Advertisement putting method and device, equipment, storage medium and program product
By acquiring and clustering set-top box user behavior metrics, generating household profiles, and combining location information and a GIS database, the problem of low accuracy in set-top box advertising has been solved, enabling precise targeting and cost reduction of local business advertising.
Patent Information
- Application Number
- CN202411490206.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-24
- Publication Date
- 2025-12-16
AI Technical Summary
Existing technologies for advertising delivery based on set-top boxes have low accuracy and cannot meet the needs of precise advertising delivery to local businesses.
By acquiring the behavioral metrics of set-top box users, clustering models are used to generate household profiles. Combined with the location information of set-top boxes and the geographic information system (GIS) database, advertising placement rules for business districts are compiled to deliver shop advertisements to the target set-top box group in real time with precision.
It enables precise targeting of local businesses' advertisements, reduces advertising costs, and enhances businesses' initiative in promoting to their target customer groups and the effectiveness of advertising.
Smart Images
Figure CN121146841A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of communication technology, and in particular to an advertising delivery method, apparatus, device, storage medium, and program product. Background Technology
[0002] TV set-top boxes have evolved from traditional live TV playback devices into intelligent network playback devices, becoming a crucial entry point for the smart home industry and providing greater business development opportunities for smart homes. Currently, in addition to traditional live TV services, smart set-top boxes also support the installation of online games, music, and educational applications. Therefore, advertising can be delivered via set-top boxes.
[0003] However, the accuracy of current set-top box-based advertising is low and cannot meet the needs of local businesses for precise advertising. Summary of the Invention
[0004] This invention provides an advertising delivery method, apparatus, device, storage medium, and program product to address the shortcomings of low accuracy in existing set-top box-based advertising delivery. It enables the analysis of household profiles of set-top boxes based on their location information, the compilation of advertising delivery rules for commercial areas to generate delivery strategies, and the real-time and accurate delivery of shop advertisements to target set-top box groups based on location. This achieves precise delivery of local shop advertisements while reducing advertising delivery costs.
[0005] This invention provides an advertising delivery method, comprising the following steps: Statistical values of behavioral indicators for multiple set-top box users are obtained; the statistical values are represented in the form of bit strings; each set-top box user includes multiple behavioral indicators; The statistical values of each behavioral indicator are input into the clustering model to obtain K clusters for each behavioral indicator output by the clustering model; the clustering model performs clustering based on the bit interval distance of the statistical values. Based on the labels of each cluster and the address labels of each set-top box, a family profile of each set-top box is generated. Advertising is delivered based on the store's location information, the store's advertising information, the household profile of each set-top box, the characteristic groups to which the set-top box users belong, and the set-top box geographic information system (GIS) database.
[0006] According to an advertising delivery method provided by the present invention, the characteristic group to which the set-top box user belongs is determined based on the following method: Convert the statistical values in each of the clusters into label feature values; The multiple label feature values of each behavioral indicator are concatenated to obtain the full feature value of each behavioral indicator. A bit string feature classifier is used to perform a bitwise AND operation on the full set of feature values and the set feature values; Based on the calculation results, the characteristic group to which the set-top box user belongs is determined.
[0007] According to an advertising delivery method provided by the present invention, the clustering model includes a clustering module; The clustering module is used to determine the statistical value with the most bits and the statistical value with the fewest bits in the initial cluster, and to take the difference in the number of bits between the statistical value with the most bits and the statistical value with the fewest bits as the maximum diameter of the initial cluster; to determine the average bit interval of each statistical value in the initial cluster based on the maximum diameter; and to cluster the initial cluster based on the average bit interval to obtain the clustered cluster.
[0008] According to an advertising delivery method provided by the present invention, the set-top box GIS database is determined based on the following method: Receive the location information of the set-top box reported by the installation and maintenance application of the set-top box; Based on the location information of the set-top box, determine the address label of the set-top box; The address tags are encoded based on the latitude and longitude of the center point of the cell to which the set-top box belongs, so as to generate an address tag library for each set-top box; Based on the address tag library, the location information of each set-top box is normalized; The set-top box GIS database is constructed based on the normalized location information.
[0009] According to an advertising placement method provided by the present invention, the latitude and longitude of the center point of the community are determined based on the following method: Determine the set of edge location points for each of the aforementioned cells; Based on the distance between each set of edge location points, a similarity value is determined between every two sets of edge location points; Based on the similarity value, the set of edge location points is divided into multiple discrete point sets; Density clustering is performed on the multiple discrete point sets to obtain multiple subsets; The latitude and longitude of the center point of each cell are determined based on the distance to the core point in each subset.
[0010] According to an advertising delivery method provided by the present invention, the step of delivering advertisements based on the location information of shops, advertising delivery information of shops, household profiles of each set-top box, the characteristic groups to which the set-top box users belong, and a set-top box geographic information system (GIS) database includes: Based on the location information of the shops, the characteristic groups to which the set-top box users belong, and the set-top box GIS database, the target set-top box group to be advertised is determined. Based on the target set-top box group, the household profile of each set-top box, and the advertising information of the shops, an advertising strategy is dynamically generated. Ads are delivered according to the aforementioned ad delivery strategy.
[0011] According to an advertising delivery method provided by the present invention, the step of determining the target set-top box group to which the advertisement is to be delivered, based on the location information of the store, the characteristic group to which the set-top box users belong, and a set-top box GIS database, includes: Based on the characteristic groups to which the set-top box users belong, determine at least one target characteristic group for the advertisement to be delivered; Based on the set-top box GIS database, obtain the location information of the center point of each cell; Based on the location information of the center point of each community and the location information of the shops, determine the first distance between each community and the shops; Determine that the first distance is less than or equal to a set distance, and that there is a first cell within the first distance that belongs to the target feature group of set-top box users, and add the tag code of the first cell to the advertising delivery group; If the first distance is greater than the set distance, then determine the second distance between each edge point in the edge location point set of each community and the shop; Determine a second cell where the second distance is less than or equal to a set distance, and where there is a set-top box user belonging to the target feature group within the second distance, and add the tag code of the second cell to the advertising delivery group; The set-top boxes corresponding to the tag codes of each community in the advertising distribution group are identified as the target set-top box group.
[0012] The present invention also provides an advertising delivery device, comprising the following modules: The acquisition module is used to acquire statistical values of behavioral indicators of multiple set-top box users; the statistical values are represented in the form of bit strings; each set-top box user includes multiple behavioral indicators; The clustering module is used to input the statistical values of each of the behavioral indicators into the clustering model to obtain K clusters for each of the behavioral indicators output by the clustering model; the clustering model performs clustering based on the bit interval distance of the statistical values; The address tag and family profile determination module is used to generate a family profile for each set-top box based on the tags of each cluster and the address tags of each set-top box. The advertising delivery module is used to deliver advertisements based on the location information of the shops, the advertising delivery information of the shops, the household profiles of each set-top box, the characteristic groups to which the set-top box users belong, and the set-top box geographic information system (GIS) database.
[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the advertising delivery method described above.
[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the advertising delivery method as described above.
[0015] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the advertising delivery method as described above.
[0016] The advertising delivery method, apparatus, equipment, storage medium, and program products provided by this invention acquire statistical values of behavioral indicators from multiple set-top box users; input the statistical values of each behavioral indicator into a clustering model to obtain K clusters for each behavioral indicator output by the clustering model; generate a household profile for each set-top box based on the labels of each cluster and the address labels of each set-top box; and deliver advertisements based on the location information of shops, the advertising delivery information of shops, the household profiles of each set-top box, the characteristic groups to which the set-top box users belong, and a set-top box GIS database. This application analyzes the household profiles of set-top boxes based on their location information, compiles advertising delivery rules for business districts to generate delivery strategies, and accurately delivers shop advertisements to the target set-top box group in real time based on location, achieving precise delivery of local shop advertisements while reducing advertising delivery costs. Attached Figure Description
[0017] To more clearly illustrate the technical solutions in this 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 some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.
[0018] Figure 1 This is a flowchart illustrating the advertising delivery method provided by the present invention.
[0019] Figure 2 This is a schematic diagram of the feature values of active users provided by the present invention.
[0020] Figure 3 This is a flowchart illustrating the process of determining the characteristic group to which a set-top box user belongs, provided by the present invention.
[0021] Figure 4 This is a schematic diagram of the location information reporting process of the set-top box provided by the present invention.
[0022] Figure 5 This is a schematic diagram of the process for constructing a five-level address tag library provided by the present invention.
[0023] Figure 6 This is a schematic diagram of the overall architecture for set-top box advertising delivery provided by the present invention.
[0024] Figure 7 This is a schematic diagram of the set-top box advertising delivery process provided by the present invention.
[0025] Figure 8 This is a flowchart illustrating the process of creating a family profile and grouping for a set-top box provided by the present invention.
[0026] Figure 9 This is a flowchart illustrating the advertising delivery method provided by the present invention.
[0027] Figure 10 This is a schematic diagram of the advertising delivery device provided by the present invention.
[0028] Figure 11 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation
[0029] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.
[0030] Currently, various advertising methods and their drawbacks are as follows: 1) The method of storing advertising data in the storage module of the front-end set-top box for delivery. This method increases the manufacturing cost of the set-top box device. Smart set-top boxes typically come with a small amount of cache storage for temporary storage of configuration data, etc. Advertising delivery directly loads data from the set-top box's storage module, which cannot update advertising content in a timely manner and also cannot deliver local area advertisements.
[0031] 2) Targeted advertising delivery based on ad viewership analysis. This method requires typical samples of viewership data before targeted delivery. Only by analyzing typical samples of each type of ad data can the effectiveness of ad viewership characteristics be guaranteed. Without viewership sample analysis, it is difficult to achieve this personalized delivery effect on the first delivery of a new ad, and it is also impossible to deliver ads in local areas.
[0032] 3) Methods that directly store and control ad delivery in the display's intelligent control module. This method directly stores and controls ad delivery in the display's storage. The display's intelligent storage control module must communicate with the backend system and also interface with the set-top box's Electronic Program Guide (EPG). Different set-top box EPGs have different open interface implementation mechanisms, and with numerous display manufacturers and models on the market, comprehensive coverage is difficult to achieve. Furthermore, the high cost of repairing and updating home TV displays makes it impossible to support localized ad delivery.
[0033] 4) Methods for precise ad targeting based on user viewing behavior data analysis. This type of method collects and statistically analyzes data such as users' TV viewing records, interface operations, and ordering behavior to accurately target ads with corresponding content based on user interests and consumption habits. However, users' interests and consumption behaviors change over time and require timely analysis and updates. This method does not consider that set-top box user analysis is actually family behavior analysis, which involves multiple family members and also has regional characteristics.
[0034] To address the aforementioned problems, this invention proposes an advertising delivery method, specifically a smart local business district advertising delivery method based on set-top box location. By constructing a set-top box Geographic Information System (GIS) database and a five-level address tag database, this method enables precise delivery of local advertisements to set-top boxes based on factors such as distance and customer characteristics, thereby achieving real-time intelligent delivery of local business district advertisements. This invention primarily solves the pain points of local business district advertising delivery: 1. Due to the regional characteristics of local business district advertising, which generally serves customers within a specific area, this invention primarily utilizes set-top box location data analysis to address this regional limitation and improve the effectiveness of local business district advertising.
[0035] 2. Based on the location information of set-top boxes, analyze the household profile characteristics of set-top boxes, compile advertising placement rules for business districts to generate placement strategies, and accurately place shop advertisements to the target set-top box group in real time based on location, solving the problem of difficult accurate placement of local shop advertisements, while reducing advertising placement costs.
[0036] 3. By using the location information of shops to search for nearby set-top box users who meet the advertising rules, the traditional passive discovery method of searching for nearby shops centered on users is broken. This solves the problem that it is difficult for shops to actively seek out target customers for promotion, effectively improving the initiative of local shops to promote to target customer groups and improving the advertising effect.
[0037] The following is combined Figures 1-11 This invention describes the advertising delivery method, apparatus, equipment, storage medium, and program product.
[0038] Figure 1 This is a flowchart illustrating the advertising delivery method provided by the present invention, as shown below. Figure 1 As shown, the method includes the following: Step 101: Obtain statistical values of behavioral metrics for multiple set-top box users.
[0039] It should be noted that the implementing entity of this application is the back-end management system, which includes the set-top box management back-end and the big data platform.
[0040] Collect household user behavior data reported by each set-top box and user attribute information uploaded by the business hall. Household user behavior data includes users' viewing history on the set-top box, channel switching frequency, program preferences, viewing time, set-top box power on / off logs, set-top box viewing data, set-top box subscription data, etc.; user attribute information includes basic information of family members (such as age, gender), residential area, consumption habits, etc.
[0041] Then, based on the collected household user behavior data and user attribute information, statistical values of behavioral indicators for each set-top box user are obtained. Each set-top box user includes multiple behavioral indicators, which may include viewing time, subscription preferences, content preferences, activity level, viewing frequency, etc.
[0042] Statistical values are represented in the form of bit strings. In other words, statistical values are stored in the form of binary bits. Each statistical value is encoded as a bit string. Specifically, each statistical value is converted into a binary code consisting of multiple bits for comparison and analysis during the clustering process. The number of bits in the binary code reflects the different characteristics of the statistical value, allowing the similarity between data points to be evaluated by the differences between the bits, and clustering operations to be performed accordingly.
[0043] Step 102: Input the statistical values of each behavioral indicator into the clustering model to obtain K clusters for each behavioral indicator output by the clustering model.
[0044] The clustering model used in this application adopts the DIANA (Divisive Analysis) model, which is a split hierarchical clustering algorithm model. This model clusters data based on the bit interval distance of statistical values. It is understood that the traditional DIANA algorithm terminates clustering by obtaining the number of trees in a cluster, while this application terminates it by obtaining the number of label categories for a certain dimension indicator. Furthermore, since the raw statistical values of all behavioral indicators are stored as bit strings, in the cluster measurement method, the cluster diameter is set to the maximum bit distance between any two statistical values of a certain dimension indicator, and the average cluster distance is set to the average bit distance between any two statistical values. The formula for calculating the average cluster distance is: ; in, Indicators Statistical values , Indicators Statistical values , Represents statistical values The number of bits, Represents statistical values The number of bits, Represents statistical values Quantity, Represents statistical values The quantity.
[0045] Understandably, the cluster diameter reflects the maximum difference between data points within the cluster. When the cluster diameter is too large, it means the cluster is not compact enough, and the algorithm will consider splitting this cluster into smaller clusters to improve the clustering quality. The average cluster distance is used to measure the overall similarity of data points within the cluster. A smaller average distance indicates that the differences between data points within the cluster are smaller, and the consistency within the cluster is higher. If the average cluster distance is large, it means that the heterogeneity within the cluster is higher, and the algorithm needs to adjust the cluster definition or further split the cluster to improve the cluster quality.
[0046] The clustering model includes a clustering module, which is used to determine the statistical value with the most and least bits in the initial cluster. The difference in the number of bits between the statistical value with the most and least bits is used as the maximum diameter of the initial cluster. Based on the maximum diameter, the average bit interval of each statistical value in the initial cluster is determined. Based on the average bit interval, the initial cluster is clustered to obtain the clustered clusters.
[0047] For example, first, iterate through all statistical values of a certain behavioral indicator to find the statistical value with the most bits and the statistical value with the fewest bits. Then, calculate the difference in the number of bits between the statistical value with the most bits and the statistical value with the fewest bits, and use this difference as the maximum diameter of the initial cluster. Further, calculate the average distance of each statistical value in the split cluster according to the above average distance formula. The implementation of the DIANA algorithm provided in this application embodiment is as follows: Input: A hash map of statistical values for all household users corresponding to a certain behavioral indicator, with the termination condition being that the indicator label classifies into K clusters; Output: K cluster categories that meet the termination condition, and a set of hash map clusters for home users; Initial conditions: Use a hash map of all user statistics as an initial cluster, with the user ID as the key and the statistical value of the behavioral metric as the value; For (i=1;i<=total;i++) Do Begin; MaxLength(C j )=len(C j )>len(C i Find the value with the most bits. MinLength(C i )=len(C j ) <len(C i Find the value with the fewest bits. MaxLen = MaxLength(C j )-MinLength(C i Calculate the maximum bit spacing, i.e., the maximum diameter; For ( i=1;i!=k;i++) Do Begin; Since the initial cluster is a single cluster, the cluster with the largest diameter is the initial cluster during the initial split. Then, the user hash map of the statistical value with the largest bit interval of the average difference between the bits of the initial cluster and other statistical values is put into the splinter group, and the rest are put into the old party. After splitting to obtain multiple clusters, select the cluster with the largest diameter from all clusters, and then put the user hash map of the statistical value with the largest bit interval of the average difference between the bits of the statistical value and other statistical values in the cluster into the splinter group, and put the rest into the old party; Repeat, in the old party, find the statistical values in the splinter group whose minimum bit interval distance is not greater than the minimum bit interval of the statistical values in the old party, and add the user hash map corresponding to the statistical value to the splinter group; Until no new old party statistics are assigned to the hash map for the splinter group; The splinter group and the old party are two clusters split from the selected cluster, which together with other clusters form a new cluster set.
[0048] In the above algorithm, the DIANA algorithm, which uses the bit interval distance as the base, ultimately clusters the statistical values of all users corresponding to a certain indicator into K clusters, i.e. K categories.
[0049] Step 103: Generate a family profile for each set-top box based on the labels of each cluster and the address labels of each set-top box.
[0050] A set-top box household profile is a comprehensive and descriptive image of user characteristics generated based on the behavioral data of set-top box users. This profile can help operators, advertisers, and others to better understand and analyze the behavioral patterns, preferences, and needs of household users.
[0051] After the DIANA model outputs K clusters for each behavioral metric, it assigns K labels to each behavioral metric. For example, assuming the behavioral metric is viewing time, the corresponding labels would include "active viewing," "moderately active viewing," "inactive viewing," and "silent viewing." Then, based on the K labels corresponding to each behavioral metric and the address labels of each set-top box, a household profile for each set-top box is generated.
[0052] Understandably, a set-top box's family profile is a multi-dimensional feature set representing the user's behavioral characteristics in various aspects. Specifically, a set-top box's family profile consists of multiple tags, categorized based on users' viewing behavior, interests, and habits to describe the family's characteristics. For example, it might include tags such as "sports enthusiast," "family viewer," and "news follower," which help operators and advertisers better understand user preferences. If a family frequently watches children's programs, educational programs, and family movies, their profile might include the "family-oriented" tag. If another family primarily focuses on financial news and business programs, their profile might be labeled "business / financial focus."
[0053] Step 104: Advertisements are placed based on the store's location information, the store's advertising information, the household profile of each set-top box, the characteristic groups to which the set-top box users belong, and the set-top box geographic information system (GIS) database.
[0054] Merchants register, fill in, and submit their store information via a local business district APK, including basic store information, location information, and licenses. They then select and upload the advertising materials to be displayed, and filter advertising rules, advertising region level, targeting scope, time period, frequency, and target audience characteristics on the APK. The backend management system reviews the merchant's advertising materials and rules. Once approved, the system determines the target set-top box group for advertising based on the store's location information, the characteristic groups of set-top box users, and the set-top box GIS database. Then, based on the target set-top box group, the household profiles of each set-top box, and the store's advertising information (such as advertising rules), it dynamically generates an advertising strategy. Finally, the ads are delivered according to the strategy. The system allows for advertising on EPG ad slots and APK (Android application package) boot animations, and a payment list is generated for timely payment by the merchant.
[0055] The advertising delivery method provided in this application involves acquiring statistical values of behavioral indicators from multiple set-top box users. These statistical values are represented as bit strings. Each set-top box user includes multiple behavioral indicators. The statistical values of each behavioral indicator are input into a clustering model to obtain K clusters for each behavioral indicator. The clustering model clusters based on the bit intervals of the statistical values. A household profile for each set-top box is generated based on the labels of each cluster and the address labels of each set-top box. Advertising is then delivered based on the location information of shops, the advertising delivery information of shops, the household profiles of each set-top box, the characteristic groups to which the set-top box users belong, and the set-top box geographic information system (GIS) database. This application analyzes the household profiles of set-top boxes based on their location information, compiles advertising delivery rules for business districts to generate delivery strategies, and accurately delivers shop advertisements to the target set-top box group in real time based on location, achieving precise delivery of local shop advertisements while reducing advertising costs.
[0056] Based on the above embodiments, the characteristic group to which the set-top box users belong is determined in the following way: Step 110: Convert the statistical values in each cluster into label feature values; Step 111: Concatenate multiple label feature values of each behavioral indicator to obtain the full feature value of each behavioral indicator; Step 112: Perform a bitwise AND operation on the full set of feature values and the set feature values using a bit string feature classifier; Step 113: Based on the calculation results, determine the characteristic group to which the set-top box user belongs.
[0057] After the DIANA model outputs K categories for each behavioral indicator, it assigns K labels to each indicator. These K labels are then arranged into a K-bit string based on degree adverbs. A 1 in a bit indicates that a household user possesses the corresponding feature label. Finally, the statistical values of all households for that category are compressed into feature values. For example, the K labels are arranged from most active to least active, and each label is assigned a bit, representing a specific label. If a household user possesses a label, the corresponding bit is set to 1; otherwise, it is 0. For instance, assuming the labels are arranged as "Active Viewing," "Moderately Active Viewing," "Inactive Viewing," and "Silent Viewing," the resulting bit string would be 4 bits. Figure 2 As shown, the feature value for the "active viewers" metric dimension is 1000, where the first bit is 1, indicating active viewership; the second bit is 1, indicating moderately active viewership; the third bit is 1, indicating inactive viewership; and the fourth bit is 1, indicating silent viewership.
[0058] Finally, the raw statistical values of all dimensions of household users are transformed into bit strings of corresponding label feature values. The feature bit strings of all household users are concatenated to form the full feature values of household users. Then, a Bloom filter (BF) is used to filter the bit model, i.e., the bit string feature classifier of the corresponding label. The full feature values are bitwise ANDed with the feature values of the corresponding indicators of all users. If the corresponding feature classifiers remain unchanged after the bitwise AND operation with the feature values of a specific user, then the user belongs to that user group. Finally, the set of all eligible household users is output.
[0059] Understandably, the BF algorithm uses K hash functions to map data to bits to determine if the target data is not in the target set. This application proposes using a corresponding feature classifier (feature bit string) and the target user's feature bit data to perform bitwise AND operations to determine if the user belongs to the feature group. Bitwise operations effectively improve the computation speed while saving storage space for temporary variables.
[0060] like Figure 3As shown, the characteristic value for active viewers is 1000, the characteristic value for moderately active viewers is 0100, and the characteristic value for inactive viewers is 0010. Concatenating these characteristic values yields 100001000010. Assuming the characteristic value for the active viewers group is 100010001000, performing a bitwise AND operation between 100001000010 and 100010001000 yields 100000000000. Since the result is different from 100010001000, it can be determined that this set-top box user does not belong to the active viewers group.
[0061] This application's embodiments improve data processing speed and reduce storage space by combining the efficient computational features of bit string representation and Bloom filters. By converting user feature values into bit strings and performing bitwise AND operations, it is possible to quickly determine whether a user belongs to a specific group, thereby effectively improving the efficiency of processing large-scale user data.
[0062] Based on the above embodiments, the set-top box GIS database is determined in the following manner: Step 120: Receive the location information of the set-top box reported by the installation and maintenance application of the set-top box; Step 121: Determine the address tag of the set-top box based on its location information; Step 122: Encode the address tag according to the latitude and longitude of the center point of the cell to which the set-top box belongs, so as to generate an address tag library for each set-top box; Step 123: Normalize the location information of each set-top box according to the address tag library; Step 124: Construct the set-top box GIS library based on the normalized location information.
[0063] Receive the location information of the set-top box reported by the set-top box installation and maintenance application. (Reference) Figure 4 Installation and maintenance personnel visit set-top boxes to install and repair them. They scan the set-top box's QR code using the installation and maintenance APK to obtain the STBID, and simultaneously use their mobile phone's location services to obtain detailed location information, such as latitude and longitude. They confirm and report the set-top box's location information. The backend management system receives the data and performs address cleaning. Based on a previously generated provincial administrative address tag library, the address information is standardized and normalized according to five levels of tags: province / city / district / street (town) / community (village). The standardized address information and latitude and longitude data of the set-top box are then saved and updated. After the set-top box powers on and reports its heartbeat, the backend management system sends the address information to the set-top box. The set-top box refreshes and caches the location information in a local file, supporting real-time display of address information on the EPG user information details page.
[0064] Based on the set-top box's location information, its address tag is determined; then, based on the latitude and longitude of the center point of the cell to which the set-top box belongs, the address tag is encoded to generate an address tag library for each set-top box. For example, refer to... Figure 5 Based on existing provincial administrative address labels (province / city / district / county), the historical set-top box installation and maintenance address information in the existing data warehouse is first cleaned and standardized, such as removing redundant punctuation. Then, a map positioning crawler is used based on the set-top box's latitude and longitude to obtain location data to complete the set-top box address. Finally, intelligent analysis generates a five-level address label tree (province / city / district / town / community / village) and a set of approximate names for each label node. The location information is categorized into five levels: province / city / district / town / community / village. All location information is segmented. Street / town names within the same province / city / district / county are normalized to generate standard names and a set of similar names. For community names belonging to the same street / town, prefix clustering is first performed to calculate prefix edit distance and generate a set of similar community names. Based on latitude and longitude values, the Vincent algorithm is used to calculate the actual geographical distance of each community name in the set. Community names with an actual geographical location greater than 3KM and a prefix similarity less than 2 are directly removed. The remaining similar community names are normalized to generate standard names. Iterative processing constructs a five-level tag tree and a set of similar names for location nodes. Simultaneously, the address tag library is continuously expanded by adding new set-top boxes based on the development of existing set-top box services.
[0065] In one embodiment, the latitude and longitude of the cell's center point are determined as follows: a set of edge location points for each cell is determined; a similarity value is determined between any two edge location point sets based on their distances; the edge location point sets are divided into multiple discrete point sets based on their similarity values; density clustering is performed on the multiple discrete point sets to obtain multiple subsets; and the latitude and longitude of the cell's center point are determined based on the reach distance to the core point in each subset. For example, all address tags are uniformly encoded and supplemented with latitude and longitude information to ultimately construct a five-level tag library. Center points are aggregated based on the latitude and longitude sets of all child nodes under each address tag. Then, based on the center point's latitude and longitude, address tags at the same level are sorted by proximity and sequentially encoded numerically. Address tags at the same level have the same parent node prefix, and neighboring addresses have similar tag codes. Finally, the cell code is the numerical code of each parent node level plus the cell's letter code. In this process, the latitude and longitude of the regional center point aggregation are determined based on a point aggregation method that combines shape and density, since the actual shape of the area covered by each address tag varies. The specific implementation process is as follows: The Canny algorithm is used to extract edge features of the set-top box cell distribution, and the gradient strength and direction of each cell location are obtained: ; in, Indicates the longitude value. Indicates latitude value, Indicates gradient strength. Indicates the gradient direction. Indicates longitude direction. Indicates latitude direction.
[0066] Then, a single threshold is applied based on the gradient strength and direction of each cell location. Finally, connectivity analysis is used to fit and connect the edges, determining the set of edge location points. Based on the global features of the set-top box location distribution, the Vincent algorithm is used to calculate the bidirectional Hausdorff distance between two locations, which measures the similarity between the two point sets (i.e., the edge location point sets). The formula for calculating the bidirectional Hausdorff distance is as follows: ; in, Indicates the two-way Hausdorff distance. Let represent the one-way Hausdorff distance from point set A to point set B. Let represent the one-way Hausdorff distance from point set B to point set A.
[0067] All location distribution points are divided into N discrete point sets based on similarity. The total distance of each point = local density × local distance. The local distance is calculated using the OPTICS algorithm. The reachable distance to the core point is: ; in, Indicates distance from the core point The distance between the nearest point and it. This represents the distance from point O to the core point P.
[0068] Remove edge points, filter out the core objects in the dataset, and then calculate the distance to each core object: Input: Data sample D, initialize the reachability distance and core distance of all points to MAX, and the radius is... And the minimum number of points, MinPts.
[0069] Step 1: Establish two queues: an ordered queue (core points and directly reachable points of the core points) and a results queue (to store sample outputs and processing order).
[0070] Step 2: If all the data in D has been processed, the algorithm ends. Otherwise, select an unprocessed point in D that is not a core object, put the core point into the result queue, put the directly density reachable points of the core point into the ordered queue, and sort the directly density reachable points in ascending order of reachability distance.
[0071] Step 3: If the ordered sequence is empty, return to step 2; otherwise, remove the first point from the ordered queue.
[0072] Step 3.1: Determine if the point is a core point. If not, return to Step 3. If it is, store the point in the result queue. If the point is not in the result queue...
[0073] Step 3.2: If the point is a core point, find all its directly density-reachable points and put these points into an ordered queue. Reorder the points in the ordered queue according to their reachability. If the point is already in the ordered queue and the new reachability is small, update the reachability of the point.
[0074] Step 3.3: Repeat step 3 until the sorted queue is empty.
[0075] Finally, based on the reach distance of the core point of each subset, the reach distance of the core of the entire region is calculated, thus obtaining the latitude and longitude of the center point of the entire region.
[0076] Furthermore, based on the constructed five-level tag address database, location information of existing and newly added set-top boxes is cleaned and normalized to generate standard address information and five-level tags, determining the regional groups to which set-top boxes belong. Simultaneously, a set-top box GIS network is constructed based on the latitude and longitude data of the set-top boxes. Existing set-top boxes in the data warehouse are grouped according to the five-level tags, establishing a relationship between address tags and set-top box groups. Ultimately, five-level hierarchical regional set-top box groups are dynamically generated, supporting cross-interaction, merging, and complement operations on different set-top box groups based on tags. New set-top boxes report location data in real time, address data is cleaned and normalized, the address tag database and set-top box GIS database are updated, and the set-top box groupings are dynamically updated.
[0077] This application proposes a point aggregation method based on a combination of shape and density. This method effectively determines the center point latitude and longitude and the set of edge points of the area to which the set-top box belongs. Simultaneously, the Canny algorithm is used to calculate the set of edge points of the area, the OPTICS algorithm is used to calculate local distances, and finally, the reach distance of the entire area's core is calculated based on the reach distance of the local core points, thus obtaining the center point latitude and longitude of the entire area. Furthermore, a five-level address tag library for the set-top box is constructed through similarity calculation and normalization processing, and a set-top box GIS network relation library is intelligently generated using the set-top box latitude and longitude data. Based on this, the accuracy of advertising delivery is improved.
[0078] Based on the above embodiments, determining the target set-top box group for which advertising will be placed, according to the location information of the shop, the characteristic group to which the set-top box users belong, and the set-top box GIS database, includes: Step 130: Determine at least one target characteristic group to which the advertisement is to be delivered, based on the characteristic group to which the set-top box user belongs; Step 131: Obtain the center point location information of each cell based on the set-top box GIS database; Step 132: Determine the first distance between each community and the shop based on the location information of the center point of each community and the location information of the shop; Step 133: Determine that the first distance is less than or equal to a set distance, and that there is a first cell within the first distance that belongs to the target feature group of set-top box users, and add the tag code of the first cell to the advertising delivery group; Step 134: If the first distance is greater than the set distance, then determine the second distance between each edge point in the edge location point set of each community and the shop. Step 135: Determine that the second distance is less than or equal to the set distance, and that there is a second cell within the second distance that belongs to the target feature group of set-top box users, and add the tag code of the second cell to the advertising delivery group; Step 136: Identify the set-top boxes corresponding to the tag codes of each community in the advertising delivery group as the target set-top box group.
[0079] The back-end management system receives the advertising information reported by the merchants, and then determines at least one target characteristic group of the advertisement to be placed based on the characteristic groups of the set-top box users, such as sports enthusiasts, family viewers, etc.
[0080] Specifically, the process involves selecting the advertising delivery distance and target user group. For example, if the target audience is households with elderly people within a 5km radius, the process of generating the target set-top box group from the backend data warehouse is as follows: 1) First, the shop address information is cleaned and standardized to generate five-level address tags and latitude and longitude details. Then, the shop's five-level tag code is matched with the neighboring cell codes under the same four-level tag, and the shop's four-level tag address code is matched with the neighboring street codes to filter out all cell codes of neighboring street codes, forming the largest set of all cells that may meet the distance requirement.
[0081] 2) Based on the previously calculated and updated latitude and longitude data of the community center point and the latitude and longitude of the shop location, if the Vincent algorithm is used to calculate the distance between the shop and the community center point, if the distance (i.e. the first distance) is less than or equal to 5KM, and there are elderly people living within this distance, then the community tag code is added to the advertising group.
[0082] 3) If the distance is greater than 5KM, the distance is calculated by comparing the previously calculated and saved set of edge points of the community with the latitude and longitude of the shop location. If the distance from the edge point to the shop (i.e. the second distance) is less than or equal to 5KM, and there are elderly people in the family within this distance, the community tag code is added to the advertising group.
[0083] 4) Finally, iterate through all eligible cell sets to generate a set of cells. Perform a union operation on each cell set-top box group based on the cell code set to generate the target set-top box group for advertising.
[0084] The embodiments of this application utilize geographic information systems and algorithms to optimize the accuracy and efficiency of advertising delivery, which helps to improve advertising effectiveness and the return on investment for businesses.
[0085] To further explain the advertising delivery method proposed in this invention, please refer to the following embodiments.
[0086] Local business district advertising has regional characteristics. To intelligently improve the effectiveness and accuracy of local shop advertising in proactively targeting customer groups, this invention proposes a method for intelligent local business district advertising based on set-top box location. The overall platform architecture for this implementation is as follows: Figure 6 As shown.
[0087] refer to Figure 7 The overall implementation process of the intelligent local business district advertising delivery method based on set-top box location is as follows: (1) Location information reporting of set-top boxes: Set-top box installation and maintenance personnel use the installation and maintenance APK to locate and report the location information of set-top boxes.
[0088] (2) Real-time update of set-top box GIS database: The background management system cleans and normalizes the location data and puts it into the database, and updates the set-top box GIS database and address tag database in real time.
[0089] (3) Set-top box family profile and grouping: The big data platform analyzes the family user behavior data reported by the set-top box and the user attribute information uploaded by the business hall in sync, generates the corresponding set-top box family profile, and then performs basic grouping of set-top boxes based on the user family profile.
[0090] refer to Figure 8 The back-end management system analyzes the household user behavior data of the set-top boxes based on the set-top box power-on / off logs, viewing data, and subscription behavior data. Then, it combines the set-top box address tags to generate a household profile of the set-top box, supporting dynamic grouping of household profile features of the set-top box.
[0091] (4) Location-based dynamic and precise targeting of store advertisements: Merchants register using the local business district APK and configure store advertisement placement rules, while uploading advertisement image data. The backend management system reviews and dynamically generates advertising placement strategies and placement costs, etc. (Reference) Figure 9 After the merchant confirms payment for the APK, the backend big data platform dynamically generates an advertising strategy for the set-top box based on set-top box location data, store location data, store advertising rules, and set-top box household profile characteristics. Then, it selects set-top boxes that meet the requirements according to the advertising strategy, generates the merchant's target customer group, and dynamically delivers store advertisements to the corresponding EPG ad slots of the target set-top box group according to the advertising strategy. At the same time, it supports selecting the set-top box to actively request the backend system to pull the boot-up advertisement animation when the application APK in the set-top box is turned on, and different advertisement boot-up animations are played when the set-top box user launches different APKs.
[0092] This application embodiment, based on the location of the set-top box and the analysis of the household profile of the set-top box, enables shops to accurately deliver local advertisements to the set-top box based on distance, customer characteristics and other targeting strategies, thereby solving the problem of regional characteristics in local business district advertising and improving the effectiveness of local business district advertising.
[0093] The advertising delivery device provided by the present invention is described below. The advertising delivery device described below and the advertising delivery method described above can be referred to in correspondence.
[0094] refer to Figure 10 The advertising delivery device provided by the present invention includes an acquisition module 1001, a clustering module 1002, an address tag and family profile determination module 1003, and an advertising delivery module 1004.
[0095] The acquisition module 1001 is used to acquire statistical values of behavioral indicators of multiple set-top box users; the statistical values are represented in the form of bit strings; each set-top box user includes multiple behavioral indicators; Clustering module 1002 is used to input the statistical values of each of the behavioral indicators into the clustering model to obtain K clusters of each of the behavioral indicators output by the clustering model; the clustering model performs clustering based on the bit interval distance of the statistical values; The address tag and family profile determination module 1003 is used to generate a family profile for each set-top box based on the tags of each cluster and the address tags of each set-top box. The advertising delivery module 1004 is used to deliver advertisements based on the location information of the store, the advertising delivery information of the store, the household profile of each set-top box, the characteristic group to which the set-top box user belongs, and the set-top box geographic information system (GIS) database.
[0096] The advertising delivery device provided in this application acquires statistical values of behavioral indicators of multiple set-top box users; the statistical values are represented in the form of bit strings; each set-top box user includes multiple behavioral indicators; the statistical values of each behavioral indicator are input into a clustering model to obtain K clusters for each behavioral indicator output by the clustering model; the clustering model clusters based on the bit interval of the statistical values; a household profile of each set-top box is generated based on the labels of each cluster and the address labels of each set-top box; advertising is delivered based on the location information of shops, the advertising delivery information of shops, the household profiles of each set-top box, the characteristic groups to which the set-top box users belong, and the set-top box geographic information system (GIS) database. This application analyzes the household profiles of set-top boxes based on their location information, compiles advertising delivery rules for business districts to generate delivery strategies, and accurately delivers shop advertisements to the target set-top box group in real time based on location, achieving precise delivery of local business advertisements while reducing advertising costs.
[0097] In one embodiment, the advertising delivery module 1004 is further configured to: Convert the statistical values in each of the clusters into label feature values; The multiple label feature values of each behavioral indicator are concatenated to obtain the full feature value of each behavioral indicator. A bit string feature classifier is used to perform a bitwise AND operation on the full set of feature values and the set feature values; Based on the calculation results, the characteristic group to which the set-top box user belongs is determined.
[0098] In one embodiment, clustering module 1002 is also used for; The clustering module is used to determine the statistical value with the most bits and the statistical value with the fewest bits in the initial cluster, and to take the difference in the number of bits between the statistical value with the most bits and the statistical value with the fewest bits as the maximum diameter of the initial cluster; to determine the average bit interval of each statistical value in the initial cluster based on the maximum diameter; and to cluster the initial cluster based on the average bit interval to obtain the clustered cluster.
[0099] In one embodiment, the advertising delivery module 1004 is further configured to: Receive the location information of the set-top box reported by the installation and maintenance application of the set-top box; Based on the location information of the set-top box, determine the address label of the set-top box; The address tags are encoded based on the latitude and longitude of the center point of the cell to which the set-top box belongs, so as to generate an address tag library for each set-top box; Based on the address tag library, the location information of each set-top box is normalized; The set-top box GIS database is constructed based on the normalized location information.
[0100] In one embodiment, the advertising delivery module 1004 is further configured to: Determine the set of edge location points for each of the aforementioned cells; Based on the distance between each set of edge location points, a similarity value is determined between every two sets of edge location points; Based on the similarity value, the set of edge location points is divided into multiple discrete point sets; Density clustering is performed on the multiple discrete point sets to obtain multiple subsets; The latitude and longitude of the center point of each cell are determined based on the distance to the core point in each subset.
[0101] In one embodiment, the advertising delivery module 1004 is specifically used for: Based on the location information of the shops, the characteristic groups to which the set-top box users belong, and the set-top box GIS database, the target set-top box group to be advertised is determined. Based on the target set-top box group, the household profile of each set-top box, and the advertising information of the shops, an advertising strategy is dynamically generated. Ads are delivered according to the aforementioned ad delivery strategy.
[0102] In one embodiment, the advertising delivery module 1004 is specifically used for: Based on the characteristic groups to which the set-top box users belong, determine at least one target characteristic group for the advertisement to be delivered; Based on the set-top box GIS database, obtain the location information of the center point of each cell; Based on the location information of the center point of each community and the location information of the shops, determine the first distance between each community and the shops; Determine that the first distance is less than or equal to a set distance, and that there is a first cell within the first distance that belongs to the target feature group of set-top box users, and add the tag code of the first cell to the advertising delivery group; If the first distance is greater than the set distance, then determine the second distance between each edge point in the edge location point set of each community and the shop; Determine a second cell where the second distance is less than or equal to a set distance, and where there is a set-top box user belonging to the target feature group within the second distance, and add the tag code of the second cell to the advertising delivery group; The set-top boxes corresponding to the tag codes of each community in the advertising distribution group are identified as the target set-top box group.
[0103] Figure 11An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 11 As shown, the electronic device may include: a processor 1110, a communications interface 1120, a memory 1130, and a communications bus 1140, wherein the processor 1110, the communications interface 1120, and the memory 1130 communicate with each other through the communications bus 1140. The processor 1110 can call logical instructions in the memory 1130 to execute an advertising delivery method, which includes: acquiring statistical values of behavioral indicators of multiple set-top box users; the statistical values are represented in the form of bit strings; each set-top box user includes multiple behavioral indicators; inputting the statistical values of each behavioral indicator into a clustering model to obtain K clusters of each behavioral indicator output by the clustering model; the clustering model clusters based on the bit interval distance of the statistical values; generating a family profile of each set-top box based on the labels of each cluster and the address labels of each set-top box; and delivering advertisements based on the location information of the store, the advertising delivery information of the store, the family profiles of each set-top box, the characteristic groups to which the set-top box users belong, and the set-top box geographic information system (GIS) database.
[0104] Furthermore, the logical instructions in the aforementioned memory 1130 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part 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 to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0105] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the advertising delivery method provided by the above methods. The method includes: obtaining statistical values of behavioral indicators of multiple set-top box users; the statistical values are represented in the form of bit strings; each set-top box user includes multiple behavioral indicators; inputting the statistical values of each behavioral indicator into a clustering model to obtain K clusters of each behavioral indicator output by the clustering model; the clustering model clusters based on the bit interval of the statistical values; generating a household profile of each set-top box based on the labels of each cluster and the address labels of each set-top box; and delivering advertisements based on the location information of the shops, the advertising delivery information of the shops, the household profiles of each set-top box, the characteristic groups to which the set-top box users belong, and the set-top box geographic information system (GIS) database.
[0106] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the advertising delivery method provided by the above methods. The method includes: acquiring statistical values of behavioral indicators of multiple set-top box users; the statistical values being represented in the form of bit strings; each set-top box user comprising multiple behavioral indicators; inputting the statistical values of each behavioral indicator into a clustering model to obtain K clusters for each behavioral indicator output by the clustering model; the clustering model performing clustering based on the bit intervals of the statistical values; generating a household profile for each set-top box based on the labels of each cluster and the address labels of each set-top box; and delivering advertisements based on the location information of shops, the advertising delivery information of shops, the household profiles of each set-top box, the characteristic groups to which the set-top box users belong, and a set-top box geographic information system (GIS) database.
[0107] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. 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 the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.
[0108] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.
[0109] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. An advertising placement method, characterized in that, include: Statistical values of behavioral indicators for multiple set-top box users are obtained; the statistical values are represented in the form of bit strings; each set-top box user includes multiple behavioral indicators; The statistical values of each behavioral indicator are input into the clustering model to obtain K clusters for each behavioral indicator output by the clustering model; the clustering model performs clustering based on the bit interval distance of the statistical values. Based on the labels of each cluster and the address labels of each set-top box, a family profile of each set-top box is generated. Advertising is delivered based on the store's location information, the store's advertising information, the household profile of each set-top box, the characteristic groups to which the set-top box users belong, and the set-top box geographic information system (GIS) database.
2. The advertising placement method according to claim 1, characterized in that, The characteristic group to which the set-top box users belong is determined based on the following method: Convert the statistical values in each of the clusters into label feature values; The multiple label feature values of each behavioral indicator are concatenated to obtain the full feature value of each behavioral indicator. A bit string feature classifier is used to perform a bitwise AND operation on the full set of feature values and the set feature values; Based on the calculation results, the characteristic group to which the set-top box user belongs is determined.
3. The advertising placement method according to claim 1, characterized in that, The clustering model includes a clustering module; The clustering module is used to determine the statistical value with the most bits and the statistical value with the fewest bits in the initial cluster, and to take the difference in the number of bits between the statistical value with the most bits and the statistical value with the fewest bits as the maximum diameter of the initial cluster; to determine the average bit interval of each statistical value in the initial cluster based on the maximum diameter; and to cluster the initial cluster based on the average bit interval to obtain the clustered cluster.
4. The advertising placement method according to claim 1, characterized in that, The set-top box GIS database was determined based on the following method: Receive the location information of the set-top box reported by the installation and maintenance application of the set-top box; Based on the location information of the set-top box, determine the address label of the set-top box; The address tags are encoded based on the latitude and longitude of the center point of the cell to which the set-top box belongs, so as to generate an address tag library for each set-top box; Based on the address tag library, the location information of each set-top box is normalized; The set-top box GIS database is constructed based on the normalized location information.
5. The advertising placement method according to claim 4, characterized in that, The latitude and longitude of the center point of the community are determined based on the following method: Determine the set of edge location points for each of the aforementioned cells; Based on the distance between each set of edge location points, a similarity value is determined between every two sets of edge location points; Based on the similarity value, the set of edge location points is divided into multiple discrete point sets; Density clustering is performed on the multiple discrete point sets to obtain multiple subsets; The latitude and longitude of the center point of each cell are determined based on the distance to the core point in each subset.
6. The advertising placement method according to any one of claims 1 to 5, characterized in that, The process of placing advertisements based on the location information of the shops, the advertising information of the shops, the household profiles of each set-top box, the characteristic groups to which the set-top box users belong, and the set-top box geographic information system (GIS) database includes: Based on the location information of the shops, the characteristic groups to which the set-top box users belong, and the set-top box GIS database, the target set-top box group to be advertised is determined. Based on the target set-top box group, the household profile of each set-top box, and the advertising information of the shops, an advertising strategy is dynamically generated. Ads are delivered according to the aforementioned ad delivery strategy.
7. The advertising placement method according to claim 6, characterized in that, The step of determining the target set-top box group for advertising based on the store's location information, the characteristic group to which the set-top box users belong, and the set-top box GIS database includes: Based on the characteristic groups to which the set-top box users belong, determine at least one target characteristic group for the advertisement to be delivered; Based on the set-top box GIS database, obtain the location information of the center point of each cell; Based on the location information of the center point of each community and the location information of the shops, determine the first distance between each community and the shops; Determine that the first distance is less than or equal to a set distance, and that there is a first cell within the first distance that belongs to the target feature group of set-top box users, and add the tag code of the first cell to the advertising delivery group; If the first distance is greater than the set distance, then determine the second distance between each edge point in the edge location point set of each community and the shop; Determine a second cell where the second distance is less than or equal to a set distance, and where there is a set-top box user belonging to the target feature group within the second distance, and add the tag code of the second cell to the advertising delivery group; The set-top boxes corresponding to the tag codes of each community in the advertising distribution group are identified as the target set-top box group.
8. An advertising delivery device, characterized in that, include: The acquisition module is used to acquire statistical values of behavioral indicators of multiple set-top box users; the statistical values are represented in the form of bit strings; each set-top box user includes multiple behavioral indicators; The clustering module is used to input the statistical values of each of the behavioral indicators into the clustering model to obtain K clusters for each of the behavioral indicators output by the clustering model; the clustering model performs clustering based on the bit interval distance of the statistical values; The address tag and family profile determination module is used to generate a family profile for each set-top box based on the tags of each cluster and the address tags of each set-top box. The advertising delivery module is used to deliver advertisements based on the location information of the shops, the advertising delivery information of the shops, the household profiles of each set-top box, the characteristic groups to which the set-top box users belong, and the set-top box geographic information system (GIS) database.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the advertising delivery method as described in any one of claims 1 to 7.
10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the advertising delivery method as described in any one of claims 1 to 7.
11. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the advertising delivery method as described in any one of claims 1 to 7.