Multi-platform advertisement putting effect analysis method and system based on artificial intelligence

Abstract

The invention relates to the technical field of computers, and provides a multi-platform advertisement putting effect analysis method and system based on artificial intelligence, and the method comprises the steps: constructing an advertisement putting structure tree and an advertisement putting knowledge graph; traversing the advertisement putting data along different paths by starting from a root node of the advertisement putting structure tree and combining semantic information of each node mapped by the association mapping relationship in the advertisement putting knowledge graph, and analyzing an association relationship and a data change trend between each level of nodes to obtain hierarchical effect information; based on the relationship between entities in the advertisement putting knowledge graph, performing putting effect reasoning in combination with the hierarchy of each entity mapped by the association mapping relationship in the advertisement putting structure tree to obtain potential effect information; and performing delivery effect decision analysis based on the hierarchical effect information and the potential effect information to obtain delivery optimization information for multi-platform advertisement delivery. According to the embodiment of the invention, the timeliness and effectiveness of advertisement putting strategy adjustment are improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular, to a method and system for analyzing the advertising placement effect based on artificial intelligence across multiple platforms. Background Art

[0002] In the field of advertising placement, an existing method for analyzing the advertising placement effect is the analysis method based on a rule engine. This method pre-sets a series of rules, such as matching and analyzing advertising placement data according to fixed conditions such as the time of advertising placement, the placement platform, and the target audience, to determine whether the advertising placement effect meets the standard. For example, if the number of displays of an advertisement on a certain platform for a specific age group of audiences reaches the set threshold within a specific time period, it is considered that the advertising placement effect in this regard is good.

[0003] However, once the advertising placement effect analysis method based on a rule engine is set, it is relatively fixed and difficult to adapt to the complex and changeable advertising placement environment and diverse advertising placement needs. As the number of advertising placement platforms continues to increase and the audience behavior patterns become increasingly complex, new factors affecting the advertising placement effect continuously emerge. Fixed rules cannot incorporate these new factors in a timely manner, resulting in the inability to accurately and comprehensively analyze the advertising placement effect, making it difficult to discover potential optimization spaces, and reducing the timeliness and effectiveness of advertising placement strategy adjustment. Summary of the Invention

[0004] The present invention provides a method and system for analyzing the advertising placement effect based on artificial intelligence across multiple platforms, so as to improve the timeliness and effectiveness of advertising placement strategy adjustment.

[0005] In a first aspect, the present invention provides a method for analyzing the advertising placement effect based on artificial intelligence, including: Constructing an advertising placement structure tree and an advertising placement knowledge graph; the advertising placement structure tree takes a preset advertising campaign as the root node, different advertising placement platforms as the sub-nodes under the root node, and advertising placement elements of different advertising placement platforms as different hierarchical nodes under its sub-nodes; the advertising placement knowledge graph is constructed with different advertising placement platforms and advertising placement entities of different advertising placement platforms as entities and the relationships between the entities as edges; Starting from the root node of the advertising placement structure tree, combining the semantic information of each node mapped in the advertising placement knowledge graph through an association mapping relationship, traversing the advertising placement data along different paths, analyzing the association relationships and data change trends between each hierarchical node, and obtaining hierarchical effect information of the advertising placement; the association mapping relationship represents the mapping relationship between the entities in the advertising placement knowledge graph and the nodes in the advertising placement structure tree; Based on the relationships between entities in the advertising placement knowledge graph, combined with the levels of each entity mapped in the advertising placement structure tree according to the association mapping relationship, perform placement effect reasoning to obtain potential effect information for advertising placement; Based on the hierarchical effect information and the potential effect information, perform placement effect decision-making analysis to obtain placement optimization information for multi-platform advertising placement.

[0006] In a second aspect, the present invention also provides a multi-platform advertising placement effect analysis system based on artificial intelligence, which is applied to the multi-platform advertising placement effect analysis method based on artificial intelligence as described in the first aspect; the multi-platform advertising placement effect analysis system based on artificial intelligence includes: A construction module, configured to construct an advertising placement structure tree and an advertising placement knowledge graph; the advertising placement structure tree takes a preset advertising campaign as the root node, different advertising placement platforms as sub-nodes under the root node, and advertising placement elements of different advertising placement platforms as different-level nodes of its sub-nodes; the advertising placement knowledge graph takes different advertising placement platforms and advertising placement entities of different advertising placement platforms as entities, and the relationships between entities as edges to construct; An effect trend analysis module, configured to start from the root node of the advertising placement structure tree, combined with the semantic information of each node mapped in the advertising placement knowledge graph according to the association mapping relationship, traverse the advertising placement data along different paths, analyze the association relationships and data change trends between nodes at each level, and obtain hierarchical effect information for advertising placement; the association mapping relationship represents the mapping relationship between entities in the advertising placement knowledge graph and nodes in the advertising placement structure tree; An effect reasoning module, configured to perform placement effect reasoning based on the relationships between entities in the advertising placement knowledge graph, combined with the levels of each entity mapped in the advertising placement structure tree according to the association mapping relationship, to obtain potential effect information for advertising placement; An effect decision-making analysis module, configured to perform placement effect decision-making analysis based on the hierarchical effect information and the potential effect information, to obtain placement optimization information for multi-platform advertising placement.

[0007] In a third aspect, the present invention also provides an electronic device, including: a memory, configured to store a computer software program; a processor, configured to read and execute the computer software program, and further implement the multi-platform advertising placement effect analysis method based on artificial intelligence as described in any one of the above.

[0008] In a fourth aspect, the present invention further provides a non-transitory computer-readable storage medium, wherein a computer software program is stored in the storage medium, and when the computer software program is executed by a processor, the multi-platform advertising delivery effect analysis method based on artificial intelligence as described above is implemented.

[0009] In a fifth aspect, the present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned artificial intelligence-based multi-platform advertising delivery effect analysis methods.

[0010] The artificial intelligence-based multi-platform advertising effect analysis method provided by the embodiment of the present invention can clearly present the hierarchical architecture of multi-platform advertising by constructing a multi-platform advertising structure tree, intuitively display the internal correlation and data change trend of advertising on each platform, and make the analysis no longer limited to simple matching under fixed rules. Generating an advertising knowledge graph and performing semantic reasoning can dig out the potential complex relationships and semantic information between entities in advertising, making up for the defect that fixed rules cannot incorporate new factors. Therefore, the association mapping between the structure tree and the knowledge graph and the subsequent comprehensive analysis can comprehensively and deeply analyze the advertising effect from the two dimensions of hierarchy and semantics, and timely discover new factors and potential optimization space that affect the advertising effect, thereby generating more accurate and effective optimization suggestions, adapting to the complex and changeable advertising environment, and improving the timeliness and effectiveness of the adjustment of advertising strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 It is a flowchart of a multi-platform advertising delivery effect analysis method based on artificial intelligence provided by an embodiment of the present invention; Figure 2 It is a structural diagram of a multi-platform advertising delivery effect analysis system based on artificial intelligence provided by an embodiment of the present invention; Figure 3 An embodiment diagram of an electronic device provided by an embodiment of the present invention; Figure 4 An embodiment diagram of a computer-readable storage medium provided for an embodiment of the present invention. DETAILED DESCRIPTION

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

[0013] In the description of the present invention, the terms "first" and "second" are used only for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the said features. In the description of the present invention, "a plurality of" means two or more unless otherwise specifically defined.

[0014] In the description of the present invention, the term "for example" is used to mean "serving as an example, illustration, or explanation". Any embodiment described as "for example" in the present invention is not necessarily construed as being more preferred or having more advantages than other embodiments. The following description is provided to enable any person skilled in the art to implement and use the present invention. In the following description, details are set forth for purposes of explanation. It should be understood that those of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other instances, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed herein.

[0015] Optionally, refer to Figure 1 , Figure 1 is a schematic flowchart of the method for analyzing the advertising placement effect on multiple platforms based on artificial intelligence provided by the present invention. In the embodiments of the present invention, the execution subject of the method for analyzing the advertising placement effect on multiple platforms based on artificial intelligence is an advertising analysis system. Therefore, the method for analyzing the advertising placement effect on multiple platforms based on artificial intelligence includes: Step 10, constructing an advertising placement structure tree and an advertising placement knowledge graph. The advertising placement structure tree is constructed with a preset advertising campaign as the root node, different advertising placement platforms as the sub-nodes under the root node, and different levels of nodes of the advertising placement elements of different advertising placement platforms as its sub-nodes. The advertising placement knowledge graph is constructed with different advertising placement platforms and the advertising placement entities of different advertising placement platforms as entities and the relationships between the entities as edges.

[0016] Optionally, the construction of the advertising placement structure tree takes a preset advertising campaign as the root node, where the root node represents the core of the entire advertising placement project. Further, the advertising analysis system takes different advertising placement platforms, such as the X-In platform, the X-Wei platform, the X-Du platform, the X-Book platform, etc., as the sub-nodes under the root node. Each advertising placement platform has its own advertising placement elements, such as placement time, placement budget, target audience, etc. These elements are then used as different levels of nodes under the sub-nodes and are constructed layer by layer to form a hierarchical tree-like structure.

[0017] Furthermore, the construction of the advertising placement knowledge graph takes different advertising placement platforms and the advertising placement entities on these platforms (such as advertisers, agents, user groups, etc.) as entities. Among them, there are various relationships between entities. For example, an advertiser places an advertisement on the Xinyin platform. "Advertiser" and "Xinyin platform" are two entities, and "placing an advertisement on the Xinyin platform" is the relationship between them. Using these relationships as edges, each entity is connected to construct a knowledge graph.

[0018] In one embodiment, for example, an online education course advertising placement activity is in progress. First, use "online education course advertising placement activity" as the root node of the advertising placement structure tree. Then, take "Xinyin platform", "Xinxin platform", and "Xindu platform" as the child nodes under the root node. Under the "Xinyin platform" child node, continue to add advertising placement elements such as "placement time", "placement budget", "target audience (aged 18 - 35, people interested in education)" as child nodes; under the "Xinxin platform" child node, add elements such as "placement period", "copy content", "targeted population (with a bachelor's degree or above, people who follow education-related public accounts)"; under the "Xindu platform" child node, add elements such as "keyword setting", "bidding strategy", "search population (people who have searched for education-related keywords)" to complete the construction of the advertising placement structure tree. When constructing the advertising placement knowledge graph, take "Xinyin platform", "Xinxin platform", "Xindu platform", as well as "online education institution (advertiser)", "potential students (user group)", etc. as entities. Establish relationships between entities, such as there is a relationship of "placing an advertisement on the Xinyin platform" between the "online education institution" and the "Xinyin platform"; there is a relationship of "viewing an advertisement on the Xinyin platform" between the "potential students" and the "Xinyin platform"; there is a relationship of "setting keyword advertisements on the Xindu platform" between the "online education institution" and the "Xindu platform", etc., to construct the advertising placement knowledge graph in this way.

[0019] Step 20: Starting from the root node of the advertising placement structure tree, combined with the semantic information of each node mapped in the advertising placement knowledge graph according to the association mapping relationship, traverse the advertising placement data along different paths, analyze the association relationships and data change trends between nodes at each level, and obtain the hierarchical effect information of the advertising placement. The association mapping relationship represents the mapping relationship between the entities in the advertising placement knowledge graph and the nodes in the advertising placement structure tree.

[0020] Furthermore, the advertising analysis system maps the entities in the advertising placement knowledge graph to the nodes in the advertising placement structure tree to obtain the association mapping relationship. Therefore, it can be understood that the association mapping relationship represents the mapping relationship between the entities in the advertising placement knowledge graph and the nodes in the advertising placement structure tree.

[0021] Further, starting from the root node of the advertising placement structure tree, according to the associated mapping relationship, find the corresponding nodes in the advertising placement knowledge graph and obtain the semantic information of these nodes. Then, traverse the advertising placement data along different paths of the structure tree. During the traversal process, analyze the association relationships between nodes at each level, such as the impact of the change in the placement budget of a certain advertising placement platform on the advertising exposure; at the same time, observe the data change trend, such as how the conversion rates of different platforms change over time, to obtain the hierarchical effect information of the advertising placement, specifically as described in steps 201 to 204.

[0022] Step 30: Based on the relationships between entities in the advertising placement knowledge graph, combined with the levels of each entity mapped in the advertising placement structure tree according to the associated mapping relationship, perform placement effect reasoning to obtain the potential effect information of the advertising placement.

[0023] Further, the advertising analysis system, based on the relationships between entities in the advertising placement knowledge graph and combined with the associated mapping relationship, finds the corresponding levels of each entity in the advertising placement structure tree. Through the analysis of these hierarchical relationships and the relationships between entities, perform placement effect reasoning. For example, according to the relationships between advertisers and different platforms in the knowledge graph, as well as the relationships between the user groups of the platforms, reason out the possible effects of increasing advertising placement for a specific user group on a certain platform, so as to obtain the potential effect information of the advertising placement, specifically as described in steps 301 to 304.

[0024] Step 40: Based on the hierarchical effect information and the potential effect information, perform placement effect decision-making analysis to obtain the placement optimization information for multi-platform advertising placement.

[0025] Further, the advertising analysis system synthesizes the hierarchical effect information and the potential effect information to conduct a comprehensive placement effect decision-making analysis. During the analysis process, consider factors such as the actual effect performance, potential improvement space, and placement cost of different platforms, and formulate the placement optimization information for multi-platform advertising placement, including adjusting the placement budget allocation of each platform, optimizing the advertising placement strategy, adjusting the target audience, etc., specifically as described in steps 401 to 404.

[0026] In an embodiment of the present invention, by constructing a multi-platform advertising placement structure tree, the hierarchical structure of multi-platform advertising placement can be clearly presented, the internal relationships and data change trends of advertising placement on each platform can be intuitively displayed, and the analysis is no longer limited to simple matching under fixed rules. Generating an advertising placement knowledge graph and performing semantic reasoning can uncover potential complex relationships and semantic information among entities in advertising placement, making up for the defect that new factors cannot be incorporated by fixed rules. Therefore, the associated mapping between the structure tree and the knowledge graph and subsequent comprehensive analysis can comprehensively and deeply analyze the advertising placement effect from two dimensions of hierarchy and semantics, timely discover new factors affecting the advertising placement effect and potential optimization space, thereby generating more accurate and effective optimization suggestions, adapting to the complex and changeable advertising placement environment, and improving the timeliness and effectiveness of advertising placement strategy adjustment.

[0027] In one embodiment, the descriptions of steps 201 to 204 are as follows: Step 201, starting from the root node of the advertising placement structure tree, locate the first entity associated with the root node in the advertising placement knowledge graph according to the associated mapping relationship, and generate an initial traversal path starting from the root node entity based on the similarity and relevance between the semantic information of the first entity.

[0028] Optionally, the advertising analysis system starts from the root node "online education course advertising placement activity" of the advertising placement structure tree, and locates the first entity associated with it in the advertising placement knowledge graph according to the pre-set associated mapping relationship. The associated mapping relationship defines the corresponding rules between the structure tree nodes and the knowledge graph entities. Therefore, the entity corresponding to the root node can be found through these rules, such as "online education course promotion project". Further, the advertising analysis system analyzes the similarity and relevance between the semantic information of the first entity and other entities in the knowledge graph. For example, through a semantic analysis algorithm (such as the cosine similarity algorithm), entities with similar semantics or business associations to the "online education course promotion project" are identified, such as "online education market promotion" and "education and training advertising placement". Based on these associated relationships, an initial traversal path starting from the root node entity is generated.

[0029] Continuing in the scenario of "online education course advertising placement activity", the advertising analysis system locates the first entity "online education course promotion project" in the advertising placement knowledge graph according to the associated mapping relationship. Through semantic analysis, it is found that the semantics of the "online education market promotion" entity are highly similar to those of the "online education course promotion project", both focusing on online education promotion business; at the same time, the "education and training advertising placement" entity has a business association with the "online education course promotion project", both belonging to the category of advertising placement. Based on these analyses, the advertising analysis system generates an initial traversal path, that is, starting from the "online education course promotion project", connecting "online education market promotion" and "education and training advertising placement" in sequence.

[0030] Step 202: Find the child nodes corresponding to different advertising platforms in the advertising placement structure tree along the initial traversal path, find the second entity corresponding to the child nodes in the advertising placement knowledge graph according to the association mapping relationship, and expand the initial traversal path based on the semantic information of the second entity to obtain the expanded traversal path.

[0031] Further, along the initial traversal path, find the child nodes corresponding to different advertising platforms in the advertising placement structure tree, such as "Xin Yin Platform", "Xin Xin Platform", and "Xin Du Platform". Further, according to the association mapping relationship, the advertising analysis system finds the second entities corresponding to these child nodes in the advertising placement knowledge graph, such as "Xin Yin Platform", "Xin Xin Platform Advertising Platform", and "Xin Du Platform". Further, the advertising analysis system mines the association with other entities in the existing path based on the semantic information of the second entity. For example, analyze the relationship between the "Xin Yin Platform" entity and the existing "Online Education Market Promotion" entity in the initial traversal path, and find that the Xin Yin Platform has an advantage in short video dissemination in online education market promotion. Based on this, add the entity path related to and semantically associated with the "Xin Yin Platform" to the initial traversal path to expand the initial traversal path and obtain a more comprehensive traversal path.

[0032] Continuing on the basis of the initially generated traversal path, the advertising analysis system finds the "Xin Yin Platform" child node in the advertising placement structure tree. According to the association mapping relationship, it locates the second entity of the "Xin Yin Platform" in the knowledge graph. Through the analysis of the semantic information of the "Xin Yin Platform", it is found that there is a close connection between it and "Online Education Market Promotion". The short video promotion form of the Xin Yin Platform is an important channel for online education market promotion. Therefore, the advertising analysis system adds the "Xin Yin Platform" and the related entity paths, such as "Xin Yin Platform Education Short Video Promotion" and "Xin Yin Platform Education Advertising Audience Group", to the initial traversal path to obtain the expanded traversal path.

[0033] Step 203: For the advertising placement element nodes of the advertising platform, find the element nodes related to the current path in the advertising placement structure tree, and fuse the corresponding element semantic information with the entity semantic information corresponding to it in the advertising placement knowledge graph to obtain the fused semantic information vector.

[0034] Further, for the advertising placement element nodes of the advertising platform, the advertising analysis system finds the element nodes related to the current expanded traversal path in the advertising placement structure tree. For example, for the "Xin Yin Platform", find the element nodes such as "Placement Time", "Placement Budget", and "Target Audience".

[0035] Furthermore, the advertisement analysis system fuses the element semantic information corresponding to these element nodes, such as "the placement time on the X Sound platform is from 7 to 9 pm on weekdays", "the placement budget on the X Sound platform is 800 yuan per day", and "the target audience on the X Sound platform is people aged 18 - 35", with the entity semantic information corresponding to them in the advertisement placement knowledge graph. By using a semantic fusion algorithm (such as the word vector splicing algorithm), the element semantic information in the structure tree and the semantic information of the corresponding entities in the knowledge graph are transformed into vectors in a unified format, and combined to form a fused semantic information vector, providing a more comprehensive semantic information basis for subsequent data traversal and analysis.

[0036] Continuing in the scenario of the "X Sound platform", find element nodes such as "placement time", "placement budget", and "target audience" in the advertisement placement structure tree. The semantic information corresponding to "placement time" is "from 7 to 9 pm on weekdays", and the semantic information of the entity "X Sound platform placement time" in the knowledge graph includes data on the active periods of X Sound platform users; the semantic information corresponding to "placement budget" is "800 yuan per day", and the entity "X Sound platform advertisement budget allocation" in the knowledge graph has information on relevant budget usage; the semantic information corresponding to "target audience" is "people aged 18 - 35", and the entity "X Sound platform education advertisement audience portrait" in the knowledge graph has a detailed description of the audience characteristics. In the embodiment of the present invention, the word vector splicing algorithm is used to transform the semantic information from the structure tree and the knowledge graph into vector form and splice them to obtain a fused semantic information vector, which integrates multi-dimensional semantic information about the advertisement placement elements on the X Sound platform.

[0037] Step 204: Based on the extended traversal path and the fused semantic information vector, traverse the advertisement placement data, analyze the association relationships and data change trends between nodes at each level, and obtain the hierarchical effect information of the advertisement placement.

[0038] Furthermore, the advertisement analysis system traverses the advertisement placement data according to the extended traversal path and the fused semantic information vector, analyzes the association relationships and data change trends between nodes at each level, and obtains the hierarchical effect information of the advertisement placement, as specifically described in Steps 2041 to 2044.

[0039] The embodiments of the present invention can comprehensively sort out the advertising data, construct a complete traversal path starting from the root node, integrate multi-dimensional semantic information, accurately analyze the correlation relationships and data change trends among nodes at each level, and obtain clear advertising hierarchical effect information. Therefore, it is possible to deeply understand the specific impacts of different advertising platforms and various advertising elements on the advertising effect, provide strong data support and decision-making basis for subsequent advertising strategy adjustment and resource optimization allocation, make the advertising more in line with market demands and user characteristics, thus adapting to the complex and changeable advertising environment, and improving the timeliness and effectiveness of advertising strategy adjustment.

[0040] In one embodiment, the descriptions of steps 2041 to 2044 are as follows: Step 2041: Based on the total difference degree of the semantic information of each node on each traversal path in the expanded traversal path, determine the semantic information difference weight of each traversal path, and filter each traversal path based on the semantic information difference weight to obtain the target traversal path in the expanded traversal path.

[0041] Optionally, for each path in the expanded traversal path, the advertising analysis system calculates the total difference degree of the semantic information of each node on each traversal path. By comparing the characteristics of the node semantic information (such as keywords, topic categories, etc.), the embodiments of the present invention adopt a suitable difference degree calculation method (such as the Jaccard distance algorithm) to quantify the difference degree. Determine the semantic information difference weight for each traversal path according to the total difference degree. The higher the difference degree of a path, the greater the weight, indicating that this path contains more unique and valuable information.

[0042] Furthermore, the advertising analysis system sets a filtering threshold according to the semantic information difference weight, and filters out the paths with higher weights as the target traversal paths to focus on the key information for subsequent analysis.

[0043] Continuing with the "online education course advertising campaign", the expanded traversal paths obtained by the advertising analysis system include multiple paths. For example, Path A: "Online education course promotion project → Online education market promotion → X - sound platform → X - sound platform educational short - video promotion → X - sound platform delivery time"; Path B: "Online education course promotion project → Educational training advertising → X - letter platform advertising platform → X - letter platform copy content → X - letter platform targeted population". The system calculates the total difference degree of the semantic information of each node on Path A. By comparing the node keywords and topic categories through the Jaccard distance algorithm, it is found that the difference degree of its node semantic information is relatively high, and a relatively high semantic information difference weight is assigned; after calculation, the difference degree of Path B is relatively low, and the weight is also low. The system sets a threshold and filters out the paths with higher weights such as Path A as the target traversal paths, excluding the paths with low weights to reduce redundant information for subsequent analysis.

[0044] Step 2042: Traverse the advertisement delivery data along the target traversal path, determine the relationship edges between the entities corresponding to each hierarchical node in the advertisement delivery knowledge graph, and construct an association relationship matrix between each hierarchical node based on the relationship strength and semantic information of the relationship edges.

[0045] Furthermore, the advertisement analysis system traverses the advertisement delivery data along the target traversal path. During the traversal process, it determines the relationship edges between the entities corresponding to each hierarchical node in the advertisement delivery knowledge graph. For example, for the "X Sound Platform" node and the "X Sound Platform Delivery Time" node, it finds the relationship edge between the "X Sound Platform" entity and the "X Sound Platform Delivery Time" entity in the knowledge graph. Then, it evaluates the relationship strength of the relationship edge, which can be determined based on factors such as the frequency of the relationship occurrence and the tightness of relevant data. At the same time, combining the semantic information of the relationship edge, it quantifies the relationship strength and semantic information and constructs an association relationship matrix between each hierarchical node. Among them, the elements in the matrix represent the association degree and relationship type between different nodes, intuitively presenting the association structure between the nodes.

[0046] Continuing the above embodiment, along the target traversal path "Online Education Course Promotion Project → Online Education Market Promotion → X Sound Platform → X Sound Platform Education Short Video Promotion → X Sound Platform Delivery Time", the advertisement analysis system determines that the relationship edge between the "X Sound Platform" entity and the "X Sound Platform Education Short Video Promotion" entity is "Provide short video promotion channels", and this relationship frequently appears in past data, and the relationship strength is set to a relatively high value; the relationship edge between the "X Sound Platform Education Short Video Promotion" entity and the "X Sound Platform Delivery Time" entity is "Place short video advertisements at a specific time", and the relationship strength is determined to be a medium value according to the actual delivery effect data. Quantify these relationship strengths and semantic information to construct an association relationship matrix. In the matrix, the element corresponding to the "X Sound Platform" and the "X Sound Platform Education Short Video Promotion" represents a strong association relationship and relationship semantics between the two; the element corresponding to the "X Sound Platform Education Short Video Promotion" and the "X Sound Platform Delivery Time" represents a medium association relationship and corresponding semantics, clearly showing the association relationship between each hierarchical node.

[0047] Step 2043: For the advertisement delivery data corresponding to each hierarchical node, determine the data change trend vector between each hierarchical node based on the change in semantic information of the entities corresponding to each hierarchical node in the advertisement delivery knowledge graph at different time points.

[0048] Further, for the advertising placement data corresponding to each level of nodes, the advertising analysis system pays attention to the changes in semantic information of the entities corresponding to each level of nodes in the advertising placement knowledge graph at different time points. In the embodiments of the present invention, by analyzing the evolution of entity semantic information over time (such as the increase or decrease of keywords, the change of themes, etc.), the data change trend between each level of nodes is determined. The data change trend is quantitatively represented and transformed into a data change trend vector. The dimension and element values of the vector reflect the direction and degree of data change, so as to clearly present the dynamic change of data in the time dimension.

[0049] Continuing with the example of the "Douyin platform placement time" node, the semantic information of the corresponding "Douyin platform placement time" entity in the advertising placement knowledge graph changes over time. Initially, the placement time was mainly concentrated on weekdays from 7 to 9 pm; after a period of time, placements were added on weekends from 2 to 4 pm. The advertising analysis system analyzes these semantic information changes and combines them with the corresponding advertising placement data (such as the advertising exposure volume and click-through rate in different time periods) to determine the data change trend. This change trend is quantified into a data change trend vector. For example, one dimension of the vector represents the change trend of exposure volume over time, and the element value is determined according to the increase or decrease amplitude of the exposure volume; another dimension represents the change trend of the click-through rate, and the element value is also set according to the change of the click-through rate, so as to form a complete data change trend vector.

[0050] Step 2044, based on the association relationship matrix and the data change trend vector between each level of nodes, perform fusion to obtain the hierarchical effect information of the advertising placement.

[0051] Further, the advertising analysis system fuses the association relationship matrix and the data change trend vector between each level of nodes. In the embodiments of the present invention, through a specific fusion method (such as corresponding integration of matrix elements and vector elements), the association relationship between nodes and the data change trend are combined to form a comprehensive information set. This information set is sorted and structured, and classified according to levels such as the advertising placement platform and placement elements to obtain the hierarchical effect information of the advertising placement, clearly showing the advertising placement effect in different levels and different dimensions.

[0052] Continuing with the above embodiment, the association relationship matrix related to the "X Sound Platform" and the data change trend vector are fused. The association relationship elements of nodes such as the "X Sound Platform" and the "X Sound Platform Education Short Video Promotion" and the "X Sound Platform Placement Time" in the association relationship matrix are integrated with the corresponding data change elements such as the exposure volume and click-through rate in the data change trend vector. For example, the strong association relationship between the "X Sound Platform" and the "X Sound Platform Education Short Video Promotion" is combined with the growth trend of the exposure volume brought by the short video promotion to form a comprehensive information unit. All relevant information units are sorted out and classified according to the levels of the "X Sound Platform" and the placement elements (such as placement time, promotion form, etc.) to obtain the hierarchical effect information of the "X Sound Platform" advertising placement. The same method is used to process other platforms, and finally the hierarchical effect information of the entire advertising placement activity is summarized to comprehensively present the impact of each platform and each element on the advertising placement effect.

[0053] The embodiment of the present invention can deeply explore the association relationship between each hierarchical node, accurately grasp the data change trend, effectively fuse the association relationship and the data change trend, and form the hierarchical effect information of the advertising placement with clear levels. Therefore, it can comprehensively cover the effect situations of different advertising placement platforms and each placement element in different time dimensions, provide a basis for the optimization and adjustment of the advertising placement strategy, so as to adapt to the complex and changeable advertising placement environment, and improve the timeliness and effectiveness of the adjustment of the advertising placement strategy.

[0054] In one embodiment, the descriptions of steps 301 to 304 are as follows: Step 301, determine the entity level of each entity in the advertising placement knowledge graph in the entity hierarchy of the advertising placement structure tree based on the association mapping relationship.

[0055] Optionally, the advertising analysis system analyzes each entity in the advertising placement knowledge graph according to the association mapping relationship. Among them, the association mapping relationship clarifies the corresponding rules between the knowledge graph entities and the nodes of the advertising placement structure tree. Therefore, through these rules, the corresponding position of each entity in the advertising placement structure tree is found, so as to determine its entity level. For example, if the "X Sound Platform" entity in the knowledge graph corresponds to the "X Sound Platform" sub-node under the root node of the "Online Education Course Advertising Placement Activity" in the structure tree, then the level of the "X Sound Platform" entity is the first-level sub-node level; if the "X Sound Platform Placement Budget" entity in the knowledge graph corresponds to the sub-node under the "X Sound Platform" sub-node, its level is the second-level sub-node level.

[0056] Continuing in the scenario of "Online Education Course Advertising Campaign", the advertising analysis system analyzes the entities in the advertising placement knowledge graph. The entity of "Online Education Institution" corresponds to the subject represented by the root node "Online Education Course Advertising Campaign" in the advertising placement structure tree through an associated mapping relationship, and its level is the root node level; the entity of "Xin Yin Platform" corresponds to the "Xin Yin Platform" sub-node under the root node in the structure tree, and the level is determined as the first-level sub-node level; the entity of "Xin Yin Platform Placement Time" under the "Xin Yin Platform" corresponds to the sub-node under the "Xin Yin Platform" sub-node in the structure tree, and the level is the second-level sub-node level. In this way, the system determines the entity level of each entity in the advertising placement structure tree for the knowledge graph.

[0057] Step 302: Based on the relationships between entities in the advertising placement knowledge graph and the entity level of each entity, construct a relationship level matrix. The matrix elements in the relationship level matrix represent the associated characteristics of the relationships between entities corresponding to the entity levels in the advertising placement structure tree.

[0058] Furthermore, the advertising analysis system constructs a relationship level matrix based on the relationships between entities in the advertising placement knowledge graph and the entity levels that have been determined for each entity. Among them, for any two entities with a relationship in the knowledge graph, analyze the associated characteristics of their corresponding entity levels in the advertising placement structure tree. The associated characteristics include whether the levels are the same, the size of the level difference, the logical relationship between levels, etc. Quantify these associated characteristics and use them as the matrix elements in the relationship level matrix. For example, if the levels of two entities in the structure tree are the same, the matrix element can be assigned a value of 1; if the levels differ by one level, the matrix element is assigned a value of -1; if there is an inclusion relationship between levels, assign corresponding values according to the specific situation, thereby constructing a complete relationship level matrix to intuitively display the association between the relationships between entities and the levels of the structure tree.

[0059] Continuing in the advertising placement knowledge graph, there is a relationship of "the platform includes the placement time setting" between the entity of "Xin Yin Platform" and the entity of "Xin Yin Platform Placement Time". The entity level of the "Xin Yin Platform" is the first-level sub-node level, and the entity level of the "Xin Yin Platform Placement Time" is the second-level sub-node level, with a level difference of one level. The advertising analysis system assigns a value of -1 to the element corresponding to the relationship between these two entities in the relationship level matrix. Another example is that there is a relationship between the entity of "Xin Xin Platform Advertising Platform" and the entity of "Xin Xin Platform Copy Content", and their level relationship in the structure tree is similar, and the corresponding element in the matrix is also assigned a value of -1. For "Online Education Institution" and "Xin Yin Platform", the "Online Education Institution" is at the root node level and the "Xin Yin Platform" is at the first-level sub-node level, with a relationship between the subject and the placement platform. According to the level and relationship characteristics, corresponding values are assigned in the matrix. Through the analysis of all entity relationships in the knowledge graph, a complete relationship level matrix is constructed.

[0060] Step 303: In the advertisement placement structure tree, traverse starting from the root node. If the target node traversed in the advertisement placement structure tree is the node corresponding to the mapped entity in the advertisement placement knowledge graph, construct a hierarchical path based on the target node.

[0061] Further, in the advertisement placement structure tree, the advertisement analysis system starts traversing from the root node "Online Education Course Advertisement Placement Activity". During the traversal process, when the encountered target node is the node corresponding to the mapped entity in the advertisement placement knowledge graph, construct a hierarchical path based on this target node. The hierarchical path starts from the root node, records each node passed through until the target node, clearly presenting the hierarchical position and subordination relationship of this node in the structure tree. Continuing with the above embodiment, starting from the root node "Online Education Course Advertisement Placement Activity" of the advertisement placement structure tree, when traversing to the "Xin Yin Platform" node, since the "Xin Yin Platform" node corresponds to the "Xin Yin Platform" entity in the advertisement placement knowledge graph, taking the "Xin Yin Platform" node as the end point, construct a hierarchical path: "Online Education Course Advertisement Placement Activity → Xin Yin Platform". Continuing the traversal, when encountering the "Placement Time" sub-node under the "Xin Yin Platform" node, and this sub-node corresponds to the "Xin Yin Platform Placement Time" entity in the knowledge graph, construct a new hierarchical path: "Online Education Course Advertisement Placement Activity → Xin Yin Platform → Placement Time", and corresponding hierarchical paths are constructed for all nodes in the structure tree that are mapped to knowledge graph entities.

[0062] Step 304: Based on the relationship hierarchical matrix, combine the hierarchical path and the relationship path in the advertisement placement knowledge graph to perform placement effect reasoning, and obtain the potential effect information of the advertisement placement.

[0063] Further, the advertisement analysis system performs placement effect reasoning according to the relationship hierarchical matrix, combining the hierarchical path and the relationship path in the advertisement placement knowledge graph, and obtains the potential effect information of the advertisement placement, as specifically described in Steps 3041 to 3043.

[0064] The embodiment of the present invention can sort out the relationship between the entities in the advertisement placement knowledge graph and the levels of the advertisement placement structure tree, construct a matrix reflecting the entity relationship and hierarchical association, so as to construct a hierarchical path by traversing the structure tree, combine the relationship hierarchical matrix and the knowledge graph relationship path, deeply analyze the connections between various elements, and infer the potential effect information of the advertisement placement. Therefore, it covers the possible effect improvement directions brought by different platforms and different combinations of placement elements, helps to plan and optimize the advertisement placement strategy in advance, thus adapting to the complex and changeable advertisement placement environment, and improving the timeliness and effectiveness of the advertisement placement strategy adjustment.

[0065] In one embodiment, the descriptions of Steps 3041 to 3043 are as follows: Step 3041, if it is determined that the hierarchical path matches the relationship path based on the relationship hierarchy matrix, then fuse the hierarchical path and the relationship path to obtain a fused path.

[0066] Optionally, the advertising analysis system analyzes and judges the hierarchical path and the relationship path based on the relationship hierarchy matrix. Among them, the hierarchical path is the node path constructed in the advertising placement structure tree, and the relationship path is the relationship connection between entities in the advertising placement knowledge graph. Therefore, by comparing the structures, node correspondence relationships, etc. of the hierarchical path and the relationship path, it is judged whether the two match. If they match, that is, the node hierarchical relationship in the hierarchical path is consistent with the relationship between entities in the relationship path logically and structurally, then fuse the hierarchical path and the relationship path, integrate the information of the two, and obtain a fused path. Among them, the fused path synthesizes the hierarchical information of the structure tree and the relationship information of the knowledge graph.

[0067] Continuing with the "online education course advertising placement activity", the constructed hierarchical path "online education course advertising placement activity → Xinyin platform → placement time" and the relationship path in the advertising placement knowledge graph "online education course promotion project → Xinyin platform → Xinyin platform placement time setting". From the perspective of structure and logic, the nodes in the hierarchical path correspond to the entities in the relationship path, and the hierarchical relationship and entity relationship are consistent. After judging that the two match based on the relationship hierarchy matrix, fuse these two paths to obtain a fused path: "online education course advertising placement activity (online education course promotion project) → Xinyin platform (Xinyin platform) → placement time (Xinyin platform placement time setting)". In the fused path, the content in parentheses is the corresponding entity information in the knowledge graph, realizing the integration of the hierarchical information of the structure tree and the relationship information of the knowledge graph.

[0068] Step 3042, for each path in the fused path, determine the advertising placement effect inference rule of the fused path based on the path characteristics of the path and the potential influence degree of each node in the path on the advertising placement effect.

[0069] Furthermore, for the obtained fused path, the advertising analysis system analyzes the path characteristics of each path. The path characteristics include path length, node type, importance of the node in the structure tree and the knowledge graph, etc. At the same time, evaluate the potential influence degree of each node in the path on the advertising placement effect. Among them, the influence degree is determined according to factors such as the historical data performance of the advertising placement elements represented by the node and industry experience.

[0070] Furthermore, the advertisement analysis system formulates inference rules for the placement effect by synthesizing the path features and the potential influence degree of nodes. For example, if the path is short and contains nodes of key placement elements, and the influence degree of such nodes on the placement effect is high, the inference rule can be set to focus on such nodes and optimize the relevant placement strategies; if there are multiple interrelated nodes in the path that have a synergistic effect on the placement effect, the inference rule can be set to adjust the relationship between these nodes to enhance the placement effect.

[0071] Continuing with the fusion path "Online education course advertisement placement activity (Online education course promotion project) → Xinyin platform (Xinyin platform) → Placement time (Xinyin platform placement time setting)" as an example, the advertisement analysis system analyzes its path features. The path length is moderate and includes the main body of the advertisement placement activity, the placement platform, and the node of the key placement element "placement time". By analyzing historical data and industry experience, the "placement time" node has a high influence degree on the advertisement placement effect of the Xinyin platform, and the "Xinyin platform" node, as the core placement platform, is also crucial. Based on this, the inference rule for the placement effect is determined: focus on the "placement time" node, further refine and optimize the placement time arrangement of the "Online education course advertisement" on the Xinyin platform according to the active period data of Xinyin platform users; at the same time, evaluate the synergistic relationship between the "Xinyin platform" and the "placement time" node, and enhance the promotion effect of both on the advertisement placement effect by adjusting the placement time strategy.

[0072] Step 3043, perform placement effect inference on the fusion path based on the placement effect inference rule to obtain the potential effect information of the advertisement placement.

[0073] Furthermore, the advertisement analysis system adjusts and simulates the nodes and relationships in the fusion path according to the placement effect inference rule, predicts the changes in the advertisement placement effect under different operations, so as to obtain the potential effect information of the advertisement placement.

[0074] For example, according to the inference rules, the delivery parameters of a certain node are adjusted, and the impact of this adjustment on other nodes and the overall delivery effect is analyzed through the relationship transmission in the fusion path, so as to obtain potential directions for effect improvement or possible problems. Continuing with the above embodiment, based on the determined delivery effect inference rules, the advertisement analysis system performs inference on the fusion path "Online education course advertisement delivery activity (Online education course promotion project) → Xinyin platform (Xinyin platform) → Delivery time (Xinyin platform delivery time setting)". The advertisement analysis system simulates adjusting the "delivery time" to a specific time period when the user activity on the Xinyin platform is higher, and analyzes how this adjustment affects the advertisement exposure volume and user click-through rate of the "Xinyin platform", as well as the impact on the conversion rate of the entire "online education course advertisement delivery activity" according to the relationships and inference rules of each node in the fusion path. Through inference, it is obtained that if the new delivery time strategy is implemented, it is expected that the advertisement click-through rate on the Xinyin platform can be increased by 15%, and the conversion rate of the overall advertisement delivery activity on the Xinyin platform is expected to increase by 10%. These prediction results are the potential effect information of the advertisement delivery.

[0075] The embodiment of the present invention can effectively integrate the hierarchical information of the advertisement delivery structure tree and the relationship information of the advertisement delivery knowledge graph to form a fusion path containing rich information. Therefore, the delivery effect inference rules determined based on the characteristics of the fusion path and the node influence provide a clear logical basis for the advertisement delivery effect analysis, can deeply explore the potential effects that may be brought by different combinations of advertisement delivery elements and strategy adjustments, obtain targeted and forward-looking advertisement delivery potential effect information, and provide precise guidance for the optimization of the advertisement delivery strategy, so as to adapt to the complex and changeable advertisement delivery environment and improve the timeliness and effectiveness of the advertisement delivery strategy adjustment.

[0076] In one embodiment, the descriptions of steps 401 to 404 are as follows: Step 401: For each advertisement delivery platform, perform an association relationship analysis based on the first advertisement features corresponding to the hierarchical effect information and the second advertisement features corresponding to the potential effect information, and construct a feature association relationship matrix.

[0077] Optionally, for each advertisement delivery platform, the advertisement analysis system respectively extracts the corresponding first advertisement features from the hierarchical effect information, such as actual delivery effect related features such as the actual exposure volume, click-through rate, and conversion rate of the platform; and extracts the corresponding second advertisement features from the potential effect information, such as features such as the potential exposure growth space and potential conversion rate improvement points obtained through inference.

[0078] Furthermore, the advertisement analysis system analyzes the correlation between these first advertisement features and second advertisement features, including causal relationships and the degree of mutual influence between features, and quantitatively represents these correlations to construct a feature correlation matrix. The elements in the matrix reflect the correlation strength and correlation nature between different features, thus clearly presenting the internal connections between advertisement features of each platform.

[0079] Continuing with the "Online Education Course Advertisement Placement Campaign", for the "Xin Platform", the advertisement analysis system extracts the first advertisement feature from the hierarchical effect information: the current advertisement exposure is 15,000 times per day, the click-through rate is 5%, and the conversion rate is 3%; and extracts the second advertisement feature from the potential effect information: if the placement time is optimized, the potential exposure can be increased to 20,000 times per day, and the potential conversion rate is expected to increase to 5%. It is found that there is a certain correlation between the current click-through rate and the increase in the potential conversion rate. If the advertisement content can be optimized to increase the current click-through rate, it may help to achieve an increase in the potential conversion rate. After quantifying this correlation, it is filled into the corresponding position in the feature correlation matrix. Similarly, the "Xinxin Platform" and the "Xidu Platform" are analyzed, their respective first advertisement features and second advertisement features are extracted, and the correlation between them is determined to construct a complete feature correlation matrix for each platform.

[0080] Step 402: Conduct an effect difference analysis based on the first advertisement feature, the second advertisement feature, and the feature correlation matrix to obtain platform effect difference information.

[0081] Furthermore, based on the extracted first advertisement feature, the second advertisement feature, and the constructed feature correlation matrix, the advertisement analysis system conducts an effect difference analysis on each advertisement placement platform, specifically: comparing the first advertisement features of different platforms to understand the differences in the current actual placement effects of each platform; comparing the second advertisement features to analyze the differences in the potential placement effects of each platform. At the same time, combining the feature correlation matrix, considering the influence of the correlation between features on the effect difference. Through comprehensive analysis, the differences in the actual and potential effects of each advertisement placement platform are obtained, forming platform effect difference information.

[0082] Continuing with the above embodiment, the advertisement analysis system compares the first advertisement features of the three platforms, namely the "XinYin Platform", the "XinXin Platform", and the "XinDu Platform", and finds that the "XinYin Platform" has a relatively high exposure volume but a relatively low conversion rate, the "XinXin Platform" has a relatively high conversion rate but a limited exposure volume, and the "XinDu Platform" has a unique performance in the conversion of keyword search traffic. When comparing the second advertisement features, the "XinYin Platform" has a relatively large potential for increasing the exposure volume, the "XinXin Platform" has a significant potential for increasing the conversion rate by optimizing the targeted population strategy, and the potential effect of the "XinDu Platform" is expected to improve after optimizing the keyword settings. Combining with the feature correlation relationship matrix, analyze the impact of the correlation between the features of each platform on the effect, such as the correlation between the click-through rate and the potential conversion rate of the "XinYin Platform" affecting the overall effect improvement direction. Finally, the advertisement analysis system obtains the platform effect difference information, and clarifies that the "XinYin Platform" currently focuses on optimizing exposure and potentially on conversion; the "XinXin Platform" currently focuses on conversion and potentially on expanding exposure; the "XinDu Platform" needs to balance the differences in keyword traffic and conversion, etc.

[0083] Step 403: Identify the potential opportunities and potential risks of each advertisement placement platform based on the potential effect information.

[0084] Furthermore, the advertisement analysis system conducts an in-depth analysis of each advertisement placement platform based on the potential effect information to identify the potential opportunities and potential risks therein. For potential opportunities, look for directions and points that can significantly improve the advertisement placement effect by reasonably adjusting the placement strategy, optimizing the placement elements, etc. For example, it is found that a specific user group of a certain platform has not been fully covered, and there is an opportunity to expand the audience scope to improve the effect. For potential risks, pay attention to factors that may lead to a decline in the advertisement placement effect. For example, policy changes may affect the advertisement display rules of a certain platform, thereby affecting the placement effect. Through comprehensive analysis, the system clarifies the potential opportunities and potential risks of each platform.

[0085] Continuing to analyze the potential effect information of the "X Sound Platform", the advertising analysis system found that the newly launched live promotion function for education on the X Sound Platform is still in the development stage. If online education course advertisements can participate in the live promotion in a timely manner, with the help of the interactivity and real-time nature of the live broadcast, more users can be attracted to pay attention. This is a potential opportunity for the "X Sound Platform". At the same time, the requirements of X Sound Platform users for the creativity of advertisement content are constantly increasing. If the advertising analysis system fails to optimize the advertisement content in a timely manner, it may lead to an increase in users' resistance to advertisements, a decrease in click-through rate and conversion rate, which is a potential risk. For the "X Message Platform", the potential opportunity lies in the fact that X Message has recently opened more advertising display positions, which can increase the advertisement exposure. The potential risk is that the adjustment of the X Message user privacy policy may affect the accuracy of targeted advertisement placement. The potential opportunity of the "X Degree Platform" is the rising search volume of emerging education keywords, and more traffic can be obtained by expanding relevant keywords. The potential risk is that competitors increase the intensity of keyword bidding, which may lead to an increase in advertising costs. Through such analysis, the advertising analysis system identifies the potential opportunities and potential risks of each platform.

[0086] Step 404, based on the platform effect difference information, potential opportunities and potential risks of each advertising placement platform, conduct a decision-making analysis on the placement effect to obtain placement optimization information for multi-platform advertising placement.

[0087] Furthermore, the advertising analysis system conducts a decision-making analysis on the placement effect according to the platform effect difference information, potential opportunities and potential risks of each advertising placement platform to obtain placement optimization information for multi-platform advertising placement, as specifically described in steps 4041 to 4044.

[0088] In the embodiment of the present invention, from constructing a feature correlation relationship matrix to clearly sorting out the internal relationships of the advertising features of each platform, to analyzing the effect differences, identifying potential opportunities and risks, and then to making a comprehensive decision based on this information, finally obtaining highly targeted placement optimization information, enabling the placement optimization information to cover multiple aspects such as budget allocation, strategy adjustment, and risk prevention and control, making the advertising placement strategy more scientific and reasonable, thus adapting to the complex and changeable advertising placement environment and enhancing the timeliness and effectiveness of the adjustment of the advertising placement strategy.

[0089] In one embodiment, the descriptions of steps 4041 to 4044 are as follows: Step 4041, for each advertising placement platform, based on the advertising placement goal, combine the platform effect difference information, potential opportunities and potential risks to conduct strategy screening and determine the preliminary placement strategy.

[0090] Optionally, for each advertising platform, the advertising analysis system combines the advertising goals (such as increasing brand awareness, increasing course registrations, and improving user activity) with platform effect difference information, potential opportunities, and potential risks for comprehensive consideration, and evaluates the strengths and weaknesses of each platform in achieving specific goals, as well as the impact of potential opportunities and risks on the achievement of goals. For example, if the advertising goal is to increase course registrations, for platforms with low actual conversion rates but large potential for improvement, consider adopting strategies such as optimizing advertising content and adjusting delivery time; for platforms where potential risks may affect the achievement of goals, screen out risk avoidance strategies and determine preliminary delivery strategies for each platform.

[0091] In the "Online Education Course Advertising Campaign", the advertising goal is to increase the number of course registrations by 30% within one month. For the "Xyin Platform", according to the platform effect difference information, its current conversion rate is low. The potential effect information shows that optimizing the delivery time and advertising content can improve the conversion rate. At the same time, there is a potential risk of users' increased requirements for advertising content creativity. The advertising analysis system screened out the initial delivery strategy for the "Xyin Platform": adjust the advertising delivery time to 7-10 pm when users are active; invest resources to optimize the advertising video content to increase fun and interactivity; arrange special personnel to monitor user feedback and adjust the advertising strategy in time to deal with potential risks. For the "Xxin Platform", because its conversion rate is high but the exposure is limited, and there is a potential opportunity for new advertising display positions, the initial delivery strategy is determined as follows: increase the advertising delivery budget and use the new display positions to expand exposure; optimize the advertising copy to highlight the advantages of the course to attract more users to click. For the "Xdu Platform", in view of its keyword traffic conversion characteristics, as well as the potential opportunities for the increase in the search volume of emerging education keywords and the potential risks of competitors' bidding, the initial delivery strategy is: increase the delivery of emerging education keywords; formulate a flexible bidding strategy, and dynamically adjust the bid according to the situation of competitors.

[0092] Step 4042, for any two first delivery strategies and second delivery strategies in the preliminary delivery strategies, a synergy analysis is performed based on a first strategy feature corresponding to the first delivery strategy and a second strategy feature corresponding to the second delivery strategy to obtain a strategy synergy degree between the two delivery strategies.

[0093] Furthermore, for any two strategies (the first placement strategy and the second placement strategy) in the preliminary placement strategy, the advertising analysis system extracts the first strategy features corresponding to the first placement strategy (such as the advertising placement time, budget, content, and other related features involved in the strategy) and the second strategy features corresponding to the second placement strategy. Further, the advertising analysis system analyzes the interaction relationship between these strategy features to determine whether the two strategies can promote each other and synergistically enhance the effectiveness in the process of achieving the advertising placement goal. For example, one strategy is to increase the advertising budget to expand exposure, and the other strategy is to optimize the advertising content to improve the click-through rate. If increasing the exposure can enable more users to see the optimized advertising content, thereby increasing the conversion rate, it indicates that these two strategies have a high degree of synergy. Therefore, this synergy is quantitatively represented to obtain the strategy synergy degree between the two placement strategies, and the higher the value, the stronger the synergy.

[0094] Continuing with the preliminary placement strategy, for the "Xin platform", the "adjust the advertising placement time to 7 - 10 pm when users are active" (the first placement strategy) and the "allocate resources to optimize the advertising video content to increase interest and interactivity" (the second placement strategy). The first strategy features corresponding to the first placement strategy include the new placement time period, and the second strategy features corresponding to the second placement strategy include the optimized advertising content style and interaction form. It is found that placing the optimized advertising content during the user-active period can enable more users to see high-quality advertisements, increasing the user stay time and click-through rate, and the two promote each other. By analyzing the promotion relationship between the two, it is quantified as the strategy synergy degree and set to a relatively high value of 0.8. Another example is the two strategies of the "WeChat platform": "increase the advertising placement budget and use the new display positions to expand exposure" and "optimize the advertising copy to highlight the course advantages to attract more users to click". Increasing the exposure can enable the optimized copy to be seen by more users, thereby increasing the click volume. For example, after analysis, its strategy synergy degree is determined to be 0.7.

[0095] Step 4043: Based on the dynamic environmental factors of multiple platforms and the strategy synergy degree between any two placement strategies, determine the fitness of each placement strategy in the dynamic environment.

[0096] Furthermore, the advertisement analysis system takes into account the dynamic environmental factors of multiple platforms, such as changes in market trends, adjustments in competitors' strategies, shifts in user demands, updates to platform policies, etc., and combines the strategy synergy degree between any two placement strategies to evaluate the adaptability of each placement strategy in the dynamic environment. Optionally, the embodiments of the present invention analyze the impact of dynamic environmental factors on each strategy, and whether the synergy between strategies can enhance the ability to respond to environmental changes. For example, if the market trend turns to pay more attention to the interactivity of video content, and a placement strategy of a certain platform includes optimizing the advertisement video content to increase interactivity, and this strategy has a high synergy degree with other strategies, then the adaptability of this strategy in the dynamic environment is relatively high. Quantify this adaptability to obtain the adaptability of each placement strategy in the dynamic environment.

[0097] Continuing with the above embodiment, the recent market trend in the online education market shows that users are more inclined to watch live courses, and at the same time, competitors have increased their live promotion efforts on the "Xin Yin platform". Among the preliminary placement strategies of the "Xin Yin platform" such as "adjusting the advertisement placement time to 7 - 10 pm when users are active", "allocating resources to optimize the advertisement video content to increase fun and interactivity", and "arranging special personnel to monitor user feedback and timely adjust advertisement strategies to address potential risks", the strategy of "allocating resources to optimize the advertisement video content to increase fun and interactivity" is in line with the market trend, and this strategy has a high strategy synergy degree with the strategy of "adjusting the advertisement placement time", and can better display highly interactive advertisement content during the user - active period. The advertisement analysis system comprehensively considers the dynamic environmental factors and the strategy synergy degree, and determines that the adaptability of the strategy of "allocating resources to optimize the advertisement video content to increase fun and interactivity" in the dynamic environment is relatively high, set to 0.9; the adaptability of the strategy of "adjusting the advertisement placement time to 7 - 10 pm when users are active" is 0.8. For the "Xin Xin platform", if Xin Xin launches a new privacy policy that affects advertisement targeted placement, the strategy of "optimizing the advertisement copy to highlight the course advantages to attract more users to click" is less affected, and has a certain synergy with the strategy of "increasing the advertisement placement budget and using new display positions to expand exposure", and determines that the adaptability of the strategy of "optimizing the advertisement copy" is 0.7, and the adaptability of the strategy of "increasing the advertisement placement budget" is 0.6.

[0098] Step 4044, based on the adaptability of each placement strategy in the dynamic environment, conduct a placement effect decision - making analysis to obtain placement optimization information for multi - platform advertisement placement.

[0099] Furthermore, the advertisement analysis system conducts a comprehensive placement effect decision - making analysis based on the adaptability of each placement strategy in the dynamic environment. The embodiments of the present invention preferentially select strategies with high adaptability for implementation, and adjust the resource allocation and execution priority of the strategies according to the adaptability.

[0100] For strategies with lower fitness, the system considers whether optimization, adjustment, or abandonment is needed. At the same time, considering the strategy situations of each platform, a comprehensive multi-platform advertising placement optimization plan is formulated, including budget allocation, strategy combination, execution time arrangement, etc. for each platform, forming advertising placement optimization information for multi-platform advertising placement to maximize the advertising placement effect in a dynamic environment.

[0101] Continuing with the above embodiment, decision-making analysis is carried out on the "online education course advertising placement activity" according to the fitness of each platform's placement strategy in a dynamic environment. For the "Xin Yin platform", since the strategy of "investing resources to optimize the advertising video content and increase its interest and interactivity" has the highest fitness (0.9), it is decided to increase the resource investment in this strategy, adding personnel and budget to the video production team; the strategy of "adjusting the advertising placement time to 7-10 pm when users are active" has a fitness of 0.8, and it is executed as planned and the effect is continuously monitored. For the "Xin Xin platform", the strategy of "optimizing the advertising copy to highlight the course advantages to attract more users to click" has a fitness of 0.7 and is executed as a key strategy. At the same time, the budget investment in the strategy of "increasing the advertising placement budget and using new display positions to expand exposure" (fitness 0.6) is appropriately reduced according to the fitness. Considering the situations of each platform, the final advertising placement optimization information is determined: 60% of the budget of the "Xin Yin platform" is used to optimize the advertising video content, 30% is used to adjust the placement time, and 10% is used for monitoring and feedback; 70% of the budget of the "Xin Xin platform" is used to optimize the advertising copy, and 30% is used for limited exposure expansion; the strategies of each platform are executed in an orderly manner according to the priority and time node to achieve the goal of increasing the course registration volume by 30% and adapting to the changes in the dynamic market environment.

[0102] The embodiments of the present invention can fully consider the advertising placement objectives, platform characteristics, strategy synergy, and dynamic environment factors. From screening preliminary placement strategies, to analyzing strategy synergy and determining the fitness of strategies in a dynamic environment, and finally making scientific decisions based on the fitness, comprehensive and targeted multi-platform advertising placement optimization information is obtained, enabling the optimization information to ensure that the advertising placement strategy not only fits the actual situations and potential opportunities of each platform, but also maintains good adaptability and synergy in a complex and changeable market environment, adapts to the complex and changeable advertising placement environment, and improves the timeliness and effectiveness of advertising placement strategy adjustment.

[0103] Furthermore, the multi-platform advertising placement effect analysis system based on artificial intelligence provided by the present invention is described below. The multi-platform advertising placement effect analysis system based on artificial intelligence described below can be mutually corresponding and referred to with the multi-platform advertising placement effect analysis method described above.

[0104] Refer to Figure 2 , Figure 2It is a structural schematic diagram of a multi-platform advertising delivery effect analysis system based on artificial intelligence provided by the present invention. The multi-platform advertising delivery effect analysis system based on artificial intelligence includes.

[0105] The construction module 210 is used to construct an advertising placement structure tree and an advertising placement knowledge graph; the advertising placement structure tree is constructed by taking a preset advertising activity as a root node, taking different advertising placement platforms as child nodes under the root node, and taking advertising placement elements of different advertising placement platforms as different level nodes of its child nodes; the advertising placement knowledge graph is constructed by taking different advertising placement platforms and advertising placement subjects of different advertising placement platforms as entities, and taking the relationship between entities as edges; The effect trend analysis module 220 is used to start from the root node of the advertising placement structure tree, combine the semantic information of each node mapped to the advertising placement knowledge graph by the association mapping relationship, traverse the advertising placement data along different paths, analyze the association relationship and data change trend between nodes at each level, and obtain hierarchical effect information of advertising placement; the association mapping relationship represents the mapping relationship between the entity in the advertising placement knowledge graph and the node in the advertising placement structure tree; The effect reasoning module 230 is used to perform advertising effect reasoning based on the relationship between entities in the advertising knowledge graph and the level of each entity mapped in the advertising structure tree in combination with the association mapping relationship, so as to obtain the potential effect information of the advertising; The effect decision analysis module 240 is used to perform delivery effect decision analysis based on the hierarchical effect information and the potential effect information to obtain delivery optimization information for multi-platform advertising delivery.

[0106] The embodiment of the present invention can clearly present the hierarchical architecture of multi-platform advertising by constructing a multi-platform advertising structure tree, intuitively display the internal correlation and data change trend of advertising on each platform, and make the analysis no longer limited to simple matching under fixed rules. Generating an advertising knowledge graph and performing semantic reasoning can dig out the potential complex relationships and semantic information between entities in advertising, making up for the defect that fixed rules cannot incorporate new factors. Therefore, the association mapping between the structure tree and the knowledge graph and the subsequent comprehensive analysis can comprehensively and deeply analyze the advertising effect from the two dimensions of hierarchy and semantics, and timely discover new factors and potential optimization space that affect the advertising effect, thereby generating more accurate and effective optimization suggestions, adapting to the complex and changing advertising environment, and improving the timeliness and effectiveness of the adjustment of advertising strategies.

[0107] See also Figure 3 , Figure 3 FIG. 1 is an embodiment diagram of an electronic device provided by an embodiment of the present invention. Figure 3As shown in the figure, 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 operable on the processor 320. When the processor 320 executes the computer program 311, the following steps are implemented: Construct an advertising placement structure tree and an advertising placement knowledge graph; the advertising placement structure tree is constructed with a preset advertising campaign as the root node, different advertising placement platforms as the sub-nodes under the root node, and different levels of nodes with advertising placement elements of different advertising placement platforms as its sub-nodes; the advertising placement knowledge graph is constructed with different advertising placement platforms and advertising placement entities of different advertising placement platforms as entities and the relationships between the entities as edges; Starting from the root node of the advertising placement structure tree, combined with the semantic information of each node mapped in the advertising placement knowledge graph according to the association mapping relationship, traverse the advertising placement data along different paths, analyze the association relationships and data change trends between the nodes at each level, and obtain the hierarchical effect information of the advertising placement; the association mapping relationship represents the mapping relationship between the entities in the advertising placement knowledge graph and the nodes in the advertising placement structure tree; Based on the relationships between the entities in the advertising placement knowledge graph, combined with the levels of each entity mapped in the advertising placement structure tree according to the association mapping relationship, perform placement effect reasoning to obtain the potential effect information of the advertising placement; Based on the hierarchical effect information and the potential effect information, perform placement effect decision-making analysis to obtain the placement optimization information for multi-platform advertising placement.

[0108] Please refer to Figure 4 , Figure 4 This is the embodiment diagram of the computer-readable storage medium provided by the embodiment of the present invention. As Figure 4 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, the following steps are implemented: Construct an advertising placement structure tree and an advertising placement knowledge graph; the advertising placement structure tree is constructed with a preset advertising campaign as the root node, different advertising placement platforms as the sub-nodes under the root node, and different levels of nodes with advertising placement elements of different advertising placement platforms as its sub-nodes; the advertising placement knowledge graph is constructed with different advertising placement platforms and advertising placement entities of different advertising placement platforms as entities and the relationships between the entities as edges; Starting from the root node of the advertising placement structure tree, in combination with the semantic information of each node mapped in the advertising placement knowledge graph according to the association mapping relationship, traverse the advertising placement data along different paths, analyze the association relationships and data change trends among nodes at each level, and obtain the hierarchical effect information of advertising placement; the association mapping relationship represents the mapping relationship between entities in the advertising placement knowledge graph and nodes in the advertising placement structure tree; Based on the relationships between entities in the advertising placement knowledge graph, perform placement effect reasoning in combination with the levels of each entity mapped in the advertising placement structure tree according to the association mapping relationship, and obtain the potential effect information of advertising placement; Perform placement effect decision-making analysis based on the hierarchical effect information and the potential effect information, and obtain the placement optimization information for multi-platform advertising placement.

[0109] On the other hand, the present invention also provides a computer program product. The computer program product includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the artificial intelligence-based multi-platform advertising placement effect analysis method provided by each of the above methods. The method includes: Construct an advertising placement structure tree and an advertising placement knowledge graph; the advertising placement structure tree is constructed with a preset advertising campaign as the root node, different advertising placement platforms as sub-nodes under the root node, and different levels of nodes with advertising placement elements of different advertising placement platforms as their sub-nodes; the advertising placement knowledge graph is constructed with different advertising placement platforms and advertising placement entities of different advertising placement platforms as entities and the relationships between the entities as edges; Starting from the root node of the advertising placement structure tree, in combination with the semantic information of each node mapped in the advertising placement knowledge graph according to the association mapping relationship, traverse the advertising placement data along different paths, analyze the association relationships and data change trends among nodes at each level, and obtain the hierarchical effect information of advertising placement; the association mapping relationship represents the mapping relationship between entities in the advertising placement knowledge graph and nodes in the advertising placement structure tree; Based on the relationships between entities in the advertising placement knowledge graph, perform placement effect reasoning in combination with the levels of each entity mapped in the advertising placement structure tree according to the association mapping relationship, and obtain the potential effect information of advertising placement; Perform placement effect decision-making analysis based on the hierarchical effect information and the potential effect information, and obtain the placement optimization information for multi-platform advertising placement.

[0110] The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0111] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. The 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 for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0112] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for analyzing the advertising placement effect on multiple platforms based on artificial intelligence, characterized in that, Including: Constructing an advertising placement structure tree and an advertising placement knowledge graph; the advertising placement structure tree takes a preset advertising campaign as the root node, different advertising placement platforms as the child nodes under the root node, and different advertising placement elements of different advertising placement platforms as the different hierarchical nodes of its child nodes; the advertising placement knowledge graph is constructed with different advertising placement platforms and advertising placement entities of different advertising placement platforms as entities and the relationships between entities as edges; Starting from the root node of the advertising placement structure tree, traversing the advertising placement data along different paths by combining the semantic information of each node mapped in the advertising placement knowledge graph according to the association mapping relationship, and analyzing the association relationships and data change trends between nodes at each level to obtain the hierarchical effect information of advertising placement; The association mapping relationship represents the mapping relationship between the entities in the advertising placement knowledge graph and the nodes in the advertising placement structure tree; Based on the relationships between entities in the advertising placement knowledge graph and combining the levels of each entity mapped in the advertising placement structure tree according to the association mapping relationship, performing placement effect reasoning to obtain the potential effect information of advertising placement; Based on the hierarchical effect information and the potential effect information, performing placement effect decision-making analysis to obtain the placement optimization information for multi-platform advertising placement.

2. The method for analyzing the advertising placement effect on multiple platforms based on artificial intelligence according to claim 1, wherein The performing placement effect decision-making analysis based on the hierarchical effect information and the potential effect information to obtain the placement optimization information for multi-platform advertising placement includes: For each advertising placement platform, performing association relationship analysis based on the first advertising feature corresponding to the hierarchical effect information and the second advertising feature corresponding to the potential effect information, and constructing a feature association relationship matrix; Based on the first advertising feature, the second advertising feature, and the feature association relationship matrix, performing effect difference analysis to obtain the platform effect difference information; Based on the potential effect information, identifying the potential opportunities and potential risks of each advertising placement platform; Based on the platform effect difference information, potential opportunities, and potential risks of each advertising placement platform, performing placement effect decision-making analysis to obtain the placement optimization information for multi-platform advertising placement.

3. The method for analyzing the advertising placement effect on multiple platforms based on artificial intelligence according to claim 2, wherein, The performing placement effect decision-making analysis based on the platform effect difference information, potential opportunities, and potential risks of each advertising placement platform to obtain the placement optimization information for multi-platform advertising placement includes: For each advertising placement platform, based on the goal of advertising placement and combining the platform effect difference information, potential opportunities, and potential risks, performing strategy screening to determine the preliminary placement strategy; For any two first placement strategies and second placement strategies in the preliminary placement strategies, performing synergy analysis based on the first strategy feature corresponding to the first placement strategy and the second strategy feature corresponding to the second placement strategy to obtain the strategy synergy degree between the two placement strategies; Based on the dynamic environmental factors of multiple platforms and combining the strategy synergy degree between any two placement strategies, determining the fitness of each placement strategy in the dynamic environment; Based on the fitness of each placement strategy in the dynamic environment, performing placement effect decision-making analysis to obtain the placement optimization information for multi-platform advertising placement.

4. The method for analyzing the advertising placement effect on multiple platforms based on artificial intelligence according to claim 1, wherein, Starting from the root node of the advertisement placement structure tree, combining the semantic information of each node mapped in the advertisement placement knowledge graph according to the association mapping relationship, traversing the advertisement placement data along different paths, analyzing the association relationships and data change trends among nodes at each level, and obtaining the hierarchical effect information of advertisement placement, including: Starting from the root node of the advertisement placement structure tree, locating the first entity associated with the root node in the advertisement placement knowledge graph according to the association mapping relationship, and generating an initial traversal path starting from the root node entity based on the similarity and relevance among the semantic information of the first entity; Finding the sub-nodes corresponding to different advertisement placement platforms in the advertisement placement structure tree along the initial traversal path, finding the second entity corresponding to the sub-nodes in the advertisement placement knowledge graph according to the association mapping relationship, and expanding the initial traversal path based on the semantic information of the second entity to obtain an expanded traversal path; For the advertisement placement element nodes of the advertisement placement platform, finding the element nodes related to the current path in the advertisement placement structure tree, and fusing the corresponding element semantic information with the entity semantic information corresponding to it in the advertisement placement knowledge graph to obtain a fused semantic information vector; Traversing the advertisement placement data based on the expanded traversal path and the fused semantic information vector, analyzing the association relationships and data change trends among nodes at each level, and obtaining the hierarchical effect information of advertisement placement.

5. The method for analyzing the advertising placement effect on multiple platforms based on artificial intelligence according to claim 4, wherein The traversing the advertisement placement data based on the expanded traversal path and the fused semantic information vector, analyzing the association relationships and data change trends among nodes at each level, and obtaining the hierarchical effect information of advertisement placement, including: Based on the total difference degree of the semantic information of each node on each traversal path in the expanded traversal path, determining the semantic information difference weight of each traversal path, and screening each traversal path based on the semantic information difference weight to obtain the target traversal path in the expanded traversal path; Traversing the advertisement placement data along the target traversal path, determining the relationship edges between the entities corresponding to the nodes at each level in the advertisement placement knowledge graph, and constructing an association relationship matrix among the nodes at each level based on the relationship strength and semantic information of the relationship edges; For the advertisement placement data corresponding to the nodes at each level, determining the data change trend vector among the nodes at each level based on the change of the semantic information of the entities corresponding to the nodes at each level in the advertisement placement knowledge graph at different time points; Fusing based on the association relationship matrix and the data change trend vector among the nodes at each level to obtain the hierarchical effect information of advertisement placement.

6. The method for analyzing the advertising placement effect on multiple platforms based on artificial intelligence according to any one of claims 1 to 5, characterized in that, Based on the relationships between entities in the advertisement placement knowledge graph, combining the levels of each entity mapped in the advertisement placement structure tree according to the association mapping relationship to perform placement effect reasoning, and obtaining the potential effect information of advertisement placement, including: Determining the entity level of each entity in the advertisement placement knowledge graph in the advertisement placement structure tree according to the association mapping relationship; Construct a relationship hierarchy matrix based on the relationships between entities in the advertising placement knowledge graph and the entity hierarchy of each entity; the matrix elements in the relationship hierarchy matrix represent the association characteristics of the relationships between entities corresponding to the entity hierarchy in the advertising placement structure tree; In the advertising placement structure tree, start traversing from the root node. If the target node traversed in the advertising placement structure tree is the node corresponding to the mapped entity in the advertising placement knowledge graph, construct a hierarchy path based on the target node; Based on the relationship hierarchy matrix, combine the hierarchy path and the relationship path in the advertising placement knowledge graph to perform placement effect reasoning, and obtain the potential effect information of the advertising placement.

7. The method for analyzing the advertising delivery effect across multiple platforms based on artificial intelligence according to claim 6, wherein The performing placement effect reasoning based on the relationship hierarchy matrix, combining the hierarchy path and the relationship path in the advertising placement knowledge graph to obtain the potential effect information of the advertising placement includes: If it is determined based on the relationship hierarchy matrix that the hierarchy path matches the relationship path, fuse the hierarchy path and the relationship path to obtain a fused path; For each path in the fused path, determine the placement effect reasoning rule of the fused path based on the path feature of the path and the potential influence degree of each node in the path on the placement effect; Based on the placement effect reasoning rule, perform placement effect reasoning on the fused path to obtain the potential effect information of the advertising placement.

8. An artificial intelligence-based multi-platform advertising placement effect analysis system, characterized in that, Applied to the multi-platform advertising placement effect analysis method based on artificial intelligence as described in any one of claims 1 to 7; the multi-platform advertising placement effect analysis system based on artificial intelligence includes: A construction module for constructing an advertising placement structure tree and an advertising placement knowledge graph; the advertising placement structure tree takes a preset advertising campaign as the root node, different advertising placement platforms as the sub-nodes under the root node, and different levels of nodes of the advertising placement elements of different advertising placement platforms as its sub-nodes; the advertising placement knowledge graph takes different advertising placement platforms and the advertising placement entities of different advertising placement platforms as entities, and the relationships between entities as edges to construct; An effect trend analysis module for starting from the root node of the advertising placement structure tree, combining the semantic information of each node mapped in the advertising placement knowledge graph through the association mapping relationship, traversing the advertising placement data along different paths, and analyzing the association relationships and data change trends between nodes at each level to obtain the hierarchical effect information of the advertising placement; the association mapping relationship represents the mapping relationship between the entities in the advertising placement knowledge graph and the nodes in the advertising placement structure tree; An effect reasoning module for performing placement effect reasoning based on the relationships between entities in the advertising placement knowledge graph, combining the entity levels of each entity mapped in the advertising placement structure tree through the association mapping relationship, and obtaining the potential effect information of the advertising placement; An effect decision analysis module for performing placement effect decision analysis based on the hierarchical effect information and the potential effect information to obtain the placement optimization information for multi-platform advertising placement.

9. An electronic device, comprising: A memory for storing computer software programs; A processor for reading and executing the computer software program, characterized in that when the processor executes the computer software program, it implements the artificial intelligence-based multi-platform advertising delivery effect analysis method according to any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium storing a computer software program, characterized in that, When the computer software program is executed by the processor, it implements the artificial intelligence-based multi-platform advertising delivery effect analysis method according to any one of claims 1 to 7.

Images