Cross-modal data processing method for cultural enabling airport design evaluation

Abstract

The invention discloses a cross-modal data processing method for cultural enabling airport design evaluation, and relates to the technical field of cross-modal data analysis and processing, and the method comprises the steps: obtaining time alignment multi-source evaluation data through an adaptive time alignment algorithm based on preprocessed multi-source evaluation data; based on a convolutional neural network model, constructing a multi-source quality enhancement model, and inputting the time alignment multi-source evaluation data into the multi-source quality enhancement model to obtain high-quality multi-source evaluation data; based on the high-quality multi-source evaluation data, constructing a multi-dimensional semantic space, and obtaining position information and association rules in the multi-dimensional semantic space; obtaining consistency multi-source evaluation data by analyzing the position information and the association rule; and based on the consistent multi-source evaluation data, through a comprehensive scoring mechanism, evaluating culture elements of the culture enabling airport. According to the invention, through the adaptive time alignment algorithm and the convolutional neural network model, the accuracy and reliability of data processing are significantly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of cross-modal data analysis and processing, and particularly to a cross-modal data processing method for cultural empowerment of airport design evaluation. Background Art

[0002] In recent years, with the rapid development of the global aviation industry, the importance of airports as transportation hubs has become increasingly prominent. An airport is not only a transportation node but also gradually becomes an important platform for showcasing urban culture and enhancing the passenger experience. To achieve this goal, the concept of cultural empowerment of airport design has emerged, aiming to enhance the functionality and attractiveness of airports by introducing and optimizing cultural elements. In this context, the development of cross-modal data processing technology provides new tools and methods for evaluating and optimizing airport design. Cross-modal data processing technology integrates data from multiple sources (such as social media comments, online user feedback, news reports, and passenger questionnaires), which can more comprehensively reflect the actual effects of airport design and the true feelings of passengers. At the same time, the role of airports in the cultural and tourism industrial chain has become increasingly important, becoming the first impression point for tourists entering the destination and the last stop when leaving. By integrating local cultural characteristics into airport design, such as setting up art exhibitions with local characteristics, handicraft sales areas, and cultural performances, airports can become a window for spreading local culture, attracting more domestic and foreign tourists, and thus promoting the upgrading and development of the tourism industry and related industries in the entire city. In addition, the cultural empowerment of airports is also reflected in their contribution to the local economy. By optimizing the spatial layout and service facilities, introducing more diversified business forms and service models, airports can improve their economic efficiency and social influence, and promote the economic development of the surrounding areas. This design concept of cultural empowerment not only enhances the functionality of airports but also provides new impetus for the diversified development of the local economy and the optimization and upgrading of the industrial structure.

[0003] Despite the significant progress in this field, the existing technologies still face two core challenges when dealing with multi-source data: one is the issue of time misalignment among different data sources, and the other is the difficulty of automatically detecting and repairing low-quality or noisy data points. First, the timestamps of different data sources are inconsistent, making it difficult to conduct effective comprehensive analysis. For example, social media comments and news reports may have different update frequencies, which makes it difficult to directly merge these data. If the time misalignment problem cannot be effectively solved, it will seriously affect the accuracy and reliability of subsequent analysis. Second, existing methods are also ineffective in dealing with low-quality or noisy data points. Due to the uneven quality of multi-source data, some data may contain incorrect information or redundant content. If these problems cannot be identified and repaired in a timely manner, it will have a negative impact on the final evaluation results. Therefore, in the context of airport culture and tourism and industrial upgrading, how to efficiently solve the time misalignment problem and use deep learning technology to automatically detect and repair low-quality or noisy data points has become a key problem that needs to be solved urgently. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a cross-modal data processing method for cultural empowerment of airport design evaluation, which solves the problems of time misalignment of multi-source data and automatic detection and repair of low-quality data points.

[0006] To solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a cross-modal data processing method for cultural empowerment of airport design evaluation, which includes collecting multi-source evaluation data and preprocessing the multi-source evaluation data; based on the preprocessed multi-source evaluation data, obtaining time-aligned multi-source evaluation data through an adaptive time alignment algorithm; based on a convolutional neural network model, constructing a multi-source quality enhancement model, and inputting the time-aligned multi-source evaluation data into the multi-source quality enhancement model to obtain high-quality multi-source evaluation data; based on the high-quality multi-source evaluation data, constructing a multi-dimensional semantic space and obtaining position information and association rules in the multi-dimensional semantic space;

[0008] By analyzing the position information and association rules, obtaining consistent multi-source evaluation data; based on the consistent multi-source evaluation data, evaluating the cultural elements of the airport empowered by culture through a comprehensive scoring mechanism.

[0009] As a preferred solution of the cross-modal data processing method for cultural empowerment of airport design evaluation according to the present invention, wherein: the multi-source evaluation data includes social media comments, online user feedback, news reports, and passenger questionnaires;

[0010] The preprocessing of the multi-source evaluation data includes data cleaning, data annotation, normalization processing, and filling missing values.

[0011] As a preferred solution of the cross-modal data processing method for cultural empowerment airport design evaluation described in the present invention, wherein: based on the preprocessed multi-source evaluation data, the time-aligned multi-source evaluation data is obtained through an adaptive time alignment algorithm. The specific steps are as follows.

[0012] The timestamps of the preprocessed multi-source evaluation data are extracted and converted into a time series form by sorting them in chronological order.

[0013] Calculate the time difference between adjacent timestamps in the multi-source evaluation data converted into a time series form.

[0014] Based on the time difference, a control signal is obtained through PID. The expression is:

[0015]

[0016] wherein, u(t) is the adjusted control signal, D z is the proportional coefficient of PID, D o is the integral coefficient of PID, D d is the differential coefficient of PID;

[0017] Dynamically adjust the time sensitivity coefficient through the control signal.

[0018] Based on the dynamically adjusted time sensitivity coefficient, each timestamp is adaptively adjusted. The expression is:

[0019]

[0020] wherein, t' i is the adjusted timestamp, t0 is the timestamp reference point, α is the dynamically adjusted time sensitivity coefficient, L is the factor affecting time alignment, and f is the transformation function of the time difference;

[0021] Based on the adjusted timestamp, the time attributes of the multi-source evaluation data are adjusted, and the adjusted attributes are integrated to obtain the time-aligned multi-source evaluation data.

[0022] As a preferred solution of the cross-modal data processing method for cultural empowerment airport design evaluation described in the present invention, wherein: based on the convolutional neural network model, a multi-source quality enhancement model is constructed. The specific steps are as follows.

[0023] Based on the convolutional neural network model;

[0024] The input layer receives the time-aligned historical multi-source evaluation data;

[0025] The convolutional layer captures features of time-aligned historical multi-source evaluation data of different types by adding multiple convolutional layers and using convolutional kernels of different sizes;

[0026] The pooling layer gradually reduces the feature dimension through max-pooling operations;

[0027] The fully connected layer maps the features after convolution and pooling to a high-dimensional feature space;

[0028] The output layer outputs the preliminary result through an activation function;

[0029] The quality evaluation layer detects and repairs the preliminary result to obtain high-quality historical multi-source evaluation data;

[0030] Finally, a multi-source quality enhancement model is constructed.

[0031] As a preferred solution of the cross-modal data processing method for cultural empowerment airport design evaluation described in the present invention, wherein: inputting the time-aligned multi-source evaluation data into the multi-source quality enhancement model to obtain high-quality multi-source evaluation data, the specific steps are as follows,

[0032] The time-aligned multi-source evaluation data is processed through the convolutional layer, pooling layer and fully connected layer, and combined with the activation function to obtain preliminary high-quality multi-source evaluation data, and the expression is:

[0033] y' = g(σ(W H ·p(σ(W k *V + b k )) + b H ));

[0034] Wherein, y' is the preliminary high-quality multi-source evaluation data, W H is the weight matrix of the fully connected layer, W k is the weight matrix of the convolutional layer, V is the time-aligned multi-source evaluation data, b k is the bias term of the convolutional layer, b H is the weight matrix of the fully connected layer, g is the activation function softmax, σ is the activation function ReLU, p is the max-pooling operation, * is the convolution operation;

[0035] Calculate the quality comprehensive score of the preliminary high-quality multi-source evaluation data through the quality evaluation mechanism, and the expression is:

[0036]

[0037] Wherein, Q is the quality comprehensive score, G is the global average pooling operation, S is the soft weighted exponential linear unit function, m is the total number of convolutional kernels, j is the index variable of the convolutional kernel, A is the rectified linear unit, K jis the j-th convolution kernel, α is the time-sensitive coefficient, F is the additional feature vector, and β is the balance factor;

[0038] Based on the quality comprehensive score distribution, define the quality threshold θ;

[0039] When Q ≥ θ, it indicates that the preliminary high-quality multi-source evaluation data is high-quality data;

[0040] When Q < θ, it indicates that the preliminary high-quality multi-source evaluation data is low-quality data, and the low-quality data needs to be repaired until Q ≥ θ;

[0041] Based on the quality judgment result, obtain the high-quality multi-source evaluation data.

[0042] As a preferred solution of the cross-modal data processing method for cultural empowerment of airport design evaluation described in the present invention, wherein: based on the high-quality multi-source evaluation data, construct a multi-dimensional semantic space, and obtain the position information and association rules in the multi-dimensional semantic space. The specific steps are as follows:

[0043] Based on the high-quality multi-source evaluation data, generate semantic features through text analysis and TF-IDF feature extraction methods;

[0044] Map the semantic features to a low-dimensional multi-dimensional semantic space through dimensionality reduction technology to construct the multi-dimensional semantic space;

[0045] Analyze the position of each high-quality multi-source evaluation data in the multi-dimensional semantic space through a clustering algorithm and encode it to obtain the position information in the multi-dimensional semantic space;

[0046] Apply frequent itemset mining and association rule learning algorithms to extract the association rules in the multi-dimensional semantic space.

[0047] As a preferred solution of the cross-modal data processing method for cultural empowerment of airport design evaluation described in the present invention, wherein: the steps of obtaining the consistent multi-source evaluation data by analyzing the position information and association rules are as follows:

[0048] Based on the position information, analyze the similarity between each multi-source evaluation data and its neighboring points;

[0049] Based on the similarity analysis result, obtain the position information consistency score;

[0050] Apply the extracted association rules to different multi-source evaluation data, and analyze the support degree and support rate of the association rules in the multi-source evaluation data from different sources;

[0051] Based on the support degree and support rate of different multi-source evaluation data, evaluate the universality and reliability of different multi-source evaluation data;

[0052] Obtain the consistency score of association rules based on the universality and reliability of different multi-source evaluation data;

[0053] Perform weighted summation on the location information consistency score and the association rule consistency score to obtain the comprehensive consistency score;

[0054] Obtain the consistent multi-source evaluation data through the comprehensive consistency score.

[0055] As a preferred solution of a cross-modal data processing method for cultural empowerment airport design evaluation described in the present invention, wherein: based on the consistent multi-source evaluation data, evaluate the cultural elements of the cultural empowerment airport through a comprehensive scoring mechanism, and the specific steps are as follows.

[0056] Extract the cultural features of the consistent multi-source evaluation data through natural language processing technology and sentiment analysis algorithms, and convert the cultural features into numerical feature vectors using the word embedding method;

[0057] Perform standardization processing on the numerical feature vectors, and integrate the standardized numerical feature vectors into a feature matrix;

[0058] Based on the feature matrix, calculate the comprehensive score of cultural elements through a comprehensive scoring mechanism, and the expression is:

[0059]

[0060] Wherein, R is the comprehensive score of cultural elements, p is the total number of cultural elements, u is the index variable of cultural elements, q is the total number of influencing factors, v is the index variable of influencing factors, w v is the weight of the vth influencing factor, z uv is the eigenvalue of the u-th cultural element under the v-th influencing factor after standardization, f(z uv ) is the non-linear transformation function of the eigenvalue, g(z uv ) is the square root function of the eigenvalue, h(z uv ) is the absolute value function of the eigenvalue;

[0061] Set the comprehensive score threshold Ф according to historical evaluation data;

[0062] When R≥Ф, it indicates that the cultural element has a positive impact on the airport design;

[0063] When R<Ф, it indicates that the impact of the cultural element on the airport design is insufficient and needs further optimization.

[0064] Second aspect, the present invention provides a computer device, including a memory and a processor, where the memory stores a computer program, and: when the computer program is executed by the processor, any step of a cross-modal data processing method for cultural empowerment of airport design evaluation as described in the first aspect of the present invention is implemented.

[0065] Third aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, and: when the computer program is executed by the processor, any step of a cross-modal data processing method for cultural empowerment of airport design evaluation as described in the first aspect of the present invention is implemented.

[0066] The beneficial effects of the present invention are as follows: The present invention solves the problem of time misalignment of multi-source heterogeneous data through an adaptive time alignment algorithm, and uses a convolutional neural network model to automatically detect and repair low-quality or noisy data points, thereby obtaining high-quality multi-source evaluation data. Based on this, a multi-dimensional semantic space is constructed and consistent multi-source evaluation data is extracted. Finally, the influence of cultural elements on cultural empowerment of airport design is accurately evaluated through a comprehensive scoring mechanism. This method not only significantly improves the accuracy and reliability of data processing, but also provides a scientific, comprehensive and efficient evaluation tool for cultural empowerment of airport design, ensuring the high quality and credibility of evaluation results. Description of the Drawings

[0067] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for description in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0068] Figure 1 It is a flowchart of a cross-modal data processing method for cultural empowerment of airport design evaluation in Embodiment 1.

[0069] Figure 2 It is a flowchart of obtaining position information and association rules in a multi-dimensional semantic space in Embodiment 1. Detailed Embodiments

[0070] To make the above objects, features and advantages of the present invention more obvious and understandable, the detailed embodiments of the present invention will be described in detail below with reference to the drawings of the specification.

[0071] Many specific details are set forth in the following description in order to fully understand the present invention, but the present invention can also be implemented in other ways different from those described herein. Those skilled in the art can make similar promotions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0072] Second, the "one embodiment" or "embodiment" referred to herein means a specific feature, structure, or characteristic that may be included in at least one implementation manner of the present invention. The appearances of "in one embodiment" in different places in this specification do not all refer to the same embodiment, nor are they separate or alternative embodiments that exclude each other with other embodiments.

[0073] Example 1, referring to Figure 1 and Figure 2 , which is the first embodiment of the present invention. This embodiment provides a cross-modal data processing method for cultural empowerment airport design evaluation, including the following steps:

[0074] S1. The multi-source evaluation data includes social media comments, online user feedback, news reports, and passenger questionnaires.

[0075] S1.2. The preprocessing of the multi-source evaluation data includes data cleaning, data annotation, standardization processing, and filling in missing values.

[0076] Specifically, for data cleaning: use OpenRefine to identify and delete duplicate records to avoid their deviation in analysis, then remove data irrelevant to the research objective, such as advertisements, automatic replies, or other non-user-generated content, to streamline the data set, and then process incorrect data items, which may include correcting spelling mistakes, fixing incorrect formats (such as dates and timestamps), and correcting or deleting obviously abnormal or illogical data points. Finally, perform a quality check on all data to ensure it meets the expected standards, thus laying a solid foundation for subsequent analysis work;

[0077] For data annotation: classify and label the multi-source evaluation data through natural language processing algorithms. Specifically, first perform text preprocessing steps, such as word segmentation, removing stop words, and lemmatization, to prepare clean data, then apply NLP techniques such as sentiment analysis to identify positive or negative emotions, topic modeling to determine the comment topic, and use named entity recognition (NER) to extract key entities such as person names and place names. Finally, based on these analysis results, add corresponding tags to each evaluation data to provide clear reference and context information, thereby enhancing the accuracy and depth of subsequent data analysis;

[0078] For standardization processing: first define unified data formats and unit standards, for example, the timestamp adopts the ISO 8601 format, and the scores are unified to a five-star system. Then perform normalization or standardization transformation on numerical data to make data with different dimensions fall within the same range. For text data, ensure that the encoding format is consistent (such as UTF-8), and unify the language expression (such as converting all variants to standard terms). Finally, unify the field names and data types of structured data to ensure seamless docking between databases;

[0079] Filling missing values: First, evaluate the patterns and causes of missing data through correlation analysis to determine whether it is randomly missing or systematically missing. Then, select an appropriate imputation method according to the data characteristics, such as filling numerical data with the mean, median, or mode, filling categorical data based on the most frequent category, and for time series data, linear interpolation or moving average interpolation can be used with the data of adjacent time points. Finally, conduct a quality check on the filled data to ensure its integrity and usability, thereby providing a reliable basis for subsequent analysis.

[0080] S2. Obtain time-aligned multi-source evaluation data through an adaptive time alignment algorithm based on the preprocessed multi-source evaluation data.

[0081] S2.1. Convert it into a time series form by extracting the timestamps of the preprocessed multi-source evaluation data and sorting them in chronological order.

[0082] S2.2. Calculate the time difference between adjacent timestamps in the multi-source evaluation data converted into a time series form.

[0083] It should be noted that the expression for calculating the time difference between adjacent timestamps in the multi-source evaluation data converted into a time series form is:

[0084] Δt i =|(t i+1 -t i )|;

[0085] where Δt i is the time difference, t i is the current timestamp, and t i+1 is the adjacent timestamp.

[0086] S2.3. Based on the time difference, obtain the control signal through PID, and the expression is:

[0087]

[0088] where u(t) is the adjusted control signal, D z is the proportional coefficient of PID, D o is the integral coefficient of PID, and D d is the differential coefficient of PID.

[0089] S2.4. Dynamically adjust the time sensitivity coefficient through the control signal.

[0090] It should be noted that for dynamically adjusting the time sensitivity coefficient, for example, if the current error is large, the system may increase the time sensitivity coefficient to increase the sensitivity of time alignment; if the error is small, the time sensitivity coefficient is reduced to maintain stability.

[0091] S2.5. Adaptive adjustment is performed on each timestamp based on dynamically adjusting the time sensitivity coefficient, and the expression is as follows:

[0092]

[0093] where t' i is the adjusted timestamp, t0 is the timestamp reference point, α is the dynamically adjusted time sensitivity coefficient, L is the factor affecting time alignment, and f is the transformation function of the time difference.

[0094] S2.6. Based on the adjusted timestamp, adjust the time attributes of the multi-source evaluation data, and integrate the adjusted attributes to obtain time-aligned multi-source evaluation data.

[0095] S3. Build a multi-source quality enhancement model based on the convolutional neural network model.

[0096] S3.1. Based on the convolutional neural network model;

[0097] The input layer receives the time-aligned historical multi-source evaluation data;

[0098] The convolutional layer captures features of different types of time-aligned historical multi-source evaluation data by adding multiple convolutional layers and using convolutional kernels of different sizes;

[0099] The pooling layer gradually reduces the feature dimension through max pooling operations;

[0100] The fully connected layer maps the features after convolution and pooling to a high-dimensional feature space;

[0101] The output layer outputs the preliminary result through an activation function;

[0102] The quality assessment layer detects and repairs the preliminary result to obtain high-quality historical multi-source evaluation data;

[0103] Finally, a multi-source quality enhancement model is constructed.

[0104] It should be noted that by introducing the quality assessment layer, further detection and repair of the preliminary result can be performed at the output stage of the multi-source quality enhancement model to ensure that the finally obtained multi-source evaluation data has higher accuracy and reliability;

[0105] It should also be noted that the quality assessment layer can automatically identify and correct low-quality or noisy data points to ensure that the output multi-source evaluation data has higher accuracy and consistency. Through the detection and repair mechanism of this layer, the multi-source quality enhancement model not only improves the data quality but also enhances the reliability of subsequent analysis, thus providing a more solid data foundation for the comprehensive evaluation of cultural elements.

[0106] S4. Align the time of the multi-source evaluation data and input it into the multi-source quality enhancement model to obtain high-quality multi-source evaluation data.

[0107] S4.1. The time-aligned multi-source evaluation data is processed through a convolutional layer, a pooling layer, and a fully connected layer, and combined with an activation function to obtain preliminary high-quality multi-source evaluation data. The expression is:

[0108] y' = g(σ(W H ·p(σ(W k *V + b k )) + b H ));

[0109] where y' is the preliminary high-quality multi-source evaluation data, W H is the weight matrix of the fully connected layer, W k is the weight matrix of the convolutional layer, V is the time-aligned multi-source evaluation data, b k is the bias term of the convolutional layer, b H is the weight matrix of the fully connected layer, g is the activation function softmax, σ is the activation function ReLU, p is the max pooling operation, and * is the convolution operation.

[0110] It should be noted that this expression represents the process of obtaining preliminary high-quality multi-source evaluation data based on the time-aligned multi-source evaluation data, as follows:

[0111] Input the time-aligned multi-source evaluation data into the multi-source quality enhancement model. After feature extraction by the convolutional layer, dimension reduction and key information retention by the pooling layer, and mapping to a high-dimensional feature space by the fully connected layer, finally, non-linear transformation and probability distribution output are performed through activation functions (such as ReLU and softmax) to obtain preliminary high-quality multi-source evaluation data.

[0112] S4.2. Calculate the quality comprehensive score of the preliminary high-quality multi-source evaluation data through a quality evaluation mechanism. The expression is:

[0113]

[0114] where Q is the quality comprehensive score, G is the global average pooling operation, S is the soft weighted exponential linear unit function, m is the total number of convolutional kernels, j is the index variable of the convolutional kernel, A is the rectified linear unit, K j is the j-th convolutional kernel, α is the time-sensitive coefficient, F is the additional feature vector, and β is the balance factor.

[0115] It should be noted that the additional feature vector refers to the supplementary features introduced to enhance the expressiveness and accuracy of the multi-source quality enhancement model on the basis of the original data features;

[0116] It should also be noted that the additional feature vectors can be obtained from multiple sources, including but not limited to external databases, domain expertise, historical data analysis, or calculated through specific algorithms. For example, in the evaluation of airport design empowered by culture, the additional feature vectors may come from passenger flow statistics, seasonal variations, airport operation status reports, etc.

[0117] S4.3. Define the quality threshold θ based on the quality comprehensive score distribution;

[0118] When Q ≥ θ, it indicates that the preliminary high-quality multi-source evaluation data is high-quality data;

[0119] When Q < θ, it indicates that the preliminary high-quality multi-source evaluation data is low-quality data, and the low-quality data needs to be repaired until Q ≥ θ.

[0120] It should be noted that repairing low-quality data through linear interpolation repair, mean filtering repair, median filtering repair, rule-based repair, and adaptive weighted average repair can effectively improve the quality of the preliminary results and ensure that the finally output multi-source evaluation data is of high quality;

[0121] It should also be noted that the quality comprehensive scores of all preliminary high-quality multi-source evaluation data are statistically analyzed, and a distribution map of the quality comprehensive scores (such as a histogram or cumulative distribution function) is drawn. A suitable percentile (such as the 90th percentile) is selected as the quality threshold θ according to the characteristics of the score distribution.

[0122] S4.4. Obtain high-quality multi-source evaluation data based on the quality judgment result.

[0123] S5. Based on the high-quality multi-source evaluation data, construct a multi-dimensional semantic space and obtain the position information and association rules in the multi-dimensional semantic space.

[0124] S5.1. Generate semantic features based on the high-quality multi-source evaluation data through text analysis and TF-IDF feature extraction methods.

[0125] Specifically, first, the text data is preprocessed, including word segmentation, stop word removal, etc. Then, the TF-IDF feature extraction method is used to quantify the importance of each word by calculating the term frequency (TF) of each word in a single document and the inverse document frequency (IDF) in the entire document collection. High-frequency words have a higher weight in a single document, while the weight of words that appear frequently in the entire document collection is reduced to highlight the uniqueness of the document. Finally, the processed text is filled into the vector corresponding to the vocabulary index position based on the importance of each term, and the value of the term that does not appear is set to 0, thereby generating semantic features that can reflect the text content and semantic information;

[0126] Semantic features include attributes that can reflect the content and meaning of text, such as keywords, themes, sentiment tendencies, and entity recognition results, etc.

[0127] S5.2. Map the semantic features to a low-dimensional multi-dimensional semantic space through dimensionality reduction techniques to construct a multi-dimensional semantic space.

[0128] Specifically, process these features through a suitable dimensionality reduction algorithm (such as PCA, t-SNE, or autoencoder) to reduce the dimensions while retaining the main information. Then, apply the selected dimensionality reduction algorithm to obtain the low-dimensional representation and construct a multi-dimensional semantic space based on this.

[0129] S5.3. Analyze the positions of each high-quality multi-source evaluation data in the multi-dimensional semantic space through a clustering algorithm and encode them to obtain the position information in the multi-dimensional semantic space.

[0130] Specifically, cluster the low-dimensional semantic feature vectors through a clustering algorithm (such as K-means or DBSCAN), group them into different clusters according to the similarity between data points, and assign cluster labels. Then, convert these cluster labels into encoded position information.

[0131] S5.4. Apply frequent itemset mining and association rule learning algorithms to extract association rules in the multi-dimensional semantic space.

[0132] Specifically, use a frequent itemset mining algorithm (such as Apriori or FP-Growth) to find frequent itemsets that meet the minimum support. Based on these frequent itemsets, apply an association rule learning algorithm to generate all possible association rules. For example, if the frequent itemset is {ABC}, then rules such as A→BC, B→AC, C→AB, AB→C, AC→B, and BC→A can be generated, and significant rules are screened out through metrics such as support and confidence. Finally, obtain the association rules in the multi-dimensional semantic space.

[0133] S6. Obtain consistent multi-source evaluation data by analyzing the position information and association rules.

[0134] S6.1. Based on the position information, analyze the similarity between each multi-source evaluation data and its neighboring points.

[0135] Specifically, select a suitable distance metric method (such as Euclidean distance or cosine similarity), calculate the similarity score between each data point and its neighboring points, determine the neighboring points through methods such as k-NN, and normalize the similarity score to quantify the similarity degree between data points.

[0136] S6.2. Based on the similarity analysis results, obtain a position information consistency score.

[0137] Specifically, based on the calculated similarity scores, analyze the relationship between each multi-source evaluation data and its neighboring points, explore the internal structure of the data through statistical methods (such as average similarity, variance) or visualization tools, and finally assign a location information consistency score to each multi-source evaluation data point to reflect its consistency level with neighboring points.

[0138] S6.3. Apply the extracted association rules to different multi-source evaluation data, and analyze the support degree and support rate of the association rules in multi-source evaluation data from different sources.

[0139] Specifically, by counting the number of times the premise and conclusion items of each rule appear simultaneously in the same transaction, and then dividing by the total number of transactions, obtain the support degree of each rule in each multi-source evaluation data.

[0140] By counting the number of times the premise and conclusion items of each rule appear simultaneously in the same transaction, and then dividing by the number of transactions containing the premise item, obtain the confidence of each rule in each multi-source evaluation data.

[0141] S6.4. Based on the support degree and support rate of different multi-source evaluation data, evaluate the universality and reliability of different multi-source evaluation data.

[0142] S6.5. Based on the universality and reliability of different multi-source evaluation data, obtain the association rule consistency score.

[0143] It should be noted that summarize the support degree and support rate of each association rule in each dataset, identify the commonly existing and consistent rules by comparing the support degree, and evaluate its universality; then analyze the confidence to determine the reliability of the rule.

[0144] S6.6. Perform weighted summation on the location information consistency score and the association rule consistency score to obtain the comprehensive consistency score.

[0145] S6.7. Obtain the consistent multi-source evaluation data through the comprehensive consistency score.

[0146] Furthermore, based on the comprehensive score, screen out the data points with higher scores to ensure that these data are consistent and reliable in terms of location information and association rules; finally, integrate the screened data into a consistent multi-source evaluation dataset.

[0147] S7. Based on the consistent multi-source evaluation data, evaluate the cultural elements empowering the airport culture through a comprehensive scoring mechanism.

[0148] S7.1. Through natural language processing technology and sentiment analysis algorithms, extract the cultural characteristics of the consistent multi-source evaluation data, and use the word embedding method to convert the cultural characteristics into numerical feature vectors.

[0149] It should be noted that the consistent multi-source evaluation data is processed through text analysis methods in natural language processing technology, the sentiment analysis algorithm is used to evaluate the user's attitude towards cultural elements, the cultural characteristics are further refined through keyword extraction and rule matching, and word embedding models (such as Word2Vec, GloVe, BERT) are selected to convert these characteristics into numerical feature vectors.

[0150] S7.2. Standardize the numerical feature vectors and integrate the standardized numerical feature vectors into a feature matrix.

[0151] Furthermore, the standardization method (such as Min-Max standardization or Z-score standardization) is used to standardize each eigenvalue; then all the standardized feature vectors are arranged in rows to construct a feature matrix.

[0152] S7.3. Based on the feature matrix, calculate the comprehensive score of cultural elements through a comprehensive scoring mechanism. The expression is:

[0153]

[0154] where R is the comprehensive score of cultural elements, p is the total number of cultural elements, u is the index variable of cultural elements, q is the total number of influencing factors, v is the index variable of influencing factors, w v is the weight of the vth influencing factor, z uv is the eigenvalue of the uth cultural element under the vth influencing factor after standardization, f(z uv ) is the non-linear transformation function of the eigenvalue, g(z uv ) is the square root function of the eigenvalue, h(z uv ) is the absolute value function of the eigenvalue.

[0155] It should be noted that cultural elements mainly come from cultural characteristics extracted from consistent multi-source evaluation data (such as user evaluations, comments, and historical assessment data) through natural language processing technology and sentiment analysis algorithms. These characteristics are converted into numerical feature vectors, integrated into a feature matrix after standardization, and used to calculate the comprehensive score of cultural elements.

[0156] S7.4. Set the comprehensive score threshold Ф according to historical assessment data;

[0157] When R≥Ф, it indicates that the cultural element has a positive impact on airport design;

[0158] When R<Ф, it indicates that the influence of the cultural element on airport design is insufficient and needs to be further optimized.

[0159] It should be noted that by collecting and analyzing past cultural empowerment airport design evaluation data, the comprehensive score distribution of different cultural elements is statistically analyzed. Then, according to actual needs and application scenarios, considering specific goals (such as improving passenger satisfaction, enhancing cultural experience, etc.), the threshold is artificially adjusted to ensure that it meets the current strategic goals and business requirements, ensuring that it can accurately distinguish between significantly positive impacts and situations that need to be optimized, and can also adapt to the changing operating environment and passenger expectations. The lowest score for identifying well-performing cases or a reasonable threshold is determined through expert review.

[0160] It should also be noted that the specific measures for further optimization include re-evaluating the selection of cultural elements, adjusting their display methods, or enhancing the interactive experience with passengers.

[0161] This embodiment also provides a computer device, which is applicable to a cross-modal data processing method for cultural empowerment airport design evaluation, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement a cross-modal data processing method for cultural empowerment airport design evaluation as proposed in the above embodiment.

[0162] This computer device can be a terminal. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner. The wireless manner can be achieved through WIFI, a carrier network, NFC (Near Field Communication), or other technologies. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covered on the display screen, or a button, a trackball, or a touchpad set on the computer device housing, or an external keyboard, a touchpad, or a mouse, etc.

[0163] This embodiment also provides a storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the cross-modal data processing method for cultural empowerment airport design evaluation as proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, abbreviated as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, abbreviated as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, abbreviated as EPROM), programmable read-only memory (Programmable Red-Only Memory, abbreviated as PROM), read-only memory (Read-Only Memory, abbreviated as ROM), magnetic memory, flash memory, magnetic disk or optical disc.

[0164] In summary, the present invention solves the problem of time misalignment of multi-source heterogeneous data through the adaptive time alignment algorithm, and uses the convolutional neural network model to automatically detect and repair low-quality or noisy data points, so as to obtain high-quality multi-source evaluation data. Based on this, a multi-dimensional semantic space is constructed and consistent multi-source evaluation data is extracted. Finally, the influence of cultural elements on airport design empowered by culture is accurately evaluated through the comprehensive scoring mechanism. This method not only significantly improves the accuracy and reliability of data processing, but also provides a scientific, comprehensive and efficient evaluation tool for cultural empowerment airport design, ensuring the high quality and credibility of evaluation results.

[0165] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. A cross-modal data processing method for culturally empowered airport design evaluation, characterized by: include, Collect multi-source evaluation data and pre-process the multi-source evaluation data; Based on the preprocessed multi-source evaluation data, the time-aligned multi-source evaluation data is obtained through an adaptive time alignment algorithm; Based on the convolutional neural network model, a multi-source quality enhancement model is constructed, and the time-aligned multi-source evaluation data is input into the multi-source quality enhancement model to obtain high-quality multi-source evaluation data; Based on high-quality multi-source evaluation data, a multidimensional semantic space is constructed to obtain location information and association rules in the multidimensional semantic space; Obtain consistent multi-source evaluation data by analyzing location information and association rules; Based on consistent multi-source evaluation data, the cultural elements of cultural empowerment airports are evaluated through a comprehensive scoring mechanism.

2. A cross-modal data processing method for cultural empowerment airport design evaluation as claimed in claim 1, characterized in that: The multi-source evaluation data includes social media comments, online user feedback, news reports, and passenger surveys; The preprocessing of multi-source evaluation data includes data cleaning, data labeling, standardization and filling in missing values.

3. A cross-modal data processing method for cultural empowerment airport design evaluation as claimed in claim 2, characterized in that: The method of obtaining time-aligned multi-source evaluation data based on the pre-processed multi-source evaluation data by using an adaptive time alignment algorithm is as follows: By extracting the timestamps of the preprocessed multi-source evaluation data and converting them into a time series form in a chronological order; Calculate the time difference between adjacent timestamps in the multi-source evaluation data converted into a time series form; Based on the time difference, the control signal is obtained through PID, and the expression is: Among them, u(t) is the adjustment control signal, D z is the proportional coefficient of PID, D o is the integral coefficient of PID, D d is the differential coefficient of PID; Dynamically adjust the time-sensitive coefficient through a control signal; Based on the dynamic adjustment of the time sensitivity coefficient, each timestamp is adaptively adjusted. The expression is: Among them, t' i is the adjusted timestamp, t0 is the timestamp reference point, α is the dynamic adjustment time sensitivity coefficient, L is the factor affecting time alignment, and f is the transformation function of the time difference; Based on the adjusted timestamps, the time attributes of the multi-source evaluation data are adjusted, and the adjusted attributes are integrated to obtain time-aligned multi-source evaluation data.

4. A cross-modal data processing method for cultural empowerment airport design evaluation as claimed in claim 3, characterized in that: The multi-source quality enhancement model is constructed based on the convolutional neural network model. The specific steps are as follows: Based on the convolutional neural network model; The input layer receives time-aligned historical multi-source evaluation data; The convolution layer adds multiple convolution layers and uses convolution kernels of different sizes to capture different types of time-aligned historical multi-source evaluation data features; The pooling layer gradually reduces the feature dimension through the maximum pooling operation; The fully connected layer maps the features after convolution and pooling to a high-dimensional feature space; The output layer outputs preliminary results through the activation function; The quality assessment layer detects and repairs the preliminary results and obtains high-quality historical multi-source evaluation data; Finally, a multi-source quality enhancement model was constructed.

5. A cross-modal data processing method for cultural empowerment airport design evaluation as claimed in claim 4, characterized in that: The time-aligned multi-source evaluation data is input into the multi-source quality enhancement model to obtain high-quality multi-source evaluation data. The specific steps are as follows: The time-aligned multi-source evaluation data is processed through convolutional layers, pooling layers, and fully connected layers, and combined with activation functions to obtain preliminary high-quality multi-source evaluation data, which is expressed as: y'=g(σ(W H ·p(σ(W k *V+b k ))+b H )); Among them, y' is the preliminary high-quality multi-source evaluation data, W H is the weight matrix of the fully connected layer, W k is the weight matrix of the convolutional layer, V is the time-aligned multi-source evaluation data, b k is the bias term of the convolutional layer, b H is the weight matrix of the fully connected layer, g is the activation function softmax, σ is the activation function ReLU, p is the maximum pooling operation, and * is the convolution operation; The quality comprehensive score of preliminary high-quality multi-source evaluation data is calculated through the quality assessment mechanism, and the expression is: Among them, Q is the comprehensive quality score, G is the global average pooling operation, S is the soft weighted exponential linear unit function, m is the total number of convolution kernels, j is the index variable of the convolution kernel, A is the rectified linear unit, K j is the jth convolution kernel, α is the time-sensitive coefficient, F is the additional eigenvector, and β is the balancing factor; Based on the comprehensive quality score distribution, define the quality threshold θ; When Q ≥ θ, it means that the preliminary high-quality multi-source evaluation data is high-quality data; When Q<θ, it means that the initial high-quality multi-source evaluation data is low-quality data, and the low-quality data needs to be repaired to achieve Q≥θ; Based on the quality judgment results, obtain high-quality multi-source evaluation data.

6. A cross-modal data processing method for cultural empowerment airport design evaluation as claimed in claim 5, characterized in that: The method of constructing a multidimensional semantic space based on high-quality multi-source evaluation data and obtaining location information and association rules in the multidimensional semantic space is as follows: Based on high-quality multi-source evaluation data, semantic features are generated through text analysis and TF-IDF feature extraction method; The semantic features are mapped to a low-dimensional multidimensional semantic space through dimensionality reduction technology to construct a multidimensional semantic space; The location of each high-quality multi-source evaluation data in the multidimensional semantic space is analyzed and encoded by a clustering algorithm to obtain the location information in the multidimensional semantic space; Frequent item set mining and association rule learning algorithms are applied to extract association rules in multidimensional semantic space.

7. A cross-modal data processing method for cultural empowerment airport design evaluation as claimed in claim 6, characterized in that: The method of obtaining consistency multi-source evaluation data by analyzing location information and association rules is as follows: Based on the location information, analyze the similarity between each multi-source evaluation data and the neighboring points; Based on the similarity analysis results, obtain the location information consistency score; Apply the extracted association rules to different multi-source evaluation data, and analyze the support and support rate of the association rules in multi-source evaluation data from different sources; Based on the support and support rate of different multi-source evaluation data, the universality and reliability of different multi-source evaluation data are evaluated; Obtain association rule consistency scores based on the universality and reliability of different multi-source evaluation data; Perform weighted summation of the location information consistency score and the association rule consistency score to obtain a comprehensive consistency score; Obtain consistent multi-source evaluation data through comprehensive consistency scoring.

8. A cross-modal data processing method for cultural empowerment airport design evaluation as claimed in claim 7, characterized in that: The above-mentioned steps of evaluating the cultural elements of cultural empowerment airports based on consistent multi-source evaluation data and through a comprehensive scoring mechanism are as follows: Through natural language processing technology and sentiment analysis algorithms, the cultural characteristics of consistent multi-source evaluation data are extracted, and the word embedding method is used to convert the cultural characteristics into numerical feature vectors; Standardize the numerical feature vectors and integrate the standardized numerical feature vectors into a feature matrix; Based on the feature matrix, the comprehensive score of cultural elements is calculated through a comprehensive scoring mechanism, and the expression is: Among them, R is the comprehensive score of cultural elements, p is the total number of cultural elements, u is the index variable of cultural elements, q is the total number of influencing factors, v is the index variable of influencing factors, and w is the index variable of cultural elements. v is the weight of the vth influencing factor, z uv is the standardized characteristic value of the u-th cultural element under the v-th influencing factor, f(z uv ) is the nonlinear transformation function of the eigenvalue, g(z uv ) is the square root function of the eigenvalue, h(z uv ) is the absolute value function of the eigenvalue; According to historical evaluation data, set the comprehensive scoring threshold Ф; When R ≥ Ф, it means that cultural elements have a positive impact on airport design; When R<Ф, it means that the influence of cultural elements on airport design is insufficient and needs further optimization.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the processor implements the steps of a cross-modal data processing method for cultural empowerment airport design evaluation as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a cross-modal data processing method for cultural empowerment airport design evaluation as described in any one of claims 1 to 8 are implemented.

Images