Information cocoon house intelligent evaluation method based on eyeball trajectory tracking
Through the intelligent evaluation method of information cocoon based on eye trajectory tracking, and the mapping relational database is established using eye movement data, the problems of cumbersome and subjectiveness of existing advertising testing methods are solved, and efficient and objective evaluation of the effect of advertising information cocoon is achieved. Improved the accuracy and speed of evaluation results.
Patent Information
- Application Number
- CN202310754407.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-06-25
- Publication Date
- 2024-12-27
AI Technical Summary
The existing advertising testing methods lack intelligence, which leads to cumbersome, time-consuming and subjective evaluation of information cocoon effects, and is unable to effectively detect the impact of advertisements on users, resulting in high corporate costs, unsatisfactory customers, and the testing process is complex, making it difficult to achieve automation. Test.
Using the intelligent evaluation method of information cocoon based on eye trajectory tracking, by obtaining standard advertising push information and user eye movement data, establishing a mapping relational database, extracting category characteristics and key features of the advertisement, and using eye movement analysis tools and statistical analysis tools to perform data Process, establish a trend of vision changes, obtain standard evaluation results, and then conduct predictive evaluation of evaluation advertisements, reduce duplicate data collection, and improve evaluation efficiency.
It realizes objective and real-time evaluation of the cocoon effect of advertising information, improves the authenticity and accuracy of the evaluation results, reduces labor costs, simplifies the evaluation process, and improves the evaluation speed and accuracy.
Smart Images
Figure CN119200813A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to an information cocoon intelligent evaluation method based on eye trajectory tracking. Background Art
[0002] In today's era of rapid development of information technology, recommendation systems are a hot topic and technical field. Some cloud vendors also provide SaaS services for recommendation systems, such as Amazon Personalize from Amazon Web Services, to solve the pain points and difficulties of customers in quickly building recommendation systems from scratch. In our daily lives, recommendation systems can be seen everywhere, such as Amazon e-commerce shopping, iQiyi videos, Meituan takeout, Douyin short videos, Toutiao news, anchor live broadcast platforms, etc. The information cocoon describes the phenomenon that the information (news, product recommendations, ideas, etc.) received by Internet users becomes gradually monotonous under the intervention of algorithms. The result of this phenomenon is that on the one hand, we can enjoy the convenience, speed and comfort brought by the Internet, and on the other hand, it more or less limits our ability to seek knowledge, think, explore and innovate. The information cocoon problem caused by excessive exposure of the recommendation system will have a negative impact on consumers. In severe cases, it will cause consumers to have a negative impression of the APP, and even reduce consumption and uninstall the software. The current advertising evaluation method is not intelligent enough, and most of them are measured by questionnaires. It is highly subjective and requires a lot of time and manpower, which brings a high cost burden to enterprises.
[0003] Specifically, most companies currently use the testing method to test the information cocoon effect in advertising. It is divided into three stages: the first is the internal testing stage: that is, employees are asked to participate in the information cocoon effect scoring, find the more obvious defects and deficiencies in the advertising recommendation system, and make modifications; the second is the public testing stage: it is the stage form of user experiments, and the method is basically the same, that is, inviting some customers to test and ask them to answer the designed questions; the last is the continuous testing stage: that is, a certain number of advertisements are piloted, and continuous tracking is carried out to see whether the diversion of advertisements will cause the information cocoon phenomenon that makes consumers uncomfortable, and some necessary information from customers is also collected. The above advertising tests are usually unable to be implemented due to the non-cooperation of customers, and the testing process is very cumbersome and lacks the ability to intelligently provide test reports. Summary of the invention
[0004] The present invention provides an information cocoon intelligent evaluation method based on eye trajectory tracking to solve the above-mentioned technical problems, and specifically adopts the following technical solutions:
[0005] An intelligent evaluation method for information cocoon based on eye trajectory tracking, comprising:
[0006] Acquire a standard data set, where the standard data set includes different standard advertisement push information;
[0007] Obtaining a standard key area set of the standard advertisement push information;
[0008] Collecting a set of eye movement data generated when a user cluster observes the standard advertisement push information;
[0009] Establishing a mapping relationship database according to the standard key area set and the eye movement data set, wherein the mapping relationship database includes each of the standard advertising push information and the corresponding standard information cocoon evaluation result in a mapping relationship;
[0010] Acquire the advertising push information to be evaluated, extract the category characteristics and key characteristics of the standard advertising push information to be evaluated, screen the mapping relationship database according to the category characteristics and key characteristics of the advertising push information to be evaluated, and use the standard evaluation results corresponding to all the screened standard advertising push information to evaluate the advertising push information to obtain a predicted evaluation result.
[0011] Furthermore, the user cluster includes a plurality of sub-user clusters, and processing is performed on different sub-user clusters to obtain a plurality of sub-mapping relationship databases corresponding to different sub-user clusters.
[0012] Further, the prediction evaluation result includes a plurality of sub-prediction evaluation results;
[0013] Different sub-screening results are obtained by screening the advertisement push information to be evaluated with different sub-mapping relationship databases, and the advertisement push information to be evaluated is evaluated with the different sub-screening results to obtain the sub-prediction evaluation results corresponding to different sub-user clusters.
[0014] Furthermore, the user cluster is divided into a plurality of sub-user clusters according to gender.
[0015] Furthermore, the user cluster is divided into a plurality of sub-user clusters according to age groups.
[0016] Furthermore, the information cocoon intelligent evaluation method based on eye trajectory tracking also includes:
[0017] Adjust the advertising push information to be evaluated based on the evaluation results.
[0018] Furthermore, the specific method of establishing a mapping relationship database according to the standard key area set and the eye movement data set is:
[0019] Using an eye movement analysis tool to process the eye movement data set to obtain a gaze point set and a sight line change data set corresponding to the standard advertisement push information;
[0020] The standard key area set, the fixation point set and the sight line change data set corresponding to the standard advertisement push information are input into a statistical analysis tool for analysis to obtain an overlapping area set between the standard key area set and the fixation point set, and a sight line change trend set is obtained according to the sight line change data set, wherein the overlapping area set includes an overlapping area subset corresponding to each user in the user set, the sight line change trend set includes a sight line change trend corresponding to each user in the user set, and the sight line change trend is used to indicate a change trend of sight line data change for each overlapping area in the overlapping area subset when the user observes the standard advertisement push information;
[0021] Acquire a standard evaluation result corresponding to the standard advertisement push information according to the sight line change trend set;
[0022] The mapping relationship database is established.
[0023] 8. The information cocoon intelligent evaluation method based on eye trajectory tracking according to claim 7 is characterized in that:
[0024] The specific method for obtaining the prediction evaluation result is:
[0025] Obtaining the advertisement push information to be evaluated;
[0026] Extracting the category features of the advertisement push information to be evaluated, performing category screening on the mapping relationship database according to the category features, and obtaining a first set of standard advertisement push information corresponding to the category features of the advertisement push information to be evaluated, wherein the first set of standard advertisement push information includes all the standard advertisement push information corresponding to the category features;
[0027] Extracting key features of the advertisement push information to be evaluated, and performing key feature screening on the first standard advertisement push information set according to the key features, so as to obtain a second standard advertisement push information set that meets the key features in the first standard advertisement push information set;
[0028] Extract the evaluation key area of the advertising push information to be evaluated, and evaluate the advertising push information to be evaluated according to the standard evaluation result corresponding to each standard advertising push information in the second standard advertising push information set in combination with the evaluation key area to obtain the predicted evaluation result.
[0029] Furthermore, the eye movement analysis tool is DataViewer analysis software;
[0030] The statistical analysis tool is SPSS;
[0031] The eye movement data in the eye movement data set includes: fixation points, total fixation times, fixation shifts, duration of each fixation on a fixation point, and fixation order of fixation points.
[0032] Furthermore, the standard advertisement push information and the advertisement push information to be evaluated both include: any one of a static advertisement image and a dynamic advertisement video.
[0033] The benefit of the present invention lies in the information cocoon intelligent evaluation method based on eye trajectory tracking, which establishes a mapping relationship between standard advertising push information and information cocoon evaluation results through eye movement data. When evaluating advertising push information, there is no need to repeatedly obtain eye movement data. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0035] Figure 1 It is a flow chart of an information cocoon intelligent evaluation method based on eye trajectory tracking of the present invention. DETAILED DESCRIPTION
[0036] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.
[0037] like Figure 1The present application shows an information cocoon intelligent evaluation method based on eye trajectory tracking, comprising: S1: obtaining a standard data set, the standard data set including different standard advertising push information. S2: obtaining a standard key area set for standard advertising push information. S3: collecting an eye movement data set generated when a user cluster observes the standard advertising push information. S4: establishing a mapping relationship database based on the standard key area set and the eye movement data set, the mapping relationship database including each standard advertising push information in a mapping relationship and the corresponding standard information cocoon evaluation result. S5: obtaining the advertising push information to be evaluated, extracting the category features and key features of the standard advertising push information to be evaluated, screening the mapping relationship database according to the category features and key features of the advertising push information to be evaluated, and using the standard evaluation results corresponding to all the screened standard advertising push information to evaluate the advertising push information to obtain a predicted evaluation result.
[0038] The above steps are described in detail below.
[0039] Regarding step S1: obtaining a standard data set, the standard data set includes different standard advertisement push information.
[0040] Standard advertising push information includes: any one of static advertising images and dynamic advertising videos.
[0041] For step S2: obtain a standard key area set of standard advertisement push information.
[0042] For step S3: collect the eye movement data set generated when the user cluster observes the standard advertisement push information.
[0043] Specifically, an eye tracker is used to obtain an eye movement data set generated when a user cluster observes standard advertisement push information. Preferably, the eye movement data in the eye movement data set includes: fixation point, total fixation times, fixation transfer, duration of each fixation point, and fixation order of fixation points.
[0044] For step S4: a mapping relationship database is established based on the standard key area set and the eye movement data set, and the mapping relationship database includes each standard advertising push information and the corresponding standard information cocoon evaluation result in a mapping relationship.
[0045] As a preferred implementation, the specific method for establishing a mapping relationship database based on the standard key area set and the eye movement data set is:
[0046] S41: using an eye movement analysis tool to process the eye movement data set to obtain a gaze point set and a sight line change data set corresponding to the standard advertisement push information. In an embodiment of the present application, the eye movement analysis tool is DataViewer analysis software.
[0047] It is understandable that the collected eye movement data set can be preprocessed by DataViewer analysis software, and then a complete eye image and a heat map (heat map is used to obtain the trend of group attention distribution of user clusters) can be obtained. The heat map can be used to obtain the trend of each user's sight line change for standard advertising push information. The sight line change trend corresponding to each user is used to show the gaze of a single user on the overlapping area when observing the standard advertising push information. As a preferred embodiment, the gaze situation in the first 15-25 seconds can be analyzed.
[0048] S42: Input the standard key area set, fixation point set and sight line change data set corresponding to the standard advertising push information into the statistical analysis tool for analysis to obtain the overlapping area set between the standard key area set and the fixation point set, and obtain the sight line change trend set according to the sight line change data set, the overlapping area set includes the overlapping area subset corresponding to each user in the user set, and the sight line change trend set includes the sight line change trend corresponding to each user in the user set. Specifically, the sight line change trend is used to indicate the changing trend of sight line data changes for each overlapping area in the overlapping area subset when the user observes the standard advertising push information. In an embodiment of the present application, the statistical analysis tool is SPSS (Statistical Product and Service Solutions).
[0049] S43: Obtaining a standard evaluation result corresponding to the standard advertisement push information according to the sight line change trend set.
[0050] S44: Establish a mapping relationship database.
[0051] For step S5: obtain the advertising push information to be evaluated, extract the category characteristics and key characteristics of the standard advertising push information to be evaluated, filter the mapping relationship database according to the category characteristics and key characteristics of the advertising push information to be evaluated, and use the standard evaluation results corresponding to all the filtered standard advertising push information to evaluate the advertising push information to obtain a predicted evaluation result.
[0052] The advertisement push information to be evaluated includes: any one of a static advertisement image and a dynamic advertisement video.
[0053] The specific method for obtaining the prediction evaluation results is:
[0054] S51: Obtaining advertisement push information to be evaluated.
[0055] S52: extracting category features of the advertisement push information to be evaluated, performing category screening on the mapping relationship database according to the category features, and obtaining a first standard advertisement push information set corresponding to the category features of the advertisement push information to be evaluated, wherein the first standard advertisement push information set includes all standard advertisement push information corresponding to the category features.
[0056] It can be understood that the advertising push information to be evaluated is subject to category screening, specifically: category screening is performed on the mapping relationship database according to category characteristics, each advertising push information category corresponding to the category characteristics in the mapping relationship database is obtained, the advertising push information category is scored, and then the score of each advertising push information category is sorted, and the standard advertising push information corresponding to the second preset number of advertising push information categories with the highest score is selected, thereby reducing the amount of data during evaluation, and then speeding up the acquisition of predicted evaluation results and improving user experience.
[0057] S53: extracting key features of the advertisement push information to be evaluated, and screening the first standard advertisement push information set according to the key features to obtain a second standard advertisement push information set that meets the key features in the first standard advertisement push information set.
[0058] S54: Extract the evaluation key area of the advertisement push information to be evaluated, and evaluate the advertisement push information to be evaluated according to the standard evaluation result corresponding to each standard advertisement push information of the second standard advertisement push information set in combination with the evaluation key area, so as to obtain a predicted evaluation result. Specifically, the evaluation key area of the advertisement push information to be evaluated is compared with the standard key area in the screened standard advertisement push information, and the result is calculated according to the comparison of the standard evaluation result corresponding to the standard advertisement push information, so as to obtain a predicted evaluation result.
[0059] The information cocoon intelligent evaluation method based on eye trajectory tracking of the present application can objectively and real-time monitor the user's gaze point when viewing standard advertising push information by analyzing the user's eye movement data, and can obtain the real-time gaze point change trend of the user when viewing standard advertising push information, thereby improving the authenticity of the standard evaluation results and improving the accuracy of the predicted evaluation results.
[0060] In an embodiment of the present application, a standard key area set of standard advertising push information and an eye movement data set generated when a user cluster observes the standard advertising push information are associated, and the standard advertising push information is evaluated according to the data relationship obtained after the association process to obtain a standard evaluation result corresponding to the standard advertising push information, thereby realizing the combination of the eye movement data set and the standard advertising push information to obtain a standard evaluation result corresponding to the advertising push information, thereby realizing the evaluation of the effect of the advertising push information through eye movement signals, avoiding the interference of subjective factors in traditional methods. In an embodiment of the present application, when obtaining the predicted evaluation result, it is not necessary to obtain the eye movement data of the user cluster when observing the advertising push information to be evaluated again, thereby reducing labor costs and speeding up the acquisition of the evaluation results.
[0061] As a preferred implementation, the information cocoon intelligent evaluation method based on eye trajectory tracking also includes:
[0062] S6: Adjust the advertisement push information to be evaluated according to the evaluation result.
[0063] It is understandable that the relevant elements in the advertisement push information to be evaluated may be deleted, modified, and replaced according to the evaluation results.
[0064] In the embodiment of the present application, after the advertisement push information to be evaluated is adjusted according to the evaluation result, the adjusted advertisement push information to be evaluated is re-evaluated by the above method, and the above process is repeated until the evaluation result meets expectations.
[0065] As a preferred implementation, the user cluster includes multiple sub-user clusters, and multiple sub-mapping relationship databases corresponding to different sub-user clusters are obtained by processing different sub-user clusters. At the same time, the prediction evaluation result includes multiple sub-prediction evaluation results. Different sub-screening results are obtained by screening the advertisement push information to be evaluated with different sub-mapping relationship databases, and the advertisement push information to be evaluated is evaluated by different sub-screening results to obtain sub-prediction evaluation results corresponding to different sub-user clusters.
[0066] It is understandable that the aforementioned information cocoon intelligent evaluation method based on eye trajectory tracking defaults user clusters to a unified group. However, in the actual operation process, it is found that different user clusters may react differently to the same advertising push information. For example, for some advertising push information, the feedback from user clusters of different genders is different. For some advertising push information, the feedback from user clusters of different age groups is different. User clusters are divided into multiple sub-user clusters according to gender. User clusters are divided into multiple sub-user clusters according to age groups. It is understandable that user clusters can also be divided according to other division criteria, which will not be repeated here.
[0067] In order to improve the evaluation results of the information cocoon intelligent evaluation method based on eye trajectory tracking of the present invention, in this application, the user cluster is divided into several sub-user clusters as needed. In this way, for different sub-user clusters, multiple sub-mapping relationship databases corresponding to different sub-user clusters can be obtained through the aforementioned steps. Different sub-screening results are obtained by screening the advertising push information to be evaluated with different sub-mapping relationship databases, and the advertising push information to be evaluated is evaluated by different sub-screening results to obtain sub-prediction evaluation results corresponding to different sub-user clusters. In this way, more accurate evaluation results can be obtained for different sub-user clusters.
[0068] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the above embodiments do not limit the present invention in any form, and any technical solution obtained by equivalent replacement or equivalent transformation falls within the protection scope of the present invention.
Claims
1. An intelligent evaluation method for information cocoon based on eye trajectory tracking, characterized in that: Include: Acquire a standard data set, where the standard data set includes different standard advertisement push information; Obtaining a standard key area set of the standard advertisement push information; Collecting a set of eye movement data generated when a user cluster observes the standard advertisement push information; Establishing a mapping relationship database according to the standard key area set and the eye movement data set, wherein the mapping relationship database includes each of the standard advertising push information and the corresponding standard information cocoon evaluation result in a mapping relationship; Acquire the advertising push information to be evaluated, extract the category characteristics and key characteristics of the standard advertising push information to be evaluated, screen the mapping relationship database according to the category characteristics and key characteristics of the advertising push information to be evaluated, and use the standard evaluation results corresponding to all the screened standard advertising push information to evaluate the advertising push information to obtain a predicted evaluation result.
2. The information cocoon intelligent evaluation method based on eye trajectory tracking according to claim 1 is characterized in that: The user cluster includes a plurality of sub-user clusters, and processing is performed on different sub-user clusters to obtain a plurality of sub-mapping relationship databases corresponding to different sub-user clusters.
3. The information cocoon intelligent evaluation method based on eye trajectory tracking according to claim 2 is characterized in that: The prediction evaluation result includes a plurality of sub-prediction evaluation results; Different sub-screening results are obtained by screening the advertisement push information to be evaluated with different sub-mapping relationship databases, and the advertisement push information to be evaluated is evaluated with the different sub-screening results to obtain the sub-prediction evaluation results corresponding to different sub-user clusters.
4. The information cocoon intelligent evaluation method based on eye trajectory tracking according to claim 2 is characterized in that: The user cluster is divided into a plurality of sub-user clusters according to gender.
5. The information cocoon intelligent evaluation method based on eye trajectory tracking according to claim 2 is characterized in that: The user cluster is divided into a plurality of sub-user clusters according to age groups.
6. The information cocoon intelligent evaluation method based on eye trajectory tracking according to claim 1 is characterized in that: The information cocoon intelligent evaluation method based on eye trajectory tracking also includes: Adjust the advertising push information to be evaluated based on the evaluation results.
7. The information cocoon intelligent evaluation method based on eye trajectory tracking according to claim 1 is characterized in that: The specific method of establishing a mapping relationship database according to the standard key area set and the eye movement data set is: Using an eye movement analysis tool to process the eye movement data set to obtain a gaze point set and a sight line change data set corresponding to the standard advertisement push information; The standard key area set, the fixation point set and the sight line change data set corresponding to the standard advertisement push information are input into a statistical analysis tool for analysis to obtain an overlapping area set between the standard key area set and the fixation point set, and a sight line change trend set is obtained according to the sight line change data set, wherein the overlapping area set includes an overlapping area subset corresponding to each user in the user set, the sight line change trend set includes a sight line change trend corresponding to each user in the user set, and the sight line change trend is used to indicate a change trend of sight line data change for each overlapping area in the overlapping area subset when the user observes the standard advertisement push information; Acquire a standard evaluation result corresponding to the standard advertisement push information according to the sight line change trend set; The mapping relationship database is established.
8. The information cocoon intelligent evaluation method based on eye trajectory tracking according to claim 7 is characterized in that: The specific method for obtaining the prediction evaluation result is: Obtaining the advertisement push information to be evaluated; Extracting the category feature of the advertisement push information to be evaluated, performing category screening on the mapping relationship database according to the category feature, and obtaining a first standard advertisement push information set corresponding to the category feature of the advertisement push information to be evaluated, wherein the first standard advertisement push information set includes all the standard advertisement push information corresponding to the category feature; Extracting key features of the advertisement push information to be evaluated, and performing key feature screening on the first standard advertisement push information set according to the key features, so as to obtain a second standard advertisement push information set that meets the key features in the first standard advertisement push information set; Extract the evaluation key area of the advertising push information to be evaluated, and evaluate the advertising push information to be evaluated according to the standard evaluation result corresponding to each standard advertising push information in the second standard advertising push information set in combination with the evaluation key area to obtain the predicted evaluation result.
9. The information cocoon intelligent evaluation method based on eye trajectory tracking according to claim 7 is characterized in that: The eye movement analysis tool is DataViewer analysis software; The statistical analysis tool is SPSS; The eye movement data in the eye movement data set includes: fixation points, total fixation times, fixation shifts, duration of each fixation on a fixation point, and fixation order of fixation points.
10. The information cocoon intelligent evaluation method based on eye trajectory tracking according to claim 1 is characterized in that: The standard advertisement push information and the advertisement push information to be evaluated both include: any one of a static advertisement image and a dynamic advertisement video.