the embodiments of this application adopt the following technical solutions: Firstly, a predictive model acquisition method is provided. This method includes: acquiring session data for each group session by inputting at least one group session from a session list interface; extracting session feature vectors related to the target business from the session data of each group session using a potential level determination model; and determining the potential level of all contacts in the at least one group session based on the session feature vectors and potential level determination rules. The potential level determination rules include any one of the following: if the number of session feature vectors corresponding to a contact is large, then the potential level of that contact is determined as the first potential level; if the number of session feature vectors corresponding to a contact is moderate, then the potential level of that contact is determined as the second potential level; if the number of session feature vectors corresponding to a contact is small, then the potential level of that contact is determined as the third potential level; and then sending different messages related to the target business to contacts with different potential levels.

[0007] Thus, by using a potential customer identification model to extract feature vectors related to the target business from the conversation data of all contacts in at least one group session, and determining the potential level of each contact based on these feature vectors and potential level determination rules, the potential level of each contact can be determined. The feature vectors related to the target business can accurately reflect the potential level of a customer, and the potential level determination rules can divide all contacts into three potential levels based on the number of conversation feature vectors corresponding to each contact. This allows for accurate determination of the potential level of each contact, thereby improving the accuracy of identifying potential customers.

[0008] In one possible implementation of the first aspect, the predicted purchase probability of a contact at the second potential level for a product related to the target business is greater than that of a contact at the third potential level for a product related to the target business, but less than that of a contact at the first potential level for a product related to the target business.

[0009] Thus, since the predicted probability of a contact in the second potential level purchasing related products of the target business is greater than that of a contact in the third potential level, but less than that of a contact in the first potential level, the potential level of contacts with a higher predicted probability of purchasing related products of the target business can be defined as high-potential, the potential level of contacts with a moderate predicted probability of purchasing related products of the target business can be defined as medium-potential, and the potential level of contacts with a lower predicted probability of purchasing related products of the target business can be defined as low-potential. This allows for more accurate identification of potential customers.

[0010] In another possible implementation of the first aspect, messages related to the target business and matching the potential level can be sent to each contact through at least one of the aforementioned group sessions.

[0011] Thus, since messages related to the target business and matching the potential level can be sent to each contact through at least one of the aforementioned group sessions, it is possible to accurately contact potential customers through private contact groups, thereby ensuring that messages reach contacts quickly. On the other hand, messages related to the target business and matching the potential level can be sent to contacts, thus ensuring the suitability of the sent messages to the contacts and increasing the probability of customers purchasing the target business's related products.

[0012] In another possible implementation of the first aspect, the basic information of all contacts can be obtained first, and then the potential level of all contacts can be determined based on the aforementioned session feature vector, potential level determination rules, and the basic information of all contacts.

[0013] In this way, since basic information about the contact can also be obtained, and the potential level of each contact is determined based on the basic information of the contact, the potential level of the contact can be determined by referring to basic information such as customer characteristics and account information, thereby further improving the accuracy of determining the potential level of the contact.

[0014] In another possible implementation of the first aspect, behavioral data of all contacts on the received messages can be obtained; and based on this behavioral data, the potential level can be adjusted to determine the model parameters of the model.

[0015] Thus, since the model parameters of the potential level determination model can be adjusted based on user feedback data of marketing messages, the potential customer identification model can be dynamically optimized so that the accuracy of identifying potential customers can be continuously improved.

[0016] Secondly, a potential customer identification device is provided, comprising: an acquisition module, configured to acquire session data of each group session in the at least one group session in response to a first input to at least one group session in a session list interface; a processing module, configured to extract session feature vectors related to a target business from the session data of each group session through a potential level determination model, and determine the potential level of all contacts in the at least one group session based on the session feature vectors and potential level determination rules; wherein the potential level determination rules include any one of the following: if the number of session feature vectors corresponding to a contact in the session feature vectors is greater than a first threshold, then the potential level of a contact is determined as a first potential level; if the number of session feature vectors corresponding to a contact in the session feature vectors is less than or equal to the first threshold and greater than a second threshold, then the potential level of a contact is determined as a second potential level; if the number of session feature vectors corresponding to a contact in the session feature vectors is less than or equal to the second threshold, then the potential level of a contact is determined as a third potential level; and a sending module, configured to send different messages related to the target business to contacts with different potential levels.

[0017] Thirdly, an electronic device is provided, the method comprising: a memory and at least one processor. The memory is communicatively connected to the processor. The memory is used to store computer program code, the computer program code including computer instructions. When the processor executes the computer instructions, it causes the electronic device to perform the method as described in the first aspect and any possible implementation thereof.

[0018] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer instructions. When executed by a processor, these computer instructions are used to implement the method described in the first aspect and any possible implementation thereof.

[0019] Fifthly, embodiments of this application provide a computer program product that, when run on a computer or executed by a computer's processor, implements the method described in the first aspect and any possible design thereof. The computer may be the electronic device described in the third aspect and any possible implementation thereof.

[0020] It is understood that the beneficial effects achieved by the potential customer identification device described in the second aspect, the electronic device described in the third aspect, the computer-readable storage medium described in the fourth aspect, and the computer program product described in the fifth aspect can be referred to as the beneficial effects in the first aspect and any possible implementation thereof, which will not be repeated here. Attached Figure Description

[0021] Figure 1 A flowchart of a potential customer identification method provided for some embodiments of this application; Figure 2 A schematic diagram of a session list interface in a potential customer identification method provided for some embodiments of this application; Figure 3 A flowchart of a potential customer identification method provided for some embodiments of this application; Figure 4 A flowchart of a potential customer identification method provided for some embodiments of this application; Figure 5 A flowchart of a potential customer identification method provided for some embodiments of this application; Figure 6 A flowchart of a potential customer identification method provided for some embodiments of this application; Figure 7 A schematic diagram of the structure of a potential customer identification device provided for some embodiments of this application; Figure 8 This is a schematic diagram of the structure of an electronic device provided for some embodiments of this application. Detailed Implementation

[0022] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this embodiment, unless otherwise stated, "a plurality of" means two or more.

[0023] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0024] The technical solutions provided in this application, including the collection, storage, use, processing, transmission, provision, and disclosure of financial data or user data, comply with relevant laws and regulations and do not violate public order and good morals.

[0025] It should be noted that in the embodiments of this application, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, it does not mean that the applicant has used or necessarily used the solution.

[0026] The following explains some concepts and terms involved in the potential customer identification method, apparatus, equipment, medium and program products provided in the embodiments of this application.

[0027] Machine learning algorithms A technical method that enables computers to automatically learn patterns from data and use these patterns to predict or make decisions about unknown events.

[0028] I. Core Classification: By Learning Method Machine learning algorithms are mainly divided into three categories, with the core difference being whether or not a manually labeled "standard answer" is needed during the learning process.

[0029] 1. Supervised Learning Features: It is trained using a dataset with labels (i.e., standard answers).

[0030] Objective: To learn the mapping relationship between inputs and outputs for prediction or classification.

[0031] Typical algorithms include: linear regression, logistic regression, decision tree, random forest, and support vector machine (SVM).

[0032] 2. Unsupervised Learning Features: Using unlabeled datasets, computers discover hidden structures in the data.

[0033] Objective: To explore the inherent patterns in data by performing clustering, dimensionality reduction, or correlation analysis.

[0034] Typical algorithms: K-means clustering, principal component analysis, association rule learning.

[0035] 3. Reinforcement Learning Features: It learns the optimal strategy by interacting with the environment through an "agent" and using "rewards" or "punishments" as feedback signals.

[0036] Objective: To identify a set of actions that maximize long-term cumulative rewards.

[0037] Typical scenarios: AlphaGo, autonomous driving, robot control.

[0038] II. Key Application Scenarios Machine learning algorithms have permeated multiple fields, solving practical problems: Computer vision: such as target detection for obstacle recognition in autonomous driving, and image generation for artificial intelligence (AI) sessions.

[0039] Natural Language Processing (NLP): Machine translation, sentiment analysis, chatbots.

[0040] Recommendation systems include e-commerce product recommendations, video content recommendations, and music recommendations.

[0041] In the financial sector: credit risk assessment, stock market trend prediction, and fraudulent transaction detection.

[0042] Feature vector In machine learning, linear algebra, and data analysis, the eigenvector is a core concept closely related to linear transformations. It describes a special vector that remains "direction invariant" under a linear transformation.

[0043] I. Intuitive Understanding Suppose there exists a matrix A (which can be understood as a linear transformation, such as rotation or scaling) and a non-zero vector v. If the product of matrix A and vector v only scales v (rather than changing its direction), then v is an eigenvector of matrix A.

[0044] II. Core Significance The essence of an eigenvector is the "inherent direction" of a linear transformation: when an eigenvector v is transformed by matrix A, the direction of v remains unchanged, only its length is scaled by the eigenvalue λ (λ>0 scales in the same direction, λ<0 scales in the opposite direction, |λ|>1 stretches, |λ|<1 shortens).

[0045] A matrix may have multiple eigenvectors, each corresponding to a different eigenvalue. Together, they form the "basic coordinate system" that describes the transformation properties of the matrix.

[0046] III. Applications in Machine Learning Eigenvectors and eigenvalues ​​are core tools in many algorithms, and typical scenarios include: Principal Component Analysis (PCA, Dimensionality Reduction) The eigenvectors of the covariance matrix of the data are calculated, and the k eigenvectors with the largest eigenvalues ​​constitute the "principal components," representing the directions of maximum variance in the data. These principal components are then projected onto the high-dimensional data to achieve dimensionality reduction, thus preserving key information while reducing computational complexity.

[0047] Spectral Clustering By solving for the eigenvectors of the similarity matrix of the data, the data can be mapped to a low-dimensional space through the eigenvectors, and then clustering can be performed, making the processing of non-linearly separable data more efficient.

[0048] Image compression and feature extraction The pixel matrix of an image can be compressed by eigenvector decomposition, such as PCA, while retaining the main eigenvectors, or key features can be extracted for recognition.

[0049] Recommendation system In matrix factorization (such as collaborative filtering), feature vectors can represent latent features of users or items (such as "user's preference for comedy" or "the comedic attributes of a movie").

[0050] IV. Key Properties Eigenvectors are non-zero vectors (the zero vector satisfies A0=λ0 for any matrix, but has no practical significance).

[0051] An eigenvalue may correspond to multiple eigenvectors (scaling vectors in the same direction), but an eigenvector corresponds to only one eigenvalue.

[0052] If the matrix is ​​a symmetric matrix, such as the covariance matrix, its eigenvectors are mutually orthogonal, which is an important theoretical basis for algorithms such as PCA.

[0053] Simply put, eigenvectors can find the "most critical direction" in data or transformations, and are an important tool for extracting core patterns from complex information.

[0054] The potential customer identification method, apparatus, equipment, medium, and program products provided in this application will be described in detail below with reference to the accompanying drawings and through specific embodiments and application scenarios.

[0055] With the rapid development of information technology and industry, such as in financial services, the types and number of target businesses have increased dramatically, making the screening of potential customers for these target businesses increasingly important. Screening potential customers for target businesses refers to assessing and selecting customers based on their qualifications, creditworthiness, and risk tolerance to determine their suitability for providing relevant services for the target business.

[0056] Currently, the screening of potential customers for the target business mainly relies on customer information reports. This depends on the professional skills of human staff, who manually categorize customers into potential and non-potential customers based on the information in these reports, thereby selecting qualified customers for the target business.

[0057] However, when classifying potential and non-potential customers for a target business, manual processing of a large amount of customer information is required. This is influenced by subjective factors and makes it difficult to accurately classify potential and non-potential customers for the target business, resulting in poor accuracy in identifying potential customers.

[0058] For example, with the arrival of an aging society, the demand for personal pension funds is increasing. By effectively identifying potential customers for pension products and conducting targeted marketing, the market penetration rate and customer satisfaction of pension products can be improved. Current methods of manually identifying potential customers are clearly insufficient to accurately identify those who wish to purchase pension products.

[0059] To address the aforementioned problems, embodiments of this application provide a method, apparatus, device, medium, and program product for identifying potential customers. The potential customer identification method provided in this application can be applied to scenarios involving the identification of potential customers in a target business, such as the identification of potential customers in a financial business.

[0060] For example, taking the identification of potential customers for pension products as an example, the conversation data of each group conversation can be obtained by inputting at least one group conversation in the conversation list interface; and the conversation feature vectors related to pension products in the conversation data of each group conversation can be extracted by using a potential level determination model, and the potential level of all contacts in the at least one group conversation can be determined based on the conversation feature vectors and the potential level determination rules; wherein, the potential level determination rules include any of the following: if the number of conversation feature vectors corresponding to a contact is large, then the potential level of that contact is determined as the first potential level; if the number of conversation feature vectors corresponding to a contact is moderate, then the potential level of that contact is determined as the second potential level; if the number of conversation feature vectors corresponding to a contact is small, then the potential level of that contact is determined as the third potential level; and then different messages related to pension products are sent to contacts with different potential levels.

[0061] For example, taking the identification of potential customers for financial products as an example, the conversation data of each group conversation can be obtained by inputting at least one group conversation in the conversation list interface; and by using a potential level determination model, conversation feature vectors related to financial products can be extracted from the conversation data of each group conversation, and the potential level of all contacts in the at least one group conversation can be determined based on the conversation feature vectors and the potential level determination rules; wherein, the potential level determination rules include any of the following: if the number of conversation feature vectors corresponding to a contact is large, then the potential level of that contact is determined as the first potential level; if the number of conversation feature vectors corresponding to a contact is moderate, then the potential level of that contact is determined as the second potential level; if the number of conversation feature vectors corresponding to a contact is small, then the potential level of that contact is determined as the third potential level; and then different messages related to financial products are sent to contacts with different potential levels.

[0062] It should be noted that the above two examples are only illustrative of pension products and wealth management products. In actual implementation, the potential customer identification method provided in this application embodiment can be applied to the identification of potential customers of any target business, and this application embodiment does not limit it.

[0063] Thus, by using a potential customer identification model to extract feature vectors related to the target business from the conversation data of all contacts in at least one group session, and determining the potential level of each contact based on these feature vectors and potential level determination rules, the potential level of each contact can be determined. The feature vectors related to the target business can accurately reflect the potential level of a customer, and the potential level determination rules can divide all contacts into three potential levels based on the number of conversation feature vectors corresponding to each contact. This allows for accurate determination of the potential level of each contact, thereby improving the accuracy of identifying potential customers.

[0064] Furthermore, the session archive data of group sessions can be used to analyze the data content, identify the degree of customer interest in the target business, and intercept potential customers; and the private domain connection attribute of group sessions can be fully utilized to reach potential customers at low cost.

[0065] It should be noted that the potential customer identification method provided in this application can be executed by a potential customer identification device, an electronic device, or a functional module within an electronic device. Some embodiments of this application use an electronic device executing the potential customer identification method as an example to illustrate the potential customer identification method provided in this application.

[0066] Figure 1 A flowchart illustrating a potential customer identification method provided for some embodiments of this application. For example... Figure 1 As shown, the method for obtaining the prediction model includes the following steps 101 to 103.

[0067] Step 101: In response to a first input to at least one group session in the session list interface, the electronic device acquires session data for each group session in at least one group session.

[0068] In some embodiments of this application, the aforementioned session list interface can be the session list interface corresponding to an instant messaging application in an electronic device.

[0069] In some embodiments of this application, the above-mentioned session list interface can be triggered for display by user input.

[0070] In some embodiments of this application, the above-mentioned session list interface may include at least one group session identifier, each group session identifier being used to indicate one of the at least one group sessions.

[0071] In some embodiments of this application, the electronic device may, in response to a first input to the at least one group session identifier, obtain session data for each of the at least one group session.

[0072] In some embodiments of this application, any of the group session identifiers in the above-mentioned at least one group session identifier can be any possible form of identifier such as a text identifier, an image identifier, or a symbol identifier.

[0073] In some embodiments of this application, the display parameters of any group session identifier among the at least one group session identifier mentioned above can be arbitrary display parameters; wherein, the display parameters may include, but are not limited to, at least one of the following: display color, display shape, display position, display transparency, display size, etc.

[0074] For example, an electronic device can display the aforementioned conversation list interface based on user input from an instant messaging application on the device, such as... Figure 2 As shown, the conversation list interface 20 includes an identifier 21 for group conversation A, an identifier 22 for group conversation B, an identifier 23 for group conversation C, and an identifier 24 for group conversation D. Each identifier indicates a group conversation. When a user needs to obtain the conversation data of group conversation A and group conversation B, they can make a first input to the identifier 21 for group conversation A and the identifier 22 for group conversation B (i.e., at least one of the aforementioned group conversation identifiers) in the conversation list interface 20. After receiving the first input, the electronic device can respond to the first input and obtain the conversation data of group conversation A and the conversation data of group conversation B, that is, the conversation data of each group conversation in the aforementioned at least one group conversation.

[0075] In some embodiments of this application, each group session in the above-mentioned at least one group session corresponds to at least two contacts; the number of contacts corresponding to any two group sessions in the at least one group session may be the same or different.

[0076] In some embodiments of this application, the session data of a group session includes: session data corresponding to all contacts in the group session.

[0077] In some embodiments of this application, the session data described above may contain various forms of information. For example, the session data may include, but is not limited to, at least one of the following: Text messages are the most common form of communication in group chats, including text exchanges between users, emoticons, web links, etc. With the popularization of voice recognition technology, voice messages are becoming increasingly common in group chats. Voice messages record information such as the speaker's tone and speaking speed, and have rich linguistic features. Image messages allow users to share information or express emotions by sending images. Image messages may contain various types of information such as text, images, and QR codes. Video messages: Some social media platforms allow users to send video messages, which contain richer information, including voice and images. Other unstructured data, such as emojis and red envelopes, are also common forms of communication in group chats.

[0078] In some embodiments of this application, the first input is used to obtain session data of the group session.

[0079] In some embodiments of this application, the first input can be any possible form of input, such as touch input, voice input, or physical button input.

[0080] For example, taking touch input as the first input, the first input can be a single click, double click, triple click, long press, hard press, swipe, or special trajectory input by the user using a finger or stylus to input at least one of the above-mentioned group conversations. Among them, special trajectory input includes, but is not limited to, at least one of the following: circular trajectory input, rectangular trajectory input, heart-shaped trajectory input, star-shaped trajectory input, or triangle trajectory input.

[0081] For example, taking voice input as the first input, the first input can be the voice "get the conversation data of group conversation A" or the voice "extract the conversation data of group conversation B and group conversation C", etc.

[0082] For example, taking the first input as a physical button input, the first input can be the user's input to the "volume +" button, or the user's input to the "volume -" button, or the user's input to both the "volume +" and "volume -" buttons simultaneously, etc.

[0083] In some embodiments of this application, the electronic device can obtain the session data of each group session in the at least one group session in response to a first input to the at least one group session in the session list interface, provided that the contact corresponding to the at least one group session has authorized the acquisition of session data.

[0084] In some embodiments of this application, the aforementioned authorization information may include, but is not limited to, at least one of the following: permission access information, device operation instruction information, encryption key, digital signature, data usage license information, etc.

[0085] In some embodiments of this application, the electronic device may, in response to a first input to at least one group session in the aforementioned session list interface, acquire session data for each group session in the at least one group session through a session data acquisition model.

[0086] In some embodiments of this application, the aforementioned session data acquisition model can be a pre-trained AI model deployed in an electronic device, which is used to acquire session data of group sessions.

[0087] Step 102: The electronic device extracts the session feature vectors related to the target business from the session data of each group session through the potential level determination model, and determines the potential level of all contacts in at least one group session based on the session feature vectors and the potential level determination rules.

[0088] The aforementioned potential level determination rules include any one of the following: if the number of conversation feature vectors corresponding to a contact in the aforementioned conversation feature vectors is greater than a first threshold, then the potential level of that contact is determined to be the first potential level; if the number of conversation feature vectors corresponding to a contact in the aforementioned conversation feature vectors is less than or equal to the first threshold and greater than a second threshold, then the potential level of that contact is determined to be the second potential level; if the number of conversation feature vectors corresponding to a contact in the aforementioned conversation feature vectors is less than or equal to the second threshold, then the potential level of that contact is determined to be the third potential level.

[0089] In some embodiments of this application, the target business described above may include, but is not limited to, at least one of the following: financial business, communication business, medical business, etc.

[0090] In some embodiments of this application, the above-described potential level determination model is used to determine the potential level of a contact.

[0091] In some embodiments of this application, the electronic device may first construct the aforementioned potential level determination model through a machine learning algorithm, which may include, but is not limited to, logistic regression or decision trees.

[0092] It's important to note that logistic regression is a machine learning algorithm widely used for classification tasks. Despite having "regression" in its name, it is actually a binary classification model. Its core idea is to use the sigmoid function (also known as the logistic function) to map the output of linear regression to the (0,1) interval, thus obtaining the probability that a sample belongs to the positive class.

[0093] Decision tree algorithms are a typical classification and regression method that classifies or predicts data through a tree structure. Its core idea is to classify data using a series of rules, resulting in models that are highly readable and fast in classification. Decision trees generate a tree structure by recursively selecting the optimal feature for splitting: 1. Feature selection: Selecting the optimal splitting feature based on information gain (ID3 algorithm), information gain ratio (C4.5 algorithm), or Gini index (CART algorithm); 2. Tree generation: Starting from the root node, dividing the data according to feature values, recursively generating child nodes until a stopping condition is met (e.g., sample purity reaching a threshold or tree depth limitation); 3. Pruning: Pre-pruning (limiting tree depth) or post-pruning (e.g., cost complexity pruning) prevents overfitting. Among them, the ID3 algorithm selects features based on information gain, but it tends to favor multi-valued attributes and lacks a pruning strategy; the C4.5 algorithm improves upon the ID3 algorithm by using information gain ratio to avoid multi-valued bias and supports continuous feature discretization and missing value handling; the CART algorithm uses the Gini index as the splitting criterion, supports classification and regression, and improves generalization ability through pruning.

[0094] It should be noted that feature vector extraction can transform the original data into more representative low-dimensional features.

[0095] In some embodiments of this application, the potential level determination model described above can extract the session feature vector related to the target service from the session data of each group session using at least one of the following methods: Principal Component Analysis (PCA): By calculating the eigenvalues ​​and eigenvectors of the data covariance matrix, the direction with the largest variance is selected as the new coordinate axis, mapping the data to a lower-dimensional space. This method can effectively reduce dimensionality while preserving key features. Its core steps include calculating the mean vector, covariance matrix, and eigenvalue decomposition. Linear Discriminant Analysis (LDA) is a supervised dimensionality reduction technique that projects data into a lower-dimensional space by maximizing inter-class distance and minimizing intra-class distance. It is suitable for classification tasks and preserves the separability information of the data. Autoencoders: Input data is compressed into a low-dimensional representation (encoded) by an encoder, and then the original data is reconstructed by a decoder. By minimizing the reconstruction error, the encoder can capture the key features of the data for subsequent tasks. Convolutional Neural Network (CNN): It extracts local features (such as edges and textures) of an image through convolutional layers, and gradually extracts high-level features (such as shapes and object parts) as the network deepens, making it excellent in image recognition. Word embedding: Mapping words to high-dimensional vectors, so that semantically similar words are close in distance in the vector space. Commonly used models include Word2Vec, GloVe, and BERT, which are suitable for text feature extraction; Recurrent Neural Networks (RNNs) and their variants: They capture the temporal dependencies of sequential data through recurrent structures. Their variants (such as LSTM and GRU) solve the gradient vanishing problem in long sequences through gating mechanisms, thereby improving feature extraction capabilities. Feature selection: Important features are selected through filtering methods (such as correlation coefficient, chi-square test), wrapping methods (such as recursive feature elimination), or embedding methods (such as Lasso regression); Sparse coding: Projecting nonlinear data into a higher-dimensional space to achieve linear separability; K-L transform: A feature extraction method based on Karhunen-Loève expansion, applicable to stochastic processes.

[0096] In some embodiments of this application, the electronic device can input the aforementioned session feature vector into the aforementioned potential level determination model, and then the potential level determination model processes the session feature vector according to the aforementioned potential level determination rules, and outputs the potential level of all contacts in the aforementioned at least one group session.

[0097] In some embodiments of this application, the potential level determination model can first perform at least one of the following processing on the above-mentioned session feature vector: feature scaling, normalization, dimensionality reduction, etc.; then, according to the above-mentioned potential level determination rules, process the processed session feature vector to output the potential level of all contacts in the above-mentioned at least one group session.

[0098] In some embodiments of this application, the first threshold and the second threshold can be default parameters of the potential level determination model, or they can be parameters that can be arbitrarily set by the user.

[0099] In some embodiments of this application, the first threshold is greater than the second threshold.

[0100] In some embodiments of this application, the first threshold is 5 and the second threshold is 2; or the first threshold is 10 and the second threshold is 5; or the first threshold is 15 and the second threshold is 10.

[0101] It is understandable that the larger the first threshold and the second threshold are, the more conversation feature vectors are required for the contact to determine the potential level, and the more accurate the potential level is determined. Conversely, the smaller the first threshold and the second threshold are, the fewer conversation feature vectors are required for the contact to determine the potential level, and the less accurate the potential level is determined.

[0102] In some embodiments of this application, the specific values ​​of the first threshold and the second threshold can be determined according to actual usage requirements, and this application does not limit them.

[0103] For example, suppose the above-mentioned conversation feature vector includes 50 conversation feature vectors, of which 10 correspond to contact A; then if the first threshold is 15 and the second threshold is 10, then since the number of conversation feature vectors corresponding to contact A is equal to the second threshold, the potential level of contact A is determined to be the third potential level; if the first threshold is 10 and the second threshold is 5, then since the number of conversation feature vectors corresponding to contact A is equal to the first threshold, the potential level of contact A is determined to be the second potential level; if the first threshold is 5 and the second threshold is 2, then since the number of conversation feature vectors corresponding to contact A is greater than the first threshold, the potential level of contact A is determined to be the first potential level.

[0104] In some embodiments of this application, the above-mentioned potential level determination rule may also include any of the following: if the matching degree between the conversation feature vector corresponding to a contact and the preset feature vector in the above-mentioned conversation feature vector is greater than a third threshold, then the potential level of that contact is determined as a first potential level; if the matching degree between the conversation feature vector corresponding to a contact and the preset feature vector in the above-mentioned conversation feature vector is less than or equal to the third threshold and greater than a fourth threshold, then the potential level of that contact is determined as a second potential level; if the matching degree between the conversation feature vector corresponding to a contact and the preset feature vector in the above-mentioned conversation feature vector is less than or equal to the fourth threshold, then the potential level of that contact is determined as a third potential level.

[0105] In some embodiments of this application, the aforementioned preset feature vector is the feature vector corresponding to the conversation data of contacts who are more likely to purchase related products of the target business.

[0106] In some embodiments of this application, the third threshold and the fourth threshold can be default parameters of the potential level determination model, or they can be parameters that can be arbitrarily set by the user.

[0107] In some embodiments of this application, the third threshold is greater than the fourth threshold.

[0108] In some embodiments of this application, the third threshold is 60% and the fourth threshold is 40%; or the third threshold is 80% and the fourth threshold is 60%; or the third threshold is 90% and the fourth threshold is 70%.

[0109] It is understandable that the larger the third and fourth thresholds are, the greater the matching degree between the conversation feature vector corresponding to the contact and the above-mentioned preset feature vector is required when determining the potential level of a contact, and the more accurate the potential level is determined. Conversely, the smaller the third and fourth thresholds are, the smaller the matching degree between the conversation feature vector corresponding to the contact and the above-mentioned preset feature vector is required when determining the potential level of a contact, and the less accurate the potential level is determined.

[0110] In some embodiments of this application, the specific values ​​of the third threshold and the fourth threshold can be determined according to actual usage requirements, and are not limited in the embodiments of this application.

[0111] In some embodiments of this application, the above-mentioned potential level determination rules may include any combination of the two rules mentioned above.

[0112] For example, the aforementioned potential level determination rule may include any of the following: if the number of conversation feature vectors corresponding to a contact in the aforementioned conversation feature vectors is greater than a first threshold, and the matching degree between the conversation feature vector corresponding to that contact and the aforementioned preset feature vector is greater than a third threshold, then the potential level of that contact is determined to be the first potential level; if the number of conversation feature vectors corresponding to a contact in the aforementioned conversation feature vectors is less than or equal to the first threshold and greater than a second threshold, and the matching degree between the conversation feature vector corresponding to that contact and the aforementioned preset feature vector is less than or equal to the third threshold and greater than a fourth threshold, then the potential level of that contact is determined to be the second potential level; if the number of conversation feature vectors corresponding to a contact in the aforementioned conversation feature vectors is less than or equal to the second threshold, and the matching degree between the conversation feature vector corresponding to that contact and the preset feature vector is less than or equal to the fourth threshold, then the potential level of that contact is determined to be the third potential level.

[0113] It should be noted that the potential level determination rule in the above example is only a combination of the two rules mentioned above. In actual implementation, the potential level determination rule can be any possible combination of rules, and this application embodiment does not limit it.

[0114] In some embodiments of this application, the predicted purchase probability of a contact at the second potential level for a product related to the target business is greater than that of a contact at the third potential level for a product related to the target business, but less than that of a contact at the first potential level for a product related to the target business.

[0115] For example, taking financial business as the target business, the relevant products of the above-mentioned target business may include, but are not limited to, at least one of the following: pension products, housing provident fund products, wealth management products, insurance products, deposit products, loan products, etc.

[0116] In some embodiments of this application, the first potential level may also be referred to as the high potential level, the second potential level may also be referred to as the medium potential level, and the third potential level may also be referred to as the low potential level.

[0117] In some embodiments of this application, since the predicted probability of a contact at the second potential level purchasing related products of the target business is greater than that of a contact at the third potential level, but less than that of a contact at the first potential level, the potential level of contacts with a higher predicted probability of purchasing related products of the target business can be determined as high-potential level, the potential level of contacts with a moderate predicted probability of purchasing related products of the target business can be determined as medium-potential level, and the potential level of contacts with a lower predicted probability of purchasing related products of the target business can be determined as low-potential level. This allows for more accurate identification of potential customers.

[0118] In some embodiments of this application, after determining the potential level of all contacts in at least one group session, the electronic device can divide all contacts according to the determined potential level to classify all contacts into customers with different potential levels.

[0119] Step 103: The electronic device sends different messages related to the target business to contacts at different potential levels.

[0120] In some embodiments of this application, the aforementioned message may be a marketing message or a notification message, etc.

[0121] In some embodiments of this application, the electronic device may send a first message to a contact at a first potential level, a second message to a contact at a second potential level, and a third message to a contact at a third potential level.

[0122] In some embodiments of this application, any two of the first message, second message, and third message may correspond to the same or different target services.

[0123] In some embodiments of this application, when any two of the first, second, and third messages correspond to the same target service, the message types or message scripts of the two messages are different to suit customers with different potential levels.

[0124] In some embodiments of this application, the message sent by the electronic device to the contact may include a link, which is a link to an introduction interface or consultation channel for the relevant products of the aforementioned target business.

[0125] In some embodiments of this application, combined with Figure 1 ,like Figure 3 As shown, step 103 above can be specifically implemented through step 103a below.

[0126] Step 103a: The electronic device sends a message to each of all contacts through at least one group session, which is relevant to the target business and matches the potential level of the contact.

[0127] In some embodiments of this application, for each of the aforementioned contacts, the electronic device can send messages to them through their respective group sessions, thereby leveraging private domain traffic to enhance customer relationships and increase customer stickiness.

[0128] In some embodiments of this application, the electronic device can preset messages corresponding to different potential levels. After determining the potential level of a contact, the device can send the message corresponding to the potential level to the contact through the group session to which the contact belongs.

[0129] In some embodiments of this application, since messages related to the target business and matching the potential level can be sent to each contact through at least one group session, it is possible to accurately contact potential customers through private contact groups, thereby enabling messages to reach contacts quickly. On the other hand, messages related to the target business and matching the potential level can be sent to contacts, thereby ensuring the suitability of the sent messages to the contacts and increasing the probability of customers purchasing the target business's related products.

[0130] In some embodiments of this application, the potential customer identification method extracts feature vectors related to the target business from the session data of all contacts in at least one group session using a potential customer identification model. Based on these feature vectors and a potential level determination rule, the potential level of each contact is determined. The feature vectors related to the target business can accurately reflect the customer's potential level. The potential level determination rule can divide all contacts into three potential levels based on the number of session feature vectors corresponding to each contact. This allows for accurate determination of the contact's potential level, thus improving the accuracy of potential customer identification.

[0131] In some embodiments of this application, combined with Figure 1 ,like Figure 4As shown, prior to step 102 above, some embodiments of this application provide a potential customer identification method that may further include step 104 below, and step 102 above can be specifically implemented through step 102a below.

[0132] Step 104: The electronic device obtains the first information of all contacts.

[0133] The first piece of information includes at least one of the following: contact characteristic information, contact account information, data on the contact's holdings of relevant products of the target business, the contact's access frequency to the relevant product interface of the target business, and the contact's dwell time on the relevant product interface of the target business.

[0134] In some embodiments of this application, the electronic device can obtain the first information of all contacts upon receiving authorization information from all the aforementioned contacts for obtaining the first information.

[0135] Step 102a: The electronic device extracts the session feature vector related to the target business from the session data of each group session through the potential level determination model, and determines the potential level of all contacts based on the session feature vector, the potential level determination rules and the first information of all contacts.

[0136] In some embodiments of this application, the electronic device can input the aforementioned session feature vector and the first information of all the aforementioned contacts into the aforementioned potential level determination model, and then the potential level determination model processes the session feature vector and the first information according to the aforementioned potential level determination rules, and outputs the potential level of all contacts in the aforementioned at least one group session.

[0137] In some embodiments of this application, since basic information of the contact can also be obtained, and the potential level of each contact is determined based on the basic information of the contact, the potential level of the contact can be determined by referring to basic information such as customer characteristic information and account information, thereby further improving the accuracy of determining the potential level of the contact.

[0138] In some embodiments of this application, combined with Figure 1 ,like Figure 5 As shown, after step 103 above, some embodiments of this application provide a potential customer identification method that may further include steps 105 and 106 as described below.

[0139] Step 105: The electronic device acquires behavioral data of all contacts in response to the received messages.

[0140] In some embodiments of this application, the received message is a message sent by the electronic device that is related to the target service.

[0141] In some embodiments of this application, the electronic device can obtain the behavioral data of all contacts in relation to the received message upon receiving authorization information from all the aforementioned contacts for obtaining behavioral data.

[0142] In some embodiments of this application, the aforementioned behavioral data may include, but is not limited to, at least one of the following: Do you want to view the message? Do you want to click the link in the message? Should I trigger a redirect to the mobile banking page? Whether to trigger a redirect to the relevant product introduction page for the target business; Whether it triggers a redirect to the relevant product consultation channels for the target business; Should further consultation be sought? Should I block the message?

[0143] In some embodiments of this application, after acquiring the aforementioned behavioral data, the electronic device can create or update a customer behavior database based on the behavioral data.

[0144] Step 106: Based on behavioral data, the electronic device adjusts the potential level to determine the model parameters of the model.

[0145] In some embodiments of this application, the electronic device adjusts the model parameters of the potential level determination model to optimize the potential level determination model and adjust customer classification criteria and marketing strategies.

[0146] For a detailed description of how electronic devices adjust the model parameters of the potential level determination model based on the aforementioned behavioral data, please refer to the relevant descriptions in related technologies. To avoid repetition, they will not be repeated here.

[0147] In some embodiments of this application, the electronic device can periodically acquire behavioral data of all the aforementioned contacts regarding the received messages according to a preset cycle or frequency, and adjust the model parameters of the potential level determination model based on the acquired behavioral data, thereby continuously optimizing the performance of the potential level determination model and improving the accuracy of identification and marketing.

[0148] In some embodiments of this application, since the model parameters of the potential level determination model can be adjusted based on user feedback data on marketing messages, the potential customer identification model can be dynamically optimized so that the potential customer identification model can continuously improve the accuracy of identifying potential customers.

[0149] The following description, in conjunction with the accompanying drawings, exemplarily illustrates some embodiments of the potential customer identification method provided in this application.

[0150] For example, taking the identification of potential customers for pension products as an example, such as Figure 6 As shown, some embodiments of this application provide a potential customer identification method that may include the following process: I. Data Collection With the customer's authorization, the electronic device collects the following information: session data of each group session in at least one group session, and first information of all contacts in the at least one group session; wherein the first information includes: contact characteristic information, contact account information, data on the contact's holding of pension products, frequency of contact's access to the pension product interface, and duration of contact's stay on the pension product interface.

[0151] II. Identification of Potential Customers By analyzing the collected data, a potential level determination model is constructed using machine learning algorithms. This model extracts session feature vectors related to pension products from the session data of each group session. Based on these feature vectors and the aforementioned potential level determination rules, the potential level of all contacts in at least one group session is determined. This allows all contacts to be categorized into three levels: high potential, medium potential, and low potential.

[0152] It should be noted that the high potential level is the first potential level in the above embodiments, the medium potential level is the second potential level in the above embodiments, and the low potential level is the third potential level in the above embodiments.

[0153] III. Marketing Strategy Formulation For identified potential customers, marketing messages matching their potential level are sent in batches through at least one of the aforementioned group sessions for personalized promotion. These marketing messages may include marketing links, which can link to pension product introduction pages or consultation channels.

[0154] IV. Customer Feedback Monitoring Monitor the behavior of all contacts in response to received marketing messages, including: whether they open marketing links, whether they are redirected to the corresponding mobile banking page, whether they make further inquiries, and whether they block the marketing messages. The collected behavioral data will be compiled into a customer behavior database.

[0155] V. Model Optimization Based on the aforementioned behavioral data, the model parameters of the potential level determination model will be adjusted to further optimize the model and refine customer classification criteria and marketing strategies. The potential level determination model will be updated regularly to improve the accuracy of identification and marketing.

[0156] In this way, by combining bank data and customer behavior data, potential customers for purchasing pension products can be accurately identified, improving marketing efficiency. Private domain traffic, such as group chats, can be leveraged to strengthen customer relationships and increase customer loyalty. Models and marketing strategies can be continuously optimized based on customer feedback to achieve precision marketing.

[0157] The above-described method embodiments, or various possible implementations of the method embodiments, can be executed individually, or, provided there are no contradictions, they can be combined with each other. The specific implementation can be determined according to actual usage requirements, and this application embodiment does not impose any restrictions on this.

[0158] Figure 7 A schematic diagram of the structure of a potential customer identification device provided for some embodiments of this application. For example... Figure 7 As shown, the potential customer identification device 70 includes: an acquisition module 71, a processing module 72, and a sending module 73.

[0159] The acquisition module 71 can be used to acquire the session data of each group session in the at least one group session in response to a first input on the session list interface. The processing module 72 can be used to extract session feature vectors related to the target business from the session data of each group session using a potential level determination model, and determine the potential level of all contacts in the at least one group session based on the session feature vectors and potential level determination rules. The potential level determination rules include any of the following: if the number of session feature vectors corresponding to a contact in the session feature vector is greater than a first threshold, then the potential level of that contact is determined to be a first potential level; if the number of session feature vectors corresponding to a contact in the session feature vector is less than or equal to the first threshold and greater than a second threshold, then the potential level of that contact is determined to be a second potential level; if the number of session feature vectors corresponding to a contact in the session feature vector is less than or equal to the second threshold, then the potential level of that contact is determined to be a third potential level. The sending module 73 can be used to send different messages related to the target business to contacts with different potential levels.

[0160] In other embodiments, the predicted purchase probability of a contact at the second potential level for a product related to the target business is greater than that of a contact at the third potential level for a product related to the target business, but less than that of a contact at the first potential level for a product related to the target business.

[0161] In other embodiments, the sending module 73 may be specifically used to send a message that matches the potential level of the target business and is related to the relevant products of the target business to each of the above-mentioned contacts through at least one group session.

[0162] In other embodiments, the acquisition module 71 may also be used to acquire first information of all contacts before the processing module 72 determines the potential level of all contacts in the at least one group session based on the aforementioned session feature vector and potential level determination rules. The first information includes at least one of the following: contact characteristic information, contact account information, data on the contact's holdings of related products of the target business, the contact's access frequency to the related product interface of the target business, and the contact's dwell time on the related product interface of the target business. Specifically, the processing module 72 may be used to determine the potential level of all contacts based on the session feature vector, the potential level determination rules, and the first information of all contacts.

[0163] In other embodiments, the acquisition module 71 can also be used to acquire behavioral data of all contacts regarding the received messages after the sending module 73 sends different messages related to the target business to contacts at different potential levels. The processing module 72 can also be used to adjust the model parameters of the potential level determination model based on the behavioral data.

[0164] The potential customer identification device provided in this application embodiment can execute the method shown in the above method embodiment. Its implementation principle and beneficial effects can be referred to the relevant description in the method embodiment, and will not be repeated here.

[0165] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 8 As shown, the electronic device 40 includes a transceiver 401, a memory 403, and at least one processor 402.

[0166] The transceiver 401 is used to interact with other devices to send and receive data. For example, in this embodiment, the transceiver 401 can be used to send marketing messages to a contact's device.

[0167] The memory 403 is used to store computer program code, which includes computer instructions. These computer instructions run in the aforementioned electronic device to implement the method shown in the above-described method embodiments. For example, the memory may include high-speed random access memory (RAM), and may also include non-volatile memory (NVM), such as at least one disk storage device, and may also be a USB flash drive, portable hard drive, read-only memory, disk, or optical disc, etc.

[0168] Processor 402 can be a general-purpose processor, including a Central Processing Unit (CPU), a network processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. Processor 402 can also be other general-purpose processors. The general-purpose processor can be a microprocessor or any conventional processor.

[0169] Transceiver 401, memory 403, and processor 402 are communicatively connected. For example, memory 403 and transceiver 401 can be connected to processor 402 via a system bus to complete mutual communication. The system bus can be a peripheral component interconnect (PCI) bus, an extended industry standard architecture (EISA) bus, an industry standard architecture (ISA) bus, etc. The system bus can be divided into address bus, data bus, control bus, etc. For ease of representation, only one thick line is used in the figure, but this does not mean that there is only one bus or one type of bus.

[0170] Alternatively, the memory 403 can be either standalone or integrated with the processor 402. When the memory 403 is set up independently, it is connected to the processor 402 via a system bus.

[0171] This application also provides a chip for executing instructions, which is used to execute the potential customer identification method described in the above embodiments.

[0172] This application also provides a computer-readable storage medium storing computer instructions. When these computer instructions are executed by a processor, they are used to implement the technical solution of the potential customer identification method described in the above embodiments. Specifically, when the computer instructions are executed by a processor, the electronic device can perform the technical solution of the potential customer identification method described in the above embodiments.

[0173] This application also provides a computer program product, which includes a computer program stored in a computer-readable storage medium. At least one processor can read the computer program from the computer-readable storage medium, and when the at least one processor executes the computer program, it can implement the technical solution of the potential customer identification method in the above embodiments. The aforementioned computer-readable storage media can be implemented from any type of volatile or non-volatile storage device or a combination thereof, such as Static Random-Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The computer-readable storage media can be any available medium accessible to a general-purpose or special-purpose computer. An exemplary computer-readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Of course, the computer-readable storage medium can also be a component of the processor. The processor and the computer-readable storage medium can reside in an application-specific integrated circuit (ASIC). Alternatively, the processor and the computer-readable storage medium can exist as discrete components in an electronic control unit or main control device; this application does not limit this.

[0174] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms.

[0175] The modules described as separate components may or may not be physically separate. The components shown as modules 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 modules can be selected to implement the solution of this embodiment according to actual needs.

[0176] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The unit composed of the above modules can be implemented in hardware or in the form of hardware plus software functional units.

[0177] The integrated modules described above, implemented as software functional modules, can be stored in a computer-readable storage medium. These software functional modules, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods of the various embodiments of this application.

[0178] It should be understood that the steps of the method disclosed in the embodiments of this application can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules in the processor.

[0179] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0180] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for identifying potential customers, characterized in that, include: In response to a first input to at least one group session in the session list interface, obtain session data for each group session in the at least one group session; Using a potential level determination model, session feature vectors related to the target business are extracted from the session data of each group session. Based on the session feature vectors and potential level determination rules, the potential level of all contacts in the at least one group session is determined. The potential level determination rules include any of the following: if the number of session feature vectors corresponding to a contact in the session feature vectors is greater than a first threshold, then the potential level of that contact is determined as a first potential level; if the number of session feature vectors corresponding to a contact in the session feature vectors is less than or equal to the first threshold and greater than a second threshold, then the potential level of that contact is determined as a second potential level; if the number of session feature vectors corresponding to a contact in the session feature vectors is less than or equal to the second threshold, then the potential level of that contact is determined as a third potential level. Send different messages related to the target business to contacts at different potential levels.

2. The method according to claim 1, characterized in that, The predicted purchase probability of the second potential level contact for the relevant products of the target business is greater than that of the third potential level contact for the relevant products of the target business, but less than that of the first potential level contact for the relevant products of the target business.

3. The method according to claim 1, characterized in that, Sending different messages related to the target business to contacts at different potential levels includes: Through the at least one group session, send a message to each of the contacts that matches the potential level of the target business and is related to the relevant products of the target business.

4. The method according to any one of claims 1 to 3, characterized in that, Before determining the potential level of all contacts in the at least one group session based on the session feature vector and the potential level determination rule, the method further includes: Obtain first information of all the contacts, the first information including at least one of the following: contact characteristic information, contact account information, data on the contact's holding of related products of the target business, the contact's access frequency to the interface of related products of the target business, and the contact's dwell time on the interface of related products of the target business; The step of determining the potential level of all contacts in the at least one group session based on the session feature vector and the potential level determination rule includes: The potential level of all contacts is determined based on the session feature vector, the potential level determination rule, and the first information of all contacts.

5. The method according to any one of claims 1 to 3, characterized in that, After sending different messages related to the target business to contacts at different potential levels, the method further includes: Obtain behavioral data of all contacts in response to the received messages; Based on the behavioral data, the model parameters of the potential level are adjusted to determine the model parameters.

6. A potential customer identification device, characterized in that, include: The acquisition module is configured to acquire the session data of each group session in the at least one group session in response to a first input to at least one group session in the session list interface; The processing module is configured to extract session feature vectors related to the target business from the session data of each group session using a potential level determination model, and determine the potential level of all contacts in the at least one group session based on the session feature vectors and potential level determination rules; wherein the potential level determination rules include any of the following: if the number of session feature vectors corresponding to a contact in the session feature vectors is greater than a first threshold, then the potential level of the contact is determined as a first potential level; if the number of session feature vectors corresponding to a contact in the session feature vectors is less than or equal to the first threshold and greater than a second threshold, then the potential level of the contact is determined as a second potential level; if the number of session feature vectors corresponding to a contact in the session feature vectors is less than or equal to the second threshold, then the potential level of the contact is determined as a third potential level. The sending module is used to send different messages related to the target business to contacts at different potential levels.

7. The apparatus according to claim 6, characterized in that, The predicted purchase probability of the second potential level contact for the relevant products of the target business is greater than that of the third potential level contact for the relevant products of the target business, but less than that of the first potential level contact for the relevant products of the target business.

8. The apparatus according to claim 6, characterized in that, The sending module is specifically used to send a message that matches the potential level of the target business and is related to the relevant products of the target business to each of the contacts through the at least one group session.

9. The apparatus according to any one of claims 6 to 8, characterized in that, The acquisition module is further configured to acquire first information of all contacts before the processing module determines the potential level of all contacts in the at least one group session based on the session feature vector and the potential level determination rule. The first information includes at least one of the following: contact feature information, contact account information, data on the contact's holding of related products of the target business, the contact's access frequency to the interface of related products of the target business, and the contact's dwell time on the interface of related products of the target business. The processing module is specifically used to determine the potential level of all contacts based on the session feature vector, the potential level determination rule, and the first information of all contacts.

10. The apparatus according to any one of claims 6 to 8, characterized in that, The acquisition module is further configured to acquire behavioral data of all contacts in response to the received messages after the sending module sends different messages related to the target business to contacts of different potential levels; The processing module is also used to adjust the model parameters of the potential level determination model based on the behavioral data.

11. An electronic device, characterized in that, include: A memory and at least one processor; the memory is communicatively connected to the processor; the memory is used to store computer program code, the computer program code including computer instructions; when the processor executes the computer instructions, the electronic device performs the method as described in any one of claims 1-5.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, are used to implement the method as described in any one of claims 1-5.

13. A computer program product, characterized in that, When the computer program product is run on a computer / executed by the computer's processor, it implements the method as described in any one of claims 1-5.

Images