GEO optimization strategy generation method and device based on simulated webpage interaction behavior, equipment and medium

Abstract

This application discloses a method, apparatus, device, and medium for generating GEO optimization strategies based on simulated web-based interactive behavior, relating to the field of model optimization technology. The method includes: constructing an evaluation prompt word matrix based on the business domain corresponding to the user terminal; inputting the question statements in the evaluation prompt word matrix into a preset web-based terminal constructed based on GEO to simulate user inquiry behavior on the preset web-based terminal and obtain corresponding text response information; using a text processing model to perform structured parsing of the text response information to extract target structured text; the target identifier name includes the target brand identifier and competitor identifiers; determining the evaluation dimensions for GEO, and evaluating the target structured text based on the evaluation dimensions to obtain the evaluation results corresponding to each evaluation dimension, so as to generate a list of optimization strategies corresponding to GEO based on the evaluation results. In this way, realistic GEO performance evaluation results can be obtained and more targeted optimization strategies can be generated.

Description

Technical Field

[0001] This invention relates to the field of model optimization technology, and in particular to a method, apparatus, device, and medium for generating GEO optimization strategies based on simulated web page interaction behavior. Background Technology

[0002] In recent years, generative AI (Artificial Intelligence) has been systematically changing how users obtain information and make decisions. More and more users are no longer relying on traditional search engines to filter links by entering keywords, but instead directly asking questions to generative engines to obtain more direct and accurate aggregated answers. For overseas brands and cross-border teams, this change not only means the formation of new traffic entry points, but also brings an urgent challenge: when users seek decision-making references from AI, can the brand appear in the answer content, and is the brand's presentation accurate and controllable? Against this backdrop, the concept of Generative Engine Optimization (GEO) has emerged, and various brands have successively invested in GEO construction in hopes of gaining a favorable position in AI-generated content. However, in evaluating the actual effectiveness of GEO, the industry currently faces two major pain points. On the one hand, discussions about GEO within the industry still largely remain at the conceptual level, with a severe lack of tools and means to quantitatively analyze, verify, and compare AI answers, making GEO optimization work like the blind men and the elephant. On the other hand, brand optimization strategies have long relied on standardized, template-based operating procedures and experience-based judgment, making it difficult to implement data-driven, customized strategy adjustments and performance reviews tailored to different brand attributes, development stages, and competitive environments. Therefore, the industry urgently needs an evaluation technology solution that can accurately reflect the user's true perspective, objectively measure model responses, and intuitively evaluate the effectiveness of GEO implementation. Summary of the Invention

[0003] In view of this, the purpose of this invention is to provide a method, apparatus, device, and medium for generating GEO optimization strategies based on simulated webpage interaction behavior, which can obtain realistic GEO performance evaluation results and generate more targeted optimization strategies. The specific solution is as follows: In a first aspect, this application discloses a method for generating GEO optimization strategies based on simulated webpage interaction behavior, including: An evaluation prompt word matrix is ​​constructed based on the business domain corresponding to the user terminal, and the question statements in the evaluation prompt word matrix are input into a preset web page based on GEO to simulate user inquiry behavior on the preset web page and obtain corresponding text answer information. The text response information is structured and parsed using a text processing model to extract the target structured text; the target structured text includes the target identifier name and the corresponding product website link; the target identifier name includes the target brand identifier and competitor identifiers; The evaluation dimensions for the GEO are determined, and the target structured text is evaluated based on the evaluation dimensions to obtain the evaluation results corresponding to each evaluation dimension, so as to generate an optimization strategy list for the GEO based on the evaluation results.

[0004] Optionally, the step of constructing an evaluation prompt word matrix based on the business domain corresponding to the user terminal includes: Receives the target brand identifier and competitor identifier configured by the user; the target brand identifier includes the main brand name, variant name, and official domain name; Based on the business domains corresponding to the target brand identifier and the competitor identifier, a business theme is generated. Under the business theme, corresponding question statements are expanded and configured, and the question statements are bound to the target search content or target intent weight to construct an evaluation prompt word matrix.

[0005] Optionally, the step of inputting the question statements from the evaluation prompt word matrix into a preset webpage built on GEO, to simulate user inquiry behavior on the preset webpage and obtain corresponding text answer information, includes: Based on a preset monitoring frequency, a preset webpage interaction simulation engine is scheduled to perform login and initialization operations on a preset webpage, and the question statements in the evaluation prompt word matrix are input into the interaction interface of the preset webpage to simulate user inquiry behavior. The natural language text responses, reference links, query behaviors, and website sources of the interactive interface are crawled to generate corresponding text response information.

[0006] Optionally, the step of using a text processing model to perform structured parsing of the text response information to extract the target structured text includes: Entity recognition is performed on the text response information using natural language processing techniques or a lightweight information extraction model to determine whether the text response information contains a target identifier name; If the text response information contains the target identifier name, then the target identifier name is extracted, and the location and link set corresponding to the target identifier name are recorded to generate the target structured text.

[0007] Optionally, evaluating the target structured text based on the evaluation dimensions to obtain evaluation results corresponding to each evaluation dimension includes: Determine the frequency percentage of the target brand logo in the target structured text to obtain the brand exposure hit rate of the target brand logo in the business theme; Based on the frequency ratio, the target text to be counted is determined, and the sample ratio of the reference link corresponding to the target brand logo in the target text to be counted is determined, so as to obtain the corresponding traceability chain reach rate. The frequency of occurrence of entities in the target structured text that are relevant to the target brand identity is determined to obtain the brand narrative weight value; Determine the proportion of text in the target text that does not contain the competitor's identifier to obtain the brand exclusivity index; Determine the correlation between the position of the target brand identifier in the target structured text and the core intent in the question statement to obtain the corresponding core intent matching degree.

[0008] Optionally, after determining the evaluation dimensions for the GEO and evaluating the target structured text based on the evaluation dimensions to obtain the evaluation results corresponding to each evaluation dimension, the method further includes: Based on the comparison of the evaluation results under the question statement, the brand gap and source gap between the target brand logo and the competitor logo are analyzed; the brand gap is determined based on the difference in brand exposure hit rate between the target brand logo and the competitor logo, and the source gap is determined based on the difference in the official website citation ratio between the target brand logo and the competitor logo in the text answer information. The data gap between the target brand identifier and the competitor identifier is determined based on the brand gap and the source gap.

[0009] Optionally, generating the optimization strategy list corresponding to the GEO based on the evaluation results includes: Based on the preset search volume and sales funnel stage of the question statement, and using a preset weighted funnel sorting algorithm, the priority score of the question gap in the question statement is determined. Based on the data gap and the priority score, a list of optimization strategies corresponding to the GEO is generated.

[0010] Secondly, this application discloses a GEO optimization strategy generation device based on simulated webpage interaction behavior, comprising: The text information extraction module is used to construct an evaluation prompt word matrix based on the business domain corresponding to the user terminal, and input the question statements in the evaluation prompt word matrix into a preset web page based on GEO, so as to simulate user inquiry behavior on the preset web page and obtain the corresponding text answer information. The text processing module is used to perform structured parsing of the text response information using a text processing model to extract the target structured text; the target structured text includes the target identifier name and the product website link corresponding to the target identifier name; the target identifier name includes the target brand identifier and competitor identifiers; The text evaluation module is used to determine the evaluation dimensions for the GEO, and evaluate the target structured text based on the evaluation dimensions to obtain the evaluation results corresponding to each evaluation dimension, so as to generate an optimization strategy list corresponding to the GEO based on the evaluation results.

[0011] Thirdly, this application discloses an electronic device, including: Memory, used to store computer programs; A processor is used to execute the computer program to implement the aforementioned method for generating GEO optimization strategies based on simulated web page interaction behavior.

[0012] Fourthly, this application discloses a computer-readable storage medium for storing a computer program, which, when executed by a processor, implements the aforementioned method for generating GEO optimization strategies based on simulated webpage interaction behavior.

[0013] As can be seen, in this embodiment, an evaluation prompt word matrix is ​​constructed based on the business domain corresponding to the user terminal. The question statements in the evaluation prompt word matrix are input into a preset web page built based on GEO to simulate user inquiry behavior and obtain corresponding text answer information. A text processing model is used to perform structured parsing of the text answer information to extract target structured text. The target structured text includes a target identifier name and a link to the product's official website corresponding to the target identifier name. The target identifier name includes a target brand identifier and competitor identifiers. Evaluation dimensions for GEO are determined, and the target structured text is evaluated based on these dimensions to obtain evaluation results for each evaluation dimension. Based on these evaluation results, an optimization strategy list corresponding to GEO is generated. In this way, by establishing a targeted evaluation prompt word matrix for the target business and innovatively using web page simulation technology to recreate the real user's questioning environment and seen content, evaluation indicators that eliminate behavioral deviations between the underlying interface and the web page front-end are obtained. Then, the evaluation metrics were processed according to the evaluation dimensions of GEO, transforming the highly unstructured natural language responses of the large model into structured metric data, which solved the pain point that brands rely on human experience to make judgments during the optimization process and cannot verify the effects on a large scale. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0015] Figure 1 This application discloses a flowchart of a method for generating GEO optimization strategies based on simulated webpage interaction behavior. Figure 2 This application discloses a specific method for generating GEO optimization strategies based on simulated webpage interaction behavior; Figure 3 This is a schematic diagram of the structure of a GEO optimization strategy generation device based on simulated web page interaction behavior disclosed in this application; Figure 4 This is a structural diagram of an electronic device disclosed in this application. Detailed Implementation

[0016] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0017] Generative AI is changing how users search and make decisions, and brands going global urgently need to quantify GEO (Generative Evaluation) effectiveness. However, there is currently a lack of objective evaluation tools, and optimization still relies on experience-driven approaches. The industry needs an evaluation technology solution that can accurately measure the performance of AI responses by reflecting the user's perspective. Therefore, this application will specifically introduce a GEO optimization strategy generation method based on simulating web-based interaction behavior, which can solve the above problems.

[0018] See Figure 1 As shown in the figure, this application discloses a method for generating GEO optimization strategies based on simulated web page interaction behavior, including: Step S11: Construct an evaluation prompt word matrix based on the business domain corresponding to the user terminal, and input the question statements in the evaluation prompt word matrix into a preset web page based on GEO, so as to simulate user inquiry behavior on the preset web page and obtain the corresponding text answer information.

[0019] In this embodiment, the step of constructing an evaluation prompt word matrix based on the business domain corresponding to the user's terminal includes: receiving a target brand identifier and a competitor identifier configured by the user; the target brand identifier includes the main brand name, variant name, and official domain name; generating a business theme based on the business domain corresponding to the target brand identifier and the competitor identifier, and expanding and configuring corresponding question statements under the business theme, and binding the question statements with the target search content or target intent weight to construct an evaluation prompt word matrix. That is, receiving the target brand identifier (including the main brand name, variant name, official domain name, etc.) and at least one competitor identifier configured by the user; generating a theme set based on the user's business domain; expanding and configuring multiple specific question statements that conform to the questioning habits of real users under each theme, and mapping and binding the question statements with the actual search volume or intent weight to form an evaluation prompt word matrix.

[0020] In this embodiment, the step of inputting the question statements from the evaluation prompt word matrix into a preset webpage built on GEO to simulate user inquiry behavior and obtain corresponding text answer information includes: scheduling a preset webpage interaction simulation engine to perform login and initialization operations on the preset webpage based on a preset monitoring frequency, and inputting the question statements from the evaluation prompt word matrix into the interactive interface of the preset webpage to simulate user inquiry behavior; crawling the natural language text answer content, reference links, query behavior, and website source of the interactive interface to generate corresponding text answer information. According to the preset monitoring frequency, the webpage interaction simulation engine (such as an automated script based on a headless browser) is scheduled. Login or session initialization operations are automatically performed on the webpage interactive interfaces of multiple target generative AI platforms. The question statements from the aforementioned steps are input into the input box of the webpage and submission is triggered to simulate the waiting and reading behavior of real users, fully capturing the rendering output content of the large model. The captured content includes not only the final generated natural language text main answer, but also the "source citation links" component displayed on the front-end page, as well as the "derived sub-query behavior" in the thinking process displayed by the AI ​​platform and the source of the websites visited.

[0021] Step S12: Use a text processing model to perform structured parsing on the text response information to extract the target structured text; the target structured text includes the target identifier name and the product website link corresponding to the target identifier name; the target identifier name includes the target brand identifier and competitor identifiers.

[0022] In this embodiment, the step of using a text processing model to perform structured parsing of the text response information to extract target structured text includes: using natural language processing technology or a lightweight information extraction model to perform entity recognition on the text response information to determine whether the text response information contains a target identifier name; if the text response information contains the target identifier name, then the target identifier name is extracted, and the location and link set corresponding to the target identifier name are recorded to generate target structured text. Using natural language processing technology or a lightweight information extraction model, entity recognition is performed on the captured text response to determine whether the text mentions the target brand or competitors, and their location and ranking are recorded. The reference link set is extracted, and domain name matching is used to determine whether it contains links pointing to the target brand's official website.

[0023] Step S13: Determine the evaluation dimensions for the GEO, and evaluate the target structured text based on the evaluation dimensions to obtain the evaluation results corresponding to each evaluation dimension, so as to generate an optimization strategy list corresponding to the GEO based on the evaluation results.

[0024] In this embodiment, after determining the evaluation dimensions for the GEO and evaluating the target structured text based on the evaluation dimensions to obtain the evaluation results corresponding to each evaluation dimension, the method further includes: comparing the brand gap and source gap between the target brand identifier and the competitor identifier under the question statement based on each evaluation result; the brand gap is determined based on the difference in brand exposure hit rate between the target brand identifier and the competitor identifier, and the source gap is determined based on the difference in the official website citation ratio between the target brand identifier and the competitor identifier in the text answer information; the data gap between the target brand identifier and the competitor identifier is determined based on the brand gap and the source gap. Specifically, regarding the calculation of brand exposure hit rate, the frequency percentage of the target brand being effectively mentioned in all relevant generative large model returned text is statistically analyzed to characterize the basic exposure effect of the brand under the target search topic. The calculation formula is: ; in, Indicates brand exposure hit rate. This indicates the number of sample responses that effectively mentioned the target brand. This represents the total number of valid answer samples obtained for a specific question. Regarding the calculation of the source tracing reach rate, among the text samples that mention the target brand, the percentage of samples that provide links to the target brand's official website or a specified domain name is further statistically analyzed to characterize the conversion effect of the generated content on the official source. The calculation formula is as follows: ; in, Indicates the contact rate of the traceability chain. This indicates the number of sample answers that contain links to the target brand's official website or a specified domain. This indicates the number of responses that effectively mentioned the target brand.

[0025] Brand narrative weighting calculation: This involves calculating the proportion of the target brand's relevant descriptions (word count or frequency) to the total description size of all entities (including all mentioned competitors) in long-text responses containing multiple entities. This percentage represents the brand's narrative position and competitive advantage among similar competitors. The calculation formula is as follows: ; in, Indicates the weight of the brand narrative. This indicates the number of words or frequency of occurrence in the description related to the target brand. This represents the number of words or frequency of occurrence in the description of the i-th mentioned competitor, where n is the total number of mentioned competitors. Regarding the calculation of contextual sentiment bias, a pre-trained sentiment analysis algorithm model is used to extract semantic features and determine the sentiment tendency of contextual statements containing the target brand, quantifying them as positive, neutral, or negative scores to characterize the effectiveness of content generation in reputation management. The calculation formula is: ; in, This represents the overall sentiment bias score, where k is the total number of contextual fragments in which the target brand is mentioned in the responses. Set the emotional polarity of the j-th segment (e.g., assign 1 for positive, 0 for neutral, and -1 for negative). This is the sentiment confidence weight output by the sentiment analysis model for this segment. For the brand exclusivity index calculation, the proportion of responses mentioning the target brand that do not simultaneously mention any preset competitors is statistically analyzed. This is used to characterize the degree of monopolistic recommendation of the target brand under a specific question intent. The calculation formula is: ; in, Indicates the brand exclusivity index. This indicates the number of responses that only mention the target brand and do not mention any pre-set competitors. This represents the number of sample answers that effectively mention the target brand. The core intent matching score is calculated by evaluating the relevance of the target brand's appearance in the generated answer content to the user's core intent (e.g., whether it's recommended as the preferred solution), thus characterizing the depth of high-quality recommendations. The calculation formula is: ; in, This represents the core intent matching degree, where M is the total number of valid recommendation solutions / locations extracted from the answer. The weight decay coefficient for the m-th recommendation position (e.g., the first recommendation). Second place (decreasing sequentially) This is a Boolean variable (if the target brand appears in the m-th position, then...). ,otherwise ).

[0026] In this embodiment, generating the optimization strategy list corresponding to the GEO based on the evaluation results includes: determining the priority score of the question gap in the question statement based on the preset search volume and sales funnel stage of the question statement, and using a preset weighted funnel ranking algorithm; generating the optimization strategy list corresponding to the GEO based on the data gap and the priority score. That is, comparing the performance of the product and competitors on multi-dimensional indicators. For each question statement, calculate the "brand gap" (the difference between the product's hit rate and the hit rate of high-ranking competitors) and the "source gap" (the difference between the proportion of references to the product's official website and the proportion of references to competitors' official websites). Combining the preset search volume, user funnel stage, and other factors of the question statement, calculate the "priority score" of each question gap using a preset weighted funnel ranking algorithm. Output an executable optimization list, prompting users with suggestions such as "gaps with high search volume but low exposure hit rate" and "key questions that are mentioned but have low sentiment bias," thereby guiding the next stage of content planning or backlink building.

[0027] As can be seen, in this embodiment, as Figure 2As shown, an evaluation prompt word matrix is ​​constructed based on the business domain corresponding to the user end. The question statements in the evaluation prompt word matrix are input into a preset web page built based on GEO to simulate user inquiry behavior and obtain corresponding text answer information. A text processing model is used to perform structured parsing of the text answer information to extract target structured text. The target structured text includes a target identifier name and a link to the product website corresponding to the target identifier name. The target identifier name includes a target brand identifier and competitor identifiers. Evaluation dimensions for GEO are determined, and the target structured text is evaluated based on these dimensions to obtain evaluation results for each evaluation dimension. Based on these evaluation results, an optimization strategy list for GEO is generated. In this way, by establishing a targeted evaluation prompt word matrix for the target business and innovatively using web page simulation technology to recreate the real user's questioning environment and seen content, evaluation indicators that eliminate behavioral deviations between the underlying interface and the web page front-end are obtained. Then, the evaluation metrics were processed according to the evaluation dimensions of GEO, transforming the highly unstructured natural language responses of the large model into structured metric data, which solved the pain point that brands rely on human experience to make judgments during the optimization process and cannot verify the effects on a large scale.

[0028] refer to Figure 3 The present application also discloses a GEO optimization strategy generation device based on simulated web page interaction behavior, comprising: The text information extraction module 11 is used to construct an evaluation prompt word matrix based on the business domain corresponding to the user terminal, and input the question statements in the evaluation prompt word matrix into a preset web page based on GEO, so as to simulate user inquiry behavior on the preset web page and obtain corresponding text answer information. The text processing module 12 is used to perform structured parsing of the text response information using a text processing model to extract target structured text; the target structured text includes a target identifier name and a link to the product website corresponding to the target identifier name; the target identifier name includes a target brand identifier and competitor identifiers; The text evaluation module 13 is used to determine the evaluation dimensions for the GEO and evaluate the target structured text based on the evaluation dimensions to obtain the evaluation results corresponding to each evaluation dimension, so as to generate an optimization strategy list corresponding to the GEO based on the evaluation results.

[0029] As can be seen, in this embodiment, a targeted evaluation prompt word matrix is ​​established for the target business, and web-based simulation technology is innovatively used to recreate the questioning environment and content seen by real users, thereby obtaining evaluation indicators that eliminate behavioral deviations between the underlying interface and the web front-end. Then, the evaluation indicators are processed for the evaluation dimensions of GEO to transform the highly unstructured natural language answers of the large model into structured metric data, solving the pain point that brands rely on human experience to judge during the optimization process and cannot verify the effect on a large scale.

[0030] In some specific embodiments, the text information extraction module 11 may specifically include: The identifier acquisition unit is used to receive the target brand identifier and competitor identifier configured by the user terminal; the target brand identifier includes the main brand name, variant name, and official domain name. The matrix construction unit is used to generate business themes based on the business areas corresponding to the target brand identifier and the competitor identifier, and to expand the configuration of corresponding question statements under the business themes, and bind the question statements with the target search content or target intent weight to construct an evaluation prompt word matrix.

[0031] In some specific embodiments, the text information extraction module 11 may specifically include: The statement input unit is used to schedule a preset web page interaction simulation engine to perform login and initialization operations on a preset web page based on a preset monitoring frequency, and input the question statements in the evaluation prompt word matrix to the interaction interface of the preset web page to simulate user inquiry behavior. The data crawling unit is used to crawl the natural language text response content, reference links, query behavior, and website sources of the interactive interface to generate corresponding text response information.

[0032] In some specific embodiments, the text processing module 12 may specifically include: The information recognition unit is used to perform entity recognition on the text response information using natural language processing technology or a lightweight information extraction model, so as to determine whether the text response information contains a target identifier name; The text generation unit is used to extract the target identifier name if the text response information contains the target identifier name, and record the position and link set corresponding to the target identifier name to generate target structured text.

[0033] In some specific embodiments, the text evaluation module 13 may specifically include: The brand exposure hit rate determination unit is used to determine the frequency ratio of the target brand logo in the target structured text, so as to obtain the brand exposure hit rate of the target brand logo in the business topic; The traceability chain reach rate determination unit is used to determine the target text to be counted based on the frequency ratio, and to determine the sample ratio of the reference link corresponding to the target brand identifier in the target text to be counted, so as to obtain the corresponding traceability chain reach rate. The brand narrative weight value determination unit is used to determine the frequency of occurrence of entities in the target structured text that are related to the target brand identity, so as to obtain the brand narrative weight value. The brand exclusivity index determination unit is used to determine the proportion of texts in the target text to be counted that do not contain the competitor's identifier, so as to obtain the brand exclusivity index. The core intent matching degree determination unit is used to determine the association between the position of the target brand identifier in the target structured text and the core intent in the question statement, so as to obtain the corresponding core intent matching degree.

[0034] In some specific embodiments, the GEO optimization strategy generation device based on simulated web page interaction behavior may further include: The first brand comparison module is used to compare the brand gap and source gap between the target brand logo and the competitor logo under the question statement based on the evaluation results. The brand gap is determined based on the difference in brand exposure hit rate between the target brand logo and the competitor logo, and the source gap is determined based on the difference in the official website citation ratio between the target brand logo and the competitor logo in the text answer information. The second brand comparison module is used to determine the data gap between the target brand identifier and the competitor identifier based on the brand gap and the source gap.

[0035] In some specific embodiments, the text evaluation module 13 may specifically include: The priority determination unit is used to determine the priority score of the question gap in the question statement based on the preset search volume and sales funnel stage of the question statement and using a preset weighted funnel sorting algorithm. The list generation unit is used to generate an optimization strategy list corresponding to the GEO based on the data gap and the priority score.

[0036] Furthermore, embodiments of this application also disclose an electronic device, Figure 4 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application.

[0037] Figure 4 This is a schematic diagram of the structure of an electronic device 20 provided in an embodiment of this application. Specifically, the electronic device 20 may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the GEO optimization strategy generation method based on simulated webpage interaction behavior disclosed in any of the foregoing embodiments. Alternatively, the electronic device 20 in this embodiment may specifically be a computer.

[0038] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0039] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0040] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the GEO optimization strategy generation method based on simulated web page interaction behavior executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs capable of performing other specific tasks.

[0041] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned disclosed method for generating GEO optimization strategies based on simulated webpage interaction behavior. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0042] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0043] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0044] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0045] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0046] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for generating GEO optimization strategies based on simulated webpage interaction behavior, characterized in that, include: An evaluation prompt word matrix is ​​constructed based on the business domain corresponding to the user terminal, and the question statements in the evaluation prompt word matrix are input into a preset web page based on GEO to simulate user inquiry behavior on the preset web page and obtain corresponding text answer information. The text response information is structured and parsed using a text processing model to extract the target structured text; The target structured text includes the target identifier name and the corresponding product website link; the target identifier name includes the target brand identifier and competitor identifiers; The evaluation dimensions for the GEO are determined, and the target structured text is evaluated based on the evaluation dimensions to obtain the evaluation results corresponding to each evaluation dimension, so as to generate an optimization strategy list for the GEO based on the evaluation results.

2. The GEO optimization strategy generation method based on simulated webpage interaction behavior according to claim 1, characterized in that, The construction of the evaluation prompt word matrix based on the corresponding business domain of the user terminal includes: Receives the target brand identifier and competitor identifier configured by the user; the target brand identifier includes the main brand name, variant name, and official domain name; Based on the business domains corresponding to the target brand identifier and the competitor identifier, a business theme is generated. Under the business theme, corresponding question statements are expanded and configured, and the question statements are bound to the target search content or target intent weight to construct an evaluation prompt word matrix.

3. The GEO optimization strategy generation method based on simulated webpage interaction behavior according to claim 1, characterized in that, The step of inputting the question statements from the evaluation prompt word matrix into a preset webpage built on GEO, to simulate user inquiry behavior on the preset webpage and obtain corresponding text answer information, includes: Based on a preset monitoring frequency, a preset webpage interaction simulation engine is scheduled to perform login and initialization operations on a preset webpage, and the question statements in the evaluation prompt word matrix are input into the interaction interface of the preset webpage to simulate user inquiry behavior. The natural language text responses, reference links, query behaviors, and website sources of the interactive interface are crawled to generate corresponding text response information.

4. The GEO optimization strategy generation method based on simulated webpage interaction behavior according to claim 1, characterized in that, The step of using a text processing model to perform structured parsing of the text response information to extract the target structured text includes: Entity recognition is performed on the text response information using natural language processing techniques or a lightweight information extraction model to determine whether the text response information contains a target identifier name; If the text response information contains the target identifier name, then the target identifier name is extracted, and the location and link set corresponding to the target identifier name are recorded to generate the target structured text.

5. The GEO optimization strategy generation method based on simulated webpage interaction behavior according to claim 2, characterized in that, The evaluation of the target structured text based on the evaluation dimensions to obtain the evaluation results corresponding to each evaluation dimension includes: Determine the frequency percentage of the target brand logo in the target structured text to obtain the brand exposure hit rate of the target brand logo in the business theme; Based on the frequency ratio, the target text to be counted is determined, and the sample ratio of the reference link corresponding to the target brand logo in the target text to be counted is determined, so as to obtain the corresponding traceability chain reach rate. The frequency of occurrence of entities in the target structured text that are relevant to the target brand identity is determined to obtain the brand narrative weight value; Determine the proportion of text in the target text that does not contain the competitor's identifier to obtain the brand exclusivity index; Determine the correlation between the position of the target brand identifier in the target structured text and the core intent in the question statement to obtain the corresponding core intent matching degree.

6. The GEO optimization strategy generation method based on simulated webpage interaction behavior according to claim 1, characterized in that, After determining the evaluation dimensions for the GEO and evaluating the target structured text based on the evaluation dimensions to obtain the evaluation results corresponding to each evaluation dimension, the method further includes: Based on the comparison of the evaluation results under the question statement, the brand gap and source gap between the target brand logo and the competitor logo are analyzed; the brand gap is determined based on the difference in brand exposure hit rate between the target brand logo and the competitor logo, and the source gap is determined based on the difference in the official website citation ratio between the target brand logo and the competitor logo in the text answer information. The data gap between the target brand identifier and the competitor identifier is determined based on the brand gap and the source gap.

7. The GEO optimization strategy generation method based on simulated webpage interaction behavior according to claim 6, characterized in that, The process of generating the optimization strategy list corresponding to the GEO based on the evaluation results includes: Based on the preset search volume and sales funnel stage of the question statement, and using a preset weighted funnel sorting algorithm, the priority score of the question gap in the question statement is determined. Based on the data gap and the priority score, a list of optimization strategies corresponding to the GEO is generated.

8. A device for generating GEO optimization strategies based on simulated webpage interaction behavior, characterized in that, include: The text information extraction module is used to construct an evaluation prompt word matrix based on the business domain corresponding to the user terminal, and input the question statements in the evaluation prompt word matrix into a preset web page based on GEO, so as to simulate user inquiry behavior on the preset web page and obtain the corresponding text answer information. The text processing module is used to perform structured parsing of the text response information using a text processing model in order to extract the target structured text; The target structured text includes the target identifier name and the corresponding product website link; the target identifier name includes the target brand identifier and competitor identifiers; The text evaluation module is used to determine the evaluation dimensions for the GEO, and evaluate the target structured text based on the evaluation dimensions to obtain the evaluation results corresponding to each evaluation dimension, so as to generate an optimization strategy list corresponding to the GEO based on the evaluation results.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor is configured to execute the computer program to implement the GEO optimization strategy generation method based on simulated web page interaction behavior as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Used to store a computer program, which, when executed by a processor, implements the GEO optimization strategy generation method based on simulated web page interaction behavior as described in any one of claims 1 to 7.

Images