Message template determination method, apparatus, device, and medium

Abstract

This specification provides a message template determination method, apparatus, device, and medium. The method includes: determining an initial transaction message template based on the behavior prediction results of a test user group, thereby obtaining an initial transaction message; performing interactive testing on the initial transaction message, adjusting the initial transaction message template based on user comment data, and obtaining a target transaction message template; and matching the target transaction message template with a target user group to push the target transaction message to the target user group.

Description

Technical Field

[0001] This specification relates to the field of computer technology, and more specifically, to a method, apparatus, device, and medium for determining a message template in the field of computer technology. Background Technology

[0002] In the digital economy era, marketing, as a core strategy for enterprises to enhance market competitiveness, helps them promote products, increase sales, and expand market share. Traditional marketing methods focus more on user management and product promotion, neglecting the real needs of users, resulting in unsatisfactory marketing effects. Therefore, it is necessary to provide marketing methods that can meet the personalized needs of users. Summary of the Invention

[0003] This specification provides a method, apparatus, device, and medium for determining message templates. This method can accurately capture changes in user needs, meet personalized user requirements, and improve marketing efficiency and market responsiveness. It achieves precise matching and delivery of targeted transaction messages, ensuring that the target user group receives personalized messages.

[0004] Firstly, embodiments of this specification provide a method for determining a message template, the method comprising:

[0005] Obtain behavioral prediction results for the test user group, and determine the initial transaction message template based on the behavioral prediction results;

[0006] The initial transaction message is obtained by filling the initial transaction message template with user information data of the test user group;

[0007] Perform interactive tests on the initial transaction message to obtain the interaction parameter values ​​of the initial transaction message;

[0008] If the interaction parameter value is greater than or equal to the preset interaction parameter threshold, the initial transaction message template is adjusted based on the user comment data of the test user group on the initial transaction message to obtain the target transaction message template.

[0009] Secondly, embodiments of this specification provide a message push method, which includes:

[0010] Identify the first target user tag that matches the transaction tag of the target transaction message template from the user tags of the first target user group;

[0011] The first target user group corresponding to the first target user tag is identified as the second target user group, and the target user information data of the second target user group is obtained.

[0012] The target transaction message is determined based on the target user information data and the target transaction message template, and then pushed to the second target user group.

[0013] Thirdly, embodiments of this specification provide a message template determining device, the device comprising:

[0014] The initial message template acquisition unit is used to acquire the behavior prediction results for the test user group and determine the initial transaction message template based on the behavior prediction results.

[0015] The initial message acquisition unit is used to fill the initial transaction message template based on the user information data of the test user group to obtain the initial transaction message.

[0016] The message testing unit is used to perform interactive tests on the initial transaction message and obtain the interaction parameter values ​​of the initial transaction message;

[0017] The target message template acquisition unit is used to adjust the initial transaction message template based on user comment data of the test user group on the initial transaction message if the interaction parameter value is greater than or equal to the preset interaction parameter threshold, so as to obtain the target transaction message template.

[0018] Fourthly, embodiments of this specification provide a message push device, which includes:

[0019] The tag matching unit is used to determine the first target user tag that matches the transaction tag of the target transaction message template from the user tags of the first target user group;

[0020] The user information acquisition unit is used to determine the first target user group corresponding to the first target user tag as the second target user group and acquire the target user information data of the second target user group.

[0021] The message push unit is used to determine the target transaction message based on the target user information data and the target transaction message template, and push the target transaction message to the second target user group.

[0022] Fifthly, embodiments of this specification provide a computer device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the method described above.

[0023] Sixthly, embodiments of this specification provide a storage medium storing a computer program that, when executed by a processor, implements the steps of the method described above.

[0024] In a seventh aspect, embodiments of this specification provide a computer program product, comprising: a computer program that, when executed by a processor of a computer device, enables the processor to at least implement the method described in the first aspect.

[0025] In the embodiments of this specification, an initial transaction message template is determined based on the behavioral prediction results of the test user group, thereby obtaining the initial transaction message. Interactive testing is performed on the initial transaction message, and the initial transaction message template is adjusted based on user comment data to obtain the target transaction message template. The target transaction message template is matched with the target user group to push the target transaction message to the target user group. This method accurately captures changes in user needs, can meet users' personalized needs, and improves marketing efficiency and market responsiveness. Simultaneously, it achieves accurate matching and pushing of target transaction messages, ensuring that the target user group receives personally tailored target transaction messages. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This is a flowchart illustrating a message template determination method provided in the embodiments of this specification;

[0028] Figure 2 This is a flowchart illustrating a message template determination method provided in the embodiments of this specification;

[0029] Figure 3 This is a flowchart illustrating a message push method provided in an embodiment of this specification;

[0030] Figure 4 This is a flowchart illustrating a message push method provided in an embodiment of this specification;

[0031] Figure 5 This is a flowchart illustrating a method for determining a message template and pushing a message, as provided in an embodiment of this specification.

[0032] Figure 6 This is a schematic diagram of the structure of a message template determining device provided in the embodiments of this specification;

[0033] Figure 7 This is a schematic diagram of the structure of a message push device provided in the embodiments of this specification;

[0034] Figure 8 This is a schematic diagram of the structure of a computer device provided in the embodiments of this specification. Detailed Implementation

[0035] The technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this specification.

[0036] The message template determination and message push method provided in the embodiments of this specification are applicable to scenarios where enterprises promote their products to users. In order to promote products and attract customers to click and purchase, enterprises push transaction messages to users' terminal devices. Terminal devices include, but are not limited to, users' personal computers, laptops, mobile phones, wearable devices and other electronic devices. Transaction messages may include basic information such as product prices and promotional offers.

[0037] This specification provides a method for determining message templates. An initial transaction message template is determined based on the behavioral prediction results of a test user group, thereby obtaining an initial transaction message. Interactive testing is performed on the initial transaction message, and the initial transaction message template is adjusted based on user comment data to obtain a target transaction message template. The target transaction message template is matched with a target user group to push the target transaction message to that group. This method accurately captures changes in user needs, meets personalized user requirements, and improves marketing efficiency and market responsiveness. Simultaneously, it achieves precise matching and pushing of target transaction messages, ensuring that the target user group receives personalized target transaction messages, thereby increasing user engagement and purchase conversion rates.

[0038] Please see Figure 1 , Figure 1 This is a flowchart illustrating a message template determination method provided in an embodiment of this specification. Figure 1 As shown, the method in the embodiments of this specification may include the following steps S102-S108.

[0039] S102, Obtain the behavior prediction results for the test user group, and determine the initial transaction message template based on the behavior prediction results;

[0040] Specifically, the user behavior prediction model obtains behavior prediction results for a test user group, which are then input into a message content generator. The message content generator determines the initial transaction message template based on the behavior prediction results. Before pushing the target transaction message to the target user group, the initial transaction message needs to be tested and adjusted to obtain the final target transaction message, thereby reducing ineffective marketing and improving marketing efficiency. The test user group can be a representative user group from the target user group, used to test the initial transaction message so that the company can understand its effectiveness and receive feedback on the textual expression of the initial transaction message. The user behavior prediction model predicts users' future behavioral needs based on past user behavior data, outputting behavior prediction results. Along with outputting the behavior prediction results, the user behavior prediction model can also classify and tag user groups. The behavior prediction results include at least user interest tags, user demand probabilities, and user consumption type tags. The message content generator determines the initial transaction message template corresponding to the behavior prediction results based on the rule engine.

[0041] S104, Fill the initial transaction message template with user information data of the test user group to obtain the initial transaction message;

[0042] Specifically, the initial transaction message template is the framework of the initial transaction message and does not yet embed users' personal information. Therefore, user information data of the test user group is obtained, and the initial transaction message template is populated with the user information data to obtain the initial transaction message.

[0043] S106, Perform an interactive test on the initial transaction message to obtain the interaction parameter values ​​of the initial transaction message;

[0044] Specifically, interactive tests are conducted on the initial transaction message to obtain its interaction parameter values. The interaction tests are implemented using a hybrid method combining inference model prediction and small-sample user validation to test the effectiveness of the initial transaction message. The inference model is trained using historical interaction data from the test user group. The small-sample user validation method involves pushing the initial transaction message to some or all users in the test user group to obtain user comment data. Interaction behaviors include users opening the initial transaction message, users clicking on product links within the initial transaction message, and users purchasing products linked by those product links. Based on these interactions, the interaction parameters include open rate, click-through rate, and conversion rate. The open rate reflects the attractiveness of the transaction message's title or preview text; the click-through rate measures the match between the transaction message content and user needs; and the conversion rate directly relates to product attractiveness and marketing goals. The final interaction parameter value is a comprehensive value calculated based on the open rate, click-through rate, and conversion rate.

[0045] S108. If the interaction parameter value is greater than or equal to the preset interaction parameter threshold, the initial transaction message template is adjusted based on the user comment data of the test user group on the initial transaction message to obtain the target transaction message template.

[0046] Specifically, the interaction parameter values ​​are compared with preset interaction parameter thresholds. If the interaction parameter values ​​are greater than or equal to the preset thresholds, the initial transaction message template is adjusted based on user comment data from the test user group during the interaction test to obtain the target transaction message template. While the initial transaction message's effectiveness is considered to have reached the preset effectiveness threshold when the interaction parameter values ​​are greater than or equal to the preset thresholds, to prevent user churn due to the initial message's textual expression, the tone and punctuation of the initial transaction message template are adjusted using user comment data to obtain the target transaction message template. This makes the final generated target transaction message more gentle or formal, providing a better user experience and thus improving marketing conversion rates.

[0047] In the embodiments described in this specification, by predicting the behavior of user groups and accurately capturing changes in user needs, personalized user demands can be met, improving marketing efficiency and market responsiveness. The initial transaction message undergoes interactive testing, and the target transaction message template is obtained after adjustments based on user comment data.

[0048] Please see Figure 2 , Figure 2 This is a flowchart illustrating a message template determination method provided in an embodiment of this specification. Figure 2 As shown, the method in the embodiments of this specification may include the following steps S202-S218.

[0049] S202, Construct a user behavior prediction model based on the first user behavior data of the sample user group;

[0050] Specifically, the first user behavior data of the sample user group is obtained, the first user behavior data is cleaned and classified to obtain the training dataset, the training dataset is input into the initial prediction model, and the initial prediction model is trained based on the machine learning algorithm to obtain the user behavior prediction model.

[0051] The sample user group can be a subset of the target user group. User behavior data includes all user actions, such as browsing, interaction, purchasing, social interaction, access, content preferences, and loyalty behavior. When acquiring the first user behavior data, operation records from the application can be obtained through data collection code, or interfaces can be established with third-party applications such as payment platforms to obtain user transaction records and operation data. User behavior preferences can also be obtained when registering, filling out questionnaires, or participating in activities. In the embodiments of this specification, the acquired first user behavior data is relatively complex, containing some missing or erroneous values. Missing values ​​refer to excessively missing sample data, such as a user's user behavior data only containing browsing behavior. Data cleaning is used to improve data quality. Through data cleaning, missing values ​​are supplemented using values ​​such as the mean, median, or mode, while erroneous or extreme values ​​in the user behavior data are corrected. Duplicate values ​​are identified and removed using key fields in the user behavior data. Data classification categorizes user actions. For example, active actions such as searching, form filling, and commenting / feedback are grouped under "interactive behavior," while passive observations such as page views, dwell time, and scrolling are grouped under "browsing behavior." User behavior prediction models use machine learning algorithms to transform structured data and predict future user behavior or potential needs.

[0052] S204, Input the second user behavior data of the test user group into the user behavior prediction model, and obtain the behavior prediction results for the test user group based on the user behavior prediction model;

[0053] Specifically, after obtaining the trained and validated user behavior prediction model, the second user behavior data of the test user group is input into the user behavior prediction model. Based on the user behavior prediction model, behavior prediction results for the test user group are obtained. The test user group can be a representative user group from the target user group. The behavior prediction results include at least user interest tags, user demand probability, and user consumption type tags. User interest tags are used to reflect user interests and preferences, and user consumption type tags are used to reflect user consumption types. For example, if user B in the test user group searches for "smartphone" 10 times and browses high-priced products 5 times in 7 days, the behavior prediction results output by the user behavior prediction model based on user B's second user behavior data are: user interest tag "smartphone", user demand probability "85%", and user consumption type tag "high-value user".

[0054] S206, Input the behavior prediction result into the message content generator, match the behavior prediction result with the message template library in the message content generator, and determine the initial transaction message template based on the matching result;

[0055] Specifically, the user behavior prediction model outputs behavior prediction results, while the transaction message is generated by the message content generator. The behavior prediction results are input into the message content generator, where they are matched against a message template library to determine the initial transaction message template. The message content generator is a data- and rule-based automated tool used to dynamically create highly personalized transaction messages. Its core logic is to transform behavior prediction results into executable transaction messages through template design and dynamic content filling. Inputting the behavior prediction results into the message content generator triggers the rule engine, which matches the behavior prediction results against the message template library based on the tags associated with the results. For example, if the user's interest tag is "smartphone," then a smartphone-related template including features such as screen and battery life is selected. If the user's consumption type tag is "high-value user," then flagship, high-end configuration, and premium versions are selected. If the user's demand probability is 85%, then limited-time offers and limited-supply templates are selected. These templates are then integrated to obtain the initial transaction message template.

[0056] S208, Fill the initial transaction message template with user information data of the test user group to obtain the initial transaction message;

[0057] Specifically, the initial transaction message template is the framework for the initial transaction message and does not yet embed users' personal information. Therefore, user information data from the test user group is obtained and used to populate the initial transaction message template to obtain the initial transaction message. User information data includes personal information such as username, gender, and age, as well as product brands that users follow. To more accurately target user preferences when populating the initial transaction message template, product brands can be included in the initial transaction message. Optionally, the initial transaction message can also include product images. When selecting images, users can choose their preferred colors or models based on their personal information.

[0058] S210, Input the initial transaction message into the inference model and obtain the interaction parameter values ​​of the initial transaction message predicted by the inference model;

[0059] Specifically, the initial transaction message has not yet undergone interaction testing, so its effectiveness cannot be guaranteed. Therefore, the initial transaction message is input into an inference model, and the interaction parameter values ​​of the initial transaction message are obtained based on the inference model. The inference model is a model trained using historical interaction behavior data of the test user group. Interaction behaviors include users opening the initial transaction message, users clicking on product links within the initial transaction message, and users purchasing the products linked by the product links. Based on the interaction behaviors, the interaction parameters include open rate, click-through rate, and conversion rate. The open rate reflects the attractiveness of the transaction message's title or preview text: Open rate = (Number of users opening the transaction message / Number of users who received the transaction message) × 100%. The click-through rate measures the match between the transaction message content and user needs: Click-through rate = (Number of users clicking on product links in the transaction message / Number of users opening the transaction message) × 100%. The conversion rate is directly related to product attractiveness and marketing goals: Conversion rate = (Number of users purchasing the products linked by the product links / Number of users clicking on product links in the transaction message) × 100%. The final interaction parameter value is a comprehensive value calculated based on the open rate, click-through rate, and conversion rate. The first weight of the open rate, the second weight of the click-through rate, and the third weight of the conversion rate are determined according to transaction priority. Therefore, the interaction parameter value = (open rate × first weight) + (click-through rate × second weight) + (conversion rate × third weight). The sum of the first, second, and third weights is 100%. Transaction priority is determined according to the transaction objective. In this embodiment, the transaction objective is primarily to improve the purchase conversion rate; therefore, the priority of the conversion rate is higher than that of the click-through rate, and the priority of the click-through rate is higher than that of the open rate. The third weight of the conversion rate can be set to the highest weight, the first weight of the open rate to the lowest weight, and the second weight of the click-through rate lower than the third weight of the conversion rate but higher than the first weight of the open rate. For example, the first weight of the conversion rate is 50%, the second weight of the click-through rate is 30%, and the first weight of the open rate is 20%.

[0060] When training the inference model, historical interaction behavior data of the test user group over a preset time period is obtained. This includes the user's actions within that period, the transaction messages received after those actions, and the open rate, click-through rate, and conversion rate of those messages. Therefore, historical interaction behavior data includes user behavior data, the characteristics of the transaction messages received by users, and the user's interaction parameters with those messages. This historical interaction behavior data is used as the training set to train the inference model, thereby improving its prediction accuracy.

[0061] S212, Push the initial transaction message to the test user group and obtain user comment data of the test user group regarding the initial transaction message;

[0062] Specifically, since the improvements to the initial transaction message cannot be fully determined based solely on the specific interaction parameter values, this embodiment employs a hybrid method of inference model prediction and small-sample user verification during interaction testing. The small-sample user verification method involves pushing the initial transaction message to some or all users in the test user group and obtaining user comment data regarding the initial transaction message. User comment data can be obtained by sending questionnaires to the test user group. The small-sample user verification method can also obtain the open rate, click-through rate, and conversion rate of the initial transaction message, which can be used as reference data to verify the accuracy of the interaction parameters predicted by the inference model.

[0063] S214. If the interaction parameter value is greater than or equal to the preset interaction parameter threshold, the initial transaction message template is adjusted based on the user comment data of the test user group regarding the initial transaction message.

[0064] Specifically, the interaction parameter values ​​are compared with preset interaction parameter thresholds. If the interaction parameter values ​​are greater than or equal to the preset thresholds, the effectiveness of the initial transaction message is considered to have reached the preset effectiveness threshold. However, to prevent user churn due to the textual expression of the initial transaction message, the initial transaction message template is adjusted based on user comment data from the test user group regarding the initial transaction message. Optionally, since the open rate reflects the attractiveness of the transaction message's title or preview text, the click-through rate measures the match between the transaction message content and user needs, and the conversion rate is directly related to product attractiveness and marketing goals, the initial transaction message template can also be adjusted based on the interaction parameters predicted by the inference model. For example, if the open rate is lower than the preset open rate threshold, the title of the initial transaction message can be adjusted to make it more eye-catching, or the preview text can be changed to a dialect version to increase its entertainment value and attract users to open the initial transaction message.

[0065] The preset interaction parameter threshold is a comprehensive value obtained by comparing the open rate, click-through rate, and conversion rate between the control group and the experimental group, combined with the empirical value and the expected value of the interaction parameters. The control group provides a benchmark for the test group, used to compare the performance of the test group. Historical transaction messages from related technologies and user interaction parameters for those historical transaction messages are used as the control group, while the initial transaction messages from the embodiments of this specification and the user interaction parameters predicted by the inference model for those initial transaction messages are used as the test group. The empirical value is the historical parameter value determined based on the historical data of the control group, the expected value is the ideal parameter value determined based on the transaction objective, and the preset interaction parameter threshold is a compromise value determined based on the empirical value and the expected value. For example, if the first weight of open rate is 20%, the second weight of click-through rate is 30%, and the third weight of conversion rate is 50%; the test value of open rate is 70% and the empirical value is 50%; the test value of click-through rate is 60% and the empirical value is 30%; the test value of conversion rate is 30% and the empirical value is 10%; then the interaction parameter test value output by the inference model = (70% × 20%) + (60% × 30%) + (30% × 50%) = 0.47; the interaction parameter empirical value = (50% × 20%) + (30% × 30%) + (10% × 50%) = 0.24. It can be seen that the interaction parameter test value is higher, that is, compared with historical transaction messages, the initial transaction messages of the embodiments of this specification are more effective. The expected value of the interaction parameter determined by the transaction objective is 0.50. Based on the empirical value of the interaction parameter of 0.24 and the expected value of the interaction parameter of 0.50, the preset threshold of the interaction parameter is determined to be 0.40. When the test value of the interaction parameter is 0.47, it is determined that the interaction parameter value is greater than the preset threshold of the interaction parameter. The initial transaction message template is then adjusted based on the user comment data.

[0066] S216, The adjusted initial transaction message template is determined as the target transaction message template. The target transaction message template includes transaction tags, and the transaction tags include at least transaction attribute tags and transaction value tags.

[0067] Specifically, after adjusting the initial transaction message template, a target transaction message template is obtained. This template includes transaction tags, which are used to categorize the target message template. This allows for matching the target transaction message with the target user group when pushing it, improving the accuracy of the push notifications. The transaction tags include at least transaction attribute tags and transaction value tags. Transaction attribute tags indicate the domain to which the product in the target message template belongs, used to match the interests of the target user group. Transaction value tags include the product price or the product value determined by the price, used to match the consumption type of the target user group.

[0068] S218, if the interaction parameter value is less than the preset interaction parameter threshold, then adjust the model parameters of the user behavior prediction model based on the user comment data, and proceed to the step of inputting the second user behavior data of the test user group into the user behavior prediction model.

[0069] Specifically, if the interaction parameter value is less than the preset interaction parameter threshold, the validity of the initial transaction message is determined to have not reached the preset validity threshold. Adjusting the initial transaction message template will not make the target transaction message reach the preset validity threshold. Therefore, the model parameters of the user behavior prediction model are adjusted based on user comment data, and the process proceeds to the step of inputting the second user behavior data of the test user group into the user behavior prediction model. The adjusted user behavior prediction model regenerates the behavior prediction result, and the initial transaction message is generated based on the new behavior prediction result. The initial transaction message is then tested again until the interaction parameter value of the initial transaction message is greater than or equal to the preset interaction parameter threshold. Based on the user comment data of the test user group regarding the initial transaction message, the initial transaction message template is adjusted to obtain the target transaction message template.

[0070] When adjusting the model parameters of the user behavior prediction model, the adjustment goals for the initial transaction message are first determined based on the open rate, click-through rate, conversion rate, and user comment data. For example, the copywriting language may be made softer, more anime-style, more official, or incorporating dialects; multimodal elements such as images and videos may be added. Secondly, the user behavior prediction model is adjusted by combining feature optimization and algorithm parameter tuning. Feature optimization involves analyzing the Shapley Value (SHAP) of the user behavior prediction model to increase the weight of features related to the interaction parameters of the initial transaction message. SHAP provides an intuitive and consistent interpretation of feature importance for the model's predictions, helping to understand the contribution of each feature to a single prediction result. For example, if "anime-style copywriting click-through rate" significantly impacts the conversion rate, the weight of anime-style features in the user behavior prediction model is increased. Algorithm tuning involves adjusting machine learning algorithms. For example, by adjusting the weights of the attention mechanism through neural network algorithms, the ability to capture user interest tags (such as "anime") can be enhanced. By adjusting the tree depth and learning rate in random forests or gradient boosting trees through ensemble learning algorithms, the classification boundary can be optimized to balance the complexity and generalization ability of user behavior prediction models.

[0071] Optionally, the message content generator and user behavior prediction model can be coordinated to further optimize the initial transaction message template. Specifically, the message content generator also utilizes natural language processing (NLP) technology. When optimizing the message content generator, the text style of the generated transaction message template can be adjusted using prompts from NLP technology; or the parameters of the image generation model in NLP technology can be adjusted to ensure that the images in the generated transaction message template conform to user preferences.

[0072] In the embodiments of this specification, a user behavior prediction model is used to predict the behavior of user groups, accurately capturing changes in user needs, meeting personalized user demands, and improving marketing efficiency and market responsiveness. The behavior prediction results output by the user behavior prediction model are input into a message content generator to obtain an initial transaction message. Highly personalized transaction messages are then dynamically created based on the message content generator. An interactive test of the initial transaction message is conducted using a hybrid method of inference model + small-sample user verification. This method obtains both the interaction parameter values ​​of the initial transaction message and user feedback. Based on the interaction parameter values ​​and user feedback, the initial transaction message template or user behavior prediction model is adjusted to obtain the target transaction message template, improving user engagement and satisfaction, ensuring the effectiveness of the final generated target transaction message, optimizing marketing strategies, and guaranteeing marketing results.

[0073] Please see Figure 3 , Figure 3 This is a flowchart illustrating a message push method provided in an embodiment of this specification. Figure 3 As shown, the method in the embodiments of this specification may include the following steps S302-S306.

[0074] S302, determine the first target user label that matches the transaction label of the target transaction message template from the user labels of the first target user group;

[0075] Specifically, before pushing the target transaction message, the first target user group is categorized, and each category within the first target user group is tagged with user tags. When pushing the target transaction message, the transaction tag of the target transaction message template is obtained, and the transaction tag is matched with the user tags of the first target user group to determine the first target user tag that matches the transaction tag of the target transaction message template. The user tags include at least user interest tags and user consumption type tags, and the transaction tags include at least transaction attribute tags and transaction value tags. User interest tags are used to match transaction attribute tags, and user consumption type tags are used to match transaction value tags.

[0076] S304, determine the first target user group corresponding to the first target user tag as the second target user group, and obtain the target user information data of the second target user group;

[0077] Specifically, within the first target user group, the first target user group corresponding to the first target user tag is identified as the second target user group, and target user information data of the second target user group is obtained. This user information data includes users' personal information, such as username, gender characteristics, and age.

[0078] S306, determine the target transaction message based on the target user information data and the target transaction message template, and push the target transaction message to the second target user group.

[0079] Specifically, target user information data is populated into a target transaction message template to obtain a target transaction message, which is then pushed to a second target user group to ensure personalization. Optionally, the target transaction message can be pushed to users via email, SMS, or the application they are using. For example, if the transaction attribute tag of the target transaction message template is "sports watch" and the transaction value tag is "high value," then a second target user group is selected from the first user group whose user interest tag is "sports watch" and whose user consumption type tag is "high value user." After the selection is completed, the target user information data of each user in the second target user group is obtained, populated into the target transaction message template, and the target transaction message is sent to each user in the second target user group.

[0080] In the embodiments of this specification, the transaction tags of the target transaction message template are matched with the user tags of the user group to achieve accurate matching and push of the target transaction message, ensuring that the target transaction messages received by the target user group are all personalized and improving the user's purchase conversion rate.

[0081] Please see Figure 4 , Figure 4 This is a flowchart illustrating a message push method provided in an embodiment of this specification. Figure 4 As shown, the method in the embodiments of this specification may include the following steps S402-S416.

[0082] S402, classify the first target user group and mark each group in the first target user group with user tags;

[0083] Specifically, at first preset time intervals, third user behavior data of the first target user group is acquired within a second preset time interval. Based on this third user behavior data, the user group is categorized. The first preset time interval is greater than or equal to the second preset time interval. By periodically acquiring the third user behavior data of the first target user group, the changing needs of the first target user group can be captured more accurately. During categorization, the third user behavior data is input into the user behavior prediction model. This allows the model to categorize the first target user group while outputting the behavior prediction results. Each category within the first target user group is labeled with user tags, which include at least user interest tags and user consumption type tags. User interest tags reflect user interests and preferences, while user consumption type tags reflect user consumption types.

[0084] S404, determine the target user interest tag that matches the transaction attribute tag of the target transaction message template from the user interest tags of the user tags of the first target user group, and determine the user tag corresponding to the target user interest tag as the second target user tag;

[0085] Specifically, before pushing a message, target user interest tags that match the transaction attribute tags of the target transaction message template are determined from the user interest tags of the user tags of the first target user group. The target transaction message template is the template of the transaction message to be pushed to the first target user group, which has not yet been filled with users' personal information. The target transaction message template carries transaction tags, which are matched with user tags to determine the users in the first target user group corresponding to the target transaction message template. The transaction tags include at least transaction attribute tags and transaction value tags, where the transaction attribute tags are the domain to which the product in the target transaction message template belongs, and the transaction value tags include the product price or the product value determined based on the product price. The transaction attribute tags are used to match with user interest tags to filter target user interest tags from the user interest tags of the user tags, and then determine the user tags corresponding to the target user interest tags as the second target user tags for a second matching. For example, if the transaction attribute tag is "smartphone", then the second target user tag with the user interest tag "smartphone" is determined from the user tags of the first target user group.

[0086] S406, determine the target user consumption type label that matches the transaction value label of the target transaction message template from the user consumption type label of the second target user label, and determine the second target user label corresponding to the target user consumption type label as the first target user label;

[0087] Specifically, the transaction value tag is used to match the user consumption type tag. After successfully matching the user interest tag with the transaction attribute tag, the user consumption type tag and the transaction value tag are matched. For example, if the transaction value tag is "high value", or the product price reflected by the transaction value tag belongs to the high value range, then the "high value user" tag is determined as the target user consumption type tag in the user consumption type tag of the second target user tag, and the second target user tag corresponding to the target user consumption type tag is determined as the first target user tag.

[0088] S408, determine the first target user group corresponding to the first target user tag as the second target user group, and obtain the target user information data of the second target user group;

[0089] Please refer to step S304 for the specific process, which will not be repeated here.

[0090] S410, determine the target transaction message based on the target user information data and the target transaction message template, and push the target transaction message to the second target user group;

[0091] Please refer to step S306 for the specific process, which will not be repeated here.

[0092] S412, Obtain the interaction behavior data of the second target user group regarding the target transaction message;

[0093] Specifically, after pushing the target transaction message to the second target user group, the interaction behavior data of the second target user group in response to the target transaction message is obtained. These interactions include users opening the initial transaction message, users clicking on product links within the initial transaction message, and users purchasing the products linked by those product links. Based on these interactions, interaction parameters include open rate, click-through rate (CTR), and conversion rate. The open rate reflects the attractiveness of the transaction message's title or preview text: Open rate = (Number of users opening the transaction message / Number of users who received the transaction message) × 100%. The CTR measures the match between the transaction message content and user needs: CTR = (Number of users clicking on product links within the transaction message / Number of users opening the transaction message) × 100%. The conversion rate directly relates to product attractiveness and marketing goals: Conversion rate = (Number of users purchasing the products linked by those product links / Number of users clicking on product links within the transaction message) × 100%.

[0094] S414, Determine the target interaction parameter values ​​for the second target user group based on interaction behavior data;

[0095] Specifically, the target interaction parameter value for the second target user group is determined based on the interaction behavior data. The target interaction parameter value is a comprehensive value calculated based on the target open rate, target click rate and target conversion rate. The first weight of open rate, the second weight of click rate and the third weight of conversion rate are determined according to the transaction priority. Then the interaction parameter value = (open rate × first weight) + (click rate × second weight) + (conversion rate × third weight).

[0096] S416, Adjust the target transaction message or user behavior prediction model based on the target interaction parameter value.

[0097] Specifically, before pushing the target transaction message, a target interaction parameter threshold is pre-set. The target interaction parameter value is compared with the target interaction parameter threshold, and the target transaction message template or user behavior prediction model is adjusted based on the comparison result. If the target interaction parameter value is greater than or equal to the target interaction parameter threshold, the target transaction message template is adjusted; if the target interaction parameter value is less than the target interaction parameter threshold, the user behavior prediction model is adjusted.

[0098] Optionally, while pushing the target transaction message, a questionnaire can also be pushed to a second target user group to obtain target user comment data on the target transaction message from the second target user group. The target user comment data can be used to more accurately adjust the target transaction message template or user behavior prediction model.

[0099] Optionally, since open rate reflects the attractiveness of the title or preview text of a transaction message, click-through rate measures the match between the content of the transaction message and user needs, and conversion rate is directly related to product attractiveness and marketing goals, when adjusting the target transaction message template or user behavior prediction model, target open rate threshold, target click-through rate threshold, and target conversion rate threshold can be preset. Based on the first comparison result of the target open rate and the target open rate threshold, the second comparison result of the target click-through rate and the target click-through rate threshold, and the third comparison result of the target conversion rate and the target conversion rate threshold, some content in the target transaction message template or model parameters in the user behavior prediction model corresponding to the open rate, click-through rate, or conversion rate can be adjusted.

[0100] In the embodiments of this specification, the first target user group is periodically classified and tagged to more accurately capture changes in their needs. The transaction tags of the target transaction message template are matched with the user tags of the user group to achieve precise matching and delivery of target transaction messages, ensuring that the target user group receives personalized target transaction messages and improving user purchase conversion rates. After the target transaction message is pushed to the second target user group, the interaction behavior data of the second target user group in response to the target transaction message is monitored to adjust the target transaction message template or user behavior prediction model, thereby improving the conversion rate of the target transaction message and ensuring marketing effectiveness.

[0101] Please see Figure 5 , Figure 5 This is a flowchart illustrating a method for determining a message template and pushing a message, as provided in an embodiment of this specification. Figure 5 As shown, the method in the embodiments of this specification may include the following steps S502-S524.

[0102] S502, collects user behavior data;

[0103] It can obtain operation records in the application through data collection code; it can also establish interfaces with third-party applications such as payment platforms to obtain users' transaction records, operation data, etc.; it can also detect the behavioral preferences of mobile phone users when they register, fill out questionnaires, or participate in activities.

[0104] S504 processes user behavior data;

[0105] The collected user behavior data is cleaned and classified.

[0106] S506, Build a user behavior prediction model;

[0107] The processed user behavior data is used as the training dataset to build a user behavior prediction model.

[0108] S508, Generate the initial transaction message;

[0109] The user behavior prediction results are obtained based on the user behavior prediction model. The behavior prediction results are then input into the message content generator, where the initial transaction message is generated based on the user's user information data.

[0110] S510 performs interactive testing on the initial transaction message;

[0111] Interaction testing was conducted using a hybrid approach combining inference model prediction and small-sample user verification to obtain interaction parameter values.

[0112] S512, Determine whether the value of the interaction parameter is greater than or equal to the preset interaction parameter threshold;

[0113] If yes, proceed to step S514; otherwise, proceed to step S516.

[0114] S514, Adjust the initial transaction message template;

[0115] Adjust the initial transaction message template to obtain the target transaction message.

[0116] S516, Adjust the user behavior prediction model;

[0117] Proceed to step S508, where the user behavior prediction results are re-obtained based on the adjusted user behavior prediction model.

[0118] S518 matches the target transaction message with the target user group;

[0119] Match the target transaction message with the target user group based on the transaction tag carried by the target transaction message and the user tag carried by the target user group.

[0120] S520, push target transaction message;

[0121] Targeted information is pushed to the target user group via email, SMS, or the user's application.

[0122] S522, monitors the interactive behavior of the target user group;

[0123] S524, System Performance Analysis.

[0124] Analyze the entire process of message template determination and message push to improve system response time and push accuracy.

[0125] In the embodiments of this specification, user behavior prediction models are used to predict user behavior, accurately capturing changes in user needs, meeting personalized user demands, and improving marketing efficiency and market responsiveness. Highly personalized transaction messages are dynamically created based on a message content generator. An interactive test of the initial transaction message is conducted using a hybrid method of inference model + small-sample user verification. This obtains both the interaction parameter values ​​of the initial transaction message and user feedback. Based on these parameters and feedback, the initial transaction message template or user behavior prediction model is adjusted to obtain the target transaction message template, improving user engagement and satisfaction, ensuring the effectiveness of the final generated target transaction message, optimizing marketing strategies, and guaranteeing marketing results. The transaction tags of the target transaction message template are matched with user tags of user groups to achieve precise matching and push of target transaction messages, ensuring that the target user groups receive personalized target transaction messages, thus improving user purchase conversion rates. After pushing the target transaction message to a second target user group, the interaction behavior data of the second target user group is monitored to adjust the target transaction message template or user behavior prediction model, improving the conversion rate of the target transaction message and guaranteeing marketing results.

[0126] The following will combine Figure 6 This specification provides a detailed description of the message template determination device provided in the embodiments. It should be noted that... Figure 6 The message template determining device in the present application is used to perform the following tasks. Figures 1-2 The methods shown in the embodiments are illustrated for ease of explanation, showing only the parts related to the embodiments of this specification. For specific technical details not disclosed, please refer to this application. Figures 1-2 The example shown.

[0127] Please see Figure 6 , Figure 6 This is a schematic diagram of the structure of a message template determining device provided in the embodiments of this specification. Figure 6 As shown, the message template determination device 1 in the embodiments of this specification may include: an initial message template acquisition unit 11, an initial message acquisition unit 12, a message testing unit 13, and a target message template acquisition unit 14.

[0128] The initial message template acquisition unit 11 is used to acquire the behavior prediction results for the test user group and determine the initial transaction message template based on the behavior prediction results.

[0129] The initial message acquisition unit 12 is used to fill the initial transaction message template based on the user information data of the test user group to obtain the initial transaction message.

[0130] The message testing unit 13 is used to perform interactive tests on the initial transaction message and obtain the interaction parameter values ​​of the initial transaction message.

[0131] The target message template acquisition unit 14 is used to adjust the initial transaction message template based on the user comment data of the test user group on the initial transaction message if the interaction parameter value is greater than or equal to the preset interaction parameter threshold, so as to obtain the target transaction message template.

[0132] Optionally, the message template determining device 1 is specifically used to construct a user behavior prediction model based on the first user behavior data of the sample user group.

[0133] Optionally, the message template determining device 1 is specifically used to acquire the first user behavior data of the sample user group;

[0134] The first set of user behavior data is cleaned and classified to obtain the training dataset.

[0135] The training dataset is input into the initial prediction model, and the initial prediction model is trained based on machine learning algorithms to obtain the user behavior prediction model.

[0136] Optionally, the initial message template acquisition unit 11 is specifically used to input the second user behavior data of the test user group into the user behavior prediction model;

[0137] Based on the user behavior prediction model, obtain the behavior prediction results for the test user group;

[0138] Input the behavior prediction results into the message content generator;

[0139] In the message content generator, the behavior prediction results are matched with the message template library, and the initial transaction message template is determined based on the matching results.

[0140] Optionally, the message testing unit 13 is specifically used to input the initial transaction message into the inference model, obtain the interaction parameter values ​​of the initial transaction message predicted by the inference model, and the inference model is trained using historical interaction behavior data of the test user group; and

[0141] The initial transaction message is pushed to the test user group, and user comments on the initial transaction message are obtained from the test user group.

[0142] Optionally, the target message template acquisition unit 14 is specifically used to adjust the initial transaction message template based on the user comment data of the test user group regarding the initial transaction message if the interaction parameter value is greater than or equal to the preset interaction parameter threshold.

[0143] The adjusted initial transaction message template is determined as the target transaction message template. The target transaction message template includes transaction tags, which include at least transaction attribute tags and transaction value tags.

[0144] Optionally, the message template determining device 1 is specifically used to adjust the model parameters of the user behavior prediction model based on user comment data if the interaction parameter value is less than the preset interaction parameter threshold, and then proceed to the step of inputting the second user behavior data of the test user group into the user behavior prediction model.

[0145] In the embodiments of this specification, a user behavior prediction model is used to predict the behavior of user groups, accurately capturing changes in user needs, meeting personalized user demands, and improving marketing efficiency and market responsiveness. The behavior prediction results output by the user behavior prediction model are input into a message content generator to obtain an initial transaction message. Highly personalized transaction messages are then dynamically created based on the message content generator. An interactive test of the initial transaction message is conducted using a hybrid method of inference model + small-sample user verification. This method obtains both the interaction parameter values ​​of the initial transaction message and user feedback. Based on the interaction parameter values ​​and user feedback, the initial transaction message template or user behavior prediction model is adjusted to obtain the target transaction message template, improving user engagement and satisfaction, ensuring the effectiveness of the final generated target transaction message, optimizing marketing strategies, and guaranteeing marketing results.

[0146] It should be noted that the message template determination device provided in the above embodiments is only illustrated by the division of the above functional modules when executing the message template determination method. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the message template determination device and the message template determination method embodiments provided in the above embodiments belong to the same concept, and the implementation process is detailed in the method embodiments, which will not be repeated here.

[0147] The embodiment numbers in this specification are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments. In some cases, the actions or steps described in the claims can be performed in a different order than that shown in the embodiments and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0148] The following will combine Figure 7 This specification provides a detailed description of the message push device provided in the embodiments. It should be noted that... Figure 7The message push device in the middle is used to execute the present application. Figures 3-4 The methods shown in the embodiments are illustrated for ease of explanation, showing only the parts related to the embodiments of this specification. For specific technical details not disclosed, please refer to this application. Figures 3-4 The example shown.

[0149] Please see Figure 7 , Figure 7 This is a schematic diagram of the structure of a message push device provided in an embodiment of this specification. Figure 7 As shown, the message push device 2 in the embodiments of this specification may include: a tag matching unit 21, a user information acquisition unit 22, and a message push unit 23.

[0150] Tag matching unit 21 is used to determine the first target user tag that matches the transaction tag of the target transaction message template from the user tags of the first target user group;

[0151] User information acquisition unit 22 is used to determine the first target user group corresponding to the first target user tag as the second target user group and acquire the target user information data of the second target user group.

[0152] The message push unit 23 is used to determine the target transaction message based on the target user information data and the target transaction message template, and push the target transaction message to the second target user group.

[0153] Optionally, the message push device 2 is specifically used to classify the first target user group;

[0154] Each group within the primary target user group is tagged using user tags, which include at least user interest tags and user consumption type tags.

[0155] Optionally, the message push device 2 is specifically used to acquire third user behavior data of the first target user group within a second preset time period at intervals of a first preset time period;

[0156] User groups are categorized based on third-party user behavior data.

[0157] Optionally, the tag matching unit 21 is specifically used to determine the target user interest tag that matches the transaction attribute tag of the target transaction message template from the user interest tags of the user tags of the first target user group, and to determine the user tag corresponding to the target user interest tag as the second target user tag.

[0158] In the user consumption type label of the second target user label, determine the target user consumption type label that matches the transaction value label of the target transaction message template, and determine the second target user label corresponding to the target user consumption type label as the first target user label.

[0159] Optionally, the message push device 2 is specifically used to acquire interactive behavior data of the second target user group regarding the target transaction message;

[0160] Determine the target interaction parameter values ​​for the second target user group based on interaction behavior data;

[0161] Adjust the target transaction message template or user behavior prediction model based on the target interaction parameter values.

[0162] In the embodiments of this specification, the first target user group is periodically classified and tagged to more accurately capture changes in their needs. The transaction tags of the target transaction message template are matched with the user tags of the user group to achieve precise matching and delivery of target transaction messages, ensuring that the target user group receives personalized target transaction messages and improving user purchase conversion rates. After the target transaction message is pushed to the second target user group, the interaction behavior data of the second target user group in response to the target transaction message is monitored to adjust the target transaction message template or user behavior prediction model, thereby improving the conversion rate of the target transaction message and ensuring marketing effectiveness.

[0163] It should be noted that the message push device provided in the above embodiments is only illustrated by the division of the above functional modules when executing the message push method. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the message push device and message push method embodiments provided in the above embodiments belong to the same concept, and the implementation process is detailed in the method embodiments, which will not be repeated here.

[0164] The embodiment numbers in this specification are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments. In some cases, the actions or steps described in the claims can be performed in a different order than that shown in the embodiments and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0165] Please see Figure 8 , Figure 8 This is a schematic diagram of the structure of a computer device provided in the embodiments of this specification.

[0166] For example, such as Figure 8 As shown, the computer device 800 includes a processor 801 and a memory 802, wherein the processor 801 is electrically connected to the memory 802.

[0167] The processor 801 is the control center of the computer device 800 and may include one or more processing cores. The processor 801 connects to various parts of the computer device using various interfaces and lines. By running or calling computer programs stored in the memory 802, and by calling data stored in the memory 802, it executes various functions of the computer device and processes data, thereby providing overall control of the computer device 800. Optionally, the processor 801 may be implemented using at least one of the following hardware forms: Digital Signal Processing (DSP), Field Programmable Gate Array (FPGA), and Programmable Logic Array (PLA). The processor 801 may integrate one or more of the following: CPU, Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user page, and applications; the GPU is responsible for rendering and drawing the displayed content; and the modem handles wireless communication. It is understood that the modem may also not be integrated into the processor 801 and may be implemented separately using a communication chip.

[0168] The memory 802 can be used to store software programs and modules. The processor 801 executes various functional applications and data processing by running the computer programs and modules stored in the memory 802. The memory 802 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, computer programs required for at least one function, etc.; the data storage area may store data created based on the use of the computer device 800, etc.

[0169] Furthermore, memory 802 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, memory 802 may also include a memory controller to provide processor 801 with access to memory 802.

[0170] In a first feasible embodiment of the embodiments of this specification, the processor 801 in the computer device 800 loads the instructions corresponding to the processes of one or more computer programs into the memory 802 according to the following steps, and the processor 801 runs the computer programs stored in the memory 802 to realize various functions, as follows:

[0171] Obtain behavioral prediction results for the test user group, and determine the initial transaction message template based on the behavioral prediction results;

[0172] The initial transaction message is obtained by filling the initial transaction message template with user information data of the test user group;

[0173] Perform interactive tests on the initial transaction message to obtain the interaction parameter values ​​of the initial transaction message;

[0174] If the interaction parameter value is greater than or equal to the preset interaction parameter threshold, the initial transaction message template is adjusted based on the user comment data of the test user group on the initial transaction message to obtain the target transaction message template.

[0175] Optionally, before executing the process of obtaining the behavior prediction results for the test user group and determining the initial transaction message template based on the behavior prediction results, the processor 801 also executes:

[0176] A user behavior prediction model is built based on the first user behavior data of the sample user group.

[0177] Optionally, when processor 801 executes the following when building a user behavior prediction model based on the first user behavior data of the sample user group:

[0178] Obtain primary user behavior data from the sample user group;

[0179] The first set of user behavior data is cleaned and classified to obtain the training dataset.

[0180] The training dataset is input into the initial prediction model, and the initial prediction model is trained based on machine learning algorithms to obtain the user behavior prediction model.

[0181] Optionally, when processor 801 executes the process of obtaining behavior prediction results for the test user group and determining the initial transaction message template based on the behavior prediction results, it specifically performs the following:

[0182] Input the second user behavior data of the test user group into the user behavior prediction model;

[0183] Based on the user behavior prediction model, obtain the behavior prediction results for the test user group;

[0184] Input the behavior prediction results into the message content generator;

[0185] In the message content generator, the behavior prediction results are matched with the message template library, and the initial transaction message template is determined based on the matching results.

[0186] Optionally, when processor 801 performs an interaction test on the initial transaction message to obtain the interaction parameter values ​​of the initial transaction message, it specifically executes the following:

[0187] The initial transaction message is input into the inference model to obtain the interaction parameter values ​​of the initial transaction message predicted by the inference model. The inference model is trained using historical interaction behavior data of the test user group; and

[0188] The initial transaction message is pushed to the test user group, and user comments on the initial transaction message are obtained from the test user group.

[0189] Optionally, when processor 801 executes the step of adjusting the initial transaction message template based on user comment data of the test user group on the initial transaction message if the interaction parameter value is greater than or equal to a preset interaction parameter threshold, to obtain the target transaction message template, the specific execution is as follows:

[0190] If the interaction parameter value is greater than or equal to the preset interaction parameter threshold, the initial transaction message template will be adjusted based on the user comment data of the test user group regarding the initial transaction message;

[0191] The adjusted initial transaction message template is determined as the target transaction message template. The target transaction message template includes transaction tags, which include at least transaction attribute tags and transaction value tags.

[0192] Optionally, after performing an interaction test on the initial transaction message and obtaining the interaction parameter values ​​of the initial transaction message, the processor 801 also executes:

[0193] If the interaction parameter value is less than the preset interaction parameter threshold, the model parameters of the user behavior prediction model are adjusted based on the user comment data, and the process proceeds to the step of inputting the second user behavior data of the test user group into the user behavior prediction model.

[0194] In a second feasible embodiment of the embodiments of this specification, the processor 801 in the computer device 800 loads the instructions corresponding to the processes of one or more computer programs into the memory 802 according to the following steps, and the processor 801 runs the computer programs stored in the memory 802 to realize various functions, as follows:

[0195] Identify the first target user tag that matches the transaction tag of the target transaction message template from the user tags of the first target user group;

[0196] The first target user group corresponding to the first target user tag is identified as the second target user group, and the target user information data of the second target user group is obtained.

[0197] The target transaction message is determined based on the target user information data and the target transaction message template, and then pushed to the second target user group.

[0198] Optionally, before executing the process to determine the first target user label that matches the transaction label of the target transaction message template in the user labels of the first target user group, the processor 801 also executes:

[0199] Classify the primary target user group;

[0200] Each group within the primary target user group is tagged using user tags, which include at least user interest tags and user consumption type tags.

[0201] Optionally, when performing the classification of the first target user group, the processor 801 specifically executes the following:

[0202] At each first preset time interval, acquire the third user behavior data of the first target user group within a second preset time interval;

[0203] User groups are categorized based on third-party user behavior data.

[0204] Optionally, when the processor 801 determines the first target user label that matches the transaction label of the target transaction message template from the user labels of the first target user group, it specifically executes:

[0205] In the user interest tags of the first target user group, identify the target user interest tags that match the transaction attribute tags of the target transaction message template, and determine the user tags corresponding to the target user interest tags as the second target user tags;

[0206] In the user consumption type label of the second target user label, determine the target user consumption type label that matches the transaction value label of the target transaction message template, and determine the second target user label corresponding to the target user consumption type label as the first target user label.

[0207] Optionally, after pushing the target transaction message to the second target user group, processor 801 also executes:

[0208] Acquire interactive behavior data of the second target user group regarding the target transaction message;

[0209] Determine the target interaction parameter values ​​for the second target user group based on interaction behavior data;

[0210] Adjust the target transaction message template or user behavior prediction model based on the target interaction parameter values.

[0211] In the embodiments of this specification, user behavior prediction models are used to predict the behavior of user groups, accurately capturing changes in user needs, meeting personalized user demands, and improving marketing efficiency and market responsiveness. The behavior prediction results output by the user behavior prediction model are input into a message content generator to obtain initial transaction messages. Highly personalized transaction messages are then dynamically created based on the message content generator. An interactive test of the initial transaction messages is conducted using a hybrid method of inference model + small-sample user validation. This obtains both the interaction parameter values ​​of the initial transaction messages and user feedback. Based on the interaction parameter values ​​and user feedback, the initial transaction message template or user behavior prediction model is adjusted to obtain the target transaction message template, improving user engagement and satisfaction, ensuring the effectiveness of the final generated target transaction messages, optimizing marketing strategies, and guaranteeing marketing results. The first target user group is periodically classified and tagged to more accurately capture changes in their needs. The transaction tags of the target transaction message template are matched with the user tags of the user group to achieve precise matching and push of target transaction messages, ensuring that the target user group receives personalized target transaction messages, thereby improving user purchase conversion rates. After pushing the target message to the second target user group, the interaction data of the second target user group in response to the target message is monitored to adjust the target message template or user behavior prediction model, thereby improving the conversion rate of the target message and ensuring marketing effectiveness.

[0212] It should be understood that the apparatus provided in the embodiments of this specification is used to execute the above-described message template determination method, and therefore can achieve the same effect as the above-described implementation method.

[0213] When using integrated units, the device may include a processing module and a storage module. When applied to a computer device, the processing module can be used to control and manage the operations of the computer device. The storage module can be used to support the computer device in executing relevant program code, etc.

[0214] The processing module may be a processor or a controller, which can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. The processor may also be a combination of functions that implement computing capabilities, such as a combination of one or more microprocessors, a combination of digital signal processing (DSP) and a microprocessor, etc., and the storage module may be a memory.

[0215] In addition, the device provided in the embodiments of this specification may specifically be a chip, component or module. The chip may include a connected processor and a memory. The memory is used to store instructions. When the processor calls and executes the instructions, the chip can execute a message template determination method provided in the above embodiments.

[0216] This specification also provides a computer-readable storage medium storing computer program code. When the computer program code is run on a computer, it causes the computer to execute the aforementioned method steps to implement the message template determination method provided in the above embodiments.

[0217] This embodiment also provides a computer program product that, when run on a computer, causes the computer to perform the aforementioned steps to implement the message template determination method provided in the above embodiment.

[0218] In this embodiment, the device, computer-readable storage medium, computer program product, or chip are all used to execute the corresponding methods provided above. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0219] Through the above description of the embodiments, those skilled in the art will understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0220] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0221] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for determining a message template, the method comprising: Obtain behavioral prediction results for the test user group, and determine the initial transaction message template based on the behavioral prediction results; The initial transaction message is obtained by filling the initial transaction message template with the user information data of the test user group; Perform an interactive test on the initial transaction message to obtain the interaction parameter values ​​of the initial transaction message; If the interaction parameter value is greater than or equal to the preset interaction parameter threshold, the initial transaction message template is adjusted based on the user comment data of the test user group on the initial transaction message to obtain the target transaction message template. The step of performing an interactive test on the initial transaction message to obtain the interaction parameter values ​​of the initial transaction message includes: The initial transaction message is input into the inference model to obtain the interaction parameter values ​​of the initial transaction message predicted by the inference model. The inference model is trained using historical interaction behavior data of the test user group. The interaction parameter values ​​include open rate, click-through rate, and conversion rate. The initial transaction message is pushed to the test user group, and user comment data of the test user group regarding the initial transaction message is obtained.

2. The method according to claim 1, before obtaining the behavior prediction results for the test user group and determining the initial transaction message template based on the behavior prediction results, further comprising: A user behavior prediction model is built based on the first user behavior data of the sample user group.

3. The method according to claim 2, wherein constructing a user behavior prediction model based on the first user behavior data of the sample user group includes: Obtain primary user behavior data from the sample user group; The first user behavior data is cleaned and classified to obtain a training dataset. The training dataset is input into the initial prediction model, and the initial prediction model is trained based on the machine learning algorithm to obtain the user behavior prediction model.

4. The method according to claim 2, wherein obtaining the behavior prediction results for the test user group and determining the initial transaction message template based on the behavior prediction results includes: The second user behavior data of the test user group is input into the user behavior prediction model; Based on the user behavior prediction model, obtain the behavior prediction results for the test user group; The behavior prediction results are input into the message content generator; In the message content generator, the behavior prediction result is matched with the message template library, and the initial transaction message template is determined based on the matching result.

5. The method according to claim 1, wherein if the interaction parameter value is greater than or equal to a preset interaction parameter threshold, adjusting the initial transaction message template based on user comment data of the test user group on the initial transaction message to obtain a target transaction message template includes: If the interaction parameter value is greater than or equal to the preset interaction parameter threshold, the initial transaction message template is adjusted based on the user comment data of the test user group regarding the initial transaction message; The adjusted initial transaction message template is determined as the target transaction message template. The target transaction message template includes transaction tags, and the transaction tags include at least transaction attribute tags and transaction value tags.

6. The method according to claim 4, after performing an interaction test on the initial transaction message to obtain the interaction parameter value of the initial transaction message, further comprising: If the interaction parameter value is less than the preset interaction parameter threshold, the model parameters of the user behavior prediction model are adjusted based on the user comment data, and the process proceeds to the step of inputting the second user behavior data of the test user group into the user behavior prediction model.

7. A message template determining device, the device comprising: An initial message template acquisition unit is used to acquire behavior prediction results for a test user group and determine an initial transaction message template based on the behavior prediction results. An initial message acquisition unit is used to fill the initial transaction message template based on the user information data of the test user group to obtain an initial transaction message. The message testing unit is used to perform interactive tests on the initial transaction message and obtain the interaction parameter values ​​of the initial transaction message. The target message template acquisition unit is used to adjust the initial transaction message template based on the user comment data of the test user group on the initial transaction message if the interaction parameter value is greater than or equal to a preset interaction parameter threshold, so as to obtain the target transaction message template. The message testing unit is specifically used to input the initial transaction message into the inference model, obtain the interaction parameter values ​​of the initial transaction message predicted by the inference model, and the inference model is trained using historical interaction behavior data of the test user group. The interaction parameter values ​​include open rate, click-through rate, and conversion rate. The initial transaction message is pushed to the test user group, and user comment data of the test user group regarding the initial transaction message is obtained.

8. A computer device, comprising: Processor and memory; The memory stores a computer program adapted to be loaded by the processor and to execute the steps of the method as described in any one of claims 1 to 6.

9. A storage medium storing a computer program that, when executed by a processor, implements the steps of the method as claimed in any one of claims 1 to 6.

10. A computer program product, comprising: A computer program, when executed by a processor of a computer device, causes the processor to perform the steps of the method as described in any one of claims 1 to 6.

Images