Manus-based financial services product marketing approach

Abstract

The present invention relates to the field of marketing technology, specifically to a marketing method for financial service products based on manus, which includes the following steps: analyzing behavioral data, calculating the response frequency of product attributes, optimizing the call sequence of portrait tags, identifying tag confidence, screening marketing content, calculating the push frequency, adjusting the push time window, adjusting push parameters in combination with user feedback, and obtaining product marketing push records. In the present invention, by identifying user cross-platform account behavior and combining time period behavior density, the timeliness and representativeness of behavioral data are improved, and the sensitivity matching of product attributes and portrait tags is adopted to achieve accurate adaptation between user characteristics and financial products. The tag confidence is used to adjust the push content priority, improve the content relevance and push effectiveness, adjust the push window according to the interaction frequency, optimize the push rhythm, and optimize the push parameters in combination with the feedback content click and stay characteristics, thereby improving the marketing accuracy and user response rate.

Description

Technical Field

[0001] The present invention relates to the field of marketing technology, and in particular to a marketing method based on manus financial service products. Background Art

[0002] The field of marketing technology includes using information technology and data analysis to assist enterprises in product promotion, customer management, sales improvement and user relationship maintenance. It involves using multi-source data such as user behavior data, transaction information, demographic information, etc. to identify, classify and model potential customers, and realize customer portrait construction and precise marketing strategy formulation through model deduction. Through the organic integration of system processes such as data collection, behavior analysis, user tag construction, and personalized content push, it covers multiple sub-directions such as customer relationship management, market segmentation, and advertising push strategy optimization. Combined with the development of digital finance and mobile Internet, it transforms from traditional population coverage communication to data-driven individual precision interaction, realizing user response behavior prediction and marketing strategy iteration.

[0003] Among them, the marketing method of financial service products based on manus refers to the use of the business processing flow embedded in the manus system platform to conduct label modeling and behavioral identification of the audience characteristics, service scenarios, and transaction processes of financial products, and use customer segmentation data, financial preference information, and risk tolerance records to build a product adaptation model and marketing push logic system, match user portraits and product features through a rule screening engine, formulate a triggered push strategy through the task scheduling link, and combine transaction log generation, account interaction frequency tracking, and speech template configuration to form an automated marketing execution method. In data interaction, combined judgment is made based on the classification attributes of financial products, customer authorization behavior, and time series preference rules to achieve multi-path push condition judgment and dynamic combination of reach content.

[0004] Traditional financial service product marketing technology lacks in-depth analysis of user behavior intensity and product attribute sensitivity, resulting in actual deviations between marketing content and user needs. Label sensitivity and stability are not fully considered when constructing customer portraits, resulting in deviations between portrait labels and users' actual preferences, reducing the effectiveness of precision marketing, causing pushed content to frequently deviate from user interests, reducing user participation enthusiasm, and lacking dynamic analysis of the relationship between user interaction frequency and push response rate. Static fixed patterns are often used for push frequency and timing, and are unable to respond to changes in user behavior rhythm in a timely manner, resulting in inappropriate marketing timing, reduced user responsiveness, and reduced overall conversion effects of marketing activities. Summary of the Invention

[0005] In order to solve the technical problems existing in the prior art, the embodiment of the present invention provides a marketing method for manus financial service products. The technical solution is as follows:

[0006] To achieve the above objectives, the present invention adopts the following technical solution, based on the Manus financial services product marketing method, including the following steps:

[0007] S1: Use Manus to identify user account operation behavior data on multiple platforms, establish user behavior data sequences based on chronological order, identify user behavior density in multiple time periods, and obtain behavior response segment labels;

[0008] S2: Call the behavioral response segment tags, analyze the user's interactive behavior towards various financial service products, calculate the user's behavioral concentration and response frequency under each product attribute, use manus to match the portrait tags, compare the correlation strength between each portrait tag and the corresponding product attribute, optimize the tag calling order, and generate a tag-sensitive sorting sequence;

[0009] S3: Based on the tag-sensitive sorting sequence, by analyzing the stability of customer profile tags in user interaction behaviors, identifying tag confidence, and combining the tag calling order, adjusting the push priority of various marketing content, and obtaining push content screening records;

[0010] S4: Call the push content screening record, analyze the changing trend of the time interval between adjacent user operation behaviors, calculate the correlation between the user behavior interaction frequency and the push response rate, identify the optimal push frequency, combine the user's interaction frequency data in multiple time periods, identify the matching degree of multiple push time periods, and generate a push time window sequence.

[0011] As a further solution of the present invention, the behavior response segment label includes a behavior-dense segment, a behavior-sparse segment, and a behavior switching boundary; the label-sensitive sorting sequence specifically includes a sorting label name, a sorting level identifier, and a call priority number; the push content screening record includes a candidate content identifier, a trigger policy number, and a push level allocation; the push time window sequence specifically refers to a candidate time period index, a recommendation trigger time point, and a window priority number.

[0012] As a further solution of the present invention, the steps of obtaining the behavior response segment label are specifically as follows:

[0013] S101: Use Manus to identify user account operation behavior data on multiple platforms, including clicks, visits, browsing, and transactions, and establish user behavior data sequences based on operation time and behavior type;

[0014] S102: Analyze the distribution of clicks, visits, browsing, and transactions in each time period based on the user behavior sequence values, calculate the number of occurrences and distribution ratios of each behavior in multiple time periods, and obtain behavior distribution information;

[0015] S103: Identify the user's behavior density in multiple time periods based on the behavior distribution information, detect the distribution density of each behavior in the multiple time periods, and establish a behavior response segment label.

[0016] As a further solution of the present invention, the step of obtaining the label-sensitive sorting sequence is specifically as follows:

[0017] S201: Invoking the behavior response segment tag to analyze the user behavior distribution corresponding to each product attribute based on the user's click, access, browsing, and transaction behavior data for multiple financial service products in each time segment, identifying the behavior concentration and response frequency under each product attribute, and generating a product attribute behavior concentration coefficient;

[0018] S202: Identify the behavior pattern characteristics of the target account based on the product attribute behavior concentration coefficient, use manus to match the portrait tags, build a customer portrait tag set for the account, and analyze the correlation strength between each portrait tag and the corresponding product attribute to obtain the portrait attribute correlation coefficient;

[0019] S203: According to the portrait attribute correlation coefficient, compare the difference in correlation strength between each customer portrait label and product attribute, adjust the priority order of the customer portrait labels, optimize the calling order of the labels, and establish a label-sensitive sorting sequence.

[0020] As a further solution of the present invention, the specific formula for analyzing the association strength between each portrait label and the corresponding product attribute is:

[0021] ;

[0022] Calculate the portrait attribute correlation coefficient and obtain the portrait attribute correlation coefficient;

[0023] in, Represents the normalized value of the correlation strength between the j-th profile label and the corresponding financial product attribute in the i-th account, Represents the normalized value of the interaction frequency of the jth tag matching financial product attribute of the i-th account in time period t, Represents the normalized value of the access frequency of non-matching product attributes under the jth tag by the i-th account in time period t, Represents the interaction weight coefficient of financial product attributes in time period t, Represents the weight coefficient of the label feature call for time period t, Represents the normalized value of the number of profile tags for the i-th account, where i represents the target account's ID index, j represents the customer profile tag's ID index, T represents the total number of time periods divided within the profile tag association cycle, and t represents the t-th specific time period in the profile tag association cycle.

[0024] As a further solution of the present invention, the steps for obtaining the push content screening record are specifically as follows:

[0025] S301: Calling the tag-sensitive sorting sequence, analyzing the stability of the customer portrait tag in the user's interactive behavior based on the response frequency and variation range, and generating a tag behavior stability coefficient;

[0026] S302: Calculate the accuracy of each tag's current expression of the user's interests and preferences based on the tag behavior stability coefficient, evaluate the confidence of each tag, and obtain tag confidence information;

[0027] S303: Based on the tag confidence information and in combination with the tag calling order, the push priority of each type of marketing content is screened and adjusted, and a push content screening record is obtained.

[0028] As a further solution of the present invention, the specific formula for calculating the accuracy of each tag's current corresponding user interests and preferences is:

[0029] ;

[0030] Calculate label expression accuracy parameters;

[0031] Among them, s represents the current user, k represents the label number, and n represents the time window number. represents the behavioral stability coefficient of the kth tag to the current user s, represents the number of forward operations of the k-th label in the n-th time window, Indicates the number of negative operations of the k-th label in the n-th time window, represents the normalized value of the standard deviation of the operation frequency of the k-th tag in the n-th time window, represents the normalized value of the standard deviation of the operation duration of the kth label in the nth time window, m represents the total number of time windows, represents the expression accuracy parameter of the k-th tag for user s.

[0032] As a further solution of the present invention, the step of acquiring the push time window sequence is specifically as follows:

[0033] S401: Calling the push content screening record, analyzing the change direction and fluctuation trend of the operation intervals according to the time intervals between adjacent user operation behaviors, and generating change trend information;

[0034] S402: Calculate the correlation between the interaction frequency of each user behavior and the number of push responses based on the change trend information, identify the optimal push frequency for the target user, and obtain a push frequency coefficient;

[0035] S403: In combination with the push frequency coefficient, by analyzing the user's interaction frequency in multiple time periods, comparing the adaptability of each push time period to the user's active behavior, adjusting the time interval of each push task, and generating a push time window sequence.

[0036] As a further embodiment of the present invention, the method further comprises:

[0037] S5: Based on the push time window sequence, collect and analyze the user's click and stay behavior characteristics for each marketing content category, calculate the actual user attraction and response trend of each content type, adjust the push parameters of the marketing content, and obtain product marketing push records;

[0038] The product marketing push record includes a content type number, a push cycle configuration, and a response behavior summary.

[0039] As a further solution of the present invention, the steps for obtaining the product marketing push record are specifically as follows:

[0040] S501: According to the push time window sequence, collect the number of clicks and page dwell time of users on different marketing content categories in each time window, analyze the user behavior data of each content category, and generate behavior feedback data;

[0041] S502: Based on the behavioral feedback data, the actual attractiveness of each type of marketing content to the user is calculated. Combined with the changing trends of the number of clicks and the length of stay, the user's response trend to the content is analyzed to obtain content responsiveness information;

[0042] S503: Adjust push parameters of financial service product marketing content based on the content responsiveness information, including push frequency, push time point, and push content category, and obtain product marketing push records.

[0043] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0044] By identifying users' cross-platform account behavior and combining it with the behavior density in different time periods, the timeliness and representativeness of behavioral data can be improved. By using sensitivity matching between product attributes and portrait tags, accurate adaptation between user characteristics and financial products can be achieved. The tag confidence is used to adjust the priority of pushed content, improve the content relevance and effectiveness of push, adjust the push window according to the frequency of interaction, optimize the push rhythm, and combine the feedback content click and stay characteristics to optimize push parameters, thereby improving marketing accuracy and user response rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0046] Figure 1 Schematic diagram of the workflow of the present invention. DETAILED DESCRIPTION

[0047] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0048] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0049] In the embodiments of the present invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same. The terms "of," "corresponding," and "corresponding" may sometimes be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same.

[0050] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0051] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0052] See also Figure 1 The present invention provides a technical solution based on the manus financial service product marketing method, comprising the following steps:

[0053] S1: Use Manus to identify user account operation behavior data on multiple platforms, establish user behavior data sequences based on chronological order, identify user behavior density in multiple time periods, and obtain behavior response segment labels;

[0054] S2: Calling behavior response segment tags, analyzing user interactions with various financial service products, calculating the concentration of user behavior and response frequency under each product attribute, using Manus to match profile tags, comparing the correlation strength between each profile tag and the corresponding product attribute, optimizing the tag calling order, and generating a tag-sensitive sorting sequence;

[0055] S3: Based on the tag-sensitive sorting sequence, by analyzing the stability of customer profile tags in user interaction behaviors, identifying tag confidence, and combining the tag call order, adjusting the push priority of various marketing content, and obtaining push content screening records;

[0056] S4: Call push content screening records, analyze the time interval trends between adjacent user actions, calculate the correlation between user interaction frequency and push response rate, identify the optimal push frequency, combine user interaction frequency data across multiple time periods, identify the matching degree of multiple push time periods, and generate a push time window sequence;

[0057] S5: Based on the push time window sequence, collect and analyze the user's click and stay behavior characteristics for each marketing content category, calculate the actual attractiveness and response trend of each content type to users, adjust the push parameters of marketing content, and obtain product marketing push records.

[0058] Behavior response segment labels include behavior-dense segments, behavior-sparse segments, and behavior switching boundaries. The label-sensitive sorting sequence specifically includes the sorting label name, sorting level identifier, and call priority number. The push content screening record includes the candidate content identifier, trigger strategy number, and push level allocation. The push time window sequence specifically refers to the candidate time period index, recommendation trigger time point, and window priority number. The product marketing push record includes the content type number, push cycle configuration, and response behavior summary.

[0059] The steps for obtaining the behavior response segment label are as follows:

[0060] S101: Use Manus to identify user account operation behavior data on multiple platforms, including clicks, visits, browsing, and transactions, and establish user behavior data sequences based on operation time and behavior type;

[0061] Using manus, we can identify the user's account operation behavior data on multiple platforms, including clicks, visits, browsing, and transactions. We collect the operation logs of user A on three financial platforms. Each log records the operation type and specific operation time. For example, user A completed a product browsing at 9:32 on July 5, 2024, completed a financial product click at 9:34, and completed a transaction operation at 9:45. By arranging these logs in the order of the specific time when the operations occurred, we can construct a complete user behavior data sequence. When traversing each log, we can extract the platform, operation type, and operation time. The three elements of type and operation time are used to classify the specific operation times of click behavior, visit behavior, browsing behavior, and transaction behavior into independent sequences. Correspondingly, the operation type numbers are set to 1, 2, 3, and 4 respectively. Assume that user A collected a total of 10 operation logs between 9:00 and 10:00 on July 5, which are arranged in sequence as 9:02 click, 9:12 visit, 9:18 browse, 9:22 transaction, 9:27 visit, 9:32 click, 9:34 browse, 9:40 transaction, 9:43 click, and 9:45 transaction. After numbering the sequence, a behavior data sequence is formed. When analyzing the operation type of each log, the operation time is converted into minutes for subsequent calculations. The specific time of each operation is used as the index of the data sequence, which is arranged in ascending order by minutes to form a time-ordered data array. Using the above numbering rules, the behavior type array is established. , operation time array According to the above data combination, each piece of data contains three items: operation platform, operation type, and operation time. For example, the first piece of data records user A's click behavior on platform A at the 2nd minute, and the second piece of data records the access behavior on platform B at the 12th minute. And so on. If the total number of operations on different platforms by the same user in one day is N, and all operations are numbered and arranged in ascending order by time, a complete user behavior data sequence can be obtained.

[0062] S102: Analyze the distribution of clicks, visits, browsing, and transactions in each time period based on the user behavior sequence values, calculate the number of occurrences and distribution ratios of each behavior in multiple time periods, and obtain behavior distribution information;

[0063] Based on the user behavior sequence values, analyze the distribution of clicks, visits, browsing, and transaction behaviors in each time period. Assume that user A's operation period in a day is divided into three time periods: 8:00-9:00, 9:00-10:00, and 10:00-11:00. Count the number of clicks, visits, browsing, and transaction operations in each time period. For example, from 9:00 to 10:00, there are 3 clicks, 2 visits, 2 browsings, and 3 transactions. For each behavior, calculate the number of times it occurs in the time period. , and compare the total number of occurrences of this behavior in all operation logs , the distribution ratio of the behavior in this period is For example, user A's click behavior is between 9:00 and 10:00. , total number of click behaviors , get the distribution ratio of click behavior in this period , perform the same operation on visit, browse, and transaction behaviors in turn, assuming that the distribution ratio of each behavior is , , , forming a distribution ratio table of different behaviors in different time periods. The table is arranged by the two dimensions of behavior and time period, and the time period distribution ratio matrix of each behavior is formed. , analyze the extreme values ​​of the distribution ratio matrix, determine which behavior is most concentrated in which period, and sort by ratio size. Combined with actual operation data, if the transaction behavior is distributed between 10:00 and 11:00, It is higher than other time periods, indicating that trading operations are intensive during this period, and thus behavioral distribution information is obtained.

[0064] S103: Identify the user's behavior density in multiple time periods based on the behavior distribution information, detect the distribution density of each behavior in multiple time periods, and establish a behavior response segment label;

[0065] Based on the behavior distribution information, we identify the behavior density of users in multiple time periods, detect the distribution density of each behavior in multiple time periods, and define the base ratio of density determination as 0.7, that is, when When the behavior is determined to be high-density in this period, combined with the above example, if the click ratio between 9:00-10:00 is 0.5, which is lower than the intensive judgment benchmark, and the transaction ratio between 10:00-11:00 is 0.8, which is higher than the judgment benchmark, the transaction behavior between 10:00-11:00 is marked as a high-density behavior segment. The distribution ratio of each behavior in all time periods is compared with the benchmark ratio item by item to form a density judgment result table. Each row in the table represents a time period, the column represents the behavior type, and the cell is filled with high density or general. Combined with the judgment results, the high-density behavior segment is selected as the behavior response segment label. For example, the output is 9:00-10:00 transactions, 10:00-11:00 transactions, 8:00-9:00 clicks, and then the behavior response segment label is established.

[0066] The specific steps for obtaining the label-sensitive sorting sequence are:

[0067] S201: Invoke the behavior response segment tag to analyze the user behavior distribution corresponding to each product attribute based on the user's click, access, browsing, and transaction behavior data for various financial service products within each time segment, identify the behavior concentration and response frequency under each product attribute, and generate a product attribute behavior concentration coefficient;

[0068] Call the behavior response segment tag. According to the user's click, visit, browse and transaction behavior data on various financial service products in each time segment, first obtain all the user's behavior logs related to financial product attributes in different time segments, and identify the behavior type, operation time and corresponding product attributes one by one. For example, the number of clicks, visits, browsing and transaction behaviors of product A in the three segments of 8:00-9:00, 9:00-10:00 and 10:00-11:00 are 3, 5, 2, 4, 6, 1, 2, 4, 5 and 2 respectively. Set the segment number to , the product attribute number to , and the behavior number in each segment to . For each product attribute, count the total number of each behavior type under the attribute in each segment. For example, the number of clicks of product A in the segment of 9:00-10:00 is counted, and then the total number of occurrences of each behavior type in the entire segment is counted. Normalization processing is performed on the rows, that is, the number of times in each segment is divided by the total number of times the behavior appears in all segments under the same product attribute. For example, the total number of click behaviors of product A in the three segments is , then the normalized value of the 9:00-10:00 segment is , and the normalized distribution coefficients of all behavior types and all segments are calculated in turn to construct a three-dimensional distribution matrix. Then, it is determined whether the distribution coefficient of each behavior in each segment is greater than the set behavior concentration benchmark coefficient. This benchmark coefficient is set to 0.4. If the normalization coefficient is greater than 0.4, it is considered that the behavior distribution of the product attribute in the segment is concentrated. For example, the click behavior of product A is concentrated from 9:00 to 10:00. The distribution concentration of each behavior type and segment under all product attributes is determined in turn, and all results are analyzed. The normalized distribution coefficients of each segment, each behavior, and each product attribute are summarized into the behavior distribution matrix. Finally, the product attribute behavior concentration coefficient is sorted and output.

[0069] S202: Based on the product attribute behavior concentration coefficient, identify the target account's behavior pattern characteristics, use manus to match the portrait tags, build the account's customer portrait tag set, and analyze the correlation strength between each portrait tag and the corresponding product attribute to obtain the portrait attribute correlation coefficient;

[0070] The specific formula for analyzing the association strength between each portrait label and the corresponding product attribute is:

[0071] ;

[0072] Calculate the portrait attribute correlation coefficient and obtain the portrait attribute correlation coefficient;

[0073] in, Represents the normalized value of the correlation strength between the j-th profile label and the corresponding financial product attribute in the i-th account, Represents the normalized value of the interaction frequency of the jth tag matching financial product attribute of the i-th account in time period t, Represents the normalized value of the access frequency of non-matching product attributes under the jth tag by the i-th account in time period t, Represents the interaction weight coefficient of financial product attributes in time period t, Represents the weight coefficient of the label feature call for time period t, Represents the normalized value of the number of profile tags for the i-th account, where i represents the target account's ID index, j represents the customer profile tag's ID index, T represents the total number of time periods divided within the profile tag association cycle, and t represents the t-th specific time period in the profile tag association cycle.

[0074] formula:

[0075] ;

[0076] Detailed explanation of the formula and the process of formula calculation and derivation:

[0077] The formula is used to calculate the normalized correlation strength between the j-th portrait label in the i-th account and the associated financial product attributes. The result is used to measure the matching ability between the current label and the target product characteristics, and is used to drive the generation of subsequent push strategies.

[0078] Parameter meaning and setting value:

[0079] The normalized value of the interaction frequency of the product attribute associated with the jth tag by the i-th account in the t-th time period;

[0080] The normalized value of the access frequency of the i-th account to the product attributes not associated with the tag in the t-th time period;

[0081] is the interaction weight coefficient of financial product attributes in time period t. This coefficient varies with the complexity of the product type and is set to 0.8 for complex products, 0.6 for ordinary products, and 0.4 for simple products;

[0082] The weight coefficient for the time period t label call is set to 0.75;

[0083] The normalized value of the number of portrait tags of the i-th account during the portrait modeling cycle is set to 0.6;

[0084] T is the number of time periods in the associated cycle, which is divided into 3 time periods, where t=1, , , , , when t=2, , , , , when t=3, , , , ;

[0085] Substitute the parameters into the formula for calculation:

[0086] ;

[0087] ;

[0088] ;

[0089] ;

[0090] =(0.6+0.3)+(0.4+0.5)+(0.5+0.4)=0.9+0.9+0.9=2.7;

[0091] ;

[0092] 2.7+0.7746=3.4746;

[0093] ;

[0094] The result of 0.0777 indicates that the correlation strength between tag j and the target financial product attributes for account i is relatively low. Values ​​closer to 1 indicate a stronger match for the tag with the current product attributes, while values ​​closer to 0 indicate a weaker match. Therefore, this tag should be assigned a lower content matching priority in the push strategy. This result, used as the profile attribute correlation coefficient, contributes to tag sorting and priority assignment, ultimately generating a tag-sensitive ranking sequence.

[0095] S203: Based on the portrait attribute correlation coefficient, compare the correlation strength difference between each customer portrait tag and the product attribute, adjust the priority order of the customer portrait tags, optimize the tag calling order, and establish a tag-sensitive sorting sequence;

[0096] Based on the correlation coefficients of profile attributes, the correlation strengths between each customer profile tag and product attribute are compared. The priority order of customer profile tags is adjusted, the tag call order is optimized, and a tag-sensitive sorting sequence is established. First, based on the correlation coefficient matrix in the previous section, all correlation coefficients are sorted by tag and product attribute. For the same account, the correlation coefficients between different tags and the same product attribute are compared in descending order. Tags with correlation coefficients above 0.5 are prioritized, those between 0.2 and 0.5 are standard tags, and those below 0.2 are candidate tags. This prioritizes the tags and arranges them in order to form a tag call queue. Tags with high correlations with product attributes are prioritized for marketing content triggering. For example, if the correlation coefficient between account C's "High-Frequency Wealth Management" tag and product C is 0.7, the "Medium Preference" tag is 0.4, and the "Risk Sensitive" tag is 0.15, then the priority order is "High-Frequency Wealth Management" > "Medium Preference" > "Risk Sensitive." All priority queues form the tag-sensitive sorting sequence, forming the final sorting structure.

[0097] The specific steps for obtaining push content filtering records are as follows:

[0098] S301: Calling the tag-sensitive sorting sequence, analyzing the stability of customer profile tags in user interaction behavior based on response frequency and variation range, and generating a tag behavior stability coefficient;

[0099] Call the tag-sensitive sorting sequence and analyze the stability of customer portrait tags in user interaction behaviors based on the response frequency and variation range. First, call the tag-sensitive sorting sequence obtained in the previous process. For each tag, obtain the actual number of responses of the tag in different interaction behaviors. Count the frequency of tag triggering within the daily, weekly, or monthly intervals. For example, the "high-risk financial management preference" tag is triggered 15 times, 20 times, 18 times, 16 times, and 17 times within 30 days to form a frequency array. Calculate the maximum and minimum values ​​in the array to obtain the variation range. The formula for calculating the variation coefficient is: ,in is the maximum number of responses, is the minimum number of responses. In this example, , ,have to , combined with the average response frequency of the label, such as the average is 17.2, set the label behavior stability judgment standard to be stable when the coefficient of variation is less than 0.3 and the average is greater than 10, and count all labels in turn to form a label behavior stability coefficient matrix, where the matrix elements are the behavioral stability values ​​of each label.

[0100] S302: Calculate the accuracy of each tag's current expression of the user's interests and preferences based on the tag behavior stability coefficient, evaluate the confidence of each tag, and obtain tag confidence information;

[0101] The specific formula for calculating the accuracy of each tag's current corresponding user interests and preferences is:

[0102] ;

[0103] Calculate label expression accuracy parameters;

[0104] Among them, s represents the current user, k represents the label number, and n represents the time window number. represents the behavioral stability coefficient of the kth tag to the current user s, represents the number of forward operations of the k-th label in the n-th time window, Indicates the number of negative operations of the k-th label in the n-th time window, represents the normalized value of the standard deviation of the operation frequency of the k-th tag in the n-th time window, represents the normalized value of the standard deviation of the operation duration of the kth label in the nth time window, m represents the total number of time windows, represents the expression accuracy parameter of the k-th tag for user s.

[0105] formula:

[0106] ;

[0107] Detailed explanation of the formula and the process of formula calculation and derivation:

[0108] The formula is used to calculate the accuracy of tags in expressing user interest preferences, and the results are used to quantify the matching weight of tags to users in the recommendation system.

[0109] Parameter meaning and setting value:

[0110] : The stability coefficient of label k for user s's behavior is obtained by inversely quantifying the standard deviation of the proportion of positive operations of the label in multiple time windows. If the fluctuation range of the positive operation ratio of a label is less than 10% within 10 consecutive monitoring cycles, the stability score result is 0.85;

[0111] : The number of positive operations of tag k in the nth time window, which comes from behaviors such as clicks, favorites, and likes. Set =30, =25, =28;

[0112] : The number of negative operations of label k in the nth time window, which comes from skipping, disinterest, etc., set =10, =15, =12;

[0113] : The normalized value of the standard deviation of the operation frequency of label k in the nth time window is normalized by the standard deviation of the number of daily operations, with a normalized upper limit of 24 times. =0.21, =0.17, =0.19.

[0114] : The normalized value of the standard deviation of the operation duration of label k in the nth time window is normalized by the standard deviation of the duration of a single operation, with a normalized upper limit of 60 seconds. =0.25, =0.22, =0.23.

[0115] m=3: total number of time windows;

[0116] Substitute the parameters into the formula for calculation:

[0117] ;

[0118] ;

[0119]

[0120] ;

[0121] This result indicates that tag k has a high degree of accuracy in expressing user s's historical behavior. A value of 1.871 indicates that the tag has a strong signal strength in expressing user interest. When it exceeds the recommendation system's threshold of 1.2, it is marked as a high-confidence tag and included in the main ranking weight matrix to optimize personalized recommendations. This parameter is then used as a weight factor in the subsequent tag priority evaluation model to improve decision accuracy.

[0122] S303: Based on the tag confidence information and the tag calling order, the push priority of each type of marketing content is screened and adjusted, and push content screening records are obtained;

[0123] Based on the tag confidence information and the tag calling order, the push priority of each type of marketing content is screened and adjusted, and the push content screening record is obtained. The tag confidence information of the previous process is sorted into three levels: high, medium, and low. The marketing content corresponding to the high-confidence tag is set as the priority push category. The content corresponding to the medium-confidence and low-confidence tags is screened in turn as supplementary and backup push content. For each round of push tasks, high-confidence tags are given priority according to the tag calling order, and the push plan is arranged according to the priority. A corresponding relationship between marketing content and tag confidence and calling order is formed. The tag, content, time and other information of each round of push tasks are recorded and organized into a push content screening record data table.

[0124] The specific steps for obtaining the push time window sequence are as follows:

[0125] S401: Calling the push content screening record, analyzing the change direction and fluctuation trend of the operation interval based on the time interval between adjacent user operation behaviors, and generating change trend information;

[0126] Call the push content screening record, and analyze the change direction and fluctuation trend of the operation interval based on the time interval between adjacent user operation behaviors. First, extract the operation time of each push and the user's actual response in the push content screening record, calculate the time interval between two adjacent operations, and construct a time interval sequence. For example, the time of the user's 5 responses are 10:01, 10:10, 10:18, 10:21, and 10:35, respectively. The intervals are 9 minutes, 8 minutes, 3 minutes, and 14 minutes, respectively. According to the sequence analysis trend, the first-order difference analysis method is used to subtract the current interval from the previous interval to obtain the difference, which is -1, -5, and 11 respectively. A positive value indicates that the interval is increasing, and a negative value indicates that the interval is decreasing. Count the number of positive and negative values ​​in the sequence to determine whether the overall interval tends to shorten or lengthen. If there are many alternating positive and negative values, it means that the fluctuation is large, and the volatility can be further calculated. The volatility formula is , where B is the volatility and n is the number of responses. is the time interval of the ith operation, and the volatility in this case is , and finally output the change trend information.

[0127] S402: Calculate the correlation between the interaction frequency of each user behavior and the number of push responses based on the change trend information, identify the optimal push frequency for the target user, and obtain the push frequency coefficient;

[0128] Based on the change trend information, the correlation between the interaction frequency of each user behavior and the number of push responses is calculated to identify the optimal push frequency for the target user and obtain the push frequency coefficient. First, the interaction frequency of the same behavior category is extracted from the push content screening records and change trend information. and the corresponding push response number , all behaviors and push response numbers are paired. For example, in browsing behavior, a user pushes 10 times and responds 7 times; in click behavior, a user pushes 8 times and responds 5 times; in transaction behavior, a user pushes 4 times and responds 2 times. The Pearson correlation coefficient is used to calculate the correlation between interaction frequency and response number. The correlation coefficient formula is ,in, is the correlation coefficient between interaction frequency and the number of responses, is the interaction frequency of the i-th behavior, is the number of responses to the ith behavior, is the mean interaction frequency, is the response mean, n is the number of behavior types, if If it is greater than 0.7, it is considered a strong correlation. The number of responses at each frequency is analyzed, and the frequency with the highest response rate is selected as the optimal push frequency. The push frequency coefficient is output. For example, when the browsing behavior push frequency is once every 10 minutes, the response rate reaches 70%. This frequency is the optimal frequency, and the push frequency coefficient is 0.7.

[0129] S403: By analyzing the user's interaction frequency in multiple time periods based on the push frequency coefficient, and comparing the degree of fit between each push time period and the user's active behavior, the time interval of each push task is adjusted to generate a push time window sequence;

[0130] Combined with the push frequency coefficient, by analyzing the user's interaction frequency in multiple time periods, comparing the adaptability of each push time period and the user's active behavior, the time interval of each push task is adjusted. First, the push frequency coefficient is compared with the user's actual interaction frequency distribution in each time period, and the time segments close to the optimal push frequency coefficient are extracted. For example, a user's interaction frequency from 12:00-13:00 and from 18:00-19:00 is once every 10 minutes and once every 15 minutes respectively. The time period with an interaction frequency of once every 10 minutes is prioritized as the push window. The degree of proximity between the interaction frequency of each push time period and the optimal push frequency coefficient is compared in turn, and the push task time interval is arranged according to the proximity. Finally, the push time window sequence is output.

[0131] The specific steps for obtaining product marketing push records are as follows:

[0132] S501: Based on the push time window sequence, collect the number of clicks and page dwell time of users on different marketing content categories within each time window, analyze the user behavior data for each content category, and generate behavior feedback data;

[0133] According to the push time window sequence, the number of clicks and page dwell time of users on different marketing content categories in each time window are collected, and the user behavior data of each content category is analyzed. First, for each push time window, the marketing content categories that users browsed, clicked, and interacted with in this time period are classified and counted. The content category number is set to j, and the number of clicks for each content category in the same time window is collected. Length of time spent on the page For example, in the 12:00-13:00 time window, the user clicked on Category A content 4 times, with a total stay time of 300 seconds, and clicked on Category B content 2 times, with a stay time of 180 seconds. The total number of clicks and stay time of each content category in each window are collected, and multi-window statistics are performed on the data in the entire cycle to form a two-dimensional behavior feedback table of content category and time window, and the behavior feedback data is output.

[0134] S502: Based on the behavioral feedback data, the actual attractiveness of each type of marketing content to the user is calculated. Combined with the changing trends of the number of clicks and the length of stay, the user's response trend to the content is analyzed to obtain content responsiveness information;

[0135] Based on behavioral feedback data, calculate the actual attractiveness of each type of marketing content to users. Combined with the changing trends of click counts and dwell time, analyze the user response trends to the content. First, for each content category, calculate the average number of clicks over the entire cycle. and the mean length of stay For example, the number of clicks on the A-type content in the four windows is 4, 5, 6, and 3 respectively, and the total average is The length of stay is 300, 320, 340 and 310 respectively, and the total average Seconds, using comprehensive content responsiveness scoring ,in, Score the responsiveness of the j-th category of content, 、 is the weight (can be 0.5), 、 The maximum values ​​of the mean click and dwell time of all categories are taken as input, and the content responsiveness score of category A is obtained. The responsiveness of all categories is sorted, and the trend of click and dwell time is analyzed to output the content responsiveness information. is the mean click value of content category j, is the mean duration of stay of content category j, is the maximum click mean, is the mean maximum stay time.

[0136] S503: Adjust push parameters for financial service product marketing content based on the content responsiveness information, including push frequency, push time, and push content category, and obtain product marketing push records;

[0137] Based on the content responsiveness information, adjust the push parameters of financial service product marketing content, including push frequency, push time and push content category, obtain product marketing push records, and set the content with the highest responsiveness score as the priority push object based on the responsiveness score of each content category. Combined with the content preferences in different time windows, the push frequency is tilted towards high-responsive content categories, and the push time points are aligned with the responsiveness peak window. Different push frequencies and push time points are assigned to different content categories, and the content push order is adjusted from high to low priority. Finally, the content category, time window, push frequency and other parameters in each round of push tasks are recorded to form product marketing push records.

[0138] The above embodiments can be implemented in whole or in part via software, hardware (e.g., circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product comprises one or more computer instructions or computer programs. When loaded or executed on a computer, the processes or functions described in accordance with the embodiments of the present invention are fully or partially performed. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired means (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server or data center that contains a collection of one or more available media. The available medium can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0139] It should be understood that the term "and / or" as used herein simply describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the associated objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0140] In this disclosure, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, "at least one of a, b, or c" can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or plural.

[0141] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0142] Those skilled in the art will appreciate that the units and algorithm steps of each example 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 performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0143] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0144] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of the device or unit, which can be electrical, mechanical or other forms.

[0145] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0146] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0147] If the functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or the portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage media include various media capable of storing program code, such as USB flash drives, mobile hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0148] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. Based on the manus financial services product marketing method, characterized by, The method comprises: S1: Use Manus to identify user account operation behavior data on multiple platforms, establish user behavior data sequences based on chronological order, identify user behavior density in multiple time periods, and obtain behavior response segment labels; S2: Call the behavioral response segment tags, analyze the user's interactive behavior towards various financial service products, calculate the user's behavioral concentration and response frequency under each product attribute, use manus to match the portrait tags, compare the correlation strength between each portrait tag and the corresponding product attribute, optimize the tag calling order, and generate a tag-sensitive sorting sequence; The steps for obtaining the label-sensitive sorting sequence are specifically as follows: S201: Invoking the behavior response segment tag to analyze the user behavior distribution corresponding to each product attribute based on the user's click, access, browsing, and transaction behavior data for multiple financial service products in each time segment, identifying the behavior concentration and response frequency under each product attribute, and generating a product attribute behavior concentration coefficient; S202: Identify the behavior pattern characteristics of the target account based on the product attribute behavior concentration coefficient, use manus to match the portrait tags, build a customer portrait tag set for the account, and analyze the correlation strength between each portrait tag and the corresponding product attribute to obtain the portrait attribute correlation coefficient; The specific formula for analyzing the association strength between each portrait label and the corresponding product attribute is: ； Calculate the portrait attribute correlation coefficient and obtain the portrait attribute correlation coefficient; in, Represents the normalized value of the correlation strength between the j-th profile label and the corresponding financial product attribute in the i-th account, Represents the normalized value of the interaction frequency of the jth tag matching financial product attribute of the i-th account in time period t, Represents the normalized value of the access frequency of non-matching product attributes under the jth tag by the i-th account in time period t, Represents the interaction weight coefficient of financial product attributes in time period t, Represents the weight coefficient of the label feature call for time period t, Represents the normalized value of the number of profile tags for the i-th account, where i represents the target account's ID index, j represents the ID index of the customer profile tag, T represents the total number of time periods within the profile tag association cycle, and t represents the t-th specific time period within the profile tag association cycle; S203: comparing the difference in correlation strength between each customer portrait tag and the product attribute based on the portrait attribute correlation coefficient, adjusting the priority order of the customer portrait tags, optimizing the tag calling order, and establishing a tag-sensitive sorting sequence; S3: Based on the tag-sensitive sorting sequence, by analyzing the stability of customer profile tags in user interaction behaviors, identifying tag confidence, and combining the tag calling order, adjusting the push priority of various marketing content, and obtaining push content screening records; S4: Call the push content screening record, analyze the changing trend of the time interval between adjacent user operation behaviors, calculate the correlation between the user behavior interaction frequency and the push response rate, identify the optimal push frequency, combine the user's interaction frequency data in multiple time periods, identify the matching degree of multiple push time periods, and generate a push time window sequence.

2. The marketing method of manus financial service products according to claim 1, characterized in that: The behavior response segment label includes a behavior-dense segment, a behavior-sparse segment, and a behavior switching boundary. The label-sensitive sorting sequence specifically includes the sorting label name, the sorting level identifier, and the call priority number. The push content screening record includes the candidate content identifier, the trigger policy number, and the push level allocation. The push time window sequence specifically refers to the candidate time period index, the recommendation trigger time point, and the window priority number.

3. The marketing method of manus financial service products according to claim 1, characterized in that: The steps for obtaining the behavior response segment label are specifically as follows: S101: Use Manus to identify user account operation behavior data on multiple platforms, including clicks, visits, browsing, and transactions, and establish user behavior data sequences based on operation time and behavior type; S102: Analyze the distribution of clicks, visits, browsing, and transactions in each time period based on the user behavior sequence values, calculate the number of occurrences and distribution ratios of each behavior in multiple time periods, and obtain behavior distribution information; S103: Identify the user's behavior density in multiple time periods based on the behavior distribution information, detect the distribution density of each behavior in the multiple time periods, and establish a behavior response segment label.

4. The marketing method of manus-based financial service products according to claim 3, characterized in that: The steps for obtaining the push content screening record are specifically as follows: S301: Calling the tag-sensitive sorting sequence, analyzing the stability of the customer portrait tag in the user's interactive behavior based on the response frequency and variation range, and generating a tag behavior stability coefficient; S302: Calculate the accuracy of each tag's current expression of the user's interests and preferences based on the tag behavior stability coefficient, evaluate the confidence of each tag, and obtain tag confidence information; S303: Based on the tag confidence information and in combination with the tag calling order, the push priority of each type of marketing content is screened and adjusted, and a push content screening record is obtained.

5. The marketing method of manus financial service products according to claim 4, characterized in that: The specific formula for calculating the accuracy of each tag's current corresponding user interests and preferences is: ； Calculate label expression accuracy parameters; Among them, s represents the current user, k represents the label number, and n represents the time window number. represents the behavioral stability coefficient of the kth tag to the current user s, represents the number of forward operations of the k-th label in the n-th time window, Indicates the number of negative operations of the k-th label in the n-th time window, represents the normalized value of the standard deviation of the operation frequency of the k-th tag in the n-th time window, represents the normalized value of the standard deviation of the operation duration of the kth label in the nth time window, m represents the total number of time windows, represents the expression accuracy parameter of the k-th tag for user s.

6. The marketing method of manus-based financial service products according to claim 4, characterized in that: The specific steps for acquiring the push time window sequence are: S401: Calling the push content screening record, analyzing the change direction and fluctuation trend of the operation intervals according to the time intervals between adjacent user operation behaviors, and generating change trend information; S402: Calculate the correlation between the interaction frequency of each user behavior and the number of push responses based on the change trend information, identify the optimal push frequency for the target user, and obtain a push frequency coefficient; S403: In combination with the push frequency coefficient, by analyzing the user's interaction frequency in multiple time periods, comparing the adaptability of each push time period to the user's active behavior, adjusting the time interval of each push task, and generating a push time window sequence.

7. The marketing method of manus-based financial service products according to claim 1, characterized in that: The method further comprises: S5: Based on the push time window sequence, collect and analyze the user's click and stay behavior characteristics for each marketing content category, calculate the actual user attraction and response trend of each content type, adjust the push parameters of the marketing content, and obtain product marketing push records; The product marketing push record includes a content type number, a push cycle configuration, and a response behavior summary.

8. The marketing method of manus-based financial service products according to claim 7, characterized in that: The specific steps for obtaining the product marketing push record are as follows: S501: According to the push time window sequence, collect the number of clicks and page dwell time of users on different marketing content categories in each time window, analyze the user behavior data of each content category, and generate behavior feedback data; S502: Based on the behavioral feedback data, the actual attractiveness of each type of marketing content to the user is calculated. Combined with the changing trends of the number of clicks and the length of stay, the user's response trend to the content is analyzed to obtain content responsiveness information; S503: Adjust push parameters of financial service product marketing content based on the content responsiveness information, including push frequency, push time point, and push content category, and obtain product marketing push records.

Images