A Big Data-Based Method for Deduplicating Marketing and Promotion Data

Abstract

This invention belongs to the field of data deduplication technology and provides a method for deduplicating marketing promotion data based on big data. The invention obtains a sample space by describing the marketing status within the corresponding time period of a marketing action. Then, it analyzes the relative modulus of the marketing action node relative to the hash ring and the number of equal divisions of the hash ring to obtain the update mapping position of the marketing action node on the hash ring in different unit time periods. Based on the update mapping position, it analyzes the influence heat of the marketing action node on different unit time periods. According to the relative modulus and influence heat, combined with the number of unit time periods contained within the corresponding time period of the marketing action, it obtains the priority score of the marketing action in different unit time periods, and deduplicates the marketing data. This invention achieves change analysis for the marketing promotion process, effectively avoiding the loss of marketing promotion data, thereby improving the accuracy of marketing promotion status analysis and the timeliness of collaborative promotion data analysis between data storage nodes.

Description

Technical Field

[0001] This invention relates to the field of data deduplication technology, and in particular to a method for deduplicating marketing and promotion data based on big data. Background Technology

[0002] Marketing promotion data is often tailored to different regions with varying marketing strategies. The different levels of implementation of these strategies in different regions result in varying amounts of data feedback, leading to redundancy in marketing promotion data and consequently increasing the resource consumption of the overall server storage space. However, since duplicate results are distributed across different promotion feedback results, directly segmenting the data based on the changing characteristics of these feedback results may result in incomplete integration of marketing strategies.

[0003] Existing methods often use consistent hashing algorithms for data similarity matching and block partitioning. However, for the update mapping position of the hash ring in consistent hashing, highly similar data are often mapped to positions close to each other in the hash ring. This leads to the problem that storage node-centered partitioning methods tend to split highly similar but gap-like data mapping points into different parts, resulting in the loss of marketing and promotion data, as well as the problem of low timeliness of collaborative promotion data analysis between data storage nodes. Summary of the Invention

[0004] To address the above technical problems, this invention provides a method for deduplicating marketing and promotion data based on big data.

[0005] According to the present invention, a method for deduplicating marketing promotion data based on big data is provided, the method comprising:

[0006] Extract marketing data to obtain a data form, and obtain the marketing action node and the corresponding number of duplicate results for any marketing action;

[0007] Based on the data form, the marketing status within the corresponding time period of the marketing action is described to obtain the sample space;

[0008] Based on the sample space and the data form, the relative modulus of the marketing action node relative to the hash ring and the number of equal divisions of the hash ring are analyzed in different time periods to obtain the update mapping position of the marketing action node on the hash ring in different time periods.

[0009] Based on the updated mapping location, analyze the impact of the marketing action node on the popularity of different time periods;

[0010] Based on the relative modulus and the influence intensity, and combined with the number of time periods contained within the corresponding time period of the marketing action, the priority score of the marketing action in different time periods is obtained;

[0011] Based on the priority score and the number of duplicate results, the marketing data is deduplicated.

[0012] In some embodiments of the present invention, the data form includes: the time of user click, the webpage tab clicked, and the amount spent during a single webpage browsing session.

[0013] In some embodiments of the present invention, based on the data form, the marketing status within the time period corresponding to the marketing action is described to obtain a sample space, including:

[0014] Based on the start and end points of the marketing action, the corresponding time period of the marketing action is obtained;

[0015] Based on the user click time and the spending amount in the data form, analyze the percentage increase in page views and the percentage increase in conversion within the corresponding time period of the marketing action;

[0016] Based on the growth rate of page views and the growth rate of conversion, a coordinate system is established to describe the marketing status within the corresponding time period of the marketing action, thus obtaining a sample space.

[0017] In some embodiments of the present invention, based on the sample space and the data form, the relative modulus of the marketing action node relative to the hash ring and the number of equal divisions of the hash ring in different unit time periods are analyzed to obtain the updated mapping position of the marketing action node on the hash ring in different unit time periods, including:

[0018] Based on the sample space, analyze the node weights of the marketing action nodes in different time units;

[0019] Based on the node weight, combined with the number of user visits to the marketing action node in different time periods, and the time period occupancy rate in different time periods, the relative modulus of the marketing action node relative to the hash ring and the number of equal divisions of the hash ring in different time periods are obtained.

[0020] Based on the relative modulus and the number of equal divisions of the hash ring, the update mapping position of the marketing action node on the hash ring is obtained for different time periods.

[0021] In some embodiments of the present invention, based on the sample space, the node weights of the marketing action nodes are analyzed in different time units, including:

[0022] Retrieve the relevant marketing action nodes that are related to the same webpage tag as the aforementioned marketing action node;

[0023] Within the sample space, the unit time period in which the relevant marketing action node is located is obtained;

[0024] Calculate the time interval between the unit time interval where the marketing action node is located in the sample space and the unit time interval corresponding to the nearest related marketing action node, and use it as the node weight of the marketing action node in different unit time intervals.

[0025] In some embodiments of the present invention, the updated mapping position of the marketing action node on the hash ring for different unit time periods is obtained according to the relative modulus, including:

[0026] Using the relative modulus of different time periods as the step number and the equal division of the hash ring as the step size, the position on the hash ring is selected along the growth order of the hash ring, which serves as the update mapping position of the marketing action node on the hash ring for different time periods.

[0027] In some embodiments of the present invention, based on the updated mapping position, the influence of the marketing action node on the popularity of different time periods is analyzed, including:

[0028] For different time periods, calculate the first hash distance between the updated mapping position and the original mapping position of the marketing action node;

[0029] For different time periods, determine the original mapping position of other marketing action nodes on the update mapping position side corresponding to the marketing action node, and denot it as the reference original mapping position. Calculate the second hash distance between the original mapping position of the marketing action node and the reference original mapping position.

[0030] For different time periods, calculate the standard deviation of the distance between the original mapping position of other marketing action nodes and the original mapping position of the marketing action node between the original mapping position and the updated mapping position of the marketing action node;

[0031] By combining the first hash distance, the second hash distance, and the standard deviation, the influence of the marketing action node on different time periods is obtained.

[0032] In some embodiments of the present invention, based on the relative modulus and the influence intensity, and combined with the number of unit time periods contained within the corresponding time period of the marketing action, a priority score for the marketing action within different unit time periods is obtained, including:

[0033] Based on the relative modulus corresponding to different time units, analyze the consistency deviation between the marketing action nodes and the target nodes;

[0034] Based on the impact of the marketing action nodes on different time periods, analyze the time period focus characteristics brought about by the marketing action nodes;

[0035] Based on the consistency deviation and the focus characteristics, combined with the number of time periods corresponding to the marketing action, the priority score of the marketing action in different time periods is obtained.

[0036] In some embodiments of the present invention, the consistency deviation between the marketing action node and the target node is analyzed based on the relative modulus corresponding to different unit time periods, including:

[0037] Obtain the maximum relative modulus among the relative moduli of the marketing action nodes relative to the hash ring in different time periods;

[0038] Based on the relative modulus corresponding to different time periods and the maximum relative modulus, the consistency deviation between the marketing action node and the target node is obtained.

[0039] In some embodiments of the present invention, the focus characteristics brought about by the marketing action nodes during different time periods are analyzed based on the influence intensity of the marketing action nodes on different time periods, including:

[0040] Calculate the sum of the influence heat of all the marketing action nodes on the unit time period to obtain the time period focus characteristics brought about by the marketing action nodes.

[0041] As can be seen from the above embodiments, the marketing promotion data deduplication method based on big data provided by the embodiments of the present invention has the following beneficial effects:

[0042] This invention describes the marketing status within the corresponding time period of a marketing action based on extracted data forms, obtaining a sample space. Based on the sample space and data forms, it analyzes the relative modulus of marketing action nodes relative to the hash ring and the number of equal divisions of the hash ring in different unit time periods, obtaining the update mapping position of marketing action nodes on the hash ring in different unit time periods. Based on the update mapping position, it analyzes the influence heat of marketing action nodes on different unit time periods. According to the relative modulus and influence heat, combined with the number of unit time periods contained within the corresponding time period of the marketing action, it obtains the priority score of the marketing action in different unit time periods, and deduplicates the marketing data. This invention achieves change analysis of the marketing promotion process by combining the stage state transformation of marketing promotion for marketing status decomposition and evaluation, and judges the progress of marketing promotion by the mapping of marketing status. Furthermore, it combines marketing promotion results with data heterogeneous decomposition, which can effectively avoid the problem of marketing promotion data loss, thereby improving the accuracy of customer promotion status analysis during the marketing process and improving the timeliness of collaborative promotion data analysis between marketing data storage nodes.

[0043] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0044] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 A schematic diagram of the basic process of a marketing promotion data deduplication method based on big data provided in an embodiment of the present invention;

[0046] Figure 2 This is a schematic diagram of a hash ring mapping structure provided in an embodiment of the present invention. Detailed Implementation

[0047] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a big data-based marketing promotion data deduplication method proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0048] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. Terms such as “comprising,” “including,” or any variations thereof are intended to cover a non-exclusive inclusion, such that a circuit structure, article, or device that comprises a list of elements includes not only those elements but also elements not expressly listed, or elements inherent to such article or device. Without further limitation, an element defined by the phrase “comprising one…” does not exclude the presence of additional identical elements in the article or device that includes the element.

[0049] The following will provide a detailed description of a big data-based deduplication method for marketing and promotion data provided in this embodiment, with reference to the accompanying drawings.

[0050] Please see Figure 1 This illustrates the basic process of a big data-based deduplication method for marketing promotion data provided by an embodiment of the present invention.

[0051] like Figure 1 As shown, an embodiment of the present invention provides a method for deduplicating marketing promotion data based on big data, which specifically includes the following steps:

[0052] S100: Extract marketing data to obtain a data form, and obtain the number of marketing action nodes and the number of duplicate results corresponding to the marketing action nodes.

[0053] First, extract marketing data sources from various interfaces on the server. Among them:

[0054] API Interface: Connects to the open APIs of mainstream marketing platforms to obtain structured promotion data, such as click-through rate, conversion rate, and user behavior.

[0055] Log files: Collect server logs and event tracking logs, such as user access paths and page dwell time.

[0056] Third-party databases: Pull customer order data from CRM systems (such as Salesforce) and e-commerce platforms (such as Shopify).

[0057] Social Media: Scrape social media comments using web crawlers or open APIs (compliance must be considered).

[0058] IoT devices: collect real-time interactive data from offline devices (such as smart POS machines and sensors).

[0059] Then, marketing data is extracted from the marketing data source using ETL tools (extract, transform, load) to obtain a data form, as shown in Table 1. The data form includes: User ID. User click time Click on the webpage tab Webpage browsing time And the amount spent during a single web browsing session. .

[0060] Table 1 Data Form Description

[0061]

[0062] Additionally, for any given marketing action, obtain the marketing action node, specifically the start and end action nodes. Also, obtain the number of duplicate results within the corresponding marketing action period.

[0063] S200: Based on the data form, describe the marketing status within the corresponding time period of the marketing action to obtain the sample space.

[0064] Marketing promotion plans involve artificially influencing clicks and announcements on web pages, resulting in targeted exposure through changes in user browsing outcomes. Therefore, the first step is to map the promotional weight of the web page based on the marketing promotion plan to a data form. From this data form, the dominant impact of adjusting the web page's promotional weight through the marketing promotion plan is extracted, thus establishing a promotional state chain. Finally, the attribute changes caused by the nodes in the promotional state chain are combined with the dominant attributes of the marketing results to describe the marketing state.

[0065] Based on the above analysis, in some embodiments of the present invention, a sample space is obtained by describing the marketing status within the corresponding time period of a marketing action based on a data form. Further aspects include:

[0066] First, the corresponding time period of a marketing action is obtained based on its start and end points. Specifically, for any given marketing action, a start point and an end point are obtained, and the time interval between the start and end points is the corresponding time period of the marketing action. It should be noted that in this embodiment of the invention, the preset unit time period is 30 minutes; however, the corresponding time period of a marketing action may include one or more unit time periods.

[0067] Then, since marketing actions can increase the frequency of visits to the corresponding topic within a limited time frame, and the increased frequency of visits can lead to an increase in conversion and spending rates, the effectiveness of the current marketing action in terms of conversion is ultimately determined. Therefore, based on the user click times and spending amounts in the data form, the percentage increase in page views and conversion rates during the corresponding time period of the marketing action are analyzed. Specifically:

[0068] In the data form, based on the user's click time, retrieve each time unit within the corresponding time period of the current marketing action. Internal user tags on web pages The number of clicks on (the webpage tags targeted by the marketing campaign) is recorded as follows: ; and obtain each unit time period within the time period corresponding to the current marketing action. The total number of user clicks on all webpage tags is denoted as . ; Calculate webpage tags Corresponding number of user clicks Total number of user clicks corresponding to all webpage tags The ratio is Get each unit time period corresponding to the current marketing action. The percentage increase in pageviews within the page was used to evaluate the exposure target.

[0069] Further analysis is needed to understand the conversion rate represented by pageviews, specifically the conversion evaluation of the percentage increase in pageviews to the percentage increase in spending, and the increase in spending resulting from increased pageviews. More specifically, the data form retrieves user spending data for each unit of time without marketing activities, calculates the average spending price, and denotes it as... And obtain the time unit for each period corresponding to the marketing action; The average spending amount of internal users is calculated and denoted as [data missing]. ; Calculate the average purchase price corresponding to the marketing campaign period. Average spending price corresponding to periods without marketing activity The ratio is To obtain each unit time period corresponding to the marketing action The conversion growth rate within the period, and the conversion growth rate Establish the output characteristic space.

[0070] Finally, the state transition process exhibited by the comparative dynamics generated during the marketing and promotion process is analyzed: that is, the obtained unit time period. Internal pageview growth rate and conversion growth rate , respectively, serve as the horizontal and vertical axes of a planar coordinate system, thereby describing the marketing status of each marketing action within the corresponding time period, thus obtaining the sample space. It should be noted that one unit of time period corresponds to one point in the sample space.

[0071] S300: Based on the sample space and data form, analyze the relative modulus of marketing action nodes in different time periods relative to the hash ring, and obtain the update mapping position of marketing action nodes in different time periods on the hash ring.

[0072] In the marketing and promotion process, marketing actions may not directly produce corresponding changes in page views, but the web page tags targeted by marketing actions do not only affect the current web page tags. There are instances where certain webpage tags serve as intermediate links in the browsing process to ultimately achieve marketing objectives. However, the detours in the marketing chain caused by these actions decouple the marketing activities from their goals, leading to an accumulation of marketing budgets and consequently, a buildup of duplicate marketing results, resulting in data duplication. Therefore, by combining the dispersion of duplication characteristics in the hash ring's mapping position, we can determine the stratification of marketing data stored within a unit of time.

[0073] The marketing action is first projected into the hash ring to output the action result of the current marketing array. That is, the data of a single row of the data form is extracted and mapped into the hash ring through a hash function to obtain the change result mapped to the current row.

[0074] By distributing the same user's requests across multiple virtual nodes on a hash ring, the "data skew" of traditional hashing is avoided. Combined with the output ratio of user requests within a time period, the changes in access characteristics targeted by the current marketing action can be described in a targeted manner. Finally, the periodic adjustment of the marketing strategy can be determined by the time parameter of the time period on the occupation of the marketing process.

[0075] Based on the above analysis, in some embodiments of the present invention, based on the sample space and data form, the relative modulus of marketing action nodes in different time periods relative to the hash ring is analyzed to obtain the update mapping position of marketing action nodes in different time periods on the hash ring. Further, it includes:

[0076] First, based on the sample space, the node weights of marketing action nodes are analyzed within different time units. Specifically, relevant marketing action nodes targeting the same webpage tag are obtained; the time units of the relevant marketing action nodes are determined in the sample space; the time unit distance between the marketing action node in the sample space and the time unit corresponding to the nearest relevant marketing action node is calculated (one time unit corresponds to one point in the sample space, so the distance refers to the distance between points in the sample space), which is used as the node weight of the marketing action node within different time units. This reflects "dynamic scalability," indicating whether different user access request flows to different webpage tags occur when facing the same access strategy targeting the same goal.

[0077] Then, obtain the number of users who accessed the targeted web page tags in different time periods after the marketing action was performed. Since the more users using the hash ring, the greater the unevenness of the projection, the problem of uneven distribution of hot requests is corrected by the number of users accessing the page.

[0078] Additionally, the percentage of time spent in different time slots relative to the total duration of all marketing campaign strategies is obtained, thus determining the time slot occupancy rate. The higher the time slot occupancy rate, the lower the probability of receiving new requests during other time slots.

[0079] Based on the node weights, combined with the number of user visits to marketing action nodes within different time periods and the time period occupancy rate within different time periods, the relative modulus of marketing action nodes relative to the hash ring and the number of equal divisions of the hash ring are obtained for different time periods. Specifically, the relative modulus of marketing action nodes relative to the hash ring for different time periods is:

[0080]

[0081] In the formula, Indicates a unit of time period Marketing action nodes The relative modulus relative to the hash ring; Indicates the time period after the marketing action is performed. The number of users accessing the webpage tags targeted by the internal access; Indicates marketing action nodes During the unit time period Node weights within; Indicates a unit of time period The proportion of the total marketing campaign duration; The modulus (i.e., ring size) of the hash ring is a fixed constant. This represents the modulo function, which calculates the remainder when two numbers are divided.

[0082] The number of equal divisions of the hash ring is:

[0083]

[0084] In the formula, Indicates the number of equal divisions of the hash ring; Indicates the time period after the marketing action is performed. The number of users accessing the webpage tags targeted by the internal access; Indicates the marketing action points within a unit of time period Node weights within; Indicates a unit of time period The proportion of the total marketing campaign duration; The modulus (i.e., ring size) of the hash ring is a fixed constant. This represents the floor function.

[0085] Finally, based on the relative modulus and the number of equal segments of the hash ring, the update mapping positions of marketing action nodes on the hash ring for different time periods are obtained. Specifically, using the relative modulus corresponding to different time periods as the step number and the equal segments of the hash ring as the step size, positions on the hash ring are selected along the growth order of the hash ring as the update mapping positions of marketing action nodes on the hash ring for different time periods, such as... Figure 2 As shown. If the hash ring is divided into 8 equal parts, the unit time period... The relative modulus of the marketing action node with respect to the hash ring is 13, so the actual value is 8+5. Therefore, the 5th equally divided position is selected as the unit time period. The update mapping position of the marketing action node on the hash ring.

[0086] S400: Based on the updated mapping location, analyze the impact of marketing action nodes on the popularity of different time periods.

[0087] After the marketing action nodes are mapped on the hash ring, the quality of the marketing action nodes needs to be evaluated. Further, the characteristics of the adjacent access data are extracted based on the marketing situation of the marketing action nodes to assess the retention of the marketing action chain. The contribution of the different types and quantities of user data required by the marketing action nodes to the descriptive completeness of the current marketing action nodes is evaluated. Furthermore, the selection ratio range of marketing action nodes is adjusted based on the diversity of marketing action nodes adjacent to the current marketing action node and the importance of each marketing action node in describing its current state.

[0088] Based on the above analysis, in an embodiment of the present invention, the impact of marketing action nodes on the popularity of different time periods is analyzed based on the updated mapping position. Specifically:

[0089] First, for different time periods, calculate the first hash distance between the updated mapping position and the original mapping position of the marketing action node. The original mapping position is the mapping position mapped on the hash ring using existing methods.

[0090] Meanwhile, for different time periods, the original mapping positions of other marketing action nodes on the update mapping position side corresponding to the marketing action node are determined and denoted as the reference original mapping position. The second hash distance between the original mapping position of the marketing action node and the reference original mapping position is calculated.

[0091] Furthermore, for different time periods, calculate the standard deviation of the distance between the original mapping position of other marketing action nodes and the original mapping position of the marketing action node between the original mapping position and the updated mapping position.

[0092] Finally, combining the first hash distance, the second hash distance, and the standard deviation, the impact of marketing action nodes on the popularity of different time periods is obtained as follows:

[0093]

[0094] In the formula, Indicates marketing action nodes For a unit of time period The impact on popularity; Indicates a unit of time period Marketing action nodes The first hash distance between the updated mapping position and the original mapping position; Indicates a unit of time period Marketing action nodes The original mapping position and the first The first reference to the original mapping position (the first of other marketing action nodes on the side of the updated mapping position corresponding to the marketing action node). The second hash distance between the original mapping locations; Indicates a unit of time period Marketing action nodes The original mapping position of other marketing action nodes between the original mapping position and the updated mapping position. The standard deviation of the distance from the original mapped location; Indicates a unit of time period Marketing action nodes The number of original mapping locations of other marketing action nodes between the original mapping location and the updated mapping location.

[0095] Using the standard deviation as the distance constraint weight, the cube of the proportion of the distance difference is denoted as the distance from the current marketing action node to the unit time period. The higher the impact of the marketing activity, the greater its influence over a given time period. The closer the impact of the actions, the stronger the combined influence of marketing actions and time period.

[0096] S500: Based on the relative modulus and influence intensity, combined with the number of time periods corresponding to the marketing action, the priority score of the marketing action is obtained in different time units.

[0097] Based on the relative modulus and impact intensity, combined with the number of time slots included within the corresponding time period of the marketing action, a priority score for the marketing action within different time slots is obtained. Further, this includes:

[0098] First, based on the relative modulus corresponding to different time units, analyze the consistency deviation between marketing action nodes and target nodes. Specifically, obtain the maximum relative modulus among the relative moduli of marketing action nodes relative to the hash ring for different time units; based on the relative modulus corresponding to different time units and the maximum relative modulus, i.e., calculate the maximum relative modulus and the value of each time unit... The difference in the corresponding relative modulus yields the consistency deviation between the marketing action node and the target node.

[0099] Furthermore, based on the impact of marketing action nodes on different time periods, the focus characteristics brought about by marketing action nodes are analyzed. Specifically, the sum of the impact of all marketing action nodes on the time period is calculated to obtain the focus characteristics brought about by the marketing action nodes.

[0100] Finally, based on consistency bias and focus characteristics, and combined with the number of time units contained within the corresponding time period of the marketing action, the priority score of the marketing action within different time units is obtained as follows:

[0101]

[0102] In the formula, Indicates marketing activities During the unit time period Priority scoring within; This represents the total number of time intervals in the hash ring; Indicates marketing activities The number of time units contained within the corresponding time period; Indicates a unit of time period All marketing action nodes within a unit time period The sum of the impact and popularity indicates the key points of the marketing action. The resulting time-based focus characteristics; Indicates marketing action nodes in different time periods. The largest relative modulus among the relative moduli of the hash ring; Indicates a unit of time period Marketing action nodes The relative modulus relative to the hash ring.

[0103] Represents the most effective marketing placement. This reflects the actual strength difference in marketing effectiveness, measures the consistency deviation between marketing action points and target points, and the smaller the deviation (i.e., (The closer a value is to 0, the higher its priority.) Describe the current marketing activities For a unit of time period The combined effect of marketing efforts reflects the high level of focus during the current marketing campaign.

[0104] Therefore, the remapping result of the current marketing action node, with the marketing action link generated by the time period state representing focus, under the influence of the key description of the connection of the whole link, reflects the state synergy effect of the mapping position of the current marketing action node in the overall current marketing link, which is close to the other states. This achieves the state decomposition of the consistent impact of the current marketing action, and ultimately facilitates the accurate screening of the actual effective marketing action state and its corresponding representative data.

[0105] S600: Based on priority scoring and the number of duplicate results, deduplicatize marketing data.

[0106] Based on priority scores and the number of duplicate results, marketing data is deduplicated. Specifically, first, a backup table is created to prevent data loss due to errors in the deduplication process. Then, the priority scores of all marketing actions up to the current moment are sorted in descending order to obtain a priority score sequence. Next, a first-order difference is performed on the priority score sequence, and the two priority score positions corresponding to the largest difference position are extracted. The data in the original log corresponding to the marketing action defined from the smaller of the two priority score positions to the smallest priority score position is considered insufficient sufficiency data and is not deduplicated for the time being. Finally, for the remaining data, the percentage of data retained for the current marketing action is calculated. More specifically:

[0107] Determine the number of repetitive results within a marketing campaign. The percentage of all results from marketing activities is: ( Indicates the number of repeated results within a marketing campaign. (Indicates the total number of results in a marketing campaign); for a unit of time period Calculate marketing actions Priority rating Average priority score compared to other marketing actions The ratio is ;pass right After weighting, the percentage of data retained is: From the duplicate results, sort by timestamp in ascending order and select the first... The data was cleared.

[0108] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0109] The various embodiments in this specification are described in a progressive manner. For the same or similar parts between the various embodiments, please refer to each other. Each embodiment focuses on describing the differences from the previous embodiment.

Claims

1. A method for deduplicating marketing promotion data based on big data, characterized in that, The method includes: Extract marketing data to obtain a data form, and obtain the marketing action node and the corresponding number of duplicate results for any marketing action; Based on the data form, the marketing status within the corresponding time period of the marketing action is described to obtain the sample space; Based on the sample space and the data form, the relative modulus of the marketing action node relative to the hash ring and the number of equal divisions of the hash ring are analyzed in different time periods to obtain the update mapping position of the marketing action node on the hash ring in different time periods. Based on the updated mapping location, analyze the impact of the marketing action node on the popularity of different time periods; Based on the relative modulus and the influence intensity, and combined with the number of time periods contained within the corresponding time period of the marketing action, the priority score of the marketing action in different time periods is obtained; Based on the priority score and the number of duplicate results, the marketing data is deduplicated. The method for determining the update mapping position of the marketing action node on the hash ring for different time periods includes: Based on the sample space, analyze the node weights of the marketing action nodes in different time units; Based on the node weight, combined with the number of user visits to the marketing action node in different time periods, and the time period occupancy rate in different time periods, the relative modulus of the marketing action node relative to the hash ring and the number of equal divisions of the hash ring in different time periods are obtained. Based on the relative modulus and the number of equal divisions of the hash ring, the update mapping position of the marketing action node on the hash ring in different time periods is obtained; Based on the sample space, the node weights of the marketing action nodes are analyzed in different time units, including: Retrieve the relevant marketing action nodes that are related to the same webpage tag as the aforementioned marketing action node; Within the sample space, the unit time period in which the relevant marketing action node is located is obtained; Calculate the time interval between the unit time interval where the marketing action node is located in the sample space and the unit time interval corresponding to the nearest related marketing action node, and use it as the node weight of the marketing action node in different unit time intervals; Specifically, based on the relative modulus, the update mapping positions of the marketing action nodes on the hash ring for different time periods are obtained, including: Using the relative modulus of different time periods as the number of steps and the equal division of the hash ring as the step size, the position on the hash ring is selected along the growth order of the hash ring as the update mapping position of the marketing action node on the hash ring in different time periods; Specifically, based on the updated mapping position, the impact of the marketing action node on the popularity of different time periods is analyzed, including: For different time periods, calculate the first hash distance between the updated mapping position and the original mapping position of the marketing action node; For different time periods, determine the original mapping position of other marketing action nodes on the update mapping position side corresponding to the marketing action node, and denot it as the reference original mapping position. Calculate the second hash distance between the original mapping position of the marketing action node and the reference original mapping position. For different time periods, calculate the standard deviation of the distance between the original mapping position of other marketing action nodes and the original mapping position of the marketing action node between the original mapping position and the updated mapping position of the marketing action node; By combining the first hash distance, the second hash distance, and the standard deviation, the influence of the marketing action node on different time periods is obtained.

2. The method for deduplicating marketing promotion data based on big data according to claim 1, characterized in that, The data form includes: the time the user clicked, the webpage tab clicked, and the amount spent during a single webpage browsing session.

3. The method for deduplicating marketing promotion data based on big data according to claim 2, characterized in that, Based on the data form, the marketing status within the corresponding time period of the marketing action is described, resulting in a sample space, including: Based on the start and end points of the marketing action, the corresponding time period of the marketing action is obtained; Based on the user click time and the spending amount in the data form, analyze the percentage increase in page views and the percentage increase in conversion within the corresponding time period of the marketing action; Based on the growth rate of page views and the growth rate of conversion, a coordinate system is established to describe the marketing status within the corresponding time period of the marketing action, thus obtaining a sample space.

4. The method for deduplicating marketing promotion data based on big data according to claim 1, characterized in that, Based on the relative modulus and the influence intensity, combined with the number of time periods contained within the corresponding time period of the marketing action, the priority score of the marketing action in different time periods is obtained, including: Based on the relative modulus corresponding to different time units, analyze the consistency deviation between the marketing action nodes and the target nodes; Based on the impact of the marketing action nodes on different time periods, analyze the time period focus characteristics brought about by the marketing action nodes; Based on the consistency deviation and the focus characteristics, combined with the number of time periods corresponding to the marketing action, the priority score of the marketing action in different time periods is obtained.

5. The method for deduplicating marketing promotion data based on big data according to claim 4, characterized in that, Based on the relative modulus corresponding to different time units, analyze the consistency deviation between the marketing action nodes and the target nodes, including: Obtain the maximum relative modulus among the relative moduli of the marketing action nodes relative to the hash ring in different time periods; Based on the relative modulus corresponding to different time periods and the maximum relative modulus, the consistency deviation between the marketing action node and the target node is obtained.

6. The method for deduplicating marketing promotion data based on big data according to claim 4, characterized in that, Based on the impact of the marketing action nodes on different time periods, the focus characteristics brought about by the marketing action nodes during different time periods are analyzed, including: Calculate the sum of the influence heat of all the marketing action nodes on the unit time period to obtain the time period focus characteristics brought about by the marketing action nodes.

Images