Information triggering channel optimization method and system, electronic equipment and storage medium

Abstract

The invention provides an information triggering channel optimization method and system, electronic equipment and a storage medium. The method comprises the following steps: determining first data of a user; extracting basic features related to a touch effect from the portrait data, and constructing a feature matrix in combination with a touch channel, historical marketing information and user response data; training the feature matrix by using a machine learning algorithm to obtain a response prediction model capable of predicting user response data; the response prediction model is used for predicting user response data of each arrival channel according to the portrait data of the user, the arrival channel data and the marketing information; predicting the user according to the trained response prediction model, and determining an optimal access channel of different marketing information of the user; namely, by collecting and analyzing the response data of the user on different channels, the model capable of predicting the response data of the customer is trained, then the bank is guided to select the optimal touch channel, the touch efficiency is high, and the marketing efficiency and the customer satisfaction degree are improved.

Description

Technical Field

[0001] The present invention belongs to the field of data processing technology, and more specifically, particularly relates to an information triggering channel optimization method and system, electronic equipment, and storage medium. Background Art

[0002] In the key area of ​​banking marketing outreach, the industry has long relied on solutions based on rule engines and traditional statistical analysis methods. The core of these solutions is to use pre-defined rules and historical data to deeply analyze customer behavior and determine the most suitable marketing strategy.

[0003] In the current technical environment, there is no distinction between reach channels, and the same reach channels are used for different users. For example, all channels are used to reach marketing information to all users; or a single channel is used for all users.

[0004] However, since different users are accustomed to using different channels, the use of the above-mentioned channels results in low reach efficiency. Summary of the invention

[0005] In view of this, the purpose of the present invention is to provide an information-triggered channel optimization method and system, an electronic device, and a storage medium, which are used to collect and analyze user response data on different channels, train a model that can predict customer response data, and then guide banks to select the optimal contact channel with high contact efficiency, thereby improving marketing efficiency and customer satisfaction.

[0006] The first aspect of the present application discloses an information trigger channel optimization method, comprising:

[0007] Determine first data of the user, the first data including: portrait data, contact channel data, marketing information and user response data;

[0008] Extracting basic features related to the reach effect from the portrait data, and building a feature matrix in combination with the reach channels, historical marketing information and the user response data; training the feature matrix using a machine learning algorithm to obtain a response prediction model that can predict user response data; the response prediction model is used to predict user response data of each reach channel based on the user's portrait data, reach channel data and marketing information;

[0009] Based on the trained response prediction model, predictions are made for users to determine the optimal channels through which different marketing messages can reach them.

[0010] Optionally, after predicting the user according to the trained response prediction model and determining the optimal channel for reaching the user with different marketing information, the method further includes:

[0011] Develop personalized information reach strategies for each user based on the optimal reach channels for different marketing messages.

[0012] Optionally, after predicting the user based on the trained response prediction model and determining the optimal reach channel for different marketing messages for the user, it also includes:

[0013] Continuously collect user response data for effect evaluation;

[0014] The response prediction model is iteratively updated according to the evaluation results.

[0015] Optionally, predicting the user based on the trained response prediction model to determine the optimal channel for reaching the user with different marketing information includes:

[0016] Group existing users according to the trained response prediction model;

[0017] And for different marketing information: predict the response behavior and response enthusiasm of each group of users in different channels; the response behavior and response enthusiasm are both response data; response scores of different channels are calculated based on the response behavior and response enthusiasm; the reach channel with the highest score in each group is used as the optimal reach channel corresponding to each group.

[0018] The second aspect of the present application discloses an information triggering channel optimization system, comprising:

[0019] A determination unit, configured to determine first data of a user, wherein the first data includes: portrait data, contact channel data, marketing information and user response data thereof;

[0020] A training unit is used to extract basic features related to the reach effect from the portrait data, and to construct a feature matrix in combination with the reach channels, historical marketing information and the user response data; the feature matrix is ​​trained using a machine learning algorithm to obtain a response prediction model that can predict user response data; the response prediction model is used to predict user response data of each reach channel based on the user's portrait data, reach channel data and marketing information;

[0021] The prediction unit is used to make predictions about users based on the trained response prediction model and determine the optimal channels for reaching users with different marketing information.

[0022] Optionally, also include:

[0023] The strategy unit is used to formulate personalized information reaching strategies for each user based on the optimal reaching channels of different marketing information.

[0024] Optionally, also include:

[0025] The optimization unit is used to continuously collect user response data for effect evaluation; and iteratively update the response prediction model according to the evaluation results.

[0026] Optionally, the prediction unit is used to predict users based on the trained response prediction model and determine the optimal reach channels for different marketing information of users, specifically for:

[0027] Group existing users according to the trained response prediction model;

[0028] And for different marketing information: predict the response behavior and response enthusiasm of each group of users in different channels; the response behavior and response enthusiasm are both response data; response scores of different channels are calculated based on the response behavior and response enthusiasm; the reach channel with the highest score in each group is used as the optimal reach channel corresponding to each group.

[0029] The third aspect of the present application discloses an electronic device, comprising:

[0030] one or more processors;

[0031] a storage device having one or more programs stored thereon;

[0032] When one or more programs are executed by one or more processors, the one or more processors implement the information-triggered channel optimization method as described in any one of the first aspects of the present application.

[0033] A fourth aspect of the present application discloses a storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the information-triggered channel optimization method as described in any one of the first aspects of the present application is implemented.

[0034] It can be seen from the above technical scheme that the present invention provides an information trigger channel optimization method, including: determining the user's first data; extracting basic features related to the reach effect from the portrait data, and constructing a feature matrix in combination with the reach channels, historical marketing information and its user response data; training the feature matrix using a machine learning algorithm to obtain a response prediction model that can predict user response data; the response prediction model is used to predict user response data of each reach channel based on the user's portrait data, reach channel data and marketing information; predicting the user based on the trained response prediction model to determine the optimal reach channel for different marketing information of the user; that is, by collecting and analyzing the user's response data on different channels, a model that can predict customer response data is trained, thereby guiding the bank to select the optimal reach channel with high reach efficiency, thereby improving marketing efficiency and customer satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0036] Figure 1 is a flow chart of an information triggering channel optimization method provided by an embodiment of the present invention;

[0037] Figure 2 is a flow chart of another information triggering channel optimization method provided by an embodiment of the present invention;

[0038] Figure 3 is a flow chart of another information triggering channel optimization method provided by an embodiment of the present invention;

[0039] Figure 4 is a schematic diagram of an information triggering channel optimization system provided by an embodiment of the present invention;

[0040] Figure 5 It is a schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] In the present application, the term "comprise", "include" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or equipment. In the absence of more restrictions, the elements defined by the sentence "comprise one ..." do not exclude the existence of other identical elements in the process, method, article or equipment including the elements. In addition, the terms "first", "second", "third", "fourth", etc. (if present) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable in appropriate circumstances, so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here, for example.

[0043] The embodiment of the present application provides an information triggering channel optimization method, which is used to solve the problem in the prior art that different users are accustomed to using different channels, resulting in low reach efficiency by using the above-mentioned reach channels.

[0044] See also Figure 1 , which triggers channel optimization methods including:

[0045] S101: Determine first data of a user.

[0046] The first data includes: portrait data, reach channel data, marketing information and user response data.

[0047] It should be noted that marketing reach refers to the behavior of banks actively establishing contact with customers through various channels (such as SMS, WeChat, phone, corporate WeChat, small envelopes from physical banks, etc.) to convey marketing information, promote products or services, and guide customers to participate in marketing activities. Marketing reach is a key link in bank customer relationship management and marketing strategy execution, and directly affects marketing effectiveness and customer satisfaction.

[0048] Reach channels refer to the specific ways or media that banks use to establish contact with customers. Common reach channels include SMS, email, phone, social media, mobile applications, etc. Different channels have different characteristics and applicable scenarios. Banks need to choose appropriate reach channels based on factors such as customer profiles, marketing goals, and resource investment.

[0049] Customer response refers to the reaction or behavior of customers after receiving marketing information from the bank. Customer response can be manifested in reading, clicking, replying, purchasing, etc. of marketing information, or in participating in, sharing, and evaluating marketing activities. The degree, speed, and method of customer response are important indicators for evaluating the effectiveness of marketing reach and customer satisfaction.

[0050] Customer portraits provide a comprehensive, detailed, and multi-dimensional description and analysis of bank customers to form an in-depth understanding of customers. Customer portraits include basic customer information (such as name, gender, age, etc.), consumer behavior (such as purchase history, consumer preferences, etc.), credit records (such as credit scores, repayment history, etc.), social relationships, and other aspects. Through customer portraits, banks can more accurately identify customer needs, predict customer behavior, and thus develop more effective marketing strategies.

[0051] These first data will serve as the basis for subsequent model training and analysis. The portrait data includes basic information, consumption behavior, credit records, etc.; the trigger channel data includes: recording the time, channel, content, etc. of each contact; user response data includes response time, degree of enthusiasm, etc.

[0052] Among them, portrait data is the key to understanding the basic attributes of users, which covers core information such as age, gender, region, occupation, income level, education level, hobbies and consumption habits. This information helps to build user portraits and provide strong support for subsequent precision marketing.

[0053] Secondly, the reach channel data reveals the ways and preferences of users to obtain information. This includes social media platforms, search engines, emails, SMS notifications, APP push notifications, etc. that users often use. By analyzing the reach channel data, you can choose the most appropriate channel to interact with users and improve the information reach and user engagement.

[0054] Thirdly, marketing information refers to various promotional content delivered to users, such as product introductions, promotional activities, coupons, new product previews, etc. These messages need to be carefully designed to ensure that they can attract users' attention and stimulate their desire to buy. At the same time, attention should also be paid to the frequency and timing of sending marketing information to avoid excessively disturbing users or missing the best marketing opportunities.

[0055] Finally, user response data is an important basis for evaluating marketing effectiveness. It records the user's feedback behavior on marketing information, such as clicks, browsing, purchases, sharing, comments, etc. By analyzing user response data, we can understand the user's acceptance and preferences for marketing information, and then adjust the marketing strategy to improve user satisfaction and conversion rate. In addition, different contact channels have their own corresponding user response data.

[0056] In other words, determining the first data of users requires comprehensive consideration of multiple aspects such as portrait data, contact channel data, marketing information and user response data. These data are interrelated and mutually supportive, and together form a solid foundation for determining precise marketing strategies.

[0057] It should be noted that this application is committed to building a detailed and multi-dimensional customer portrait, and this process begins with extensive collection and in-depth analysis of various types of customer data. These data cover multiple levels, including basic customer information (such as name, gender, age and other key details), consumer behavior characteristics (including purchase history, consumer preference tendencies, etc.), credit history background (such as credit score, past repayment record, etc.), and social network information. These data not only intuitively reflect the obvious characteristics of customers, but also deeply explore hidden customer needs and potential value.

[0058] In the technical solution, the construction of customer portraits occupies a core position. It aims to outline a three-dimensional and personalized image for each customer through comprehensive and detailed data collection and analysis. This step is crucial for the subsequent formulation of precise marketing strategies and optimization of customer communication channels.

[0059] The following is an explanation of the process of building a user portrait:

[0060] (1) Data collection stage.

[0061] The construction of customer portraits begins with extensive data collection. The required data may include: basic information, consumer behavior data, credit information, social network data, etc., which will not be listed here one by one.

[0062] Specifically, basic information: such as name, gender, age, occupation, income level, etc., these data provide the basis for the initial classification and positioning of customers.

[0063] Consumer behavior data: Detailed records of customers’ purchase history, including purchase time, frequency, amount, and product category. At the same time, it also pays close attention to customers’ brand loyalty, price sensitivity, and acceptance of new products. These data reveal customers’ consumption habits and patterns.

[0064] Credit information: A customer’s credit score, repayment record, loan history, etc. These are key indicators for assessing their credit status and financial stability, and have a significant impact on risk management and credit decisions.

[0065] Social network data: The customer’s social circles, interactions and performance on social media. These data can reveal the customer’s social influence, interests and lifestyle.

[0066] (2) Data analysis phase.

[0067] The data collected in the above stages needs to be deeply analyzed and mined to extract valuable information to support the construction of customer portraits. The main analysis methods include data cleaning, feature extraction, correlation analysis and cluster analysis.

[0068] Specifically, data cleaning: ensure the purity and accuracy of data, and remove redundant, erroneous or incomplete data. Feature extraction: extract meaningful features from raw data, such as calculating average consumption, purchase frequency, etc. through consumer behavior data. Correlation analysis: explore potential connections between different data sets, such as analyzing the correlation between purchase history and credit record. Cluster analysis: group customers with similar characteristics into the same group to form different customer segments.

[0069] (3) Based on data analysis, customer profiles are constructed.

[0070] This is a model that integrates multi-dimensional data such as basic customer information, consumption behavior, credit record, and social network. The model not only shows the explicit characteristics of customers (such as age, gender, etc.), but also deeply explores the hidden needs and potential value of customers (such as consumption potential, credit rating, social influence, etc.).

[0071] By constructing such a comprehensive and multi-dimensional customer portrait, the present invention can more deeply understand the uniqueness and needs of each customer, and thus provide strong support for the formulation of accurate marketing strategies. For example, based on the customer's purchase history and consumption preferences, the potential demand for specific products is predicted; based on the customer's credit record and financial status, the credit risk is assessed, and risk control strategies are formulated accordingly.

[0072] S102. Extract basic features related to the reach effect from the portrait data, and build a feature matrix based on the reach channels, historical marketing information, and user response data; use a machine learning algorithm to train the feature matrix to obtain a response prediction model that can predict user response data.

[0073] Among them, the response prediction model is used to predict user response data of each reach channel based on user portrait data, reach channel data and marketing information.

[0074] The basic characteristics related to the reach effect include: at least one of age, gender, income level and consumption habits. Of course, other information may also be included, which will not be listed here one by one. It depends on the actual situation and is within the scope of protection of this application.

[0075] Machine learning is a data-driven algorithm and model that improves the accuracy of predictions or decisions through automatic learning and optimization. In marketing reach optimization, machine learning algorithms can process large amounts of customer data and marketing data, automatically discover patterns and associations in the data, and generate models that can predict customer response behaviors. These models can be applied to new data to guide banks in selecting the best reach channels and developing personalized marketing strategies. Common machine learning algorithms include decision trees, random forests, neural networks, etc.

[0076] The machine learning algorithm includes: at least one of random forest, gradient boosting tree and neural network; they will not be described one by one here.

[0077] Specifically, we conduct an in-depth analysis of the portrait data to extract basic features that are closely related to the reach effect. These features can intuitively reflect the user's potential response tendency in different reach scenarios. Then, we combine the specific reach channel information, past marketing records, and the user's actual response data to build a comprehensive and detailed feature matrix. This matrix not only covers the basic attributes of the user, but also incorporates multi-dimensional information such as channel characteristics and historical marketing effects.

[0078] Next, we use advanced machine learning algorithms to deeply train the constructed feature matrix. This process aims to explore the potential connections and patterns between data, so as to train an efficient response prediction model. This model has strong prediction capabilities and can accurately predict the user's response data on different reach channels based on the user's portrait data, the specific situation of the reach channel, and the current marketing information.

[0079] The response prediction model has a wide range of application scenarios. It can quickly generate user response prediction results for each contact channel based on the input user portrait, contact channel and marketing content. These prediction data are of great reference value for enterprises to formulate marketing strategies, optimize contact methods and improve marketing effects.

[0080] The following is an explanation of the training of the response prediction model:

[0081] Based on detailed customer portrait data, this application adopts advanced machine learning algorithms (covering decision trees, random forests, neural networks and other technologies) to build a response prediction model. This model has excellent data insight capabilities, can independently mine and identify potential patterns and correlations in the data, and then accurately estimate customers' response tendencies when facing different marketing methods, such as key indicators such as click-through rate and purchase conversion rate. With this model, banking institutions can more accurately target customer groups that show high potential for interest in specific marketing information, and tailor personalized marketing blueprints accordingly.

[0082] In the core structure of the technical solution, the response prediction model plays a pivotal role. It aims to rely on machine learning algorithms and, based on the solid foundation of existing customer portraits, proactively predict customer response patterns to diversified marketing strategies. To achieve this goal, the present invention specifically introduces a long short-term memory (LSTM) neural network model, which shows extraordinary adaptability in processing sequence data and can effectively capture the long-term dependency characteristics between data.

[0083] LSTM, as an advanced version of recurrent neural network (RNN), cleverly solves the problem of gradient vanishing or exploding faced by traditional RNN when processing long sequence data through built-in gating mechanism (input gate, forget gate and output gate) and memory unit. Its core strategy is to be able to intelligently choose to retain or forget historical information, so as to accurately capture the long-term dependencies in sequence data. In the response prediction scenario, the customer's reaction pattern is often affected by multiple factors such as their past behavior, the specific content of marketing information, and the communication channel, which together constitute a complex time series. With its powerful processing capabilities, the LSTM model can efficiently parse such time series data and reveal the hidden patterns and correlations therein.

[0084] The following is a detailed description of the implementation process of the LSTM model in the field of response prediction:

[0085] First, before using the LSTM model, data preprocessing and feature engineering are required. This includes encoding and converting information such as customer profiles, marketing information content, and release channels into numerical forms that can be interpreted by the model. At the same time, it is also necessary to mark the customer's response status and its level of positivity based on historical data to provide strong support for subsequent model training.

[0086] Secondly, in the training phase, the preprocessed data is fed into the LSTM model. The model learns deeply about the patterns and correlations in historical data, and autonomously adjusts internal weights and parameter configurations to minimize prediction errors. During the training process, optimization algorithms such as gradient descent can be used to accelerate the convergence of the model.

[0087] Then, after the model training is successfully completed, it can be put into practice for response prediction. Faced with new marketing information and customer profile data, the model will make accurate predictions based on the patterns and correlations it has learned, and output the probability of the customer's response to the marketing information or the positive evaluation.

[0088] Finally, to fully evaluate the predictive performance of the model, a test data set can be used for verification. By comparing the model's prediction results with the actual response data, key indicators such as accuracy, recall, and F1 score are calculated to scientifically measure the performance of the model. If the model performance fails to meet the expected standards, the model can be continuously optimized by adjusting model parameters, expanding the training data set, or trying other machine learning algorithms.

[0089] In other words, the successful application of the LSTM model in the field of response prediction fully demonstrates its powerful ability to utilize characteristics such as customer historical behavior information, marketing information content, and release channels. By accurately capturing hidden patterns and correlations in the data, the LSTM model can provide customers with accurate response predictions for different marketing reach strategies, thereby helping banking institutions to develop more accurate and personalized marketing strategies, significantly improving marketing effectiveness and customer satisfaction.

[0090] S103. Predict users based on the trained response prediction model and determine the optimal channels for reaching different marketing information to users.

[0091] After successfully training an efficient response prediction model, it is applied to the in-depth prediction of user behavior. This step aims to use the power of the model to find the best contact channel for users under different marketing messages, thereby maximizing the marketing effect.

[0092] Specifically, based on the user's portrait data, including their basic attributes, consumption habits, preferences, etc., combined with the marketing information content currently planned to be pushed, these data are input into the trained response prediction model. The model will quickly process and analyze these data to predict the user's response probability and possible response behavior on different contact channels (such as social media, email, SMS notifications, phone calls, APP push, etc.).

[0093] After obtaining these prediction results, the prediction effects of different reach channels are comprehensively evaluated and compared. By comparing the differences in user response probabilities and response behaviors on different channels, it is possible to determine what the optimal reach channel is for the current user and marketing information.

[0094] This step is crucial to improving marketing effectiveness. By accurately finding the best channels to reach users, you can ensure that marketing information can reach users in the most effective way, thereby increasing user attention, participation, and ultimately purchase conversion rates. At the same time, this can also help companies better allocate marketing resources and improve overall marketing efficiency and ROI.

[0095] In short, using a trained response prediction model to predict users and determine the optimal reach channels for different marketing messages to users is a key step in improving marketing effectiveness and optimizing resource allocation.

[0096] In this embodiment, the first data of the user is determined; the basic features related to the reach effect are extracted from the portrait data, and a feature matrix is ​​constructed in combination with the reach channels, historical marketing information and user response data; the feature matrix is ​​trained using a machine learning algorithm to obtain a response prediction model that can predict user response data; the response prediction model is used to predict the user response data of each reach channel based on the user's portrait data, reach channel data and marketing information; based on the trained response prediction model, the user is predicted to determine the optimal reach channel for different marketing information of the user; that is, by collecting and analyzing the user's response data on different channels, a model that can predict customer response data is trained, thereby guiding the bank to select the optimal reach channel with high reach efficiency, thereby improving marketing efficiency and customer satisfaction.

[0097] It should be noted that although existing marketing reach technologies can work to a certain extent, they still face a series of significant problems and limitations. Solutions based on preset rules are often limited by limited personalized options and cannot fully meet the unique and changing needs and preferences of each customer. At the same time, traditional statistical analysis methods are powerless when dealing with complex and multi-dimensional customer data, and it is difficult to accurately capture and predict customer response behaviors. In addition, due to the lack of a scientific and effective channel effect evaluation system, banks are often confused when deciding which channel is most effective for specific customers. More importantly, in today's era of rapidly changing market environment and customer behavior, solutions based on static rules and historical data often find it difficult to quickly adapt to these changes, resulting in a significant reduction in the effectiveness of marketing strategies.

[0098] In response to the above challenges, this application aims to fundamentally solve the shortcomings of existing solutions in personalization, response prediction, channel selection and adaptability. Specifically, first, by comprehensively collecting and analyzing the diversified data of customers (covering basic information, consumption behavior, social relationships and other dimensions), a detailed and rich customer portrait is constructed; this portrait enables marketing strategies to more accurately fit customer needs and preferences, thereby achieving a significant improvement in personalization. Secondly, using advanced machine learning algorithms, massive historical data is deeply processed and analyzed; the algorithm automatically mines hidden patterns and associations in the data and generates a model that can accurately predict customer response behavior; the introduction of this model has achieved a qualitative leap in response prediction. Furthermore, based on the prediction results of the model, dynamic selection of marketing channels is achieved; for each customer, the channel with the best reach effect can be found, thereby greatly improving marketing efficiency and customer satisfaction; this move marks an important breakthrough in channel selection optimization. Finally, because the machine learning model has the ability to self-learn and update, the solution can flexibly adapt to changes in the market environment and customer behavior. This enables marketing strategies to always maintain their effectiveness and competitiveness, and achieves a significant enhancement of adaptability.

[0099] In summary, by introducing machine learning technology, this application has achieved significant improvement and optimization in customer portrait construction, response prediction, channel selection, strategy adaptability, etc. This innovative solution not only effectively solves the problems and limitations of the current general solutions in the industry, but also opens up a new path for the development of customer marketing reach.

[0100] Optionally, after predicting the user based on the trained response prediction model and determining the optimal reach channel for different marketing messages for the user, it also includes:

[0101] Develop personalized information reach strategies for each user based on the optimal reach channels for different marketing messages.

[0102] Based on the in-depth analysis and accurate prediction of the optimal user contact channels under different marketing information, a personalized information contact strategy can be tailored for each user. The formulation process of this strategy fully considers the unique profile characteristics of users, preferred channels and the specific content of current marketing information, aiming to ensure that marketing information can reach users in the most appropriate way and at the best time, thereby maximizing user attention and participation.

[0103] When developing a personalized information reach strategy, follow these core principles:

[0104] Accurate matching: Based on the user's profile data and model prediction results, we find the most suitable contact channel for each marketing message. This ensures that the information can be directly delivered to the user's "interest point" and improve the opening rate and reading rate of the information.

[0105] Time optimization: By combining the user's historical active time and current behavior pattern, we can select the time period when the user is most likely to be online and concerned for information push. This helps to improve the immediate visibility of information and user engagement.

[0106] Content customization: On the basis of maintaining the core elements of marketing information, fine-tuning is performed according to user preferences and interests to make the information more personalized to the user's needs. This helps to enhance the appeal of the information and the resonance of users.

[0107] Frequency control: Reasonably control the frequency of information push to avoid excessively disturbing users. Use intelligent algorithms to evaluate user response and fatigue, dynamically adjust the push rhythm, and ensure that information is continuously and effectively delivered.

[0108] Effect monitoring and feedback: After implementing a personalized information reach strategy, continuously monitor user responses and marketing results. By collecting and analyzing user feedback data, continuously optimize strategies and improve reach efficiency and conversion rate.

[0109] In summary, by formulating a personalized information reach strategy, we can ensure that marketing information reaches users in the most appropriate way, increase user attention and participation, and thus improve overall marketing effectiveness and ROI. The formulation process of this strategy not only reflects a deep understanding and respect for user needs, but also demonstrates the company's innovation and refined operational capabilities in the field of digital marketing.

[0110] It should be noted that this application emphasizes the importance of adaptive adjustment of strategies. In the face of a rapidly changing market environment and increasingly diverse customer behaviors, the flexibility of marketing strategies has become a key factor in determining their effectiveness. Therefore, this application not only builds a mechanism that can regularly update customer portrait data, but also establishes a retraining process for responsive prediction models to ensure that marketing strategies can keep up with market trends and accurately capture every subtle change in customer needs. This forward-looking dynamic adjustment mechanism ensures that marketing strategies always keep pace with market and customer needs, thereby maintaining a leading position in the competition.

[0111] In summary, this application has successfully created a highly personalized, accurate and continuously optimized marketing reach system by deeply integrating machine learning technology and multi-dimensional customer data analysis. Compared with some solutions, this application has achieved significant breakthroughs and innovations in the detailed portrayal of customer portraits, the high accuracy of response predictions, the intelligence of channel selection, and the flexibility of strategy adjustment. These advantages not only enable the present invention to better respond to the diversified needs of banks and customers, but also greatly improve the execution efficiency of marketing activities and the overall satisfaction of customers, injecting new vitality and possibilities into the development of the banking industry.

[0112] Optionally, after predicting the user based on the trained response prediction model and determining the optimal reach channel for different marketing messages for the user, it also includes:

[0113] Continuously collect user response data for effect evaluation; iteratively update the response prediction model based on the evaluation results.

[0114] Specifically, it is possible to comprehensively monitor various response behaviors of users after receiving marketing information, including but not limited to clicks, shares, comments, purchases, etc. These behavioral data provide valuable user feedback and help to gain a deeper understanding of users' acceptance and preferences for information.

[0115] In order to ensure the timeliness and accuracy of the data, a real-time data update mechanism can also be established. Whenever new user response data is generated, the system will immediately collect and organize it to provide the latest data support for subsequent analysis and evaluation.

[0116] Based on the collected user response data, in-depth analysis is conducted from multiple dimensions, including but not limited to key indicators such as user activity, conversion rate, satisfaction, etc. These analysis results provide intuitive feedback on the effectiveness of marketing strategies.

[0117] Based on the results of the effect evaluation, the response prediction model will be iteratively updated. During this process, new user response data will be incorporated into the model training to further improve the model's prediction accuracy and generalization ability. At the same time, the feature selection and parameter settings in the model will be optimized based on user feedback and market changes to ensure that the model can better adapt to the future marketing environment.

[0118] Based on the iterative update of the model, the personalized information contact strategy will be continuously optimized in combination with the latest market trends and user preferences. This includes adjusting the contact channels, optimizing the content presentation method, adjusting the push frequency, etc., aiming to continuously improve user participation and satisfaction.

[0119] In other words, by continuously collecting user response data for effect evaluation and iteratively updating the response prediction model based on the evaluation results, the effectiveness and adaptability of the personalized information reach strategy can be ensured. This approach not only helps to improve marketing effectiveness, but also brings long-term competitive advantages and sustainable development capabilities to enterprises.

[0120] Optionally, predict users based on the trained response prediction model to determine the optimal channels for reaching users with different marketing messages, including:

[0121] Group existing users according to the trained response prediction model; and for different marketing information: predict the response behavior and response positivity of each group of users in different channels; both response behavior and response positivity are response data; score the response of different channels based on response behavior and response positivity; take the highest-scoring reach channel for each group as the optimal reach channel for each group.

[0122] It should be noted that the grouping is not based solely on user portraits, but is based on a trained response prediction model; the response prediction model is trained on portrait data, reach channels, historical marketing information and user response data, that is, the grouping is related to portrait data, reach channels, historical marketing information and user response data.

[0123] After having a fully trained response prediction model, we further scientifically grouped existing users to improve marketing efficiency through refined operations. This grouping process is based on user portrait features, historical behavior data, and model prediction results, ensuring that users in the group have similar marketing response patterns and preferences.

[0124] User grouping: Using advanced clustering algorithms, users are divided into multiple groups with similarities based on their profile features, historical response behaviors, and potential response patterns predicted by the model. This step lays the foundation for subsequent precision marketing and personalized outreach.

[0125] Response prediction: For each user group, we use the trained response prediction model to make in-depth predictions on different marketing messages. The predictions include the response behavior and response positivity of each group of users on different contact channels. These response data provide direct evidence of user reactions to different marketing stimuli.

[0126] In order to quantify the response effect of users on different channels, the response score indicator is introduced. This score is based on a comprehensive evaluation of the user's response behavior and response positivity on different channels, and aims to reflect the user's preference and response intensity for different channels.

[0127] Based on the response scores, the highest scoring reach channel is determined for each group of users, which is the optimal reach channel for that group of users. This step ensures that marketing information can reach users in the most efficient way, thereby maximizing marketing effectiveness.

[0128] On the basis of determining the optimal contact channel, the personalized information contact strategy was further adjusted and optimized in combination with the characteristics and preferences of the user groups. This includes content customization, push frequency, interaction methods and other aspects, aiming to further improve user participation and satisfaction.

[0129] It should be noted that when a new user appears, the user can be directly predicted and the optimal reach channel and marketing strategy can be determined without grouping. Of course, it is not ruled out that the optimal reach channel and marketing strategy can be determined after grouping. I will not go into details here. It depends on the actual situation and is within the protection scope of this embodiment.

[0130] In this embodiment, through a series of steps such as user grouping, response prediction, response score calculation, and optimal reach channel determination, accurate prediction of user marketing responses and personalized reach are achieved. This strategy not only improves marketing efficiency, but also brings users a marketing experience that is more in line with their needs and preferences.

[0131] like Figure 2 As shown, one implementation form of the information trigger channel optimization method is described:

[0132] 1) Customer data collection. The collected data includes basic information and consumption behavior.

[0133] 2) Customer portrait construction: Conduct multi-dimensional data analysis based on the collected data to construct a multi-dimensional customer portrait.

[0134] 3) Response prediction model: Apply customer profile data to the response prediction model; a machine learning algorithm is used to build the response prediction model.

[0135] 4) Channel optimization selection: The prediction structure is obtained based on the response prediction model, and the best channel is selected based on the prediction structure.

[0136] 5) Strategy execution: Send marketing information through the selected channels.

[0137] 6) Strategy adaptive adjustment: According to the feedback after the implementation of the strategy, the strategy is adjusted; at the same time, the data (user data collection part) and model (response prediction model) are updated regularly.

[0138] That is, marketing strategies can be adjusted dynamically based on market changes and customer feedback.

[0139] The following example illustrates the information trigger channel optimization method:

[0140] To help you understand this solution, let's use a simple example to describe the entire process. Suppose you want to send different marketing messages to different types of customers through different contact channels (such as SMS, email, and App push) and predict which combination is most likely to be successful. Figure 3 shown.

[0141] ①. Data collection and preparation.

[0142] First, historical data needs to be collected, including:

[0143] Customer information: For example, the customer’s age, income, credit score, etc.

[0144] Marketing information: the specific content sent, such as preferential interest rates, cash rebates, etc.

[0145] Reach channels: such as SMS, email, App push, etc.

[0146] Response data: whether the customer clicked the link, made a purchase, etc.

[0147] Table 1: Example data

[0148] Customer ID age income Credit score Marketing Information Reach Channel Click Buy or not 001 30 high excellent Preferential interest rates Short message yes yes 002 45 middle good Cashback mail no no … … … … … … … …

[0149] ②. Feature engineering and data encoding.

[0150] Next, the data needs to be processed to convert the text information into a numerical form that the model can understand. For example:

[0151] Divide age groups into: under 30, 30-45, and over 45.

[0152] Income: low, medium, high.

[0153] Marketing information and reach channels: Use One-Hot Encoding.

[0154] Table 2: Encoded data example

[0155] Customer ID age income Credit score Marketing Information Reach Channel Click Buy or not 001 30 high excellent Preferential interest rates Short message yes yes 002 45 middle good Cashback mail no no … … … … … … … …

[0156] ③. Model training.

[0157] Use LSTM model (or other suitable model) for training.

[0158] The training data will include customer characteristics (age group, income, credit score), marketing information, and contact channels, with the goal of predicting whether the customer will click and purchase.

[0159] ④. Model evaluation and optimization.

[0160] Set aside a portion of the data as a test set and evaluate the performance of the model on this data. Based on the evaluation results, adjust the model parameters or try other models to optimize performance.

[0161] ⑤. Response prediction and decision making.

[0162] Once the model is trained and optimized, it can be used to predict how new customers will respond to different marketing messages and reach channels. For example:

[0163] For a customer under 30 years old with a medium income and a good credit score, the model may predict that he has a higher probability of clicking and purchasing the preferential interest rate information received through App push.

[0164] Based on these predictions, banks can develop personalized marketing strategies and choose the most appropriate channels and marketing messages to reach different types of customers.

[0165] ⑥. Implementation and monitoring.

[0166] Finally, implement marketing strategies based on the model’s predictions and continuously monitor the actual results. If the actual results deviate significantly from the predictions, you may need to reassess data quality, adjust the model, or update the marketing strategy.

[0167] It should be noted that in the field of bank marketing reach, the solutions commonly used in the industry are mainly based on rule engines and traditional statistical analysis methods. These solutions usually rely on pre-set rules and historical data to analyze customer behavior, and select corresponding reach channels and strategies based on the analysis results. For example, banks may formulate marketing rules based on basic information such as customer age, gender, income, and past purchase records, and push them in batches through channels such as SMS and email.

[0168] Some existing marketing reach systems are mainly based on data analysis and customer segmentation. These systems usually collect data such as basic customer information, transaction history, and purchase preferences, use statistical methods and business rules to segment customers, and formulate marketing strategies accordingly. Specific steps include: building customer portraits, grouping customers through methods such as cluster analysis, designing marketing information and reach channels for different groups, and executing marketing and evaluating the results. Although these solutions have achieved personalization and efficiency improvement in marketing to a certain extent, they are still limited by traditional statistical methods and fixed rules, making it difficult to deal with large-scale, high-dimensional data, and have deficiencies in predicting customer responses and dynamically optimizing channels. These technical solutions have achieved personalization and targeting of marketing to a certain extent, and improved marketing efficiency. However, they still rely on traditional statistical methods and fixed business rules, making it difficult to process large-scale, high-dimensional data, and have limitations in predicting customer responses and dynamically optimizing reach channels.

[0169] In this embodiment, the invention significantly improves the performance of existing technologies by leveraging machine learning technology, especially its excellent ability to analyze complex data patterns and make predictions. It can not only accurately predict customer response behaviors, but also dynamically adjust the contact channels based on real-time data, thereby greatly improving marketing efficiency and customer satisfaction.

[0170] In other words, existing technologies are limited by fixed business rules and statistical methods, and it is difficult to quickly adapt to changes in the market environment and customer behavior. Once customer preferences or market trends change, these systems often cannot adjust strategies quickly, resulting in weakened marketing effects.

[0171] In contrast, this application introduces a machine learning model that can learn and adapt to dynamic changes in market and customer data in real time. The model updates itself to ensure that marketing strategies always remain timely and accurate.

[0172] In addition, customer segmentation using existing technologies is mostly based on limited data dimensions, which makes it difficult to fully capture the subtle differences and deep-seated needs of customers, thus limiting the personalization and accuracy of marketing strategies.

[0173] This application uses rich customer portrait data and advanced machine learning algorithms to achieve more detailed and multi-dimensional customer segmentation. This not only enables marketing strategies to more accurately match customer needs, but also significantly improves marketing effectiveness.

[0174] At the same time, when it comes to predicting customer responses, existing technologies typically rely on simple statistical models or rules of thumb, with limited predictive accuracy and reliability.

[0175] This application uses advanced machine learning algorithms to build a response prediction model that can handle complex data patterns and nonlinear relationships and more accurately predict customer response behavior.

[0176] Finally, existing technologies often lack scientific basis when selecting reach channels and rely more on experience or intuition.

[0177] This application uses a machine learning model to quantify and compare the reach effects of different channels, and can scientifically select the channel that best reaches each customer, thereby improving marketing efficiency and optimizing customer experience.

[0178] In summary, this application is committed to solving the problems of inconsistent customer responses and low reach efficiency encountered by banks in the process of marketing reach. At present, banks use a variety of channels (such as SMS, WeChat, telephone, corporate WeChat, and mobile banking envelopes, etc.) to implement marketing reach, but the reach effectiveness of each channel varies greatly, and there are significant differences in the timeliness and enthusiasm of customer responses. This situation makes it difficult for banks to accurately determine which channel is most effective for specific customers, which in turn affects the efficiency of marketing activities and customer satisfaction. The root of the problem lies in the lack of in-depth analysis and accurate prediction of customer response behavior. Different customers have different preferences and response patterns for different channels, and banks currently lack effective tools to accurately identify these differences and optimize reach strategies accordingly.

[0179] See also Figure 4 Another embodiment of the present application provides an information triggering channel optimization system, including:

[0180] The determination unit 101 is used to determine the first data of the user, and the first data includes: portrait data, contact channel data, marketing information and user response data.

[0181] The training unit 102 is used to extract basic features related to the reach effect from the portrait data, and to construct a feature matrix in combination with the reach channels, historical marketing information and user response data; the feature matrix is ​​trained using a machine learning algorithm to obtain a response prediction model that can predict user response data; the response prediction model is used to predict user response data of each reach channel based on the user's portrait data, reach channel data and marketing information.

[0182] The prediction unit 103 is used to predict users based on the trained response prediction model and determine the optimal contact channels for different marketing information of users.

[0183] Optionally, also include:

[0184] The strategy unit is used to formulate personalized information reach strategies for each user based on the optimal reach channels for different marketing messages.

[0185] Optionally, also include:

[0186] The optimization unit is used to continuously collect user response data for effect evaluation; and iteratively update the response prediction model based on the evaluation results.

[0187] Optionally, the prediction unit is used to predict users based on the trained response prediction model and determine the optimal reach channels for different marketing messages for users, specifically for:

[0188] Group existing users according to the trained response prediction model;

[0189] And for different marketing information: predict the response behavior and response enthusiasm of each group of users in different channels; the response behavior and response enthusiasm are both response data; the response scores of different channels are based on the response behavior and response enthusiasm; the reach channel with the highest score in each group is used as the optimal reach channel corresponding to each group.

[0190] For details on the specific working process and principle of each of the above modules, please refer to the information-triggered channel optimization method provided in the above embodiment, which will not be described here one by one. It depends on the actual situation and is within the protection scope of this application.

[0191] In this embodiment, the determination unit 101 is used to determine the first data of the user, and the first data includes: portrait data, reach channel data, marketing information and its user response data; the training unit 102 is used to extract basic features related to the reach effect from the portrait data, and construct a feature matrix in combination with the reach channels, historical marketing information and its user response data; the feature matrix is ​​trained using a machine learning algorithm to obtain a response prediction model that can predict user response data; the response prediction model is used to predict user response data of each reach channel based on the user's portrait data, reach channel data and marketing information; the prediction unit 103 is used to predict the user based on the trained response prediction model, and determine the optimal reach channel for different marketing information of the user; that is, by collecting and analyzing the user's response data on different channels, a model that can predict customer response data is trained, thereby guiding the bank to select the optimal reach channel with high reach efficiency, thereby improving marketing efficiency and customer satisfaction.

[0192] Another embodiment of the present application provides a storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the information-triggered channel optimization method as described in any one of the above embodiments is implemented.

[0193] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0194] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0195] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0196] Another embodiment of the present invention provides an electronic device, such as Figure 5 As shown, including:

[0197] One or more processors 201.

[0198] The storage device 202 stores one or more programs.

[0199] When one or more programs are executed by one or more processors 201, one or more processors 201 implement the information-triggered channel optimization method as described in any one of the above embodiments.

[0200] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart.

[0201] Although the subject matter has been described in language specific to structural features and / or methodological logical actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims.

[0202] Although several specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Certain features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination.

[0203] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

[0204] The features recorded in the various embodiments in this specification can be replaced or combined with each other, and the same and similar parts between the various embodiments can refer to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment. The system and system embodiments described above are only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative work.

[0205] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0206] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for optimizing information triggering channels, characterized in that: include: Determine first data of the user, the first data including: portrait data, contact channel data, marketing information and user response data; Extracting basic features related to the reach effect from the portrait data, and building a feature matrix in combination with the reach channels, historical marketing information and the user response data; training the feature matrix using a machine learning algorithm to obtain a response prediction model that can predict user response data; the response prediction model is used to predict user response data of each reach channel based on the user's portrait data, reach channel data and marketing information; Based on the trained response prediction model, predictions are made for users to determine the optimal channels through which different marketing messages can reach them.

2. The information trigger channel optimization method according to claim 1, characterized in that: After predicting the user based on the trained response prediction model and determining the optimal channel for reaching the user with different marketing information, the method further includes: Develop personalized information reach strategies for each user based on the optimal reach channels for different marketing messages.

3. The information trigger channel optimization method according to claim 1, characterized in that: After predicting users based on the trained response prediction model and determining the optimal channels for reaching users with different marketing messages, it also includes: Continuously collect user response data for effect evaluation; The response prediction model is iteratively updated according to the evaluation results.

4. The information trigger channel optimization method according to claim 1, characterized in that: The method predicts the user based on the trained response prediction model and determines the optimal channel for reaching the user with different marketing information, including: Group existing users according to the trained response prediction model; And for different marketing information: predict the response behavior and response enthusiasm of each group of users in different channels; the response behavior and response enthusiasm are both response data; response scores of different channels are calculated based on the response behavior and response enthusiasm; the reach channel with the highest score in each group is used as the optimal reach channel corresponding to each group.

5. An information trigger channel optimization system, characterized in that: include: A determination unit, configured to determine first data of a user, wherein the first data includes: portrait data, contact channel data, marketing information and user response data thereof; A training unit is used to extract basic features related to the reach effect from the portrait data, and to construct a feature matrix in combination with the reach channels, historical marketing information and the user response data; the feature matrix is ​​trained using a machine learning algorithm to obtain a response prediction model that can predict user response data; the response prediction model is used to predict user response data of each reach channel based on the user's portrait data, reach channel data and marketing information; The prediction unit is used to make predictions about users based on the trained response prediction model and determine the optimal channels for reaching users with different marketing information.

6. The information trigger channel optimization system according to claim 5, characterized in that: Also includes: The strategy unit is used to formulate personalized information reaching strategies for each user based on the optimal reaching channels of different marketing information.

7. The information trigger channel optimization system according to claim 5, characterized in that: Also includes: The optimization unit is used to continuously collect user response data for effect evaluation; and iteratively update the response prediction model according to the evaluation results.

8. The information trigger channel optimization system according to claim 5, characterized in that: The prediction unit is used to predict users based on the trained response prediction model and determine the optimal channel for reaching users with different marketing information. Specifically, it is used to: Group existing users according to the trained response prediction model; And for different marketing information: predict the response behavior and response positivity of each group of users in different channels; the response behavior and response positivity are both response data; response scores of different channels based on the response behavior and response positivity; The reach channel with the highest score in each group is used as the optimal reach channel for each group.

9. An electronic device, characterized in that: include: one or more processors; a storage device having one or more programs stored thereon; When one or more programs are executed by one or more processors, the one or more processors implement the information triggering channel optimization method as described in any one of claims 1-4.

10. A storage medium, characterized in that: A computer program is stored thereon, wherein when the computer program is executed by a processor, the information triggering channel optimization method as described in any one of claims 1-4 is implemented.

Images