Method, device and system for analyzing correlation of attention link based on user search, travel, consumption and arrival destination

Abstract

This invention discloses a method, apparatus, and system for analyzing the correlation between user retrieval, travel, consumption, and destination attention links, belonging to the field of consumer behavior prediction. The method includes: structuring and preprocessing legally and compliantly collected user data; identifying cognitive dwell nodes in the structured and preprocessed data to construct an attention link model module; constructing a cognitive behavior evolution prediction and path interpretability model module based on the attention link model module; and finally, dividing the attention link model module and the cognitive behavior evolution prediction and path interpretability model module into independent modules, integrating them through a unified data interface and communication protocol. During cross-scenario deployment and model migration, adjustments are made according to the requirements of new scenarios. This invention enables efficient deployment and rapid migration across different application scenarios, effectively improving the overall system's engineering adaptability, scalability, and commercialization efficiency.

Description

Technical Field

[0001] This invention relates to the field of consumer behavior prediction, and more specifically, to a method, apparatus, and system based on the analysis of the correlation between user search, travel, consumption, and destination attention links. Background Technology

[0002] Existing algorithm systems for user behavior analysis, cognitive modeling, and other scenarios typically suffer from the following problems: (1) High coupling between scenarios and high migration costs. Existing solutions are often customized for specific application scenarios (such as e-commerce recommendation and content distribution), and the algorithm model is highly bound to the scenario data features and business logic. When the same algorithm capabilities need to be migrated to different scenarios (such as from online shopping to offline retail, or from video recommendation to information distribution), it is often necessary to redesign the model structure, adjust the parameter configuration, or even reconstruct the code framework, resulting in long development cycles, low deployment efficiency, and difficulty in meeting the business needs of rapid iteration.

[0003] (2) Poor system scalability and low reusability. In the existing solution, the modules are tightly coupled, making it difficult to upgrade or replace them independently. When business requirements change (such as adding data sources, adjusting model algorithms, or adapting to different hardware environments), a complete reconstruction is often required. It is impossible to achieve a quick response through local adjustments, which leads to difficulties in the accumulation and reuse of technical assets.

[0004] (3) Insufficient engineering adaptability limits the efficiency of commercialization. Existing technologies lack a unified and configurable deployment framework, making it difficult to achieve seamless migration across different hardware platforms, operating systems, or cloud environments. This affects the efficiency of technology solutions being transformed from R&D to the production environment and restricts large-scale commercial applications. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method, device and system based on the analysis of the relationship between user retrieval, travel, consumption and destination attention links. Through module decoupling, parameter remapping, strategy template switching and dynamic adaptation mechanisms, it can achieve efficient deployment and rapid migration across different application scenarios, effectively improving the overall system's engineering adaptability, scalability and commercialization efficiency.

[0006] The objective of this invention is achieved through the following solution: A method based on the analysis of the relationship between user retrieval, travel, consumption, and destination attention links includes the following steps: The legally and compliantly collected user data is structured and preprocessed; the user data includes data on searches, travel, consumption, and arrival at destinations from online platforms. Identify cognitive dwelling nodes in the structured and preprocessed data, and construct an attention link model module; Based on the attention link model module, a cognitive behavior evolution prediction and path interpretability model module is constructed; Finally, the attention link model module and the cognitive behavior evolution prediction and path interpretability model module are divided into their own independent modules, and then integrated through a unified data interface and communication protocol. When deploying across scenarios and migrating models, adjustments are made according to the requirements of new scenarios.

[0007] Furthermore, the process of cross-scenario deployment and model migration, adjusted according to the requirements of the new scenario, specifically includes the following sub-steps: To address the differences in behavioral feature distribution and cognitive evolution paths across different scenarios, the following parameter remapping mechanism is provided: By initializing or fine-tuning the parameters of the feature encoding layer and state evolution layer in the original model, it adapts to the new input distribution and behavioral semantics; the transfer is performed by freezing some shared parameters and only fine-tuning key sub-modules. The parameter transfer is expressed as follows: ; Where, θ base The original model parameters, D new For new scenario data, M remap This indicates remapping and fine-tuning operations.

[0008] Furthermore, the remapping and fine-tuning operation includes retraining.

[0009] Furthermore, the remapping and fine-tuning operations include incremental adaptation.

[0010] Furthermore, the process of cross-scenario deployment and model migration, adjusted according to the requirements of the new scenario, specifically includes the following sub-steps: To address the characteristics of cognitive tasks in different scenarios, multiple strategy templates are pre-set. During deployment, the appropriate template is selected or combined according to the scenario to adjust the inference objective and strategy generation logic. The strategy templates are as follows: ; Among them, T k This represents the k-th strategy template.

[0011] Furthermore, after making adjustments according to the requirements of the new scenario, the process also includes the following steps: By combining real-time collected behavioral data from new scenarios, and employing online learning or incremental fine-tuning strategies, we continuously optimize the cognitive link modeling and path inference performance of the cognitive behavior evolution prediction and path interpretability model module.

[0012] An apparatus for analyzing the correlation between user retrieval, travel, consumption and destination attention links includes a processor and a memory, wherein the memory stores a computer program that, when loaded by the processor, executes the method described in any of the preceding descriptions.

[0013] A system based on the analysis of the relationship between user search, travel, consumption and destination attention links includes the apparatus described above.

[0014] The beneficial effects of this invention include: The method of this invention improves adaptability and migration efficiency in different business scenarios, cognitive domains or application systems. Based on the cross-scenario deployment and model migration mechanism of module decoupling, parameter remapping and strategy template switching, and by designing transferable module and scenario adaptation strategies, it can quickly respond to the cognitive modeling and path inference needs in new scenarios while maintaining the stability of the core link modeling framework, reduce deployment and expansion costs, and improve the dynamic adaptability and long-term migration effect of the model. Attached Figure Description

[0015] 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 some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating the steps of the method in an embodiment of the present invention. Detailed Implementation

[0017] All features disclosed in all embodiments of this specification, or steps in all methods or processes implied in the disclosure, may be combined and / or extended or replaced in any way, except for mutually exclusive features and / or steps.

[0018] In a preferred embodiment, the present invention specifically proposes a method based on the analysis of the correlation between user retrieval, travel, consumption, and destination attention links, including the following steps: The legally and compliantly collected user data is structured and preprocessed; the user data includes data on searches, travel, consumption, and arrival at destinations from online platforms, and this data complies with relevant laws and regulations.

[0019] The structured and preprocessed data is used to identify cognitive residency nodes and construct an attention link model module. Specifically, the technical solution in patent application CN121438395A can be used to process legitimate and compliant data sources to obtain the corresponding attention link model. It should be noted that the construction of the attention link model in this embodiment is not limited to this method; attention link models obtained through variations or other substantially equivalent methods can also be applied to subsequent steps.

[0020] A cognitive behavior evolution prediction and path interpretability model module is constructed based on the attention link model module. Specifically, any cognitive behavior evolution prediction and path interpretability model based on the above attention link model can be applied to the subsequent steps, including but not limited to the technical solution in the patent application with publication number CN121434398A.

[0021] Finally, the attention link model module and the cognitive behavior evolution prediction and path interpretability model module are divided into independent modules, and then integrated through a unified data interface and communication protocol. When deploying and migrating models across different scenarios, adjustments are made according to the requirements of the new scenario. Specifically, the cross-scenario deployment and model migration mechanism includes the following sub-steps: To improve the adaptability and migration efficiency of the method of this invention in different business scenarios, cognitive domains, or application systems, this step provides a cross-scenario deployment and model migration mechanism based on module decoupling, parameter remapping, and strategy template switching. By rationally designing transferable basic units and scenario adaptation strategies, this invention can maintain the stability of the core link modeling framework while quickly responding to the cognitive modeling and path inference needs in new scenarios, reducing deployment and expansion costs.

[0022] (1) Module decoupling and interface standardization The overall cognitive link modeling system is divided into several functionally independent, standardized sub-modules (such as behavior chain construction units, cognitive state evolution prediction units, path inference units, and policy generators). Each module is integrated through a unified data interface and communication protocol. When deployed across scenarios, some modules can be replaced or adjusted according to the needs of new scenarios without overall reconstruction, improving the system's flexibility and reusability.

[0023] (2) Parameter remapping and fine-tuning mechanism To address the differences in behavioral feature distribution and cognitive evolution paths across different scenarios, a parameter remapping mechanism is provided: By initializing or fine-tuning the parameters of the feature encoding layer and state evolution layer in the original model, it can adapt to the new input distribution and behavioral semantics. Efficient migration can be achieved by freezing some shared parameters and only fine-tuning key sub-modules (such as the state continuity predictor and path inferrer). Parameter migration can be formally expressed as: ; Where, θ base The original model parameters, D new For new scenario data, M remap This indicates remapping and fine-tuning operations (such as retraining, incremental adaptation, etc.).

[0024] (3) Strategy template switching and scenario adaptation For different cognitive task characteristics in various scenarios (such as travel, consumption, and healthcare), multiple strategy templates are pre-set (such as goal preference-oriented, path optimization-oriented, and motivation evolution-oriented). During deployment, the appropriate template is selected or combined according to the scenario to quickly adjust the inference goal and strategy generation logic. A strategy template example is shown below: ; Among them, T k This represents a specific strategy template, such as prioritizing goal-oriented paths or encouraging exploratory paths.

[0025] (4) Dynamic adaptation and online fine-tuning After cross-scenario deployment, the model can be continuously optimized for cognitive link modeling and path inference performance by combining real-time collected behavioral data from new scenarios and adopting online learning or small-step fine-tuning strategies, thereby further improving the model's dynamic adaptability and long-term transfer effect.

[0026] In summary, through module decoupling, parameter remapping, strategy template switching, and dynamic adaptation mechanisms, this invention can achieve efficient deployment and rapid migration across different application scenarios, effectively improving the overall system's engineering adaptability, scalability, and commercialization efficiency.

[0027] It should be noted that, within the scope of protection defined in the claims of this invention, the following embodiments can be combined and / or extended or replaced in any logical manner from the above specific embodiments, such as the disclosed technical principles, disclosed technical features or implicitly disclosed technical features.

[0028] Example 1 A method based on the analysis of the relationship between user retrieval, travel, consumption, and destination attention links includes the following steps: The legally and compliantly collected user data is structured and preprocessed; the user data includes data on searches, travel, consumption, and arrival at destinations from online platforms. Identify cognitive dwelling nodes in the structured and preprocessed data, and construct an attention link model module; Based on the attention link model module, a cognitive behavior evolution prediction and path interpretability model module is constructed; Finally, the attention link model module and the cognitive behavior evolution prediction and path interpretability model module are divided into their own independent modules, and then integrated through a unified data interface and communication protocol. When deploying across scenarios and migrating models, adjustments are made according to the requirements of new scenarios.

[0029] Example 2 Based on Example 1, the adjustment to cross-scenario deployment and model migration according to the requirements of the new scenario specifically includes the following sub-steps: To address the differences in behavioral feature distribution and cognitive evolution paths across different scenarios, the following parameter remapping mechanism is provided: By initializing or fine-tuning the parameters of the feature encoding layer and state evolution layer in the original model, it adapts to the new input distribution and behavioral semantics; the transfer is performed by freezing some shared parameters and only fine-tuning key sub-modules. The parameter transfer is expressed as follows: ; Where, θ base The original model parameters, D new For new scenario data, M remap This indicates remapping and fine-tuning operations.

[0030] Example 3 Based on Example 2, the remapping and fine-tuning operation includes retraining.

[0031] Example 4 Based on Example 2, the remapping and fine-tuning operation includes incremental adaptation.

[0032] Example 5 Based on Example 1, the adjustment to cross-scenario deployment and model migration according to the requirements of the new scenario specifically includes the following sub-steps: To address the characteristics of cognitive tasks in different scenarios, multiple strategy templates are pre-set. During deployment, the appropriate template is selected or combined according to the scenario to adjust the inference objective and strategy generation logic. The strategy templates are as follows: ; Among them, T k This represents the k-th strategy template.

[0033] Example 6 Based on Example 1, after making adjustments according to the new scenario requirements, the following steps are also included: By combining real-time collected behavioral data from new scenarios, and employing online learning or incremental fine-tuning strategies, we continuously optimize the cognitive link modeling and path inference performance of the cognitive behavior evolution prediction and path interpretability model module.

[0034] Example 7 An apparatus for analyzing the correlation between user retrieval, travel, consumption and destination attention links includes a processor and a memory, wherein the memory stores a computer program that, when loaded by the processor, executes the method described in any one of Embodiments 1 to 6.

[0035] Example 8 A system based on the analysis of the relationship between user retrieval, travel, consumption and arrival at the destination attention link includes the device described in Example 7.

[0036] The units described in the embodiments of the present invention can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0037] According to one aspect of the present invention, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the methods provided in the various optional implementations described above.

[0038] In another aspect, embodiments of the present invention also provide a computer-readable medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the methods described in the above embodiments.

Claims

1. A method based on the analysis of the correlation between user retrieval, travel, consumption, and destination attention links, characterized in that, Includes the following steps: The legally and compliantly collected user data is structured and preprocessed; the user data includes data on searches, travel, consumption, and arrival at destinations from online platforms. Identify cognitive dwelling nodes in the structured and preprocessed data, and construct an attention link model module; Based on the attention link model module, a cognitive behavior evolution prediction and path interpretability model module is constructed; Finally, the attention link model module and the cognitive behavior evolution prediction and path interpretability model module are divided into their own independent modules, and then integrated through a unified data interface and communication protocol. When deploying across scenarios and migrating models, adjustments are made according to the requirements of new scenarios.

2. The method based on the analysis of the relationship between user retrieval, travel, consumption, and destination attention links as described in claim 1, is characterized in that, When performing cross-scenario deployment and model migration, adjustments are made according to the requirements of the new scenario, specifically including the following sub-steps: To address the differences in behavioral feature distribution and cognitive evolution paths across different scenarios, the following parameter remapping mechanism is provided: By initializing or fine-tuning the parameters of the feature encoding layer and state evolution layer in the original model, it adapts to the new input distribution and behavioral semantics; the transfer is performed by freezing some shared parameters and only fine-tuning key sub-modules. The parameter transfer is expressed as follows: ； Where, θ base The original model parameters, D new For new scenario data, M remap This indicates remapping and fine-tuning operations.

3. The method based on the analysis of the relationship between user retrieval, travel, consumption, and destination attention links as described in claim 2, is characterized in that... The remapping and fine-tuning operations include retraining.

4. The method based on the analysis of the relationship between user retrieval, travel, consumption, and destination attention links as described in claim 2, is characterized in that, The remapping and fine-tuning operations include incremental adaptation.

5. The method based on the analysis of the relationship between user retrieval, travel, consumption, and destination attention links as described in claim 1, characterized in that, When performing cross-scenario deployment and model migration, adjustments are made according to the requirements of the new scenario, specifically including the following sub-steps: To address the characteristics of cognitive tasks in different scenarios, multiple strategy templates are pre-set. During deployment, the appropriate template is selected or combined according to the scenario to adjust the inference objective and strategy generation logic. The strategy templates are as follows: ； Among them, T k This represents the k-th strategy template.

6. The method based on the analysis of the relationship between user retrieval, travel, consumption, and destination attention links according to claim 1, characterized in that, After making adjustments according to the new scenario requirements, the following steps are also included: By combining real-time collected behavioral data from new scenarios, and employing online learning or incremental fine-tuning strategies, we continuously optimize the cognitive link modeling and path inference performance of the cognitive behavior evolution prediction and path interpretability model module.

7. A device for analyzing the correlation between user retrieval, travel, consumption, and destination attention links, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program that, when loaded by the processor, executes the method as described in any one of claims 1 to 6.

8. A system based on the analysis of the correlation between user retrieval, travel, consumption, and destination attention links, characterized in that, The apparatus includes the device described in claim 7, which analyzes the relationship between user retrieval, travel, consumption, and destination attention links.

Images