Digital scene design and compiling drawing system based on AI features

Abstract

The invention relates to the technical field of AI, in particular to a digital scene design and compiling drawing system based on AI features, which comprises a feature extraction and user portrait construction module, a tourism scene planning module, a user experience feedback module and an intelligent evaluation and optimization adjustment module. A basis is provided for accurate recommendation through user data feature extraction and user portrait construction; the scene recommendation list is generated by calculating the user similarity and establishing the similarity matrix, and the requirement of the user for obtaining recommendation information is met; a satisfaction degree analysis result is obtained by analyzing the satisfaction degree of the user on the generated scene recommendation list, and a direction is provided for subsequent optimization and adjustment; by identifying the problems and improvement points existing in the scene recommendation list and constructing the optimization target function optimization parameters, intelligent optimization is realized, and the user experience is improved. The method and the device are used for solving the technical problem that tourism scene recommendation in current virtual scene management is inaccurate.

Description

Technical Field

[0001] The present invention relates to the field of AI technology, and in particular to a digital scene design, compilation and rendering system based on AI features. Background Art

[0002] AI technology refers to the technology of simulating, extending and expanding human intelligence through computer systems. It involves multiple disciplines, such as computer science, biology, psychology, linguistics, mathematics and engineering. Its essence is to create a system that can perceive, understand, think, judge and make decisions like humans.

[0003] With the continuous advancement of technology and the continuous expansion of application fields, the tourism field has also made new progress. However, in virtual scene management, the following problems still exist: user data analysis and processing are not comprehensive. When analyzing and processing user data, traditional methods fail to make full use of multi-dimensional user characteristics (such as the frequency of on-site visits to scenic spots, click behavior, search behavior, purchase behavior, and social behavior, etc.), resulting in inaccurate user portrait construction, which in turn affects the accuracy of scene recommendation; the scene recommendation algorithm is not smart enough. The existing virtual scene recommendation algorithm may lack intelligence and adaptability, and cannot be dynamically adjusted and optimized according to user feedback and satisfaction analysis results; the quality of virtual scene experience is uneven. In some virtual scene applications, users may encounter problems such as the scenery is not beautiful enough, the interaction is not smooth enough, the cultural experience is insufficient, or the navigation is not convenient enough. It is impossible to automatically adjust the recommendation strategy to adapt to the changes of users, affecting the overall user experience. To this end, the present invention proposes a digital scene design and compilation drawing system based on AI features. Summary of the invention

[0004] The purpose of the present invention is to solve the problems in the background technology and to propose a digital scene design and compilation rendering system based on AI features.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions: A digital scene design and compilation drawing system based on AI features, including: feature extraction and user portrait construction module, tourism scene planning module, user experience feedback module and intelligent evaluation and optimization adjustment module; The feature extraction and user portrait construction module is used to collect user data, preprocess the user data, extract features from the preprocessed data, and construct user portraits; The tourism scene planning module is used to set feature vectors for the basic scenes of the tourist destination, calculate user similarity using the cosine similarity formula, and establish a similarity matrix based on user similarity to generate a scene recommendation list; The user experience feedback module is used to collect user experience feedback, analyze the user's satisfaction with the generated scene recommendation list, and obtain satisfaction analysis results; The intelligent evaluation and optimization adjustment module is used to analyze the results of satisfaction, identify the problems and improvement points of the scene recommendation list, and construct an optimization objective function to optimize the parameters of the scene recommendation list.

[0006] It should be noted that the application object of a digital scene design and compilation and drawing system based on AI features in an embodiment of the present invention can be virtual scene management in the tourism field. Specifically, it can be for comprehensive analysis and data processing of user behavior data through artificial intelligence and data analysis technology to achieve accurate generation and optimization of tourism scene recommendation lists. Through feature extraction and user portrait construction, tourism scene planning, user experience feedback, intelligent evaluation and optimization adjustment and other links, it can improve the accuracy of recommendations and user satisfaction, meet the user's needs for obtaining recommendation information, and automatically adjust the parameters of the recommendation list to achieve intelligent optimization and improve the performance of the recommendation system and user experience; at the same time, the system also promotes the digital development of the tourism industry, which is of great significance to improving the quality of tourism services, user experience and the overall competitiveness of the tourism industry.

[0007] Furthermore, the feature extraction and user portrait construction module collects user data, preprocesses the user data, and extracts features from the preprocessed data, including: Collect user data from multiple channels, including historical travel records and online behavior data; Preprocess the collected user data, including missing value processing, outlier detection, and data normalization; Obtain a preprocessed user data set, namely, historical travel records and online behavior data; wherein the preprocessed user data set contains multiple features of multiple users, specifically, extracting the frequency features of users' on-site visits to scenic spots from historical travel records, and extracting click behavior features (user click count, click rate, click depth), search behavior features (user search keyword frequency, search category preference), purchase behavior features (user souvenir purchase frequency, purchase category distribution) and social behavior features (user frequency of sharing travel photos on social media) from online behavior data; Let the user data set exist Each user has five user characteristics, namely, the user's frequency of visiting attractions, click behavior characteristics, search behavior characteristics, purchase behavior characteristics, and social behavior characteristics; Then the feature matrix , where the feature matrix is a A matrix where each row represents a user and each column represents a feature; Calculate the covariance matrix using the formula , where the formula is:

[0008] In the formula, Represents the feature matrix The transpose of Covariance matrix Perform eigenvalue decomposition and obtain the eigenvalue (in descending order) and the corresponding eigenvectors ; Select the front according to the size of the eigenvalue The selection of principal components is based on the eigenvalue greater than the preset threshold or the cumulative contribution rate. The preset threshold means that the variance explained by the principal component is greater than the random noise, and the cumulative contribution rate means that the previous The proportion of the variance explained by the principal components to the total variance, and represents the principal component index; Before selection The eigenvectors form the projection matrix , the feature matrix Projecting onto the principal component, we get the reduced-dimensional data matrix ,Right now , where Represents the projection matrix; each row Indicates Users in The score on the principal component is the score on the first principal component in the feature space. The feature vector of each user is projected onto the The coordinates obtained after the subspace defined by the principal components represent the user's position in the new coordinate system formed by the principal components.

[0009] Furthermore, the process of constructing a user portrait in the feature extraction and user portrait construction module includes: Using the K-means clustering algorithm, users are divided into different groups with similar characteristics based on the extracted user features, that is, the portraits of different users; the specific analysis formula is as follows:

[0010] In the formula, represents the clustering target, represents clustering partition, Indicates The set of samples in the cluster, that is, all samples assigned to the The set of user feature vectors of clusters, Indicates The mean vector of the clusters, Indicates the first The feature vector of a user, represents the feature vector index of the user, represents the number of clusters or groups, Represents a clustering index; By dividing user data into groups with similar characteristics (i.e. Mean clustering), and calculate and analyze the characteristics of each group to build a user portrait.

[0011] Furthermore, the tourism scene planning module sets a feature vector for the basic scene of the tourism destination, and the process of calculating the user similarity using the cosine similarity formula includes: Based on the user portrait constructed by the feature extraction and user portrait construction module, multiple basic scenarios of tourist destinations are defined, and feature vectors are set for each basic scenario. Each basic scenario is represented as a multidimensional vector, namely , where Represents the basic scene In the The value of the feature, Represents the basic scene feature value index; It is understandable that the feature vector of the basic scene includes multiple dimensions such as the theme, location, type, and activity content of the basic scene; the basic scene includes natural scenery sightseeing area, historical and cultural site exploration area, leisure and entertainment experience area, etc. The specific characteristics of each scene, such as the natural scenery sightseeing area may include natural elements such as mountains, rivers, lakes, and vegetation coverage as part of the feature vector of the basic scene, while the historical and cultural site exploration area may have ancient buildings, cultural relics, etc. as its main feature vector; Select two users at random and mark them as users and users , through the user and users The cosine value of the rating vector is used to calculate the user similarity :

[0012] In the formula, Respectively represent users and users The rating vector of Based on the similarity between users, establish a similarity matrix ; where the matrix is ​​a The matrix of Represents the number of users, similarity matrix Each row and column of represents a user, and the elements in the matrix represent the similarity between users.

[0013] Furthermore, the travel scenario planning module establishes a similarity matrix according to user similarity, so that the process of generating a scenario recommendation list includes: For target users , according to the users with higher similarity in the similarity matrix (i.e. neighbor users ) based on the basic scene preferences, generating a scene recommendation list for the target user; wherein the target user represents any user currently being processed by the recommendation algorithm to generate a scene recommendation list, and can be any user in the user data set, including but not limited to user and users ; Use target users Neighbor users The weighted average of the ratings of For basic scenes Rating :

[0014] In the formula, Indicates the target user The neighbor set of Indicates neighbor users For basic scenes Ratings; Combined with the predicted target user's basic scenario scores, a scenario recommendation list is generated.

[0015] Furthermore, the user experience feedback module collects user experience feedback and analyzes user satisfaction with the generated scene recommendation list. The process of obtaining the satisfaction analysis result includes: Define a user experience dataset, which contains each user Satisfaction rating for the scene recommendation list ,and , the larger the value, the higher the satisfaction; Count the satisfaction scores of all users and obtain the mean and standard deviation of the user satisfaction scores: ,

[0016] In the formula, represents the average value of user satisfaction ratings, Indicates the number of users, Represents the user index, It represents the standard deviation of user satisfaction ratings, which is used to measure the degree of deviation between user satisfaction ratings and the average rating. When the standard deviation of satisfaction ratings is higher, the degree of deviation between user satisfaction ratings and their average value is greater. When the standard deviation of satisfaction ratings is lower, the degree of deviation between user satisfaction ratings and their average value is smaller. Combined with the mean and standard deviation of user satisfaction ratings, the skewness of user satisfaction ratings is calculated and analyzed:

[0017] In the formula, Indicates the skewness value of the user satisfaction rating; When analyzing the skewness of the user satisfaction score, if the skewness value of the user satisfaction score is a positive number, it means that the data of the user satisfaction score is right-skewed, that is, the user satisfaction score has a large number of high scores; if the skewness value of the user satisfaction score is a negative number, it means that the data of the user satisfaction score is left-skewed, that is, the user satisfaction score has a large number of low scores; set the skewness limit range, and compare the skewness value of the user satisfaction score with the skewness limit range. When the skewness value of the user satisfaction score is not within the skewness limit range, it means that the user satisfaction score is unstable and has an extreme skewness; when the skewness value of the user satisfaction score is within the skewness limit range, it means that the user satisfaction score is stable; Preset satisfaction threshold , used to judge whether the overall satisfaction meets the standard; Combine the preset satisfaction threshold to analyze user satisfaction, so as to obtain the satisfaction analysis results; if the average value of user satisfaction score Above the preset satisfaction threshold , accompanied by a low standard deviation, indicating that the satisfaction distribution is uniform, the satisfaction level is generally high, and the skewness is within the skewness limit, then the overall satisfaction is judged to have reached the standard and a basic feedback signal is generated; conversely, if any of the following situations occurs, it can be judged that the overall satisfaction has not reached the established standard and an important feedback signal is generated: the average value of the user satisfaction score Not higher than the preset satisfaction threshold ; or although the average Above the preset satisfaction threshold , but the standard deviation is high, indicating a large difference in satisfaction; or the skewness is beyond the normal range; The satisfaction analysis results including basic feedback signals and important feedback signals are sent to the intelligent evaluation and optimization adjustment module.

[0018] Furthermore, the intelligent evaluation and optimization adjustment module identifies the problems and improvement points of the scene recommendation list based on the satisfaction analysis results, and constructs an optimization objective function. The process of optimizing the parameters of the scene recommendation list includes: According to the basic feedback signal of the satisfaction analysis result, a scene recommendation list corresponding to the feedback signal is extracted to generate a high-quality recommendation template; For the important feedback signals in the satisfaction analysis results, the satisfaction scores corresponding to the important feedback signals are compared with the expected satisfaction scores, where the expected satisfaction scores are the expected values ​​obtained through expert scoring, industry standards, etc., to identify problems in the scene recommendation list, such as generally low satisfaction with a certain indicator in the scene recommendation list and large differences in user opinions. Based on the existing problems in the scene recommendation list, determine the improvement points that need to be optimized in the scene recommendation list, including the beauty of the scenery, smooth interaction, cultural experience, and navigation convenience of the scene recommendation list; It can be understood that for the beauty of the scenery, the details and color richness of the scenery in the virtual scene are increased; for the smoothness of interaction, the interaction logic and response time in the virtual scene are optimized to reduce the differences between users; for the cultural experience, the local cultural characteristics are deeply explored and integrated into the virtual scene; for the convenience of navigation, the navigation interface and path planning algorithm in the virtual scene are optimized; The improvement points that need to be optimized are used as weight parameters of the scene recommendation list, and the weight parameters of the scene recommendation list are marked as ,in, They correspond to the weight values ​​of beautiful scenery, smooth interaction, cultural experience and navigation convenience, and the sum of the weight parameters is 1; Construct the optimization objective function:

[0019] In the formula, represents the optimization objective function value, Indicates The predicted value of satisfaction of a user under the weight parameter, Indicates User satisfaction rating; Use the optimized weight parameters , adjust the scene recommendation list.

[0020] Compared with the existing technology, the present invention provides a digital scene design and compilation rendering system and method based on AI features, which have the following advantages: 1. The present invention collects user data, pre-processes the user data, extracts features from the pre-processed data, and constructs user portraits to gain an in-depth understanding of user needs and preferences, thereby meeting the needs of different user groups and improving user experience; 2. The present invention sets feature vectors for basic scenes of tourist destinations, calculates user similarity using the cosine similarity formula, and establishes a similarity matrix based on user similarity, thereby generating a scene recommendation list, which can accurately match user interests with tourist scenes, thereby improving the accuracy of recommendations and user satisfaction; 3. The present invention collects user experience feedback, analyzes the user's satisfaction with the generated scene recommendation list, and obtains satisfaction analysis results, which is conducive to subsequent optimization and adjustment; by identifying the problems and improvement points of the scene recommendation list based on the satisfaction analysis results, and constructing an optimization objective function to optimize the parameters of the scene recommendation list, the digital performance of the system and the tourism scene recommendation effect are improved.

[0021] In summary, the present invention significantly improves the efficiency of personalized tourism scene recommendation through the advantages of feature extraction and user portrait construction, tourism scene planning, user experience feedback, intelligent evaluation and optimization adjustment, and ensures the efficient and stable operation of a subsequent digital scene design and compilation and drawing system based on AI features. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a module diagram of a digital scene design and compilation and drawing system based on AI features proposed by the present invention. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the implementation regulations described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] Reference Figure 1 , a digital scene design and compilation drawing system based on AI features, which includes a feature extraction and user portrait construction module, a tourism scene planning module, a user experience feedback module, and an intelligent evaluation and optimization adjustment module; The feature extraction and user portrait construction module is used to collect user data, preprocess the user data, extract features from the preprocessed data, and construct user portraits; The tourism scene planning module is used to set feature vectors for the basic scenes of the tourist destination, calculate user similarity using the cosine similarity formula, and establish a similarity matrix based on user similarity to generate a scene recommendation list; The user experience feedback module is used to collect user experience feedback, analyze the user's satisfaction with the generated scene recommendation list, and obtain satisfaction analysis results; The intelligent evaluation and optimization adjustment module is used to analyze the results of satisfaction, identify the problems and improvement points of the scene recommendation list, and construct an optimization objective function to optimize the parameters of the scene recommendation list.

[0025] It should be noted that the application object of a digital scene design and compilation and drawing system based on AI features in an embodiment of the present invention can be virtual scene management in the tourism field. Specifically, it can be for comprehensive analysis and data processing of user behavior data through artificial intelligence and data analysis technology to achieve accurate generation and optimization of tourism scene recommendation lists. Through feature extraction and user portrait construction, tourism scene planning, user experience feedback, intelligent evaluation and optimization adjustment and other links, it can improve the accuracy of recommendations and user satisfaction, meet the user's needs for obtaining recommendation information, and automatically adjust the parameters of the recommendation list to achieve intelligent optimization and improve the performance of the recommendation system and user experience; at the same time, the system also promotes the digital development of the tourism industry, which is of great significance to improving the quality of tourism services, user experience and the overall competitiveness of the tourism industry.

[0026] The feature extraction and user portrait construction module collects user data, preprocesses the user data, extracts features from the preprocessed data, and constructs a user portrait, including the following steps: Step 101: Collect user data from multiple channels, including historical travel records and online behavior data; Step 102: preprocess the collected user data, including missing value processing, outlier detection and data normalization processing: taking mean filling as an example to process missing values, assuming that feature data The missing value set is , the set of non-missing values ​​is , then after filling For each missing value for , where represents the index of non-missing items; Take the example of outlier detection, for the feature data ,That for: , where represents the standard score value, They are The mean and standard deviation of ),judge Is it an outlier? Taking the normalization formula as an example, , where represents the feature dataset, Indicates the feature dataset feature data, Represents the normalized value; Both represent indexes; Step 103: Obtain the pre-processed user data set, namely, historical travel records and online behavior data; In step 103, the preprocessed user data set includes multiple features of multiple users. Specifically, the frequency features of users' on-site visits to scenic spots are extracted from historical travel records, and the click behavior features (user click times, click rate, click depth), search behavior features (user search keyword frequency, search category preference), purchase behavior features (user souvenir purchase frequency, purchase category distribution) and social behavior features (user sharing frequency of travel photos on social media) are extracted from online behavior data. Step 104: Set the user data set to exist Each user has five user characteristics, namely, the user's frequency of visiting attractions, click behavior characteristics, search behavior characteristics, purchase behavior characteristics, and social behavior characteristics; Step 105, then there is a feature matrix , where the feature matrix is a A matrix where each row represents a user and each column represents a feature; Step 106: Calculate the covariance matrix using the formula , where the formula is:

[0027] In the formula, Represents the feature matrix The transpose of Step 107: covariance matrix Perform eigenvalue decomposition and obtain the eigenvalue (in descending order) and the corresponding eigenvectors ; Step 108: Select the front The selection of principal components is based on the eigenvalue greater than the preset threshold or the cumulative contribution rate. The preset threshold means that the variance explained by the principal component is greater than the random noise, and the cumulative contribution rate means that the previous The proportion of the variance explained by the principal components to the total variance, and represents the principal component index; Step 109, select front The eigenvectors form the projection matrix , the feature matrix Projecting onto the principal component, we get the reduced-dimensional data matrix ,Right now , where Represents the projection matrix; each row Indicates Users in The score on the principal component is the score on the first principal component in the feature space. The feature vector of each user is projected onto the The coordinates obtained after the subspace defined by the principal components represent the user's position in the new coordinate system formed by the principal components; For example, if we select the first two principal components ( ), then the data matrix after dimensionality reduction is a A matrix where each row represents the projection of a user on two principal components, and the two principal components are represented by the eigenvectors and Definition, used to capture the most dominant direction of variation in the user dataset; Step 110: Use the K-means clustering algorithm to divide users into different groups with similar characteristics based on the extracted user characteristics, that is, portraits of different users; wherein the specific analysis formula is as follows:

[0028] In the formula, represents the clustering target, represents clustering partition, Indicates The set of samples in the cluster, that is, all samples assigned to the The set of user feature vectors of clusters, Indicates The mean vector of the clusters, Indicates the first The feature vector of a user, represents the feature vector index of the user, represents the number of clusters or groups, Represents a clustering index; Step 111: By dividing the user data into groups with similar characteristics (i.e. Mean clustering), and calculate and analyze the characteristics of each group to build a user portrait.

[0029] The tourism scene planning module sets feature vectors for basic scenes of tourist destinations, calculates user similarity using the cosine similarity formula, and establishes a similarity matrix based on user similarity, thereby generating a scene recommendation list in the following steps: Step 201: Based on the user portrait constructed by the feature extraction and user portrait construction module, define multiple basic scenarios of tourist destinations, and set a feature vector for each basic scenario. Each basic scenario is represented as a multidimensional vector, that is, , where Represents the basic scene In the The value of the feature, Represents the basic scene feature value index; In step 201, the feature vector of the basic scene includes multiple dimensions such as the theme, location, type, and activity content of the basic scene; the basic scene includes a natural scenery tour area, a historical and cultural site exploration area, a leisure and entertainment experience area, etc., and the specific features of each scene, such as a natural scenery tour area may include natural elements such as mountains, rivers, lakes, and vegetation coverage as part of the feature vector of the basic scene, while a historical and cultural site exploration area may have ancient buildings, cultural relics, etc. as its main feature vector; Step 202: arbitrarily select two users and mark them as users and users , through the user and users The cosine value of the rating vector is used to calculate the user similarity :

[0030] In the formula, Respectively represent users and users The rating vector of Step 203: Create a similarity matrix based on the similarity between users. ; where the matrix is ​​a The matrix of Represents the number of users, similarity matrix Each row and column represents a user, and the elements in the matrix represent the similarity between users; Step 204: For target users , according to the users with higher similarity in the similarity matrix (i.e. neighbor users ) based on the basic scene preferences, generating a scene recommendation list for the target user; wherein the target user represents any user currently being processed by the recommendation algorithm to generate a scene recommendation list, and can be any user in the user data set, including but not limited to user and users ; Step 205: Use target user Neighbor users The weighted average of the ratings of For basic scenes Rating :

[0031] In the formula, Indicates the target user The neighbor set of Indicates neighbor users For basic scenes Ratings; Step 206: Generate a scene recommendation list based on the predicted target user's basic scene scores.

[0032] The user experience feedback module collects user experience feedback and analyzes the user's satisfaction with the generated scene recommendation list. The steps for obtaining the satisfaction analysis result include: Step 301: Define a user experience dataset, which includes each user Satisfaction rating for the scene recommendation list ,and , the larger the value, the higher the satisfaction; Step 302: Count the satisfaction scores of all users and obtain the average and standard deviation of the user satisfaction scores: ,

[0033] In the formula, represents the average value of user satisfaction ratings, Indicates the number of users, Represents the user index, It represents the standard deviation of user satisfaction ratings, which is used to measure the degree of deviation between user satisfaction ratings and the average rating. When the standard deviation of satisfaction ratings is higher, the degree of deviation between user satisfaction ratings and their average value is greater. When the standard deviation of satisfaction ratings is lower, the degree of deviation between user satisfaction ratings and their average value is smaller. Step 303: Calculate and analyze the skewness of the user satisfaction rating by combining the mean value and standard deviation of the user satisfaction rating:

[0034] In the formula, Indicates the skewness value of the user satisfaction rating; Step 304: when analyzing the skewness of the user satisfaction score, if the skewness value of the user satisfaction score is a positive number, it means that the data of the user satisfaction score is right-skewed, that is, the user satisfaction score has a large number of high scores; if the skewness value of the user satisfaction score is a negative number, it means that the data of the user satisfaction score is left-skewed, that is, the user satisfaction score has a large number of low scores; set a skewness limit range, and compare the skewness value of the user satisfaction score with the skewness limit range. When the skewness value of the user satisfaction score is not within the skewness limit range, it means that the user satisfaction score is unstable and has an extreme skewness; when the skewness value of the user satisfaction score is within the skewness limit range, it means that the user satisfaction score is stable; Step 305: Preset satisfaction threshold , used to judge whether the overall satisfaction meets the standard; Step 306: Perform user satisfaction analysis based on the preset satisfaction threshold to obtain satisfaction analysis results; if the average value of the user satisfaction score is Above the preset satisfaction threshold , accompanied by a low standard deviation, indicating that the satisfaction distribution is uniform, the satisfaction level is generally high, and the skewness is within the skewness limit, then the overall satisfaction is judged to have reached the standard and a basic feedback signal is generated; conversely, if any of the following situations occurs, it can be judged that the overall satisfaction has not reached the established standard and an important feedback signal is generated: the average value of the user satisfaction score Not higher than the preset satisfaction threshold ; or although the average Above the preset satisfaction threshold , but the standard deviation is high, indicating a large difference in satisfaction; or the skewness is beyond the normal range; Step 307: Send the satisfaction analysis result including the basic feedback signal and the important feedback signal to the intelligent evaluation and optimization adjustment module.

[0035] The intelligent evaluation and optimization adjustment module identifies the problems and improvement points of the scene recommendation list based on the satisfaction analysis results, and constructs an optimization objective function. The steps for optimizing the parameters of the scene recommendation list include: Step 401: extracting a scene recommendation list corresponding to the feedback signal according to the satisfaction analysis result, thereby generating a high-quality recommendation template; Step 402: for important feedback signals of the satisfaction analysis results, compare the satisfaction scores corresponding to the important feedback signals with the expected satisfaction scores, where the expected satisfaction scores are expected values ​​obtained through expert scoring, industry standards, etc., to identify existing problems; Step 403: Based on the problems existing in the scene recommendation list, determine the improvement points that need to be optimized in the scene recommendation list, including the scene beauty, interaction fluency, cultural experience, and navigation convenience of the scene recommendation list; In step 403, for the beauty of the scenery, the details and color richness of the scenery in the virtual scene are increased; for the smoothness of the interaction, the interaction logic and response time in the virtual scene are optimized to reduce the differences between users; for the cultural experience, the local cultural characteristics are deeply explored and integrated into the virtual scene; for the convenience of navigation, the navigation interface and path planning algorithm in the virtual scene are optimized; Step 404: The improvement point to be optimized is used as the weight parameter of the scene recommendation list, and the weight parameter of the scene recommendation list is marked as ,in, They correspond to the weight values ​​of beautiful scenery, smooth interaction, cultural experience and navigation convenience, and the sum of the weight parameters is 1; Step 405: construct an optimization objective function:

[0036] In the formula, represents the optimization objective function value, Indicates The predicted value of satisfaction of a user under the weight parameter, Indicates User satisfaction rating; Step 406: Use the optimized weight parameters , adjust the scene recommendation list.

[0037] In the embodiment of the present invention, by collecting user data, preprocessing the user data, extracting features from the preprocessed data, and constructing a user portrait, the accuracy and consistency of the data are ensured, and a basis for accurate recommendation is provided; by setting feature vectors for the basic scenes of the tourist destination, using the cosine similarity formula to calculate the user similarity, and establishing a similarity matrix based on the user similarity, a scene recommendation list is generated, the accuracy and satisfaction of the recommendation are improved, and the user's need for obtaining recommendation information is met; by collecting user experience feedback, the user's satisfaction with the generated scene recommendation list is analyzed, and the satisfaction analysis results are obtained to provide a direction for optimization and adjustment; by identifying the problems and improvement points of the scene recommendation list according to the satisfaction analysis results, and constructing an optimization objective function to optimize the parameters of the scene recommendation list, the parameters of the recommendation list can be automatically adjusted, intelligent optimization is achieved, and the performance and user experience of the recommendation system are improved. In summary, the example of the present invention solves the problem of inaccurate tourist scene recommendation in the current virtual scene management. In actual situations, more data and context information may be required to make specific decisions and optimization plans.

[0038] In addition, the formulas involved in the above are all calculated by removing dimensions and taking their numerical values. They are a formula that is closest to the actual situation obtained by collecting a large amount of data and performing software simulation. The proportional coefficient in the formula and the various preset thresholds in the analysis process are set by technical personnel in this field according to actual conditions or obtained by simulating a large amount of data; the size of the proportional coefficient is to quantify each parameter to obtain a specific value for subsequent comparison. The size of the proportional coefficient depends on the amount of sample data and the preliminary setting of the corresponding processing coefficient for each group of sample data by technical personnel in this field; as long as it does not affect the proportional relationship between the parameter and the quantized value.

[0039] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically based on the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0040] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing the present application, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0041] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0042] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0043] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0044] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0045] Secondly: In the drawings of the embodiments disclosed in the present invention, only the structures related to the embodiments disclosed in the present invention are involved, and other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of the present invention can be combined with each other; Finally: The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A digital scene design and compilation rendering system based on AI features, characterized by: It includes feature extraction and user portrait construction module, tourism scenario planning module, user experience feedback module and intelligent evaluation and optimization adjustment module; The feature extraction and user portrait construction module is used to collect user data, preprocess the user data, extract features from the preprocessed data, and construct user portraits; The tourism scene planning module is used to set feature vectors for the basic scenes of the tourist destination, calculate user similarity using the cosine similarity formula, and establish a similarity matrix based on user similarity to generate a scene recommendation list; The user experience feedback module is used to collect user experience feedback, analyze the user's satisfaction with the generated scene recommendation list, and obtain satisfaction analysis results; The intelligent evaluation and optimization adjustment module is used to analyze the results of satisfaction, identify the problems and improvement points of the scene recommendation list, and construct an optimization objective function to optimize the parameters of the scene recommendation list.

2. According to the AI ​​feature-based digital scene design and compilation rendering system of claim 1, it is characterized by: The feature extraction and user portrait construction module collects user data, preprocesses the user data, and extracts features from the preprocessed data. The process includes: Collect user data from multiple channels, including historical travel records and online behavior data; Preprocess the collected user data, including missing value processing, outlier detection, and data normalization; Obtain a preprocessed user data set, namely, historical travel records and online behavior data; wherein the preprocessed user data set contains multiple features of multiple users: extract the frequency features of users' on-site visits to scenic spots from historical travel records, and extract click behavior features, search behavior features, purchase behavior features, and social behavior features from online behavior data; Let the user data set exist Each user has five user characteristics, namely, the user's frequency of visiting attractions, click behavior characteristics, search behavior characteristics, purchase behavior characteristics, and social behavior characteristics; Then the feature matrix , where the feature matrix is a A matrix where each row represents a user and each column represents a feature; Calculate the covariance matrix using the formula , where the formula is: In the formula, Represents the feature matrix The transpose of Covariance matrix Perform eigenvalue decomposition and obtain the eigenvalue and the corresponding eigenvector ; Select the front according to the size of the eigenvalue The selection of principal components is based on the eigenvalue greater than the preset threshold or the cumulative contribution rate. The preset threshold means that the variance explained by the principal component is greater than the random noise, and the cumulative contribution rate means that the previous The proportion of the variance explained by the principal components to the total variance, and represents the principal component index; Before selection The eigenvectors form the projection matrix , the feature matrix Projecting onto the principal component, we get the reduced-dimensional data matrix ,Right now , where Represents the projection matrix; each row Indicates Users in The score on the principal component is the score on the first principal component in the feature space. The feature vector of each user is projected onto the The coordinates obtained after the subspace defined by the principal components represent the user's position in the new coordinate system formed by the principal components.

3. According to the AI ​​feature-based digital scene design and compilation rendering system of claim 1, it is characterized by: The process of constructing user portraits in the feature extraction and user portrait construction module includes: Using the K-means clustering algorithm, users are divided into different groups with similar characteristics based on the extracted user features, that is, the portraits of different users; the specific analysis formula is as follows: In the formula, represents the clustering target, represents clustering partition, Indicates The set of samples in the cluster, that is, all samples assigned to the The set of user feature vectors of clusters, Indicates The mean vector of the clusters, Indicates the first The feature vector of each user, represents the feature vector index of the user, represents the number of clusters or groups, Represents a clustering index; By dividing user data into groups with similar characteristics, and calculate and analyze the characteristics of each group to build a user portrait.

4. According to the AI ​​feature-based digital scene design and compilation rendering system of claim 1, it is characterized by: The tourism scenario planning module sets feature vectors for the basic scenarios of the tourism destination and uses the cosine similarity formula to calculate the user similarity process, which includes: Based on the user portrait constructed by the feature extraction and user portrait construction module, multiple basic scenarios of tourist destinations are defined, and feature vectors are set for each basic scenario. Each basic scenario is represented as a multidimensional vector, namely , where Represents the basic scene In the The value of the feature, Represents the basic scene feature value index; Select two users at random and mark them as users and users , through the user and users The cosine value of the rating vector is used to calculate the user similarity : In the formula, Respectively represent users and users The rating vector of Based on the similarity between users, establish a similarity matrix ; where the matrix is ​​a The matrix of Represents the number of users, similarity matrix Each row and column of represents a user, and the elements in the matrix represent the similarity between users.

5. According to the AI ​​feature-based digital scene design and compilation rendering system of claim 1, it is characterized by: The travel scenario planning module establishes a similarity matrix based on user similarity, thereby generating a scenario recommendation list in the following process: For target users , based on the basic scene preferences of users with high similarity in the similarity matrix, a scene recommendation list is generated for the target user; wherein the target user represents any user currently being processed by the recommendation algorithm to generate a scene recommendation list, and users with high similarity are neighbor users of the target user; Use target users Neighbor users The weighted average of the ratings of For basic scenes Rating : In the formula, Indicates the target user The neighbor set of Indicates neighbor users For basic scenes Ratings; Combined with the predicted target user's basic scenario scores, a scenario recommendation list is generated.

6. The digital scene design and compilation rendering system based on AI features according to claim 1, characterized in that: The user experience feedback module collects user experience feedback and analyzes the user's satisfaction with the generated scenario recommendation list. The process of obtaining the satisfaction analysis results includes: Define a user experience dataset, which contains each user Satisfaction rating for the scene recommendation list ,and , the larger the value, the higher the satisfaction; Count the satisfaction scores of all users and obtain the mean and standard deviation of the user satisfaction scores: ， In the formula, represents the average value of user satisfaction ratings, Indicates the number of users, Represents the user index, It represents the standard deviation of user satisfaction ratings, which is used to measure the degree of deviation between user satisfaction ratings and the average rating. When the standard deviation of satisfaction ratings is higher, the degree of deviation between user satisfaction ratings and their average value is greater. When the standard deviation of satisfaction ratings is lower, the degree of deviation between user satisfaction ratings and their average value is smaller. Combined with the mean and standard deviation of user satisfaction ratings, the skewness of user satisfaction ratings is calculated and analyzed: In the formula, Indicates the skewness value of the user satisfaction rating; When analyzing the skewness of the user satisfaction score, if the skewness value of the user satisfaction score is a positive number, it means that the data of the user satisfaction score is right-skewed, that is, the user satisfaction score has a large number of high scores; if the skewness value of the user satisfaction score is a negative number, it means that the data of the user satisfaction score is left-skewed, that is, the user satisfaction score has a large number of low scores; set the skewness limit range, and compare the skewness value of the user satisfaction score with the skewness limit range. When the skewness value of the user satisfaction score is not within the skewness limit range, it means that the user satisfaction score is unstable and has an extreme skewness; when the skewness value of the user satisfaction score is within the skewness limit range, it means that the user satisfaction score is stable; Preset satisfaction threshold , used to judge whether the overall satisfaction meets the standard; Combine the preset satisfaction threshold to analyze user satisfaction, so as to obtain the satisfaction analysis results; if the average value of user satisfaction score Above the preset satisfaction threshold , accompanied by a low standard deviation, indicating that the satisfaction distribution is uniform, the satisfaction level is generally high, and the skewness is within the skewness limit, then the overall satisfaction is judged to have reached the standard and a basic feedback signal is generated; conversely, if any of the following situations occurs, it can be judged that the overall satisfaction has not reached the standard and an important feedback signal is generated: the average value of the user satisfaction score Not higher than the preset satisfaction threshold ; or although the average Above the preset satisfaction threshold But the standard deviation is high; or the skewness is beyond the normal range; The satisfaction analysis results including basic feedback signals and important feedback signals are sent to the intelligent evaluation and optimization adjustment module.

7. A digital scene design and compilation rendering system based on AI features according to claim 1 or 6, characterized in that: The intelligent evaluation and optimization adjustment module identifies the problems and improvements of the scene recommendation list based on the satisfaction analysis results, and constructs an optimization objective function. The process of optimizing the parameters of the scene recommendation list includes: According to the basic feedback signal of the satisfaction analysis result, a scene recommendation list corresponding to the feedback signal is extracted to generate a high-quality recommendation template; For important feedback signals in the satisfaction analysis results, compare the satisfaction scores corresponding to the important feedback signals with the expected satisfaction scores to identify problems in the scene recommendation list; Based on the existing problems in the scene recommendation list, determine the improvement points that need to be optimized in the scene recommendation list, including the beauty of the scenery, smooth interaction, cultural experience, and navigation convenience of the scene recommendation list; The improvement points that need to be optimized are used as weight parameters of the scene recommendation list, and the weight parameters of the scene recommendation list are marked as ,in, They correspond to the weight values ​​of beautiful scenery, smooth interaction, cultural experience and navigation convenience, and the sum of the weight parameters is 1; Construct the optimization objective function: In the formula, represents the optimization objective function value, Indicates The predicted value of satisfaction of each user under the weight parameter, Indicates User satisfaction rating; Use the optimized weight parameters , adjust the scene recommendation list.

Images