Cross-channel intelligent marketing decision system and method based on dynamic user behavior mining

Abstract

The application relates to a cross-channel intelligent marketing decision system and method based on dynamic user behavior mining, and relates to the technical field of marketing data processing.The method comprises the following steps: acquiring a multi-channel user behavior data sequence, constructing a multi-dimensional channel priority scoring system, and realizing accurate classification of core channels and long-tail channels.Based on user conversion path data, the correlation coefficient of core channels and main channels is quantified, and a dynamic budget allocation strategy and a content delivery strategy are generated in combination with the priority score.Finally, based on the budget allocation strategy and the content delivery strategy of each channel, the operation parameters of the corresponding channel are automatically adjusted.The method effectively solves the problems of one-sided channel value evaluation, unquantified synergistic effect and rigid strategy in traditional marketing, and improves the utilization rate of marketing resources and the user conversion efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of marketing data processing, and relates to a cross-channel intelligent marketing decision system and method based on dynamic user behavior mining. BACKGROUND

[0002] In today's digital marketing environment, enterprises usually reach users through multiple online channels, including social media, search engines, emails and various platforms. With the increase in the number of channels and the complexity of user behavior, in the multi-platform marketing environment, enterprises face problems such as difficulty in data integration, lag in strategy adjustment, and opaque decision-making process, resulting in low efficiency of marketing resource allocation.

[0003] Existing marketing budget allocation schemes mostly use fixed proportion allocation based on historical experience or simple ROI sorting method.

[0004] However, such methods have obvious limitations. On the one hand, they lack quantitative evaluation of the synergistic effect between channels and cannot identify channel combinations with complementary effects. On the other hand, they use a static allocation mode and cannot dynamically adjust according to changes in user behavior. SUMMARY

[0005] The technical problem to be solved by the present application is to provide a cross-channel intelligent marketing decision system and method based on dynamic user behavior mining, aiming to solve at least one of the above technical problems.

[0006] The technical solution of the present application to solve the above technical problems is as follows:

[0007] In a first aspect, the present application provides a cross-channel intelligent marketing decision method based on dynamic user behavior mining, which adopts the following technical solution:

[0008] A cross-channel intelligent marketing decision method based on dynamic user behavior mining, comprising:

[0009] Obtaining a user behavior data sequence of multiple channels, the user behavior data including exposure, click volume, user dwell time, interaction data and conversion event data corresponding to each channel, the conversion event data representing events completed by users on the channel and defined by the business as having value;

[0010] Based on the user behavior data sequence, calculating a priority score for each channel;

[0011] Based on the priority score of each channel, sorting all channels, and based on the sorting results of all channels, determining a core channel set and a long-tail channel set;

[0012] Acquire user conversion path data, which records the sequence information of all channels through which the user interacts from the first contact to the completion of the conversion;

[0013] Based on the conversion path data, calculate the correlation coefficient between each channel in the core channel set and the channel with the highest priority score;

[0014] Based on the priority scores and relevance coefficients of each channel in the core channel set, a budget allocation strategy and content delivery strategy are generated for each channel.

[0015] Based on the budget allocation and content delivery strategies of each channel, the corresponding operational parameters are automatically adjusted.

[0016] The beneficial effects of this invention are: by calculating channel priority scores based on user behavior data, core channels and long-tail channels can be accurately identified, avoiding resource dispersion; by combining conversion path data to calculate the correlation coefficient between core channels, the synergistic effect of channels can be quantified, and the order of placement and content adaptation can be optimized; finally, through the strategic linkage of budget and content and automatic parameter adjustment, the efficient allocation of marketing resources and the maximization of conversion effect are achieved.

[0017] Based on the above technical solution, the present invention can be further improved as follows.

[0018] Furthermore, the calculation of the priority score for each channel based on the user behavior data sequence includes:

[0019] Based on the exposure, clicks and user dwell time of each channel, determine the first weight value of each channel;

[0020] Based on interactive data and preset feature extraction rules, keyword information is determined;

[0021] Based on keyword information, determine user satisfaction with each channel;

[0022] Based on user engagement and satisfaction levels from each channel, a second weight value is determined for each channel.

[0023] Based on the traffic volume, cost data, and conversion event data of each channel, a third weight value is determined for each channel;

[0024] Based on the historical performance and trend information of each channel within a preset time period, a fourth weight value is assigned to each channel according to the stability and growth of the channel performance.

[0025] Based on the first, second, third, and fourth weight values ​​of each channel, calculate the performance index value of each channel within the predetermined time period;

[0026] Based on the performance indicator values ​​corresponding to the current time period and the historical time period, determine the historical comprehensive score of each channel;

[0027] The trend coefficient for each channel is determined based on the ratio of the performance indicator values ​​within the near-term time window to the long-term time window.

[0028] The priority score for each channel is obtained by multiplying the historical composite score of each channel by the trend coefficient.

[0029] The beneficial effects of adopting the above-mentioned further solutions are as follows: By integrating basic metrics such as exposure and clicks with in-depth behavioral data such as user dwell time and interactive keywords, it ensures both an objective assessment of traffic volume and reveals users' true interests and needs; by introducing a product mechanism of historical comprehensive scores and trend coefficients, it avoids the interference of short-term fluctuations through long-term data accumulation and captures the channel's development potential through recent trend coefficients; finally, by integrating key business elements such as traffic costs, conversion efficiency, and user engagement, priority scoring is directly linked to marketing, providing enterprises with a basis for resource allocation.

[0030] Furthermore, based on the conversion path data, calculating the correlation coefficient between each channel in the core channel set and the channel with the highest priority score includes:

[0031] Based on the conversion path data, calculate the total number of paths that have appeared in the main channel among all conversion paths, and record the total number of paths that have appeared in the main channel as the first value. The main channel represents the channel with the highest priority score in the core channel set.

[0032] Based on the conversion path data, calculate the total number of paths in all conversion paths that appear in either the main channel or any channel in the core channel set, and record the total number of paths that appear in either the main channel or any channel in the core channel set as the second value;

[0033] For any channel in the core channel set, the correlation coefficient between the channel in the core channel set and the channel with the highest priority score is determined based on the ratio of the first value to the second value of the corresponding channel in the core channel set.

[0034] The beneficial effects of adopting the above-mentioned further solutions are: by quantitatively analyzing the synergistic occurrence patterns of different channels in the user conversion path, the correlation coefficient between each marketing channel and the core main channel can be accurately calculated. This not only effectively solves the limitation of isolated evaluation of channel effects in traditional marketing strategies, but also provides data-driven decision-making basis for cross-channel collaborative deployment, thereby significantly improving the overall conversion efficiency and resource utilization of multi-channel marketing.

[0035] Furthermore, the step of generating a budget allocation strategy for each of the channels based on the priority scores and relevance coefficients of each channel in the core channel set includes:

[0036] For the core channel set, a budget allocation scheme for each channel in the core channel set is generated based on the genetic algorithm, the total budget constraint, the upper and lower limits of the budget for each channel in the core channel set, and the objective function of maximizing the comprehensive efficiency index. The objective function is constructed based on the priority score and the correlation coefficient.

[0037] For the long-tail channel set, a reserved exploration budget is allocated to the long-tail channel set, and the budget allocation strategy for each channel in the long-tail channel set is determined based on the Multi-Armed Slots (MAB) algorithm, the priority score of each channel in the long-tail channel set, and the reserved exploration budget.

[0038] The beneficial effects of adopting the above-mentioned further solutions are as follows: By constructing a multi-constraint optimization model using a genetic algorithm for core channels, and incorporating priority scores and relevance coefficients into the objective function, resource allocation to high-value channels is ensured, while overall conversion efficiency is improved through the quantification of inter-channel synergies. For long-tail channels, the MAB algorithm is introduced to allocate exploration budgets, and the testing intensity is dynamically adjusted in conjunction with priority scores, thereby controlling risks while identifying potential high-value channels and reducing resource waste.

[0039] Furthermore, the process of generating a budget allocation scheme for each channel in the core channel set based on genetic algorithms, total budget constraints, upper and lower budget constraints for each channel in the core channel set, and maximizing the overall efficiency index as the objective function includes:

[0040] The parameters to be optimized for all channels in the core channel set are encoded into a chromosome, and an initial population consisting of multiple chromosomes is generated. The chromosome represents a budget allocation scheme for the core channel set.

[0041] Based on channel performance function and user behavior data, a fitness function is constructed, and the fitness of each chromosome in the current population is calculated. The fitness is used to evaluate the comprehensive performance index of the budget allocation scheme corresponding to the chromosome. The channel performance function is constructed based on priority score, correlation coefficient and budget allocated to the channel.

[0042] Based on the fitness of each chromosome in the current iteration, multiple parent pairs are selected from the population in the current iteration using the roulette wheel selection method;

[0043] Based on all parent pairs, the crossover and mutation algorithm, the upper and lower limits of budget for each channel in the core channel set, and the total budget constraint, multiple offspring chromosomes are generated.

[0044] Based on multiple new offspring chromosomes, the population for the next iteration is determined and cyclical until the current iteration number reaches the set iteration number. The chromosome with the highest fitness in the previous generation population is decoded to obtain the budget allocation scheme for each channel in the core channel set.

[0045] The beneficial effects of adopting the above-mentioned further scheme are as follows: the budget allocation problem is transformed into an evolutionary process of chromosome coding. By simulating the natural selection mechanism, the optimal solution is automatically searched under multiple constraints such as total budget and channel upper and lower limits, avoiding the subjectivity of human experience allocation; the fitness function integrates channel priority scores, correlation coefficients and budget allocation amounts to scientifically quantify the comprehensive effectiveness of each scheme and ensure that resources are concentrated on high-value and highly collaborative channels; the crossover and mutation operations and iterative evolution mechanism ensure population diversity, quickly converge to a high-quality solution, and achieve global optimization and precise coordination of core channel budget allocation.

[0046] Furthermore, the step of generating content delivery strategies for each of the channels based on their priority scores and relevance coefficients in the core channel set includes:

[0047] Based on user behavior data sequences from multiple channels, user profiles for each channel are determined.

[0048] Based on each channel's priority score, relevance coefficient, marketing objectives, and user profile, multiple content delivery strategies matching each channel are selected from the marketing content database;

[0049] For each channel, the selected multiple content delivery strategies are grouped according to the type and content of the content delivery strategy, and the correlation between each group and the channel is determined;

[0050] For each channel, the group order is determined based on the correlation between each group and the channel;

[0051] For each channel, obtain historical performance data of each content delivery strategy corresponding to that channel in different vertical industries;

[0052] For each channel, multiple content delivery strategies are sorted according to historical performance data of each content delivery strategy in different vertical industries to generate the first ranking result;

[0053] For each channel, multiple content delivery strategies are sorted according to the performance data of the content delivery strategy used by the currently associated channel within the first preset time period, and a second sorting result is generated;

[0054] For each channel, multiple content delivery strategies are sorted according to the performance data of the content delivery strategy used by the currently associated channel within a second preset time period, and a third sorting result is generated; the second preset time period is located before the first preset time period and is adjacent to the first preset time period.

[0055] For each channel, a priority ranking is determined based on the first ranking result, the second ranking result, and the third ranking result;

[0056] For each channel, the content delivery strategy for that channel is determined based on the priority ranking and grouping of multiple content delivery strategies that match that channel.

[0057] The beneficial effects of adopting the above-mentioned further solutions are as follows: By generating channel-specific user profiles based on user behavior data, and combining priority scoring and relevance coefficients to select content delivery strategies, the matching between content and channels is ensured, and the targeting of content is improved through accurate user profiling; by introducing historical effects, recent effects, and effects of adjacent time periods, the stability of content delivery strategies is verified through long-term data accumulation, and market trend changes are captured through recent effects, avoiding deviations in content delivery strategies caused by a single time dimension; and the optimal content delivery strategy for each channel is determined through group correlation ranking and priority ranking.

[0058] Secondly, this application provides a cross-channel intelligent marketing decision-making system based on dynamic user behavior mining, which adopts the following technical solution:

[0059] A cross-channel intelligent marketing decision-making system based on dynamic user behavior mining includes:

[0060] The first acquisition module is used to acquire user behavior data sequences from multiple channels. The user behavior data includes exposure, clicks, user dwell time, interaction data, and conversion event data for each channel. The conversion event data represents events that users complete on the channels and that are defined as valuable by the business.

[0061] The first calculation module is used to calculate the priority score of each channel based on the user behavior data sequence;

[0062] The sorting module is used to sort all channels based on the priority score of each channel, and to determine the core channel set and the long-tail channel set based on the sorting results of all channels.

[0063] The second acquisition module is used to acquire the user's conversion path data, which records the sequence information of all channels through which the user interacts from the first contact to the completion of the conversion;

[0064] The second calculation module is used to calculate the correlation coefficient between each channel in the core channel set and the channel with the highest priority score based on the conversion path data.

[0065] The generation module is used to generate budget allocation strategies and content delivery strategies for each of the channels in the core channel set based on the priority scores and relevance coefficients of each channel.

[0066] The adjustment module is used to automatically adjust the operational parameters of the corresponding channels based on the budget allocation strategy and content delivery strategy of each channel.

[0067] Furthermore, the first calculation module is specifically used for:

[0068] Based on the exposure, clicks and user dwell time of each channel, determine the first weight value of each channel;

[0069] Based on interactive data and preset feature extraction rules, keyword information is determined;

[0070] Based on keyword information, determine user satisfaction with each channel;

[0071] Based on user engagement and satisfaction levels from each channel, a second weight value is determined for each channel.

[0072] Based on the traffic volume, cost data, and conversion event data of each channel, a third weight value is determined for each channel;

[0073] Based on the historical performance and trend information of each channel within a preset time period, a fourth weight value is assigned to each channel according to the stability and growth of the channel performance.

[0074] Based on the first, second, third, and fourth weight values ​​of each channel, calculate the performance index value of each channel within the predetermined time period;

[0075] Based on the performance indicator values ​​corresponding to the current time period and the historical time period, determine the historical comprehensive score of each channel;

[0076] The trend coefficient for each channel is determined based on the ratio of the performance indicator values ​​within the near-term time window to the long-term time window.

[0077] The priority score for each channel is obtained by multiplying the historical composite score of each channel by the trend coefficient.

[0078] Thirdly, this application provides an electronic device that adopts the following technical solution:

[0079] An electronic device includes a memory and a processor, wherein the memory stores a computer program capable of being loaded by the processor and executing the cross-channel intelligent marketing decision-making method based on dynamic user behavior mining as described in any of the first aspects.

[0080] Fourthly, this application provides a computer-readable storage medium, which adopts the following technical solution:

[0081] A computer-readable storage medium storing a computer program capable of being loaded by a processor and executing the cross-channel intelligent marketing decision-making method based on dynamic user behavior mining as described in any one of the first aspects.

[0082] Additional aspects and advantages of this application will be set forth in part in the description which follows, and will become apparent from the description or may be learned by practice of this application. Attached Figure Description

[0083] Figure 1 A flowchart illustrating a cross-channel intelligent marketing decision-making method based on dynamic user behavior mining, as provided in one embodiment of the present invention;

[0084] Figure 2 A schematic diagram of the structure of a cross-channel intelligent marketing decision-making system based on dynamic user behavior mining, provided as an embodiment of the present invention;

[0085] Figure 3 This is a schematic diagram of the structure of an electronic device provided in one embodiment of the present invention. Detailed Implementation

[0086] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0087] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article, unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0088] This application provides a cross-channel intelligent marketing method based on dynamic user behavior mining. The method can be executed by an electronic device, which can be a server or a mobile terminal device. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services. The mobile terminal device can be a laptop, a desktop computer, etc., but is not limited to these.

[0089] like Figure 1 As shown, a cross-channel marketing decision-making method based on dynamic user behavior mining mainly includes:

[0090] S1, acquire user behavior data sequences from multiple channels. The user behavior data includes exposure, clicks, user dwell time, interaction data, and conversion event data for each channel. The conversion event data represents events that users complete on the channels and that are defined as valuable by the business.

[0091] In this embodiment, techniques such as real-time API capture, server log parsing, and third-party SDK tracking are used to collect user behavior data sequences from various channels. For example, "exposure" and "clicks" are obtained through e-commerce platform APIs; "user dwell time," such as the number of seconds the page is spent, and "interaction data," such as comments, sharing, and adding items to the shopping cart, are collected through APP tracking.

[0092] Conversion event data is collected based on valuable events defined by the business, such as user registration and product purchase. Conversion event data can be recorded in log files or message queues by the business database through transaction logs or directly by the application calling log interfaces at key business logic points, and can then be retrieved from these files or message queues.

[0093] S2, Based on the user behavior data sequence, calculate the priority score for each channel;

[0094] In this embodiment of the application, calculating the priority score for each channel based on the user behavior data sequence includes:

[0095] Based on the exposure, clicks and user dwell time of each channel, determine the first weight value of each channel;

[0096] Based on interactive data and preset feature extraction rules, keyword information is determined;

[0097] Based on keyword information, determine user satisfaction with each channel;

[0098] Based on user engagement and satisfaction levels from each channel, a second weight value is determined for each channel.

[0099] Based on the traffic volume, cost data, and conversion event data of each channel, a third weight value is determined for each channel;

[0100] Based on the historical performance and trend information of each channel within a preset time period, a fourth weight value is assigned to each channel according to the stability and growth of the channel performance.

[0101] Based on the first, second, third, and fourth weight values ​​of each channel, calculate the performance index value of each channel within the predetermined time period;

[0102] Based on the performance indicator values ​​corresponding to the current time period and the historical time period, determine the historical comprehensive score of each channel;

[0103] The trend coefficient for each channel is determined based on the ratio of the performance indicator values ​​within the near-term time window to the long-term time window.

[0104] The priority score for each channel is obtained by multiplying the historical composite score of each channel by the trend coefficient.

[0105] In this embodiment of the application, the acquired user behavior data sequence is cleaned to obtain a cleaned user behavior data sequence. The cleaning process includes handling missing values, removing outliers, and standardizing the data format.

[0106] Next, the click-through rate (CTR) of each channel is calculated based on the exposure and click-through rates corresponding to each channel. The average user dwell time within a preset time period is calculated based on the multiple user dwell times corresponding to each channel. Finally, a first weight value is determined for each channel based on its CTR and average user dwell time. For example, if a channel has high exposure, indicating a wide audience reach, and also good click-through rates and dwell time, then its first weight value will be relatively large.

[0107] Using natural language processing tools, preset keywords are extracted from interaction data, and the sentiment polarity of text containing keywords is judged to determine user satisfaction with each channel. For example, if the keywords are mostly positive, then user satisfaction is high.

[0108] Based on the sharing rate, like rate, and comment interaction rate in the interaction data, the user's participation level for each channel is determined; after normalizing the two indicators of satisfaction and user participation, they are weighted and combined again to obtain the second weight value.

[0109] Based on the traffic volume, cost data, and conversion event data of each channel, a third weight value is determined for each channel. Channels with large traffic volume, low cost, and many conversion events have a high third weight value.

[0110] Based on the historical performance and trend information of each channel within a preset time period, a fourth weight value is assigned to each channel according to the stability and growth of its performance. For example, channels with stable historical performance and growth trends have a higher fourth weight value. Based on the first, second, third, and fourth weight values ​​of each channel, the performance indicator value of each channel within the predetermined time period is calculated.

[0111] Then, different weights are assigned to the data from different time periods, the historical weighted average performance index value of each channel is calculated, and the historical comprehensive score of each channel is determined based on the performance index value corresponding to the current time period and the historical time period.

[0112] Then, linear regression is used to compare recent data with long-term data to determine whether the effectiveness of each channel is increasing, stable, or decreasing, thereby determining the trend coefficient of each channel.

[0113] Finally, the historical composite score of each channel is multiplied by the trend coefficient to obtain the priority score for each channel.

[0114] By integrating basic metrics such as exposure and clicks with in-depth behavioral data such as user dwell time and interactive keywords, the system ensures both an objective assessment of traffic volume and reveals users' true interests and needs. It introduces a product mechanism of historical comprehensive scores and trend coefficients, which avoids interference from short-term fluctuations through long-term data accumulation and captures channel development potential through recent trend coefficients. Finally, by integrating key business elements such as traffic costs, conversion efficiency, and user engagement, priority scoring is directly linked to marketing, providing enterprises with a basis for resource allocation.

[0115] S3, based on the priority score of each channel, sort all channels, and based on the sorting results of all channels, determine the core channel set and the long-tail channel set;

[0116] In this embodiment of the application, when sorting all channels and determining the core channel set and the long-tail channel set, all channels are sorted from high to low according to the calculated priority score. Channels with high priority scores form the core channel set, and channels with relatively low priority scores form the long-tail channel set.

[0117] S4, Obtain the user's conversion path data, which records the sequence information of all channels the user interacts with from the first contact to the completion of the conversion;

[0118] In this embodiment of the application, the user's account information is used to track their interaction behavior on various channels, thereby obtaining complete conversion path data.

[0119] S5. Based on the conversion path data, calculate the correlation coefficient between each channel in the core channel set and the channel with the highest priority score;

[0120] In this embodiment of the application, the step of calculating the correlation coefficient between each channel in the core channel set and the channel with the highest priority score based on the conversion path data includes:

[0121] Based on the conversion path data, calculate the total number of paths that have appeared in the main channel among all conversion paths, and record the total number of paths that have appeared in the main channel as the first value. The main channel represents the channel with the highest priority score in the core channel set.

[0122] Based on the conversion path data, calculate the total number of paths in all conversion paths that appear in either the main channel or any channel in the core channel set, and record the total number of paths that appear in either the main channel or any channel in the core channel set as the second value;

[0123] For any channel in the core channel set, the correlation coefficient between the channel in the core channel set and the channel with the highest priority score is determined based on the ratio of the first value to the second value of the corresponding channel in the core channel set.

[0124] By quantitatively analyzing the patterns of synergy among different channels in the user conversion path, we can accurately calculate the correlation coefficient between each marketing channel and the core main channel. This not only effectively solves the limitations of isolated evaluation of channel performance in traditional marketing strategies, but also provides data-driven decision-making basis for cross-channel collaborative deployment, thereby significantly improving the overall conversion efficiency and resource utilization of multi-channel marketing.

[0125] S6. Based on the priority scores and relevance coefficients of each channel in the core channel set, generate budget allocation strategies and content delivery strategies for the core marketing channels.

[0126] In this embodiment of the application, the step of generating a budget allocation strategy for each of the channels based on the priority score and relevance coefficient of each channel in the core channel set includes:

[0127] For the core channel set, a budget allocation scheme for each channel in the core channel set is generated based on the genetic algorithm, the total budget constraint, the upper and lower limits of the budget for each channel in the core channel set, and the objective function of maximizing the comprehensive efficiency index. The objective function is constructed based on the priority score and the correlation coefficient.

[0128] For the long-tail channel set, a reserved exploration budget is allocated to the long-tail channel set, and the budget allocation strategy for each channel in the long-tail channel set is determined based on the Multi-Armed Slots (MAB) algorithm, the priority score of each channel in the long-tail channel set, and the reserved exploration budget.

[0129] In this embodiment of the application, the step of generating a budget allocation scheme for each channel in the core channel set based on genetic algorithms, total budget constraints, upper and lower budget constraints for each channel in the core channel set, and maximizing the overall efficiency index as the objective function includes:

[0130] The parameters to be optimized for all channels in the core channel set are encoded into a chromosome, and an initial population consisting of multiple chromosomes is generated. The chromosome represents a budget allocation scheme for the core channel set.

[0131] Based on channel performance function and user behavior data, a fitness function is constructed, and the fitness of each chromosome in the current population is calculated. The fitness is used to evaluate the comprehensive performance index of the budget allocation scheme corresponding to the chromosome. The channel performance function is constructed based on priority score, correlation coefficient and budget allocated to the channel.

[0132] Based on the fitness of each chromosome in the current iteration, multiple parent pairs are selected from the population in the current iteration using the roulette wheel selection method;

[0133] Based on all parent pairs, the crossover and mutation algorithm, the upper and lower limits of budget for each channel in the core channel set, and the total budget constraint, multiple offspring chromosomes are generated.

[0134] Based on multiple new offspring chromosomes, the population for the next iteration is determined and cyclical until the current iteration number reaches the set iteration number. The chromosome with the highest fitness in the previous generation population is decoded to obtain the budget allocation scheme for each channel in the core channel set.

[0135] In this embodiment of the application, the parameter to be optimized represents the budget amount allocated to each channel in the core channel set.

[0136] The channel performance function is:

[0137]

[0138] in, Priority score for channel i The budget amount allocated to channel i, Let be the correlation coefficient between channel i and channel j. If there is user path overlap between channel i and channel j, then... If positive, there is no direct coordination between the two channels. =0, For the synergistic effect function, quantify the budget. and The synergistic effect.

[0139] Fitness = Σ -Penalty;

[0140] Penalty is a penalty function, which is applied when the total budget exceeds a constraint or a certain condition. When the limit is exceeded, the penalty function returns a negative value, thus eliminating the invalid solution during the evolution process.

[0141] In this embodiment, for a set of long-tail channels, a small portion of the exploration budget is allocated to the set of long-tail channels. An expected value is initialized for each long-tail channel, and the initial value of the expected value is positively correlated with its priority score. The entire exploration budget for this round is allocated to the channel with the highest current expected value, and a channel is randomly selected for investment. Based on the real-time returns after the channel's deployment (such as click-through rate and conversion cost), its expected value is updated. The above steps are repeated until the current iteration number reaches the set iteration number or the exploration budget is exhausted. Finally, the budget allocation ratio for each channel is output.

[0142] In this embodiment of the application, the step of generating content delivery strategies for each of the channels based on the priority scores and relevance coefficients of each channel in the core channel set includes:

[0143] Based on user behavior data sequences from multiple channels, user profiles for each channel are determined.

[0144] Based on each channel's priority score, relevance coefficient, marketing objectives, and user profile, multiple content delivery strategies matching each channel are selected from the marketing content database;

[0145] For each channel, the selected multiple content delivery strategies are grouped according to the type and content of the content delivery strategy, and the correlation between each group and the channel is determined;

[0146] For each channel, the group order is determined based on the correlation between each group and the channel;

[0147] For each channel, obtain historical performance data of each content delivery strategy corresponding to that channel in different vertical industries;

[0148] For each channel, multiple content delivery strategies are sorted according to historical performance data of each content delivery strategy in different vertical industries to generate the first ranking result;

[0149] For each channel, multiple content delivery strategies are sorted according to the performance data of the content delivery strategy used by the currently associated channel within the first preset time period, and a second sorting result is generated;

[0150] For each channel, multiple content delivery strategies are sorted according to the performance data of the content delivery strategy used by the currently associated channel within a second preset time period, and a third sorting result is generated; the second preset time period is located before the first preset time period and is adjacent to the first preset time period.

[0151] For each channel, a priority ranking is determined based on the first ranking result, the second ranking result, and the third ranking result;

[0152] For each channel, the content delivery strategy for that channel is determined based on the priority ranking and grouping of multiple content delivery strategies that match that channel.

[0153] By generating channel-specific user profiles based on user behavior data, and combining priority scoring and relevance coefficients to select content delivery strategies, the matching of content and channels is ensured, and the precise characterization of user profiles enhances the targeting of content. By introducing historical effects, recent effects, and effects from adjacent time periods, the stability of content delivery strategies is verified through long-term data accumulation, while recent effects capture market trend changes, avoiding deviations in content delivery strategies caused by a single time dimension. The optimal content delivery strategy for each channel is determined through group correlation ranking and priority ranking.

[0154] S7 automatically adjusts the operational parameters of each channel based on the budget allocation strategy and content delivery strategy of the core marketing channels.

[0155] Based on the budget allocation and content delivery strategies of each channel, adjust the operational parameters such as the delivery budget and content for each channel.

[0156] This method calculates channel priority scores based on user behavior data, which can accurately identify core channels and long-tail channels and avoid resource dispersion. By combining conversion path data to calculate the correlation coefficient between core channels, the synergistic effect of channels can be quantified, and the order of placement and content adaptation can be optimized. Finally, through the strategic linkage of budget and content and automatic parameter adjustment, the efficient allocation of marketing resources and the maximization of conversion effect are achieved.

[0157] Figure 2 A schematic diagram of the structure of a cross-channel intelligent marketing decision-making system based on dynamic user behavior mining is shown.

[0158] like Figure 2 As shown, a cross-channel intelligent marketing decision-making system 200 based on dynamic user behavior mining mainly includes:

[0159] The first acquisition module 201 is used to acquire user behavior data sequences from multiple channels. The user behavior data includes exposure, clicks, user dwell time, interaction data, and conversion event data for each channel. The conversion event data represents events that users complete on the channels and that are defined as valuable by the business.

[0160] The first calculation module 202 is used to calculate the priority score of each channel based on the user behavior data sequence;

[0161] The sorting module 203 is used to sort all channels based on the priority score of each channel, and to determine the core channel set and the long-tail channel set based on the sorting results of all channels.

[0162] The second acquisition module 204 is used to acquire the user's conversion path data, which records the sequence information of all channels through which the user interacts from the first contact to the completion of the conversion;

[0163] The second calculation module 205 is used to calculate the correlation coefficient between each channel in the core channel set and the channel with the highest priority score based on the conversion path data.

[0164] The generation module 206 is used to generate budget allocation strategies and content delivery strategies for core marketing channels based on the priority scores and relevance coefficients of each channel in the core channel set.

[0165] Adjustment module 207 is used to automatically adjust the operational parameters of each channel based on the budget allocation strategy and content delivery strategy of the core marketing channels.

[0166] Optionally, the first calculation module 202 is specifically used for:

[0167] Based on the exposure, clicks and user dwell time of each channel, determine the first weight value of each channel;

[0168] Based on interactive data and preset feature extraction rules, keyword information is determined;

[0169] Based on keyword information, determine user satisfaction with each channel;

[0170] Based on user engagement and satisfaction from each channel, a second weight value is determined for each channel.

[0171] Based on the traffic volume, cost data, and conversion event data of each channel, a third weight value is determined for each channel;

[0172] Based on the historical performance and trend information of each channel within a preset time period, a fourth weight value is assigned to each channel according to the stability and growth of the channel performance.

[0173] Based on the first, second, third, and fourth weight values ​​of each channel, calculate the performance index value of each channel within the predetermined time period;

[0174] Based on the performance indicator values ​​corresponding to the current time period and the historical time period, determine the historical comprehensive score of each channel;

[0175] The trend coefficient for each channel is determined based on the ratio of the performance indicator values ​​within the near-term time window to the long-term time window.

[0176] The priority score for each channel is obtained by multiplying the historical composite score of each channel by the trend coefficient.

[0177] In one example, the module in any of the above devices may be one or more integrated circuits configured to implement the above methods, such as one or more application-specific integrated circuits (ASICs), or one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs), or a combination of at least two of these integrated circuit forms.

[0178] For example, when modules in a device can be implemented via a processing element scheduler, the processing element can be a general-purpose processor, such as a central processing unit (CPU) or other processor capable of calling programs. Alternatively, these modules can be integrated together as a system-on-a-chip (SOC).

[0179] In this application, various objects such as messages / information / devices / network elements / systems / apparatus / actions / operations / processes / concepts may be named. It is understood that these specific names do not constitute a limitation on the relevant objects. The names may be changed depending on the scenario, context, or usage habits. The understanding of the technical meaning of the technical terms in this application should be mainly determined from their functions and technical effects embodied / performed in the technical solution.

[0180] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0181] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0182] Figure 3 This is a structural block diagram of an electronic device 300 according to an embodiment of this application.

[0183] like Figure 3 As shown, the electronic device 300 includes a processor 301 and a memory 302, and may further include one or more of an information input / output (I / O) interface 303, a communication component 304, and a communication bus 305.

[0184] The processor 301 controls the overall operation of the electronic device 300 to complete all or part of the steps in the cross-channel intelligent marketing decision-making method based on dynamic user behavior mining described above. The memory 302 stores various types of data to support the operation of the electronic device 300. This data may include, for example, instructions for any application or method operating on the electronic device 300, as well as application-related data. The memory 302 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as one or more of Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0185] I / O interface 303 provides an interface between processor 301 and other interface modules, such as keyboards, mice, and buttons. These buttons can be virtual or physical. Communication component 304 is used to test wired or wireless communication between electronic device 300 and other devices. Wireless communication includes Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, or 4G, or a combination thereof. Therefore, the corresponding communication component 304 may include a Wi-Fi component, a Bluetooth component, and an NFC component.

[0186] The communication bus 305 may include a path for transmitting information between the aforementioned components. The communication bus 305 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The communication bus 305 may be divided into an address bus, a data bus, a control bus, etc.

[0187] The electronic device 300 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to execute the cross-channel intelligent marketing decision-making method based on dynamic user behavior mining given in the above embodiments.

[0188] The following describes the computer-readable storage medium provided in the embodiments of this application. The computer-readable storage medium described below can be referred to in correspondence with the cross-channel intelligent marketing decision-making method based on dynamic user behavior mining described above.

[0189] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the cross-channel intelligent marketing decision-making method based on dynamic user behavior mining described above.

[0190] The computer-readable storage medium may include various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0191] The terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0192] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing application concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions claimed in this application.

Claims

1. A cross-channel intelligent marketing decision-making method based on dynamic user behavior mining, characterized in that, include: Acquire user behavior data sequences from multiple channels. The user behavior data includes exposure, clicks, user dwell time, interaction data, and conversion event data for each channel. The conversion event data represents events that users complete on the channels and that are defined as valuable by the business. Based on the user behavior data sequence, calculate the priority score for each channel; Based on the priority score of each channel, all channels are ranked, and based on the ranking results of all channels, the core channel set and the long-tail channel set are determined. Acquire user conversion path data, which records the sequence information of all channels through which the user interacts from the first contact to the completion of the conversion; Based on the conversion path data, calculate the correlation coefficient between each channel in the core channel set and the channel with the highest priority score; Based on the priority scores and relevance coefficients of each channel in the core channel set, a budget allocation strategy and content delivery strategy are generated for each channel. Based on the budget allocation strategy and content delivery strategy of each channel, the corresponding operational parameters of the channel are automatically adjusted. The step of calculating the correlation coefficient between each channel in the core channel set and the channel with the highest priority score based on the conversion path data includes: Based on the conversion path data, calculate the total number of paths that have appeared in the main channel among all conversion paths, and record the total number of paths that have appeared in the main channel as the first value. The main channel represents the channel with the highest priority score in the core channel set. Based on the conversion path data, calculate the total number of paths in all conversion paths that appear in either the main channel or any channel in the core channel set, and record the total number of paths that appear in either the main channel or any channel in the core channel set as the second value; For any channel in the core channel set, the correlation coefficient between the channel in the core channel set and the channel with the highest priority score is determined based on the ratio of the first value to the second value of the corresponding channel in the core channel set. The step of generating content delivery strategies for each channel based on the priority scores and relevance coefficients of each channel in the core channel set includes: Based on user behavior data sequences from multiple channels, user profiles for each channel are determined. Based on each channel's priority score, relevance coefficient, marketing objectives, and user profile, multiple content delivery strategies matching each channel are selected from the marketing content database; For each channel, the selected multiple content delivery strategies are grouped according to the type and content of the content delivery strategy, and the correlation between each group and the channel is determined; For each channel, the group order is determined based on the correlation between each group and the channel; For each channel, obtain historical performance data of each content delivery strategy corresponding to that channel in different vertical industries; For each channel, multiple content delivery strategies are sorted according to historical performance data of each content delivery strategy in different vertical industries to generate the first ranking result; For each channel, multiple content delivery strategies are sorted according to the performance data of the content delivery strategy used by the currently associated channel within the first preset time period, and a second sorting result is generated; For each channel, multiple content delivery strategies are sorted according to the performance data of the content delivery strategy used by the currently associated channel within a second preset time period, and a third sorting result is generated; the second preset time period is located before the first preset time period and is adjacent to the first preset time period. For each channel, a priority ranking is determined based on the first ranking result, the second ranking result, and the third ranking result; For each channel, the content delivery strategy for that channel is determined based on the priority ranking and grouping of multiple content delivery strategies that match that channel.

2. The cross-channel intelligent marketing decision-making method based on dynamic user behavior mining according to claim 1, characterized in that, The calculation of the priority score for each channel based on the user behavior data sequence includes: Based on the exposure, clicks and user dwell time of each channel, determine the first weight value of each channel; Based on interactive data and preset feature extraction rules, keyword information is determined; Based on keyword information, determine user satisfaction with each channel; Based on user engagement and satisfaction levels from each channel, a second weight value is determined for each channel. Based on the traffic volume, cost data, and conversion event data of each channel, a third weight value is determined for each channel; Based on the historical performance and trend information of each channel within a preset time period, a fourth weight value is assigned to each channel according to the stability and growth of the channel performance. Based on the first, second, third, and fourth weight values ​​of each channel, calculate the performance index value of each channel within the predetermined time period; Based on the performance indicator values ​​corresponding to the current time period and the historical time period, determine the historical comprehensive score of each channel; The trend coefficient for each channel is determined based on the ratio of the performance indicator values ​​within the near-term time window to the long-term time window. The priority score for each channel is obtained by multiplying the historical composite score of each channel by the trend coefficient.

3. The cross-channel intelligent marketing decision-making method based on dynamic user behavior mining according to claim 1, characterized in that, The step of generating a budget allocation strategy for each channel based on the priority score and relevance coefficient of each channel in the core channel set includes: For the core channel set, a budget allocation scheme for each channel in the core channel set is generated based on the genetic algorithm, the total budget constraint, the upper and lower limits of the budget for each channel in the core channel set, and the objective function of maximizing the comprehensive efficiency index. The objective function is constructed based on the priority score and the correlation coefficient. For the long-tail channel set, a reserved exploration budget is allocated to the long-tail channel set, and the budget allocation strategy for each channel in the long-tail channel set is determined based on the Multi-Armed Slots (MAB) algorithm, the priority score of each channel in the long-tail channel set, and the reserved exploration budget.

4. The cross-channel intelligent marketing decision-making method based on dynamic user behavior mining according to claim 3, characterized in that, The process of generating a budget allocation scheme for each channel in the core channel set based on genetic algorithms, total budget constraints, upper and lower budget limits for each channel in the core channel set, and maximizing the overall efficiency index as the objective function includes: The parameters to be optimized for all channels in the core channel set are encoded into a chromosome, and an initial population consisting of multiple chromosomes is generated. The chromosome represents a budget allocation scheme for the core channel set. Based on channel performance function and user behavior data, a fitness function is constructed, and the fitness of each chromosome in the current population is calculated. The fitness is used to evaluate the comprehensive performance index of the budget allocation scheme corresponding to the chromosome. The channel performance function is constructed based on priority score, correlation coefficient and budget allocated to the channel. Based on the fitness of each chromosome in the current iteration, multiple parent pairs are selected from the population in the current iteration using the roulette wheel selection method; Based on all parent pairs, the crossover and mutation algorithm, the upper and lower limits of budget for each channel in the core channel set, and the total budget constraint, multiple offspring chromosomes are generated. Based on multiple new offspring chromosomes, the population for the next iteration is determined and cyclical until the current iteration number reaches the set iteration number. The chromosome with the highest fitness in the previous generation population is decoded to obtain the budget allocation scheme for each channel in the core channel set.

5. A cross-channel intelligent marketing decision-making system based on dynamic user behavior mining, characterized in that, include: The first acquisition module is used to acquire user behavior data sequences from multiple channels. The user behavior data includes exposure, clicks, user dwell time, interaction data, and conversion event data for each channel. The conversion event data represents events that users complete on the channels and that are defined as valuable by the business. The first calculation module is used to calculate the priority score of each channel based on the user behavior data sequence; The sorting module is used to sort all channels based on the priority score of each channel, and to determine the core channel set and the long-tail channel set based on the sorting results of all channels. The second acquisition module is used to acquire the user's conversion path data, which records the sequence information of all channels through which the user interacts from the first contact to the completion of the conversion; The second calculation module is used to calculate the correlation coefficient between each channel in the core channel set and the channel with the highest priority score based on the conversion path data. The generation module is used to generate budget allocation strategies and content delivery strategies for each of the channels in the core channel set based on the priority scores and relevance coefficients of each channel. The adjustment module is used to automatically adjust the operational parameters of the corresponding channels based on the budget allocation strategy and content delivery strategy of each channel. The second calculation module is specifically used to: calculate the total number of paths that have appeared in the main channel among all conversion paths based on the conversion path data, and record the total number of paths that have appeared in the main channel as the first value, wherein the main channel represents the channel with the highest priority score in the core channel set; Based on the conversion path data, calculate the total number of paths in all conversion paths that appear in either the main channel or any channel in the core channel set, and record the total number of paths that appear in either the main channel or any channel in the core channel set as the second value; For any channel in the core channel set, the correlation coefficient between the channel in the core channel set and the channel with the highest priority score is determined based on the ratio of the first value to the second value of the corresponding channel in the core channel set. The generation module, used to generate content delivery strategies for each channel based on the priority score and relevance coefficient of each channel in the core channel set, specifically includes: Based on user behavior data sequences from multiple channels, user profiles for each channel are determined. Based on each channel's priority score, relevance coefficient, marketing objectives, and user profile, multiple content delivery strategies matching each channel are selected from the marketing content database; For each channel, the selected multiple content delivery strategies are grouped according to the type and content of the content delivery strategy, and the correlation between each group and the channel is determined; For each channel, the group order is determined based on the correlation between each group and the channel; For each channel, obtain historical performance data of each content delivery strategy corresponding to that channel in different vertical industries; For each channel, multiple content delivery strategies are sorted according to historical performance data of each content delivery strategy in different vertical industries to generate the first ranking result; For each channel, multiple content delivery strategies are sorted according to the performance data of the content delivery strategy used by the currently associated channel within the first preset time period, and a second sorting result is generated; For each channel, multiple content delivery strategies are sorted according to the performance data of the content delivery strategy used by the currently associated channel within a second preset time period, and a third sorting result is generated; the second preset time period is located before the first preset time period and is adjacent to the first preset time period. For each channel, a priority ranking is determined based on the first ranking result, the second ranking result, and the third ranking result; For each channel, the content delivery strategy for that channel is determined based on the priority ranking and grouping of multiple content delivery strategies that match that channel.

6. The cross-channel intelligent marketing decision-making system based on dynamic user behavior mining according to claim 5, wherein the first calculation module is specifically used for: Based on the exposure, clicks and user dwell time of each channel, determine the first weight value of each channel; Based on interactive data and preset feature extraction rules, keyword information is determined; Based on keyword information, determine user satisfaction with each channel; Based on user engagement and satisfaction levels from each channel, a second weight value is determined for each channel. Based on the traffic volume, cost data, and conversion event data of each channel, a third weight value is determined for each channel; Based on the historical performance and trend information of each channel within a preset time period, a fourth weight value is assigned to each channel according to the stability and growth of the channel performance. Based on the first, second, third, and fourth weight values ​​of each channel, calculate the performance index value of each channel within the predetermined time period; Based on the performance indicator values ​​corresponding to the current time period and the historical time period, determine the historical comprehensive score of each channel; The trend coefficient for each channel is determined based on the ratio of the performance indicator values ​​within the near-term time window to the long-term time window. The priority score for each channel is obtained by multiplying the historical composite score of each channel by the trend coefficient.

7. An electronic device, characterized in that, Includes a processor, which is coupled to a memory; The processor is configured to execute a computer program stored in the memory, causing the electronic device to perform the method as described in any one of claims 1 to 4.

8. A computer-readable storage medium, characterized in that, Includes a computer program or instructions that, when run on a computer, cause the computer to perform the method as described in any one of claims 1-4.

Images