Digital creative content audience coverage statistical method and system

Abstract

The application provides a digital creative content audience coverage statistical method and system, which comprises the following steps: matching based on text information entropy distribution sequence and page stay duration sequence to obtain behavior deviation degree; performing cluster distribution feature identification on source tracking parameters associated with access records corresponding to the behavior deviation degree exceeding the preset normal interval to obtain confirmed machine access records; performing record clustering based on terminal features and space-time behavior trajectories to obtain an initial device association group; updating the initial device association group with the device association group corresponding to the confirmed machine access records and the abnormal behavior deviation degree to obtain a target device association group; performing segmented mapping based on the maximum content consumption progress of the target device association group to obtain effective consumption groups; and determining hierarchical audience coverage statistical results based on the number of device association groups corresponding to each content segment interval in each effective consumption group. The application improves the accuracy and reliability of audience hierarchical coverage statistics.

Description

Technical Field

[0001] This invention relates to the field of digital media dissemination analysis technology, and in particular to a method and system for statistical analysis of audience coverage of digital cultural and creative content. Background Technology

[0002] In the digital cultural and creative industries (such as short videos and online games), accurately measuring the true audience reach of content is a key basis for optimizing operational strategies. Current mainstream statistical methods typically rely on server access logs or data from third-party monitoring platforms, counting the number of unique device identifiers or account IDs and combining this with user registration profiles for stratified statistics. To eliminate fake data from bots, existing technologies often employ methods such as IP access frequency restrictions, fixed time interval detection, or simple CAPTCHA blocking.

[0003] However, with the upgrading of black market technologies, advanced automated programs can simulate the random click intervals of real users, forge dynamic IP addresses, and even complete simple interactive actions through human crowdsourcing platforms. Existing rules based on single-point frequency or duration are insufficient to distinguish between "carefully disguised machine traffic" and "real user traffic." Furthermore, existing methods often view the access behavior of individual users in isolation, ignoring the topological structure of digital cultural and creative content dissemination on social networks. Natural propagation exhibits obvious hierarchical diffusion and a long-tail effect, while machine-generated traffic often shows concentrated star-shaped distribution or linear injection characteristics. More importantly, existing methods lack verification of the correlation between the inherent complexity of content and users' micro-behavior. When humans read or watch content with high information density (high complexity), their cognitive load increases, naturally extending their dwell time; while machine scripts cannot perceive the semantic depth of content, and their dwell time is often disconnected from the complexity of the content. Due to the lack of in-depth verification based on the coupling relationship between "content and behavior" and the lack of precise definition of the "spatial uniqueness of physical devices", existing methods include a large number of machine-generated false popularity in the statistics, resulting in a distortion of the audience segmentation coverage and causing content providers to misjudge the distribution of audience interests. Summary of the Invention

[0004] This invention provides a method and system for statistical analysis of audience coverage of digital cultural and creative content, aiming to achieve accurate statistics on the true audience size and structure of digital cultural and creative content under complex dissemination links and high-fidelity data manipulation interference, thereby improving the accuracy and reliability of audience stratification and coverage statistics.

[0005] In a first aspect, the present invention provides a method for statistical analysis of audience coverage of digital cultural and creative content, including: Local correlation matching is performed on the text information entropy distribution sequence of digital cultural and creative content and the page dwell time sequence of each access record to obtain the behavior deviation degree; Based on the source tracking parameters associated with the access records corresponding to the behavior deviation degree that exceeds the preset normal range, cluster distribution feature identification is performed to obtain the confirmed machine access records. Based on the terminal characteristics and spatiotemporal behavior trajectory of each access record, record clustering is performed to obtain an initial device association group. The initial device association group is then updated with the device association group associated with the access records corresponding to the confirmed machine access records and the abnormal behavior deviation degree to obtain the target device association group. Based on the maximum content consumption progress of the target device association group, segmentation mapping is performed to obtain effective consumption groups. Based on the number of device association groups corresponding to each content segment interval in each effective consumption group, the hierarchical audience coverage statistics are determined.

[0006] Secondly, the present invention also provides a digital cultural and creative content audience coverage statistics system for implementing the digital cultural and creative content audience coverage statistics method as described in the first aspect; the digital cultural and creative content audience coverage statistics system includes: The behavior matching module is used to perform local correlation matching based on the text information entropy distribution sequence of digital cultural and creative content and the page dwell time sequence of each access record to obtain the behavior deviation degree; The feature recognition module is used to perform cluster distribution feature recognition based on the source tracking parameters associated with the access records corresponding to the behavior deviation degree that exceeds the preset normal range, so as to obtain the confirmed machine access records. The record clustering module is used to perform record clustering based on the terminal characteristics and spatiotemporal behavior trajectory of each access record to obtain an initial device association group, and update the initial device association group with the device association group associated with the access records corresponding to the confirmed machine access records and the abnormal behavior deviation degree to obtain the target device association group; The audience coverage statistics module is used to perform segmented mapping based on the maximum content consumption progress of the target device association group to obtain effective consumption groups, and to determine the hierarchical audience coverage statistics results based on the number of device association groups corresponding to each content segment interval in each effective consumption group.

[0007] Thirdly, the present invention also provides an electronic device, comprising: a memory for storing a computer program; and a processor for reading and executing the computer program, thereby realizing the digital cultural and creative content audience coverage statistics method as described above.

[0008] Fourthly, the present invention also provides a non-transitory computer-readable storage medium storing a computer program, which, when executed by a processor, implements the digital cultural and creative content audience coverage statistics method described above.

[0009] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the digital cultural and creative content audience coverage statistics method described above.

[0010] The digital cultural and creative content audience coverage statistics method provided in this invention obtains local behavioral deviation by matching the text information entropy distribution sequence of digital cultural and creative content with the user page dwell time sequence. Relying on the natural coupling relationship between content complexity and real human cognitive behavior, it overcomes the shortcomings of unrelated content semantics and the inability to identify highly simulated machine traffic-boosting behavior that is disconnected from content complexity, accurately filtering out abnormal access records. Then, for abnormal behavioral deviation access records outside the normal range, it conducts cluster distribution feature identification by combining source tracking parameters. Relying on the hierarchical diffusion of real social media propagation, long-tail features, and the differentiated topological characteristics of concentrated star-shaped distribution and linear injection features of machine traffic boosting, it confirms machine access records from the propagation link dimension, solving the problem that existing single-point rule detection cannot identify fake machine traffic. Subsequently, it integrates the terminal features and spatiotemporal behavioral trajectories of all access records to complete clustering and obtain initial device association groups. Then, iteratively updates the initial grouping through the aforementioned confirmed machine access records and the device association groups corresponding to abnormal behavioral deviations, overcoming the shortcomings of viewing single user behavior in isolation and failing to accurately associate abnormal device clusters, and eliminating scattered and batch-faked fake device data. Finally, based on the maximum content consumption progress of the updated target device association group, segmented mapping was completed to construct effective consumption groups. Audience segmentation statistics were completed by counting the number of device association groups corresponding to each content segment interval. Throughout the process, multi-dimensional verification was used to filter out high-simulation traffic interference and conform to the real content dissemination pattern. This enabled accurate statistics on the real audience size and structure of digital cultural and creative content under complex dissemination links and high-simulation traffic interference, effectively avoiding the statistical distortion problem caused by false popularity and improving the accuracy and credibility of audience segmentation coverage statistics. Attached Figure Description

[0011] Figure 1 This is a flowchart of the digital cultural and creative content audience coverage statistics method provided in an embodiment of the present invention; Figure 2 This is a structural diagram of the digital cultural and creative content audience coverage statistics system provided in this embodiment of the invention; Figure 3 An embodiment diagram of the electronic device provided in this invention; Figure 4 An embodiment diagram of a computer-readable storage medium provided in accordance with the present invention. Detailed Implementation

[0012] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0013] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified. In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. In the following description, details are set forth for illustrative purposes. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes are not described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0014] Optionally, see Figure 1 , Figure 1 This is a flowchart of the digital cultural and creative content audience coverage statistics method provided by the present invention. In this embodiment of the invention, the executing entity of the digital cultural and creative content audience coverage statistics method is the audience statistics system. Therefore, the digital cultural and creative content audience coverage statistics method includes: Step 10: Perform local correlation matching based on the text information entropy distribution sequence of digital cultural and creative content and the page dwell time sequence of each access record to obtain the behavior deviation degree.

[0015] Optionally, the audience statistics system acquires the text information entropy distribution sequence of digital cultural and creative content and the page dwell time sequence for each access record. The text information entropy distribution sequence refers to the numerical sequence obtained by calculating the information entropy value of the text in each content paragraph unit according to the order of the content paragraph units, and then arranging the paragraph units in chronological order. The information entropy value reflects the text information density of the corresponding content paragraph unit; a higher information entropy value indicates a greater text information density and higher semantic complexity. The page dwell time sequence refers to the numerical sequence obtained by arranging the dwell time of visitors on each content paragraph unit according to the chronological order of the content paragraph units. The audience statistics system performs local correlation matching between the text information entropy distribution sequence and the page dwell time sequence, that is, comparing the correspondence between the information entropy value and dwell time at each content paragraph unit position, thereby quantifying the degree of matching between the visitor's dwell behavior and the content information density in each access record, and obtaining the behavioral deviation degree, as detailed in steps 101 to 104.

[0016] Among them, the behavioral deviation degree is used to characterize the degree of deviation of the dwell behavior pattern of a single access record from the normal behavior pattern of the group. The greater the behavioral deviation degree, the less the correspondence between the dwell behavior of the access record and the content information density is consistent with the normal cognitive consumption pattern of humans.

[0017] Step 20: Based on the source tracking parameters associated with the access records corresponding to the behavior deviation degree that exceeds the preset normal range, cluster distribution feature identification is performed to obtain the confirmed machine access records.

[0018] Optionally, the audience statistics system filters out access records whose behavioral deviation exceeds a preset normal range from the behavioral deviation of each access record. The preset normal range refers to the reasonable fluctuation range of behavioral deviation determined based on historical normal access data. Behavioral deviation within the preset normal range indicates that the correspondence between the dwell behavior and the content information density of the corresponding access record is at a normal level.

[0019] For access records whose behavior deviates beyond a preset normal range, the audience statistics system obtains the source tracking parameters associated with these access records. Source tracking parameters refer to parameters recording the redirection path information taken by the visitor before reaching the current digital cultural and creative content page, including the source page address and the redirection path. Based on the source tracking parameters, the audience statistics system performs cluster distribution feature identification, that is, analyzes whether the redirection relationship between the source page addresses of these access records exhibits a centralized distribution topology, thereby identifying access records with cluster distribution characteristics as verified machine access records, as detailed in steps 201 to 204. Verified machine access records refer to access records confirmed to be generated in batches by an automated program after analysis of the propagation topology.

[0020] Step 30: Cluster the records based on the terminal characteristics and spatiotemporal behavior trajectory of each access record to obtain an initial device association group. Update the initial device association group with the device association group associated with the access records corresponding to the confirmed machine access records and the abnormal behavior deviation to obtain the target device association group.

[0021] Optionally, the audience statistics system acquires the terminal characteristics and spatiotemporal behavioral trajectory of each access record. Terminal characteristics refer to the hardware identification information and software environment information of the terminal device corresponding to the access record, including device model, operating system version, screen resolution, and browser type. Spatiotemporal behavioral trajectory refers to the combination of timestamp information and geographic location information corresponding to the access record, including the specific time the access occurred and the geographic coordinates of the terminal device at the time of access. Based on the terminal characteristics and spatiotemporal behavioral trajectory, the audience statistics system performs record clustering on all access records, grouping access records with the same or similar terminal characteristics and spatiotemporally correlated behavioral trajectories into the same group, obtaining an initial device association group. Each access record in the initial device association group is determined to originate from the same physical device or the same operating entity.

[0022] The audience statistics system uses the verified machine access records and access records corresponding to abnormal behavior deviations obtained in step 20 as clues to update the initial device association group. It excludes or removes abnormal records from the device association groups associated with the verified machine access records and the device association groups associated with the access records corresponding to abnormal behavior deviations, thus obtaining the target device association group, as detailed in steps 301 to 304. Abnormal behavior deviation refers to the degree of behavior deviation exceeding a preset normal range. The target device association group is the genuine audience device association group retained after machine traffic cleaning.

[0023] Step 40: Segment mapping is performed based on the maximum content consumption progress of the target device association group to obtain the effective consumption group, and the hierarchical audience coverage statistics are determined based on the number of device association groups corresponding to each content segment interval in each effective consumption group.

[0024] Optionally, the audience statistics system obtains the maximum content consumption progress for each target device's associated group. The maximum content consumption progress refers to the deepest consumption position reached by the access records in the target device's associated group on the digital cultural and creative content, expressed as the proportion of content consumption progress to the total duration of the digital cultural and creative content.

[0025] The audience statistics system maps the maximum content consumption progress to the content segment intervals pre-divided according to the content narrative structure, and classifies the target device association groups whose maximum content consumption progress falls into the same content segment interval into the same effective consumption group, thereby stratifying the target device association groups according to consumption depth, as in steps 401 to 403.

[0026] The audience statistics system determines the tiered audience coverage statistics for each effective consumer group based on the number of device-associated groups corresponding to each content segment interval within each effective consumer group, combined with the distribution of interaction operation records, as detailed in steps 404 to 407. The tiered audience coverage statistics include the interaction participation level and participation depth results for each effective consumer group, reflecting the coverage scale and participation quality of the audience at different consumption depth levels.

[0027] This invention obtains local behavioral deviation by matching the text information entropy distribution sequence of digital cultural and creative content with the user page dwell time sequence. Relying on the natural coupling relationship between content complexity and real human cognitive behavior, it overcomes the shortcomings of unrelated content semantics and the inability to identify highly simulated bot traffic manipulation behavior that is disconnected from content complexity, thus accurately filtering out abnormal access records. Subsequently, for abnormal behavioral deviation access records outside the normal range, cluster distribution feature identification is performed by combining source tracking parameters. Based on the hierarchical diffusion of real social media propagation, long-tail features, and the differentiated topological characteristics of concentrated star-shaped distribution and linear injection features of bot traffic manipulation, bot access records are confirmed from the propagation link dimension, solving the problem that existing single-point rule detection cannot identify fake bot traffic. Then, the terminal features and spatiotemporal behavioral trajectories of all access records are integrated to complete clustering and obtain initial device association groups. The initial grouping is then iteratively updated using the aforementioned confirmed bot access records and the device association groups corresponding to abnormal behavioral deviations, overcoming the shortcomings of viewing single user behavior in isolation and failing to accurately associate abnormal device clusters, thus eliminating scattered and batch-faked fake device data. Finally, based on the maximum content consumption progress of the updated target device association group, segmented mapping was completed to construct effective consumption groups. Audience segmentation statistics were completed by counting the number of device association groups corresponding to each content segment interval. Throughout the process, multi-dimensional verification was used to filter out high-simulation traffic interference and conform to the real content dissemination pattern. This enabled accurate statistics on the real audience size and structure of digital cultural and creative content under complex dissemination links and high-simulation traffic interference, effectively avoiding the statistical distortion problem caused by false popularity and improving the accuracy and credibility of audience segmentation coverage statistics.

[0028] Optionally, the process of steps 101 to 104 includes: Step 101: Pair the text information entropy distribution sequence with the page dwell time sequence one by one according to the position of the content paragraph unit to obtain paragraph pairing groups.

[0029] Optionally, the audience statistics system obtains the text information entropy distribution sequence of the digital cultural and creative content and the page dwell time sequence for each access record. Each information entropy value in the text information entropy distribution sequence corresponds to a content segment unit in the digital cultural and creative content, and each dwell time segment in the page dwell time sequence corresponds to the same content segment unit. A content segment unit refers to the smallest content unit in which the digital cultural and creative content is divided according to narrative logic, and each content segment unit has a clear start and end position.

[0030] The audience statistics system pairs the first entropy value in the text information entropy distribution sequence with the first dwell time segment in the page dwell time sequence according to their corresponding content paragraph unit positions. It then pairs the second entropy value with the second dwell time segment, and so on, until all entropy values ​​in the text information entropy distribution sequence are paired with all dwell time segments in the page dwell time sequence. Each pairing results constitute a paragraph pairing group, and each paragraph pairing group contains one entropy value and one dwell time segment, both corresponding to the same content paragraph unit.

[0031] Assuming that the digital cultural and creative content contains 10 content paragraph units, then each access record will generate 10 paragraph pairing groups.

[0032] Step 102: Based on the entropy range of the information entropy value in each paragraph pairing group and the time range of the dwell time segment, the paragraph behavior state of each paragraph pairing group is obtained.

[0033] Optionally, the audience statistics system pre-divides the range of information entropy values ​​into several entropy value intervals and the range of dwell time segments into several time intervals. Entropy value intervals are used to distinguish the text information density levels of content paragraph units. For example, assuming the range of information entropy values ​​is divided into three entropy value intervals, corresponding to low information density, medium information density, and high information density, respectively. Time intervals are used to distinguish the dwell time levels of visitors on content paragraph units. For example, assuming the range of dwell time segments is divided into three time intervals, corresponding to short dwell time, medium dwell time, and long dwell time, respectively.

[0034] The audience statistics system determines the entropy range within each paragraph pairing and the time interval within each paragraph pairing, then combines these ranges to obtain the paragraph behavior state of that pairing. This paragraph behavior state describes the type of visitor behavior on content paragraphs with a specific information density. For example, if the information entropy falls within the high information density range and the dwell time falls within the long dwell time interval, the paragraph behavior state of that pairing is "high information density - long dwell time," indicating that the visitor spent a relatively long time on the high information density content paragraph, consistent with normal human cognitive consumption patterns. Conversely, if the information entropy falls within the high information density range and the dwell time falls within the short dwell time interval, the paragraph behavior state is "high information density - short dwell time," indicating that the visitor only spent a short time on the high information density content paragraph, deviating from normal human cognitive consumption patterns.

[0035] Step 103: Based on the frequency of occurrence of various paragraph behavior states corresponding to each paragraph pairing group in each access record, determine the frequency distribution of behavior states for each access record, and based on the frequency of occurrence of various paragraph behavior states in each access record within the same period, obtain the frequency distribution of group behavior states for each access record.

[0036] Optionally, for each access record, the audience statistics system counts the frequency of occurrence of various paragraph behavior states corresponding to all paragraph pairings within that access record. This involves counting the number of times each paragraph behavior state appears in the access record, and then arranging the various paragraph behavior states and their corresponding occurrence counts according to the category of paragraph behavior state to obtain the behavior state frequency distribution for that access record. The behavior state frequency distribution reflects the composition of the visitor's dwell behavior patterns on content paragraph units with different information densities within a single access record. Subsequently, the audience statistics system obtains the frequency of occurrence of various paragraph behavior states in all access records within the same period. The same period refers to all access records within the same statistical time window as the current access record. The audience statistics system then summarizes the occurrence counts of each paragraph behavior state in all access records within the same period to obtain the group behavior state frequency distribution.

[0037] Among them, the frequency distribution of group behavior status reflects the composition of the overall stay behavior pattern of all visitors within the same statistical time window, and represents the normal behavior benchmark in the current period.

[0038] Step 104: Determine the behavioral deviation of each access record based on the degree of difference between the frequency distribution of behavioral states of each access record and the frequency distribution of collective behavioral states.

[0039] Optionally, the audience statistics system compares the frequency distribution of behavioral states for each access record with the frequency distribution of group behavioral states to calculate the degree of difference between the two distributions.

[0040] The calculation method for the degree of distribution difference is as follows: For each paragraph behavior state, the proportion of the occurrence frequency of that paragraph behavior state in the access record relative to the total frequency of all paragraph behavior states in that access record is calculated, and the difference is calculated between this proportion and the proportion of the occurrence frequency of that paragraph behavior state in the group behavior state frequency distribution relative to the total frequency of all paragraph behavior states in the group. This difference yields the proportional deviation value corresponding to that paragraph behavior state. Then, the proportional deviation values ​​corresponding to all paragraph behavior states are summed to obtain the behavior deviation degree of the access record. A behavior deviation degree of 0 indicates that the frequency distribution of the behavior state in the access record is completely consistent with the frequency distribution of the group behavior state. The larger the behavior deviation degree, the greater the difference between the dwell behavior pattern of the access record and the normal behavior pattern of the group. The audience statistics system associates and stores the calculated behavior deviation degree with each access record for use in subsequent steps.

[0041] This invention pairs the text information entropy distribution sequence of digital cultural and creative content with the page dwell time sequence of each access record at the content paragraph unit granularity. The pairing results are then transformed into paragraph behavior states according to a combination of information density level and dwell time level. The difference between the frequency distribution of behavior states of a single access record and the frequency distribution of behavior states of a group within the same period is used as the quantitative basis for behavior deviation. This allows for the identification of abnormal access records where dwell behavior is disconnected from content information density, based on the coupling relationship between content semantic complexity and visitor micro-dwelling behavior. This improves the accuracy and reliability of audience layering coverage statistics.

[0042] Optionally, the processes of steps 201 to 204 include: Step 201: Using each source page address in the source tracing parameters as a graph node, establish directed edges between the corresponding nodes based on the jump direction indicated by the redirection path in the source tracing parameters to obtain an initial propagation graph. Then, take the nodes in the initial propagation graph that do not have incoming edges as propagation source nodes and count the number of downstream nodes directly connected to each propagation source node through directed edges.

[0043] Optionally, the audience statistics system obtains the source tracking parameters associated with the access records whose behavioral deviation exceeds the preset normal range selected in step 10. The source tracking parameters include the source page address and the redirection path. The source page address refers to the network address of the previous page visited by the visitor before arriving at the current digital cultural and creative content page, and the redirection path refers to the redirection direction information from the source page address to the current digital cultural and creative content page.

[0044] The audience statistics system treats each source page address as a graph node and establishes directed edges between the graph node corresponding to the source page address and the graph node corresponding to the current digital cultural and creative content page, based on the jump direction indicated by the guided jump path. The direction of the directed edges is from the graph node corresponding to the source page address to the graph node corresponding to the current digital cultural and creative content page.

[0045] Using the above method, the audience statistics system constructs an initial propagation map from the source tracking parameters of all access records whose behavior deviations exceed the preset normal range.

[0046] The audience statistics system identifies nodes without incoming edges (i.e., nodes with no other directed edges pointing to them) in the initial propagation graph and designates them as propagation source nodes. A propagation source node represents the upstream originating page in the propagation chain, with no further upstream redirects. The audience statistics system then counts the number of downstream nodes directly connected to each propagation source node via directed edges; that is, it counts the number of nodes directly pointed to by directed edges originating from each propagation source node.

[0047] Step 202: Take the propagation source node with the most downstream nodes as the central distribution node, and construct a machine propagation subgraph based on the central distribution node, the downstream nodes directly connected to the central distribution node, and the directed edges between each downstream node.

[0048] Optionally, the audience statistics system selects the propagation source node with the most downstream nodes from the downstream node counts obtained in step 201 as the central distribution node. The central distribution node refers to the source page that directly distributes access traffic to the most downstream pages in the propagation topology, and it has the typical characteristics of centralized distribution.

[0049] The audience statistics system uses a central distribution node as its core. It extracts the central distribution node, all downstream nodes directly connected to the central distribution node through directed edges, and the directed edges between these downstream nodes to construct a machine propagation subgraph. The machine propagation subgraph is a local propagation topology structure with the central distribution node as the root node and the downstream nodes directly connected to the central distribution node as the main body. It is used to characterize the centralized traffic distribution mode originating from the central distribution node.

[0050] Step 203: In the machine propagation subgraph, determine whether each downstream node has a directed edge extending outward. Determine the downstream nodes without a directed edge extending outward as end receiving nodes and count the number of end receiving nodes.

[0051] Optionally, the audience statistics system examines each downstream node in the machine propagation subgraph one by one to determine whether each downstream node has directed edges pointing from that node to other nodes or external nodes in the machine propagation subgraph. Outward-extending directed edges indicate that the downstream node also undertakes the function of forwarding traffic to further downstream pages. For downstream nodes without outward-extending directed edges—that is, nodes that only receive redirected traffic from upstream nodes and do not forward traffic to other nodes—the audience statistics system identifies them as end-receiving nodes. End-receiving nodes represent the final traffic receiver in the propagation chain; they passively receive traffic without generating further propagation behavior. The audience statistics system counts the number of end-receiving nodes in the machine propagation subgraph.

[0052] Step 204: If the proportion of the number of nodes of the end receiving node to the number of downstream nodes in the machine propagation subgraph exceeds a preset proportion threshold, then the access records corresponding to all downstream nodes in the machine propagation subgraph are determined as confirmed machine access records.

[0053] Optionally, the audience statistics system calculates the ratio of the number of end-receiving nodes obtained in step 203 to the total number of all downstream nodes in the machine propagation subgraph to obtain the proportion of end-receiving nodes. The preset proportion threshold is used to determine whether the propagation topology exhibits machine distribution characteristics. It is assumed that the preset proportion threshold is 80%. If the proportion of the number of end-receiving nodes to the number of downstream nodes in the machine propagation subgraph exceeds the preset proportion threshold, it indicates that the propagation topology exhibits typical star-shaped distribution characteristics. That is, traffic is injected unidirectionally from the central distribution node to a large number of end-receiving nodes, and the end-receiving nodes no longer generate further propagation behavior. This is significantly different from the hierarchical diffusion and long-tail effect characteristics in natural propagation and is consistent with the behavior pattern of automated programs distributing access requests in batches.

[0054] In this case, the audience statistics system identifies all access records corresponding to downstream nodes in the machine propagation subgraph as confirmed machine access records. If the proportion of the number of nodes in the end receiving node to the number of downstream nodes in the machine propagation subgraph does not exceed a preset proportion threshold, it indicates that there are many intermediate forwarding nodes in the propagation topology, and the propagation link exhibits hierarchical diffusion characteristics, which does not conform to the typical pattern of centralized machine distribution. Therefore, the audience statistics system does not identify the access records in the machine propagation subgraph as confirmed machine access records.

[0055] This invention constructs a directed propagation graph by using the source tracking parameters of access records whose behavioral deviation exceeds a preset normal range. A machine propagation subgraph is then constructed centered on the node with the most downstream nodes in the propagation source node. The determination is based on whether the proportion of end-receiving nodes exceeds a preset threshold. This allows for the identification of machine traffic clusters with centralized star-shaped distribution characteristics from the perspective of propagation topology. It leverages the structural difference between the hierarchical diffusion and long-tail effect of natural propagation and the centralized unidirectional injection of machine traffic, avoiding misjudgments caused by relying solely on single-point behavioral characteristics. This improves the accuracy and reliability of audience layering coverage statistics.

[0056] Optionally, the processes of steps 301 to 304 include: Step 301: Based on the first device association group to which the access record corresponding to the abnormal behavior deviation belongs, determine the target access record in the first device association group that has the same source page address as the confirmed machine access record, and determine the device association group of the target access record and the second device association group to which each confirmed machine access record belongs as the association group to be excluded.

[0057] Optionally, the audience statistics system performs record clustering on all access records based on the terminal characteristics and spatiotemporal behavior trajectory of each access record to obtain an initial device association group.

[0058] Terminal characteristics include the device model, operating system version, screen resolution, and browser type of the terminal device corresponding to the access record. Spatiotemporal behavior trajectory includes the timestamp information and geographical location information corresponding to the access record. The specific method of record clustering is as follows: access records with the same device model, operating system version, screen resolution, and browser type in the terminal characteristics are initially grouped into the same candidate group. Within the same candidate group, access records with time intervals between timestamp information less than a preset time interval and distances between geographical location information less than a preset distance threshold are further grouped into the same initial device association group, thereby ensuring that access records within the same initial device association group come from the same physical device or the same operating entity.

[0059] The audience statistics system obtains the device association group to which the access record corresponding to the abnormal behavior deviation belongs, and defines it as the first device association group. The access record corresponding to the abnormal behavior deviation is the access record in step 10 whose behavior deviation exceeds the preset normal range.

[0060] The audience statistics system searches for access records in the first device association group that have the same source page address as the verified machine access record obtained in step 20, and defines them as target access records. The fact that target access records and verified machine access records share the same source page address indicates that they have the same traffic source entry point and are likely injected through the same distribution channel.

[0061] The audience statistics system identifies the device association group to which the target access record belongs, along with the device association group to which each verified machine access record belongs (defined as the second device association group), as the association group to be excluded. The association group to be excluded is a collection of device association groups that have a source association or direct attribution relationship with the verified machine access record, and it is necessary to further determine whether they should be excluded from the initial device association group.

[0062] Step 302: Count the total number of access records contained in each device association group in the association groups to be excluded.

[0063] Optionally, for each device association group to be excluded determined in step 301, the audience statistics system counts the total number of access records contained in that device association group to obtain the total number of access records for each device association group to be excluded. The total number of access records reflects the size of the device association group. The more access records a device association group has, the more likely that the machines it contains are generating access records, because automated programs usually generate a large number of access requests in batches on the same device.

[0064] Step 303: Remove device association groups whose total number of access records exceeds the preset record number threshold from the initial device association group. For device association groups whose total number of access records does not exceed the preset record number threshold, remove access records whose behavior deviation exceeds the normal range from the corresponding initial device association group.

[0065] Optionally, the audience statistics system compares the total number of access records for each group to be excluded, obtained in step 302, with a preset record count threshold. The preset record count threshold is a scale criterion used to distinguish between overall exclusion and item-by-item exclusion; assuming the preset record count threshold is 50 records.

[0066] For device association groups whose total number of access records exceeds the preset record number threshold, since the device association group is large in scale and has a source association with the confirmed machine access records, the audience statistics system determines that the device association group as a whole is a machine traffic aggregation group and removes the device association group from the initial device association group. That is, all access records in the device association group will no longer participate in subsequent statistics.

[0067] For device association groups whose total number of access records does not exceed the preset record number threshold, since the size of the device association group is small and may contain real access records, the audience statistics system does not remove the entire device association group. Instead, it removes the access records in the device association group whose behavior deviation exceeds the preset normal range from the corresponding initial device association group one by one, only removing the access records that are judged to be abnormal, and retaining the access records whose behavior deviation is within the preset normal range.

[0068] Step 304: Determine the remaining device association groups after the removal and removal operations as the target device association groups.

[0069] Optionally, after completing the removal and removal operations in step 303, the audience statistics system determines the target device association group as the device association group that was not removed from the initial device association group and the device association group that was retained after removing abnormal access records one by one. The access records in the target device association group are all real audience access records retained after behavioral deviation verification and device association group cleaning. The target device association group represents the real set of audience devices for digital cultural and creative content.

[0070] This invention uses the first device association group to which the access record corresponding to the abnormal behavior deviation belongs as the starting point for investigation. It confirms the machine access record by using the same source page address to determine the association group to be excluded. Based on the comparison result of the total number of access records in the association group to be excluded with the preset record number threshold, two differentiated cleaning strategies are adopted: overall removal and removal of records one by one. This ensures that the cleaning effect can be guaranteed while avoiding damage to real access records. For large-scale machine traffic clusters, overall exclusion is performed to improve cleaning efficiency. For small-scale mixed groups, only abnormal records are removed to retain real access data. Finally, the target device association group after fine cleaning is obtained, thereby improving the accuracy and reliability of audience layering coverage statistics.

[0071] Optionally, the processes of steps 401 to 404 include: Step 401: Divide the total duration of the digital cultural and creative content into content segment intervals according to the content narrative structure. The content narrative structure includes an introductory section, a main body development section, and a concluding section.

[0072] Optionally, the audience statistics system obtains the total duration of digital cultural and creative content and divides the total duration into several content segment intervals according to the content narrative structure. The content narrative structure refers to the narrative logic structure followed by digital cultural and creative content during creation, including the introductory paragraph, the main body development paragraph, and the concluding paragraph.

[0073] The digital cultural and creative content is divided into three parts: the introductory section, which introduces the theme and attracts the audience's attention; the main body, which carries the core information and unfolds the narrative; and the concluding section, which summarizes and concludes the narrative. Based on the creation annotations or content structure analysis of the digital cultural and creative content, the audience statistics system determines the time range of each of the introductory, main body, and concluding sections within the total duration, defining each time period as a content segment interval. For example, if a digital cultural and creative content has a total duration of 10 minutes, with the introductory section corresponding to minutes 0-2, the main body to minutes 2-8, and the concluding section to minutes 8-10, then three content segment intervals are generated.

[0074] Step 402: Compare the maximum content consumption progress of each target device association group with each content segment interval one by one to determine the content segment interval in which the maximum content consumption progress falls, and determine the target device association groups in which the maximum content consumption progress falls into the same content segment interval as the same consumption progress set.

[0075] Optionally, the audience statistics system obtains the maximum content consumption progress for each target device's associated group. Maximum content consumption progress refers to the deepest consumption position reached by all access records within that target device's associated group on the digital cultural and creative content, expressed as a percentage of the total duration of the digital cultural and creative content. For example, assuming that the deepest consumption of access records in a target device's associated group reaches 75% of the total duration of the digital cultural and creative content, then the maximum content consumption progress for that target device's associated group is 75%.

[0076] The audience statistics system compares the maximum content consumption progress of each target device group with each content segment interval obtained in step 401 to determine the content segment interval into which the maximum content consumption progress falls. For example, a maximum content consumption progress of 75% falls within the content segment interval corresponding to the main body expansion segment. Subsequently, the audience statistics system groups all target device groups whose maximum content consumption progress falls within the same content segment interval into the same consumption progress set. That is, the maximum content consumption progress of all target device groups in the consumption progress set is within the same content segment interval, indicating that the consumption depth of these target device groups is at the same level.

[0077] Step 403: Count the number of target device association groups in each consumption progress set, and determine the consumption progress sets with the number of target device association groups exceeding the preset number of groups as valid consumption groups.

[0078] Optionally, the audience statistics system counts the number of target device association groups included in each consumption schedule set in step 402. Since some consumption schedule sets may contain only a very small number of target device association groups, these consumption schedule sets are not statistically significant and cannot reflect the true audience stratification coverage.

[0079] The audience statistics system compares the number of target device association groups with the preset number of groups. The preset number of groups is the minimum number of device association groups used to determine whether the consumption progress set has statistical validity.

[0080] The audience statistics system identifies consumption progress sets where the number of associated groups for the target device exceeds a preset number as valid consumption groups, and excludes consumption progress sets where the number of associated groups for the target device does not exceed the preset number, excluding them from subsequent statistics. Valid consumption groups represent consumption depth levels with sufficient sample size, reliably reflecting the audience coverage scale at that consumption depth level.

[0081] This invention divides the total duration of digital cultural and creative content into content segments with clear semantic meaning according to the content narrative structure, and maps the maximum content consumption progress of each target device's associated group to the corresponding content segment interval. After filtering by a preset number of groups, it is possible to divide the real audience into several statistically valid consumption groups according to the consumption depth. This realizes the transformation from continuous consumption progress to discrete semantic stratification. At the same time, the filtering by the preset number of groups eliminates invalid groups with insufficient sample size, thereby improving the accuracy and reliability of audience stratification coverage statistics.

[0082] Optionally, the processes of steps 404 to 407 include: Step 404: For each valid consumer group, count the number of target device associated groups that have generated at least one interactive operation record in the valid consumer group, and obtain the number of interactive participations in the valid consumer group.

[0083] Optionally, for each valid consumer group, the audience statistics system iterates through all target device-related groups within that group, determining whether each target device-related group has generated at least one interactive operation record. Interactive operation records refer to records of actions initiated by visitors during the consumption of digital cultural and creative content, including liking, commenting, sharing, and saving. For target device-related groups that have generated at least one interactive operation record, the audience statistics system includes them in the interaction participation count. The interaction participation count represents the number of target device-related groups within that valid consumer group that have actually generated interactive behavior, reflecting the scale of active participation by the audience at that consumption depth level.

[0084] Step 405: Compare the number of interactive participations in each valid consumer group with the total number of target device associated groups in the valid consumer group to determine the level of interactive participation in the valid consumer group. Then, classify the interactive operation records of all target device associated groups in each valid consumer group according to the content segment intervals and count the number of target device associated groups that generated interactive operation records in each content segment interval.

[0085] Optionally, the audience statistics system calculates the interaction participation level of each effective consumer group by comparing the number of interactions with the total number of target device-related groups within that effective consumer group. This interaction participation level, expressed as a ratio, reflects the proportion of audience members who generated interactive behavior within the effective consumer group; a higher interaction participation level indicates higher audience engagement at that consumption depth level. Subsequently, the audience statistics system categorizes the interaction operation records of all target device-related groups within the effective consumer group according to the content segment interval in which the interaction occurred. Specifically, it determines the content segment interval in which the interaction occurred based on the timestamp corresponding to each interaction operation record, and then counts the number of target device-related groups that generated interaction operation records within each content segment interval. This statistical result reflects how many target device-related groups generated interactive behavior within each content segment interval.

[0086] Step 406: The content segment interval with the most target device association groups that generate interactive operation records is determined as the audience interaction concentration segment of the effective consumer group, and the content segment interval with the most device association groups corresponding to each content segment interval is determined as the audience interest concentration segment of the effective consumer group.

[0087] Optionally, the audience statistics system selects the content segment with the largest number of interactive operation records generated by target device association groups in each content segment interval obtained in step 405, and determines it as the audience interaction concentration segment of the effective consumer group.

[0088] The audience interaction concentration segment represents the content segment where audience interaction behavior is most concentrated within the effective consumer group, reflecting which content segment audiences are most likely to initiate proactive interactions in. Simultaneously, from the number of device association groups corresponding to each content segment interval determined in step 402, the content segment interval with the largest number of device association groups is selected and defined as the audience interest concentration segment for this effective consumer group. The audience interest concentration segment represents the content segment where the target device association groups are most densely distributed within the effective consumer group, reflecting which content segment has the highest concentration of audience consumption depth, and highlighting the main location where audience interests converge.

[0089] Step 407: If the audience interaction concentration segment and the audience interest concentration segment are within the same content segment, then the deep interest segment is used to determine the engagement depth result of the effective consumer group. If they are in different content segment segments, then the interaction-driven segment is used to determine the engagement depth result of the effective consumer group. The engagement depth result and the level of interaction participation are used to determine the stratified audience coverage statistics of the effective consumer group.

[0090] Optionally, the audience statistics system compares the audience interaction concentration segment with the audience interest concentration segment for each effective consumer group. If the audience interaction concentration segment and the audience interest concentration segment are within the same content segment, it means that the segment with the most concentrated audience interaction behavior overlaps with the segment with the densest audience coverage. This indicates that the audience has both the highest coverage density and the strongest willingness to interact in this content segment, and the audience's interests and interaction behavior are deeply coupled in this segment. The audience statistics system identifies the deep interest segment as the participation depth result for this effective consumer group. The deep interest segment indicates that the audience of this effective consumer group simultaneously exhibits deep participation behavior in the interest-concentrated segment, and the audience has a high degree of initiative and immersion in consuming the content in this segment. If the audience interaction concentration segment and the audience interest concentration segment are not within the same content segment, it means that the segment with the most concentrated audience interaction behavior does not overlap with the segment with the densest audience coverage. This indicates that the audience's interaction behavior is not driven by the depth of content interest, but by other factors (such as the interactive guidance mechanism set in the content). The audience statistics system identifies the interaction-driven segment as the participation depth result for this effective consumer group.

[0091] The "interaction-driven segment" indicates that the audience interaction behavior of this effective consumer group is mainly triggered by external interaction mechanisms rather than driven by the attractiveness of the content itself.

[0092] The audience statistics system combines the engagement depth result with the interaction engagement level obtained in step 405 to determine the stratified audience coverage statistics result for the effective consumer group. The stratified audience coverage statistics result comprehensively reflects the audience coverage scale, interaction engagement ratio, and engagement depth type of each effective consumer group.

[0093] This invention, through statistical analysis of the number of interactions and participation levels in each effective consumer group, calculates the degree of interaction participation. Simultaneously, it categorizes and statistically analyzes interaction records according to content segmentation intervals, identifying concentrated audience interaction segments and concentrated audience interest segments. The overlap between these segments determines the depth of participation. This allows for a refined characterization of each effective consumer group from two dimensions: breadth and type of interaction participation. It distinguishes between interest-driven and interaction-mechanism-driven audience groups, thereby improving the accuracy and reliability of audience segmentation coverage statistics. Furthermore, the segmented audience coverage statistics in this invention not only include coverage quantity information but also participation quality information, enhancing the guiding value of audience segmentation coverage statistics for operational decisions.

[0094] Optionally, steps 50 to 80 include: Step 50: Obtain the content segment interval and the total number of device-associated groups corresponding to each effective consumer group from the tiered audience coverage statistics results, and classify each effective consumer group according to the content segment interval to obtain the consumer group set corresponding to each content segment interval.

[0095] Optionally, the audience statistics system obtains the content segment interval and the total number of device-associated groups corresponding to each effective consumption group from the tiered audience coverage statistics results obtained in step 40. The content segment interval corresponding to each effective consumption group is the content segment interval into which the maximum content consumption progress of the target device-associated groups of that effective consumption group falls in step 402, and the total number of device-associated groups is the number of target device-associated groups contained in that effective consumption group. The audience statistics system categorizes all effective consumption groups according to their respective content segment intervals, and groups effective consumption groups corresponding to the same content segment interval into the same consumption group set, thus obtaining the consumption group set corresponding to each content segment interval. The maximum content consumption progress of all effective consumption groups contained in the consumption group set falls within the same content segment interval.

[0096] Step 60: Based on the total number of device-associated groups for each valid consumer group in the consumer group set of each content segment interval, sum up to obtain the total audience coverage for each content segment interval.

[0097] Optionally, the audience statistics system sums up the total number of device-associated groups for each valid consumer group in the consumer group set corresponding to each content segment interval to obtain the total audience coverage for that content segment interval.

[0098] Total audience reach represents the total number of target device association groups covered by all valid consumer groups within a given content segment, reflecting the true audience reach scale within that content segment. For example, assuming the consumer group set corresponding to the introductory segment contains 3 valid consumer groups with total device association groups of 200, 150, and 80 respectively, then the total audience reach of the introductory segment is 430.

[0099] Step 70: Based on the total audience coverage of each content segment interval, arrange them in the order of the content narrative structure to obtain the audience coverage gradient sequence. Based on the change range of the total audience coverage between two adjacent content segment intervals in the audience coverage gradient sequence, determine the degree of audience loss at the boundary of the content segment interval.

[0100] Optionally, the audience statistics system arranges the total audience coverage of each content segment obtained in step 60 according to the chronological order of the content narrative structure, that is, according to the order of the introductory paragraph, the main body development paragraph, and the concluding paragraph, to obtain the audience coverage gradient sequence. The audience coverage gradient sequence reflects the changing trend of total audience coverage as the depth of content consumption progresses.

[0101] The audience statistics system calculates the change in total audience reach between two adjacent content segment intervals in the audience reach gradient sequence. The change is calculated as follows: subtract the total audience reach of the next content segment interval from the total audience reach of the preceding content segment interval, divide the difference by the total audience reach of the preceding content segment interval, and the change is expressed as a percentage. A larger change indicates a more severe audience churn at the boundary between the preceding and following content segment intervals.

[0102] The audience statistics system determines the degree of audience churn at the boundary of each content segment interval by the magnitude of the change at that boundary.

[0103] Step 80: Using the content segment interval with the greatest audience loss as the content loss breakpoint, optimize the structure of digital cultural and creative content based on the content loss breakpoint and its corresponding audience loss level.

[0104] Optionally, the audience statistics system selects the content segment boundary with the highest audience churn rate from the audience churn rates at the boundaries of each content segment obtained in step 70, and identifies it as the content churn breakpoint. The content churn breakpoint represents the location in digital cultural and creative content where audience churn is most severe, i.e., the location where the most audience members stop consuming the content.

[0105] The audience statistics system outputs information on the location of content churn points and the corresponding degree of audience churn, enabling content creators to optimize the structure of digital cultural and creative content. Structural optimization refers to content creators adjusting the narrative rhythm, information density, or presentation style of digital cultural and creative content at the content churn point based on its location and the magnitude of audience churn, in order to reduce audience churn at that point and increase the overall depth of audience consumption of the digital cultural and creative content.

[0106] This invention categorizes effective consumer groups from layered audience coverage statistics according to content segment intervals and accumulates the total number of device-associated groups to obtain the total audience coverage for each content segment interval. Then, it arranges these segments according to the chronological order of the content narrative structure to form an audience coverage gradient sequence. The degree of audience churn is quantified by the change in the total audience coverage between adjacent content segment intervals. Finally, the boundary with the highest degree of audience churn is used as the content churn breakpoint to output the structural optimization basis. This enables the layered coverage statistics to be transformed into quantitative diagnostic information on the narrative structure of digital cultural and creative content, accurately locating the content position with the most severe audience churn, and realizing a closed-loop application from audience coverage statistics to content quality improvement.

[0107] Furthermore, the digital cultural and creative content audience coverage statistics system described below and the digital cultural and creative content audience coverage statistics method described above can be referred to in correspondence.

[0108] Optionally, refer to Figure 2 , Figure 2 This is a structural diagram of the digital cultural and creative content audience coverage statistics system provided by the present invention. The digital cultural and creative content audience coverage statistics system includes: The behavior matching module 210 is used to perform local correlation matching based on the text information entropy distribution sequence of digital cultural and creative content and the page dwell time sequence of each access record to obtain the behavior deviation degree.

[0109] The feature recognition module 220 is used to perform cluster distribution feature recognition based on the source tracking parameters associated with the access records corresponding to the behavior deviation degree that exceeds the preset normal range, so as to obtain the confirmed machine access records.

[0110] The record clustering module 230 is used to perform record clustering based on the terminal characteristics and spatiotemporal behavior trajectory of each access record to obtain an initial device association group. The device association group associated with the machine access record and the access record corresponding to the deviation degree of abnormal behavior is then updated to obtain the target device association group.

[0111] The audience coverage statistics module 240 is used to perform segmented mapping based on the maximum content consumption progress of the target device association group to obtain the effective consumption group, and to determine the hierarchical audience coverage statistics results based on the number of device association groups corresponding to each content segment interval in each effective consumption group.

[0112] This invention obtains local behavioral deviation by matching the text information entropy distribution sequence of digital cultural and creative content with the user page dwell time sequence. Relying on the natural coupling relationship between content complexity and real human cognitive behavior, it overcomes the shortcomings of unrelated content semantics and the inability to identify highly simulated bot traffic manipulation behavior that is disconnected from content complexity, thus accurately filtering out abnormal access records. Subsequently, for abnormal behavioral deviation access records outside the normal range, cluster distribution feature identification is performed by combining source tracking parameters. Based on the hierarchical diffusion of real social media propagation, long-tail features, and the differentiated topological characteristics of concentrated star-shaped distribution and linear injection features of bot traffic manipulation, bot access records are confirmed from the propagation link dimension, solving the problem that existing single-point rule detection cannot identify fake bot traffic. Then, the terminal features and spatiotemporal behavioral trajectories of all access records are integrated to complete clustering and obtain initial device association groups. The initial grouping is then iteratively updated using the aforementioned confirmed bot access records and the device association groups corresponding to abnormal behavioral deviations, overcoming the shortcomings of viewing single user behavior in isolation and failing to accurately associate abnormal device clusters, thus eliminating scattered and batch-faked fake device data. Finally, based on the maximum content consumption progress of the updated target device association group, segmented mapping was completed to construct effective consumption groups. Audience segmentation statistics were completed by counting the number of device association groups corresponding to each content segment interval. Throughout the process, multi-dimensional verification was used to filter out high-simulation traffic interference and conform to the real content dissemination pattern. This enabled accurate statistics on the real audience size and structure of digital cultural and creative content under complex dissemination links and high-simulation traffic interference, effectively avoiding the statistical distortion problem caused by false popularity and improving the accuracy and credibility of audience segmentation coverage statistics.

[0113] Please see Figure 3 , Figure 3 An embodiment diagram of an electronic device provided in accordance with the present invention. For example... Figure 3 As shown, an embodiment of the present invention provides an electronic device 300, including a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor 320. When the processor 320 executes the computer program 311, it implements the processes of steps 10 to 40.

[0114] Please see Figure 4 , Figure 4 An embodiment diagram of a computer-readable storage medium provided in accordance with an embodiment of the present invention is shown. Figure 4 As shown, this embodiment provides a computer-readable storage medium 400 on which a computer program 311 is stored. When the computer program 311 is executed by a processor, it implements the processes of steps 10 to 40.

[0115] 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, and which, when executed by a processor, implements the processes of steps 10 to 40.

[0116] The system 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.

[0117] 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.

[0118] 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 it. 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. These modifications or substitutions do not cause the essence of the corresponding technical solutions to depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for statistically analyzing audience coverage of digital cultural and creative content, characterized in that, include: Local correlation matching is performed on the text information entropy distribution sequence of digital cultural and creative content and the page dwell time sequence of each access record to obtain the behavioral deviation degree; Based on the source tracking parameters associated with the access records corresponding to the behavior deviation degree that exceeds the preset normal range, cluster distribution feature identification is performed to obtain the confirmed machine access records. Based on the terminal characteristics and spatiotemporal behavior trajectory of each access record, record clustering is performed to obtain an initial device association group. The initial device association group is then updated with the device association group associated with the access records corresponding to the confirmed machine access records and the abnormal behavior deviation degree to obtain the target device association group. Based on the maximum content consumption progress of the target device association group, segmentation mapping is performed to obtain effective consumption groups. Based on the number of device association groups corresponding to each content segment interval in each effective consumption group, the hierarchical audience coverage statistics are determined.

2. The method for statistically analyzing the audience coverage of digital cultural and creative content according to claim 1, characterized in that, The determination of tiered audience coverage statistics based on the number of device-associated groups corresponding to each content segment interval in each effective consumer group includes: For each valid consumer group, count the number of target device associated groups that have generated at least one interactive operation record in the valid consumer group to obtain the number of interactive participations in the valid consumer group; The interaction participation number of each effective consumer group is compared with the total number of target device associated groups in the effective consumer group to determine the interaction participation level of the effective consumer group. The interaction operation records of all target device associated groups in each effective consumer group are classified according to content segment intervals, and the number of target device associated groups that generate interaction operation records in each content segment interval is counted. The content segment interval with the most target device association groups that generate interactive operation records is determined as the audience interaction concentration segment of the effective consumer group, and the content segment interval with the most device association groups corresponding to each content segment interval is determined as the audience interest concentration segment of the effective consumer group. If the audience interaction concentration segment and the audience interest concentration segment are in the same content segment, then the deep interest segment is used to determine the participation depth result of the effective consumer group; if they are in different content segment segments, then the interaction-driven segment is used to determine the participation depth result of the effective consumer group; the participation depth result and the degree of interaction participation are used to determine the stratified audience coverage statistics result of the effective consumer group.

3. The method for statistically analyzing the audience coverage of digital cultural and creative content according to claim 1, characterized in that, The cluster distribution feature identification based on the source tracking parameters associated with the access records corresponding to the behavior deviation exceeding the preset normal range, to obtain the confirmed machine access records, includes: Using each source page address in the source tracing parameters as a graph node, and establishing directed edges between corresponding nodes based on the jump direction indicated by the redirection path in the source tracing parameters, an initial propagation graph is obtained. The nodes in the initial propagation graph that do not have incoming edges are the propagation source nodes. The number of downstream nodes directly connected to each propagation source node through directed edges is counted. The source node with the most downstream nodes is designated as the central distribution node. A machine propagation subgraph is constructed based on the central distribution node, the downstream nodes directly connected to the central distribution node, and the directed edges between each downstream node. In the machine propagation subgraph, it is determined whether each downstream node has a directed edge extending outward. Downstream nodes without a directed edge extending outward are identified as end receiving nodes, and the number of nodes of the end receiving nodes is counted. If the number of nodes of the terminal receiving node accounts for more than the number of downstream nodes in the machine propagation subgraph, then the access records corresponding to all downstream nodes in the machine propagation subgraph are determined as confirmed machine access records.

4. The method for statistically analyzing the audience coverage of digital cultural and creative content according to claim 1, characterized in that, The text information entropy distribution sequence based on digital cultural and creative content and the page dwell time sequence of each access record are locally correlated to obtain the behavioral deviation degree, including: The text information entropy distribution sequence and the page dwell time sequence are paired one by one according to the position of the content paragraph unit to obtain paragraph pairing groups; The paragraph behavior state of each paragraph pairing group is obtained by combining the entropy value range of the information entropy value in each paragraph pairing group and the time range of the dwell time segment. Based on the frequency of occurrence of various paragraph behavior states corresponding to each paragraph pairing group in each access record, the frequency distribution of behavior states of each access record is determined, and based on the frequency of occurrence of various paragraph behavior states in each access record within the same period, the frequency distribution of group behavior states of each access record is obtained. The behavioral deviation of each access record is determined based on the degree of difference between the frequency distribution of behavioral states of each access record and the frequency distribution of collective behavioral states.

5. The method for statistically analyzing the audience coverage of digital cultural and creative content according to claim 1, characterized in that, The process of updating the initial device association group with the device association group associated with the confirmed machine access records and the access records corresponding to the abnormal behavior deviation to obtain the target device association group includes: Based on the first device association group to which the access record corresponding to the abnormal behavior deviation degree belongs, a target access record with the same source page address as the confirmed machine access record is determined in the first device association group. The device association group of the target access record and the second device association group to which each confirmed machine access record belongs are determined as association groups to be excluded. Calculate the total number of access records contained in each device association group in the association groups to be excluded; Device association groups whose total number of access records exceeds a preset record number threshold are removed from the initial device association groups. For device association groups whose total number of access records does not exceed the preset record number threshold, access records whose behavior deviation exceeds the normal range are removed from the corresponding initial device association groups. The remaining device association groups after removal and removal operations are identified as the target device association groups.

6. The method for statistically analyzing the audience coverage of digital cultural and creative content according to claim 1, characterized in that, The segmented mapping based on the maximum content consumption progress of the target device association group yields effective consumption groups, including: The total duration of digital cultural and creative content is divided into content segment intervals according to the content narrative structure; the content narrative structure includes an introductory section, a main body development section, and a concluding section. The maximum content consumption progress of each target device group is compared with each content segment interval one by one to determine the content segment interval into which the maximum content consumption progress falls, and the target device groups into which the maximum content consumption progress falls into the same content segment interval are determined as the same consumption progress set. The number of target device associated groups in each consumption progress set is counted, and consumption progress sets with the number of target device associated groups exceeding the preset number are identified as valid consumption groups.

7. The method for statistically analyzing the audience coverage of digital cultural and creative content according to any one of claims 1 to 6, characterized in that, After determining the tiered audience coverage statistics based on the number of device association groups corresponding to each content segment interval in each effective consumer group, the method further includes: From the hierarchical audience coverage statistics, obtain the content segment intervals and the total number of device-associated groups corresponding to each effective consumer group, and classify each effective consumer group according to the content segment intervals to obtain the consumer group set corresponding to each content segment interval; The total audience coverage for each content segment is obtained by summing up the total number of device-associated groups for each valid consumer group in the consumer group set for each content segment. Based on the total audience coverage of each content segment interval, arranged in the order of the content narrative structure, an audience coverage gradient sequence is obtained. Based on the change in the total audience coverage between two adjacent content segment intervals in the audience coverage gradient sequence, the degree of audience loss at the boundary of the content segment interval is determined. The content segment with the highest audience loss is used as the content loss breakpoint, and the structure of the digital cultural and creative content is optimized based on the content loss breakpoint and the corresponding audience loss level.

8. A digital cultural and creative content audience coverage statistics system, characterized in that, The system is used to implement the digital cultural and creative content audience coverage statistics method as described in any one of claims 1 to 7; the system includes: The behavior matching module is used to perform local correlation matching based on the text information entropy distribution sequence of digital cultural and creative content and the page dwell time sequence of each access record to obtain the behavior deviation degree; The feature recognition module is used to perform cluster distribution feature recognition based on the source tracking parameters associated with the access records corresponding to the behavior deviation degree that exceeds the preset normal range, so as to obtain the confirmed machine access records. The record clustering module is used to perform record clustering based on the terminal characteristics and spatiotemporal behavior trajectory of each access record to obtain an initial device association group, and update the initial device association group with the device association group associated with the access records corresponding to the confirmed machine access records and the abnormal behavior deviation degree to obtain the target device association group; The audience coverage statistics module is used to perform segmented mapping based on the maximum content consumption progress of the target device association group to obtain effective consumption groups, and to determine the hierarchical audience coverage statistics results based on the number of device association groups corresponding to each content segment interval in each effective consumption group.

9. An electronic device, comprising: Memory, used to store computer programs; A processor for reading and executing the computer program, characterized in that, when the processor executes the computer program, it implements the digital cultural and creative content audience coverage statistics method as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the digital cultural and creative content audience coverage statistics method as described in any one of claims 1 to 7.

Images