Railway hub passenger transport demand prediction system and method based on multi-dimensional analysis

Abstract

The invention relates to the technical field of traffic transportation, in particular to a railway hub passenger transport demand prediction system and method based on multi-dimensional analysis, and the method comprises the following steps: S1, multi-dimensional data collection and preprocessing: collecting multi-dimensional data related to railway hub passenger transport demands; s2, key factor analysis: analyzing the influence of key factors on the passenger transport demand in different time and space ranges; s3, regional linkage prediction: capturing a passenger flow dependency relationship of the weighted graph in a spatial dimension through a graph neural network model, and predicting a passenger transport demand of regional linkage in combination with a key factor analysis result; and S4, prediction result verification and resource optimization suggestion: generating the resource optimization suggestion for the railway hub. According to the method, specific strategies such as train number adjustment, platform scheduling optimization and peak period resource allocation are generated for railway hub operation, passenger flow fluctuation can be flexibly adapted, and resource allocation is dynamically optimized.

Description

Technical Field

[0001] The present invention relates to the field of transportation technology, and in particular to a railway hub passenger transport demand forecasting system and method based on multi-dimensional analysis. Background Art

[0002] As an important part of the transportation system, railway hubs carry the functions of gathering, distributing and transferring large-scale passenger flows, and are of great significance to regional economic development and social travel. However, with the acceleration of urbanization and the increasing complexity of passenger demand, railway hubs face problems such as increased pressure during peak passenger flow periods and uneven resource allocation in their operation and management. How to accurately predict passenger demand and reasonably allocate resources has become a key challenge in the operation and management of railway hubs.

[0003] At present, most railway hub passenger demand forecasting methods are mainly based on simple modeling based on single-dimensional data (such as historical passenger flow or geographical characteristics), and fail to fully utilize the comprehensive impact of multi-dimensional data (such as weather, holidays, traffic flow and socio-economic indicators) on passenger demand. In addition, existing technologies do not adequately consider the spatial linkage relationship between regions and cannot accurately capture the complex passenger flow dependence characteristics of railway hubs and surrounding areas. At the same time, resource optimization strategies usually rely on static planning and lack the ability to adapt to dynamic demand changes, resulting in significant limitations in prediction accuracy and resource allocation efficiency.

[0004] The present invention aims to provide a railway hub passenger demand forecasting system and method based on multi-dimensional analysis, improve the railway hub operation efficiency and service quality, and provide strong support for scientific decision-making and dynamic management. Summary of the invention

[0005] The present invention provides a railway hub passenger transport demand forecasting system and method based on multi-dimensional analysis.

[0006] A railway hub passenger transport demand forecasting method based on multi-dimensional analysis includes the following steps:

[0007] S1, multi-dimensional data collection and preprocessing: Collect multi-dimensional data related to railway hub passenger demand, including historical passenger flow data, weather data, holiday information, surrounding traffic flow, and socio-economic indicators, and preprocess the collected multi-dimensional data, including data cleaning and normalization;

[0008] S2, key factor analysis: extract the key factors that affect the passenger demand of railway hubs from the pre-processed multi-dimensional data, and analyze the impact of key factors on passenger demand in different time and space ranges;

[0009] S3, regional linkage prediction: construct a dynamic weighted graph with railway hubs and surrounding areas as nodes and transportation networks and passenger flow relationships as edges. The graph neural network (GNN) model is used to capture the passenger flow dependency relationship of the weighted graph in the spatial dimension. Combined with the results of key factor analysis, the passenger demand of regional linkage is predicted.

[0010] S4, prediction result verification and resource optimization suggestions: The results of regional linkage prediction are compared with the actual operation data, and the prediction accuracy is evaluated using residual analysis and mean square error (MSE). Resource optimization suggestions for railway hubs are generated, including train adjustment, platform scheduling optimization, and peak resource allocation strategies.

[0011] Optionally, the multi-dimensional data collection and preprocessing in S1 includes:

[0012] S11, historical passenger flow data collection: collect passenger entry and exit data through the ticketing system and gate records of the railway hub, and generate daily passenger flow time series according to the time series;

[0013] S12, weather data collection: calling the meteorological data interface to obtain weather data in the railway hub area, including temperature, precipitation, and wind speed;

[0014] S13, holiday information collection: Combine national holidays and local festivals and activities information to generate holiday impact factors, and represent them with binary variables, where holidays are 1 and non-holidays are 0;

[0015] S14, surrounding traffic flow: obtain surrounding road traffic flow and subway transfer station passenger flow through the urban traffic management department interface;

[0016] S15, collection of social and economic indicators: collect economic indicators related to the area where the railway hub is located, including GDP growth rate, per capita income, and unemployment rate;

[0017] S16, data cleaning: missing values ​​were processed using linear interpolation, and outliers were detected and replaced using the 3-times standard deviation method;

[0018] S17, normalization processing: normalize the collected multi-dimensional data to the range of [0,1].

[0019] Optionally, the key factor analysis in S2 includes:

[0020] S21, extract key factors: extract key factors that affect the passenger demand of railway hubs from pre-processed multi-dimensional data based on the linear regression model, including:

[0021] Construct a linear regression model: The preprocessed multidimensional data X = {x 1 ,x 2,...,x n} and the target variable y (railway hub passenger demand) as input to construct a linear regression model;

[0022] Key factor screening: Based on the regression coefficient of the linear regression model, multi-dimensional data with regression coefficients exceeding the preset threshold are screened as key factors affecting y;

[0023] S22, analyze the impact of key factors in the time dimension: perform time series analysis on the extracted key factors, calculate the importance weights of key factors in different time windows, and calculate the time-weighted impact value W through the sliding window method. t (x i );

[0024] S23, analyze the impact of key factors in the spatial dimension: Based on the geographical division of railway hubs and surrounding areas, aggregate the impact of key factors by region and calculate the weighted impact value W of different regions s (x i );

[0025] S24, construction of factor feature matrix: Based on the extracted key factors, a standardized factor feature matrix F is generated by weighted analysis of the time dimension and space dimension. key .

[0026] Optionally, the regional linkage prediction in S3 includes:

[0027] S31, Dynamic Weighted Graph Construction: Construct a dynamic weighted graph based on the geographical scope and transportation network characteristics of the railway hub and its surrounding areas;

[0028] S32, spatial dependency modeling: Based on a dynamic weighted graph, the spatial dependency relationship of nodes in the graph is modeled using a graph neural network (GNN) model;

[0029] S33, Regional Interconnection Passenger Transport Demand Forecast: Utilize the node embedding features generated by the graph neural network and combine the results of key factor analysis to predict the regional interconnection passenger transport demand.

[0030] Optionally, the dynamic weighted graph construction in S31 includes:

[0031] S311, determine the basic elements of the dynamic weighted graph: take the railway hub and its surrounding areas as the node set V = {v 1 ,v 2 ,...,v n}, with the traffic network and passenger flow relationship as the edge set E = {e ij}, each edge e ij Represents the connection between two areas (such as passenger commuting paths or geographical adjacency), for each edge eij Assign dynamic weight w ij , indicating the strength of the relationship between regions;

[0032] S312, calculation of dynamic weight: dynamic weight w ij Calculation is based on the intensity of inter-regional passenger flow and the characteristics of the transportation network;

[0033] S313, the complete definition of dynamic weighted graph: construct a dynamic weighted graph G = (V, E, W), where V is the node set, E is the edge set, and W = {w ij} is the dynamic weight matrix of the edge.

[0034] Optionally, the spatial dependency modeling in S32 includes:

[0035] S321, node feature update: Based on the dynamic weighted graph, the graph convolutional network (GCN) model is used to model spatial dependencies. The node features of each layer are updated by aggregating the information of neighboring nodes.

[0036] S322, multi-layer feature propagation and embedding generation: By stacking multi-layer graph convolutional networks, the features of each node are aggregated for multiple rounds to capture spatial dependencies and finally generate node embedding features H = {h 1 ,h 2 ,...,h l}, used for regional linkage passenger transport demand forecasting.

[0037] Optionally, the regional linkage passenger transport demand forecast in S33 includes:

[0038] S331, fusion of key factor features: the node embedding feature H and the factor feature matrix F generated by key factor analysis key Fusion is performed to form the input feature matrix Z;

[0039] S332, regional passenger demand prediction: input the fused input feature matrix Z to the multi-layer perceptron (MLP) model to perform regional linkage passenger demand prediction.

[0040] Optionally, the prediction result verification and resource optimization suggestions in S4 include:

[0041] S41, compare the regional linkage prediction results with the actual operation data: compare the passenger demand results predicted by regional linkage with the actual operation data by region, and calculate the residual e for each region i ;

[0042] S42, evaluate the prediction accuracy: use residual analysis and mean square error (MSE) to quantitatively evaluate the prediction accuracy, including:

[0043] Residual analysis: Statistical residual distribution of all regions and calculate the mean of the residuals and standard deviation σ e , to assess the bias and volatility of forecast results;

[0044] Mean square error: Calculates the overall error of the prediction results, which is used to measure the prediction accuracy;

[0045] S43, generate resource optimization suggestions: Generate resource optimization suggestions for railway hubs based on the residuals of each area and the evaluation results of prediction accuracy.

[0046] Optionally, the generating resource optimization suggestion in S43 includes:

[0047] S431, Train adjustment: When the residual e i The demand in the region is higher than the forecast value, that is, e i >0, it is recommended to increase the number of trains during peak hours. i <0, the train interval is optimized to reduce the vacancy rate;

[0048] S432, platform scheduling optimization: adjust the order and frequency of train platform use based on actual operation data and regional linkage predicted passenger demand results;

[0049] S433, Peak-period resource allocation strategy: Increase resources (such as manpower and train equipment) during peak hours and in high-demand areas.

[0050] A railway hub passenger transport demand prediction system based on multi-dimensional analysis, used to implement the above-mentioned railway hub passenger transport demand prediction method based on multi-dimensional analysis, includes the following modules:

[0051] Data collection module: collects multi-dimensional data related to railway hub passenger demand, including historical passenger flow data, weather data, holiday information, surrounding traffic flow and socio-economic indicators;

[0052] Data preprocessing module: cleans and normalizes the collected multi-dimensional data;

[0053] Key factor analysis module: Analyze the key factors that affect the passenger demand of railway hubs based on the extracted multi-dimensional data, and analyze the impact of key factors in combination with time and space scope;

[0054] Regional linkage prediction module: By constructing a dynamic weighted graph with railway hubs and surrounding areas as nodes and transportation networks and passenger flow relationships as edges, the graph neural network model is used to capture passenger flow dependencies in the spatial dimension, and combined with the results of key factor analysis, the passenger demand of regional linkage is predicted;

[0055] Result verification and optimization module: Compare the results of regional linkage prediction with actual operating data, evaluate the prediction accuracy through residual analysis and mean square error, and generate resource optimization suggestions, including train adjustment, platform scheduling optimization and peak resource allocation strategy.

[0056] Beneficial effects of the present invention:

[0057] The present invention realizes the accurate prediction of passenger demand in railway hubs through comprehensive collection and preprocessing of multi-dimensional data, combined with key factor analysis and regional linkage prediction methods. It dynamically screens key factors that have a significant impact on passenger demand based on linear regression models, ensuring the scientificity and accuracy of input data. At the same time, it combines graph neural networks to model the spatial dependency relationship of dynamic weighted graphs, effectively capturing the complex passenger flow linkage characteristics of railway hubs and their surrounding areas, and providing a high-quality data foundation and accurate model input for demand forecasting.

[0058] The present invention comprehensively reflects the dynamic characteristics of passenger demand by integrating weighted analysis of key factors of time and space, verifies the prediction accuracy by using residual analysis and mean square error, dynamically adjusts the prediction model to ensure the reliability of the prediction results, and generates specific strategies such as train adjustment, platform scheduling optimization and peak resource allocation for railway hub operations based on the comparison between regional predictions and actual demand. It can flexibly adapt to passenger flow fluctuations and dynamically optimize resource allocation.

[0059] The present invention can effectively improve the operational efficiency and service quality of railway hubs through the implementation of resource optimization suggestions. The train adjustment avoids resource waste and passenger backlogs. Platform scheduling optimization realizes priority allocation of high-demand areas and enhances platform utilization efficiency. The peak resource allocation strategy specifically alleviates the operational pressure in key areas and time periods. It not only improves the refined management capability of railway hub operations, but also provides strong support for scientific decision-making, and has broad practical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0061] Figure 1 A schematic diagram of a prediction method flow chart of an embodiment of the present invention;

[0062] Figure 2 Schematic diagram of system function modules according to an embodiment of the present invention. DETAILED DESCRIPTION

[0063] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0064] It should be noted that the references to "one embodiment", "an embodiment", "an exemplary embodiment", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0065] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0066] like Figure 1 As shown, a railway hub passenger demand forecasting method based on multi-dimensional analysis includes the following steps:

[0067] S1, multi-dimensional data collection and preprocessing: Collect multi-dimensional data related to railway hub passenger demand, including historical passenger flow data, weather data, holiday information, surrounding traffic flow, and socio-economic indicators, and preprocess the collected multi-dimensional data, including data cleaning and normalization;

[0068] S2, key factor analysis: extract the key factors that affect the passenger demand of railway hubs from the pre-processed multi-dimensional data, and analyze the impact of key factors on passenger demand in different time and space ranges;

[0069] S3, regional linkage prediction: construct a dynamic weighted graph with railway hubs and surrounding areas as nodes and transportation networks and passenger flow relationships as edges. The graph neural network (GNN) model is used to capture the passenger flow dependency relationship of the weighted graph in the spatial dimension. Combined with the results of key factor analysis, the passenger demand of regional linkage is predicted.

[0070] S4, prediction result verification and resource optimization suggestions: compare the results of regional linkage prediction with actual operation data, use residual analysis and mean square error (MSE) to evaluate the prediction accuracy, and generate resource optimization suggestions for railway hubs, including train adjustment, platform scheduling optimization, and peak resource allocation strategy;

[0071] Through the above content, accurate prediction and dynamic adjustment of passenger demand in railway hubs have been achieved, which not only improves the accuracy and adaptability of predictions, but also generates resource optimization strategies based on actual needs, providing strong support for efficient management and scientific decision-making of railway hub operations.

[0072] Multi-dimensional data collection and preprocessing in S1 include:

[0073] S11, historical passenger flow data collection: collect passenger entry and exit data through the ticketing system and gate records of the railway hub, and generate daily passenger flow time series according to the time series;

[0074] S12, weather data collection: calling the meteorological data interface to obtain weather data in the railway hub area, including temperature, precipitation, and wind speed;

[0075] S13, holiday information collection: Combine national holidays and local festivals and activities information to generate holiday impact factors, and represent them with binary variables, where holidays are 1 and non-holidays are 0;

[0076] S14, surrounding traffic flow: obtain surrounding road traffic flow and subway transfer station passenger flow through the urban traffic management department interface;

[0077] S15, collection of social and economic indicators: collect economic indicators related to the area where the railway hub is located, including GDP growth rate, per capita income, and unemployment rate;

[0078] S16, data cleaning: linear interpolation is used to process missing values, and the 3 times standard deviation method is used to detect and replace outliers, expressed as:

[0079] Missing value handling: When x t When missing;

[0080] Among them, x t is the data value at time point t, x t-1 is the data value at time point t-1, x t+1 is the data value at time point t+1;

[0081] Outlier handling:

[0082] Among them, μ is the mean of the current data sequence, and σ is the standard deviation of the current data sequence;

[0083] S17, normalization processing: normalize the collected multi-dimensional data to the range of [0,1] to eliminate the influence of dimensional differences on subsequent analysis, expressed as:

[0084]

[0085] Among them, x min and x max are the minimum and maximum values ​​of the data, respectively. is the normalized data;

[0086] Through the above content, the factors affecting the passenger demand of railway hubs are fully captured, and the linear interpolation method and the three-times standard deviation method are used to effectively deal with the missing and abnormal data problems to ensure the integrity and consistency of the data. The normalization process eliminates the dimensional differences and improves the comparability of the data.

[0087] The key factor analyses in S2 include:

[0088] S21, extract key factors: extract key factors that affect the passenger demand of railway hubs from pre-processed multi-dimensional data based on the linear regression model, including:

[0089] Construct a linear regression model: The preprocessed multidimensional data X = {x 1 ,x 2 ,...,x n} and the target variable y (railway hub passenger demand) as input to construct a linear regression model, which is expressed as:

[0090]

[0091] Among them, y is the target variable (railway hub passenger demand), x i is the i-th multidimensional data, β 0 is the intercept term, β i For x i The regression coefficient of x i The impact strength on y, ∈ is the error term, and n is the total number of multi-dimensional data;

[0092] Key factor screening: According to the regression coefficient of the linear regression model, multi-dimensional data with regression coefficients exceeding the preset threshold are screened as key factors affecting y, expressed as:

[0093] |β i |>Threshold;

[0094] Among them, Threshold is the preset threshold;

[0095] The preset threshold Threshold is set based on historical data, including:

[0096] Collect historical data: Collect historical data related to passenger demand at railway hubs, including multi-dimensional feature data (such as weather data, holiday information, surrounding traffic flow, etc.) and passenger demand data;

[0097] Train linear regression model: Use historical data to train a linear regression model;

[0098] Calculate the distribution of regression coefficients: Count all regression coefficients | β i The distribution characteristics of |, including mean and standard deviation, are expressed as:

[0099]

[0100] Among them, μ β is the mean of the regression coefficient, σ β is the standard deviation of the regression coefficient, p is the total number of features (variables) used in the model;

[0101] Threshold setting: According to the distribution of regression coefficient, the threshold is set, expressed as:

[0102] Threshold=μ β +k·σ β ;

[0103] Wherein, k is a tuning parameter (take k=1 or k=2);

[0104] S22, analyze the impact of key factors in the time dimension: perform time series analysis on the extracted key factors, calculate the importance weights of key factors in different time windows, and calculate the time-weighted impact value W through the sliding window method. t (x i ), expressed as:

[0105]

[0106] Among them, t is the current time point, m is the sliding window size, β i (k) is the key factor x i The influence coefficient of time k on y;

[0107] S23, analyze the impact of key factors in the spatial dimension: Based on the geographical division of railway hubs and surrounding areas, aggregate the impact of key factors by region and calculate the weighted impact value W of different regions s (x i ), expressed as:

[0108]

[0109] Among them, R j is the jth region, l is the total number of divided regions, w j is the weight of the jth region, β i (R j ) is the key factor x i In the jth region R j Regression coefficient of impact on passenger demand;

[0110] S24, construction of factor feature matrix: Based on the extracted key factors, a standardized factor feature matrix F is generated by weighted analysis of the time dimension and space dimension. key , expressed as:

[0111]

[0112] Among them, γ t , γ s They are the adjustment parameters of time and space weights (satisfying γ t +γ s =1);

[0113]

[0114] Through the above content, the key factors that have a significant impact on the passenger demand of railway hubs are extracted based on the linear regression model, and the screening threshold is dynamically set in combination with the distribution of regression coefficients to ensure that the extracted factors are scientific and accurate. The analysis also takes into account the dynamic effects of key factors in time and space, revealing the global and local impact of factors on demand. It can significantly improve the prediction accuracy of the model and reduce the impact of redundant features on the complexity of the model, providing a high-quality data foundation for the dynamic prediction of passenger demand in railway hubs.

[0115] The regional linkage forecast in S3 includes:

[0116] S31, Dynamic Weighted Graph Construction: Construct a dynamic weighted graph based on the geographical scope and transportation network characteristics of the railway hub and its surrounding areas;

[0117] S32, spatial dependency modeling: Based on a dynamic weighted graph, the spatial dependency relationship of nodes in the graph is modeled using a graph neural network (GNN) model;

[0118] S33, Regional linkage passenger transport demand prediction: Utilize the node embedding features generated by graph neural network and combine the results of key factor analysis to predict regional linkage passenger transport demand;

[0119] Through the above content, we fully utilize the geographical and transportation network characteristics of the railway hub and its surrounding areas, combine the graph neural network model's ability to model complex relationships between regions, accurately capture the passenger flow dependency in the spatial dimension, and combine the results of key factor analysis to further optimize the input characteristics of the prediction model. This not only improves the accuracy and reliability of the prediction, but also can dynamically adapt to changes in passenger flow between regions, providing a scientific basis for resource scheduling and passenger management of railway hubs.

[0120] The dynamic weighted graph construction in S31 includes:

[0121] S311, determine the basic elements of the dynamic weighted graph: take the railway hub and its surrounding areas as the node set V = {v 1 ,v 2 ,...,v n}, with the traffic network and passenger flow relationship as the edge set E = {e ij}, each edge e ij Represents the connection between two areas (such as passenger commuting paths or geographical adjacency), for each edge e ij Assign dynamic weight w ij , indicating the strength of the relationship between regions;

[0122] S312, calculation of dynamic weight: dynamic weight w ij The calculation is based on the inter-regional passenger flow intensity and traffic network characteristics, expressed as:

[0123] w ij =α·f flow (i,j)+β·f distance (i,j);

[0124]

[0125] Among them, f flow (i,j) is region v i and v j The passenger flow intensity between distance (i,j) is region v i and v j The geographical or transportation distance between them, distance(i,j) is the distance between the two regions, d mean is the mean of the distances between all regions, α and β are the parameters for adjusting the weights;

[0126] S313, the complete definition of dynamic weighted graph: construct a dynamic weighted graph G = (V, E, W), where V is the node set, E is the edge set, and W = {w ij} is the dynamic weight matrix of the edge;

[0127] Through the above content, the geographical scope and transportation network characteristics of the railway hub and surrounding areas are comprehensively considered. By introducing dynamic weights such as passenger flow intensity and spatial distance, the actual connection and dependency relationship between regions are accurately portrayed. By updating the weights in real time, the temporal and spatial characteristics of regional changes and passenger demand can be dynamically reflected. This not only improves the adaptability of the prediction model to dynamic changes, but also enhances the prediction accuracy and application value of the model, making it more suitable for complex traffic scenarios and dynamic passenger flow environments.

[0128] Spatial dependency modeling in S32 includes:

[0129] S321, node feature update: Based on the dynamic weighted graph, the graph convolutional network (GCN) model is used to model spatial dependencies. The node features of each layer are updated by aggregating the information of neighboring nodes, which is expressed as:

[0130]

[0131] in, is the k-th layer node v i The feature vector of node v i The set of neighbor nodes, w ij For node v i and v j The edge weight between i =∑ j∈N(i) w ij For node v i Degree, W (k) is the learnable parameter matrix of the kth layer, σ is the ReLU activation function;

[0132] S322, multi-layer feature propagation and embedding generation: By stacking multi-layer graph convolutional networks, the features of each node are aggregated for multiple rounds to capture spatial dependencies and finally generate node embedding features H = {h 1 ,h 2 ,...,h l}, used for regional linkage passenger transport demand forecasting;

[0133] Through the above content, the complex spatial dependency relationship between railway hubs and surrounding areas is effectively captured, and dynamic weights are used to reflect the passenger flow intensity and spatial distance between regions. Combined with the multi-layer graph convolutional network to aggregate neighbor node features, it not only achieves accurate modeling of local regional relationships, but also can be expanded to global linkage characteristics. This method is highly adaptable and can dynamically respond to regional changes.

[0134] The regional linkage passenger demand forecast in S33 includes:

[0135] S331, fusion of key factor features: the node embedding feature H and the factor feature matrix F generated by key factor analysis key Fusion is performed to form the input feature matrix Z, which is expressed as:

[0136] Z=Concat(H,F key );

[0137] Among them, Z∈R n×(d+k) represents the final fusion feature matrix, d is the dimension of the node embedding feature, k is the dimension of the key factor feature, and Concat represents the concatenation operation of the feature;

[0138] S332, regional passenger demand prediction: input the fused input feature matrix Z to the multi-layer perceptron (MLP) model to perform regional linkage passenger demand prediction, specifically including:

[0139] The first layer of linear transformation: Z' = σ(W 1 Z+b 1 );

[0140] Among them, W 1 is the weight matrix of the first layer, b 1 is the bias vector of the first layer, σ is the ReLU activation function, and Z' is the activation output matrix after linear transformation of the first layer;

[0141] Output layer prediction: y i =W 2 Z'+b 2 ;

[0142] Among them, y i ={y 1 ,y 2 ,...,y l} is the predicted regional passenger demand, y 1 ,y 2 ,...,y l Represents node v 1 ,v 2 ,...,v l forecast demand;

[0143] Through the above content, combined with the node embedding features and key factor analysis results generated by the graph neural network, the complex spatial dependency relationship and multi-dimensional influencing factors between the railway hub and the surrounding areas are accurately modeled, and the refined prediction of regional passenger demand is achieved. It can not only capture the dynamic changes within and outside the region, but also provide clear guidance for regional optimization.

[0144] The prediction result verification and resource optimization suggestions in S4 include:

[0145] S41, compare the regional linkage prediction results with the actual operation data: compare the passenger demand results predicted by regional linkage with the actual operation data by region, and calculate the residual e for each region i , expressed as:

[0146]

[0147] Among them, e i is the residual value of the i-th region, is the actual operation data of the ith region, is the predicted passenger demand value of the i-th area;

[0148] S42, evaluate the prediction accuracy: use residual analysis and mean square error (MSE) to quantitatively evaluate the prediction accuracy, including:

[0149] Residual analysis: Statistical residual distribution of all regions and calculate the mean of the residuals and standard deviation σ e , to evaluate the bias and volatility of the forecast results, expressed as:

[0150]

[0151]

[0152] Where, l is the total number of regions;

[0153] Mean square error: Calculates the overall error of the prediction results, which is used to measure the prediction accuracy and is expressed as:

[0154]

[0155] Among them, MSE is the mean square error;

[0156] S43, generating resource optimization suggestions: generating resource optimization suggestions for railway hubs based on the residuals of each region and the evaluation results of the prediction accuracy;

[0157] Through the above content, regional-level forecast errors can be accurately located, thereby identifying deviations and dynamic changes in passenger demand, and generating targeted resource optimization suggestions to ensure efficient allocation and flexible scheduling of railway hub resources. This not only improves the reliability of the forecasting model, but also significantly enhances the ability of railway hubs to respond to changes in actual operational needs, providing strong support for improving overall operational efficiency.

[0158] Suggestions for optimizing generated resources in S43 include:

[0159] S431, Train adjustment: When the residual e i The demand in the region is higher than the forecast value, that is, e i>0, it is recommended to increase the number of trains during peak hours. i <0, the train interval is optimized to reduce the vacancy rate;

[0160] S432, platform scheduling optimization: adjust the order and frequency of train platform use based on actual operation data and regional linkage predicted passenger demand results, including:

[0161] Determine the priority of station usage: Determine the priority based on actual needs, expressed as:

[0162]

[0163] Among them, P i is the priority of station i, Actual operational data for station i services, Actual operational data serving station j;

[0164] Dynamically adjust the order of platform use: re-arrange the order of platform use based on the actual operation data and predicted demand of the trains, and give priority to high-demand trains entering and leaving the station. If the passenger demand of a platform during peak hours is significantly higher than that of other platforms, it will be dispatched to a platform close to the main entrance or exit or with a large passenger flow, thus shortening the passenger transfer time;

[0165] Adjust platform usage frequency: If the actual operating data of platform i service is higher than the predicted passenger demand value, increase the platform usage frequency, that is, allow more trains to stop;

[0166] S433, peak resource allocation strategy: increase resources (such as manpower and train equipment) during peak hours and in high-demand areas to ease operational pressure;

[0167] Through the above content, it is possible to flexibly increase or decrease the number of flights according to demand fluctuations to avoid waste of resources or passenger backlogs; platform scheduling optimization ensures that platform resources are allocated first to high-demand areas to improve platform utilization efficiency; peak period resource allocation strategies specifically increase resource investment in key areas and time periods to alleviate operational pressure and improve overall service levels.

[0168] like Figure 2 As shown, a railway hub passenger demand prediction system based on multi-dimensional analysis is used to implement the above-mentioned railway hub passenger demand prediction method based on multi-dimensional analysis, including the following modules:

[0169] Data collection module: collects multi-dimensional data related to railway hub passenger demand, including historical passenger flow data, weather data, holiday information, surrounding traffic flow and socio-economic indicators;

[0170] Data preprocessing module: cleans and normalizes the collected multi-dimensional data;

[0171] Key factor analysis module: Analyze the key factors that affect the passenger demand of railway hubs based on the extracted multi-dimensional data, and analyze the impact of key factors in combination with time and space scope;

[0172] Regional linkage prediction module: By constructing a dynamic weighted graph with railway hubs and surrounding areas as nodes and transportation networks and passenger flow relationships as edges, the graph neural network model is used to capture passenger flow dependencies in the spatial dimension, and combined with the results of key factor analysis, the passenger demand of regional linkage is predicted;

[0173] Result verification and optimization module: Compare the results of regional linkage prediction with actual operating data, evaluate the prediction accuracy through residual analysis and mean square error, and generate resource optimization suggestions, including train adjustment, platform scheduling optimization and peak resource allocation strategy.

[0174] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0175] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A railway hub passenger demand forecasting method based on multi-dimensional analysis, characterized in that: The following steps are involved: S1, multi-dimensional data collection and preprocessing: Collect multi-dimensional data related to railway hub passenger demand, including historical passenger flow data, weather data, holiday information, surrounding traffic flow, and socio-economic indicators, and preprocess the collected multi-dimensional data, including data cleaning and normalization; S2, key factor analysis: extract the key factors that affect the passenger demand of railway hubs from the pre-processed multi-dimensional data, and analyze the impact of key factors on passenger demand in different time and space ranges; S3, regional linkage prediction: construct a dynamic weighted graph with railway hubs and surrounding areas as nodes and transportation networks and passenger flow relationships as edges. The graph neural network model is used to capture the passenger flow dependency relationship of the weighted graph in the spatial dimension. Combined with the results of key factor analysis, the passenger demand of regional linkage is predicted. S4, prediction result verification and resource optimization suggestions: compare the results of regional linkage prediction with actual operation data, use residual analysis and mean square error to evaluate the prediction accuracy, and generate resource optimization suggestions for railway hubs, including train adjustment, platform scheduling optimization and peak resource allocation strategy.

2. The method for predicting passenger demand at a railway hub based on multi-dimensional analysis according to claim 1, characterized in that: The multi-dimensional data collection and preprocessing in S1 includes: S11, historical passenger flow data collection: collect passenger entry and exit data through the ticketing system and gate records of the railway hub, and generate daily passenger flow time series according to the time series; S12, weather data collection: calling the meteorological data interface to obtain weather data in the railway hub area, including temperature, precipitation, and wind speed; S13, holiday information collection: Combine national holidays and local festivals and activities information to generate holiday impact factors, and represent them with binary variables, where holidays are 1 and non-holidays are 0; S14, surrounding traffic flow: obtain surrounding road traffic flow and subway transfer station passenger flow through the urban traffic management department interface; S15, collection of social and economic indicators: collect economic indicators related to the area where the railway hub is located, including GDP growth rate, per capita income, and unemployment rate; S16, data cleaning: missing values ​​were processed using linear interpolation, and outliers were detected and replaced using the 3-times standard deviation method; S17, normalization processing: normalize the collected multi-dimensional data to the range of [0,1].

3. The method for predicting passenger demand at a railway hub based on multi-dimensional analysis according to claim 1, characterized in that: The key factor analysis in S2 includes: S21, extract key factors: extract key factors that affect the passenger demand of railway hubs from pre-processed multi-dimensional data based on the linear regression model, including: Construct a linear regression model: The preprocessed multidimensional data X = {x1, x2, ..., x n } and the target variable y as input to build a linear regression model; Key factor screening: Based on the regression coefficient of the linear regression model, multi-dimensional data with regression coefficients exceeding the preset threshold are screened as key factors affecting y; S22, analyze the impact of key factors in the time dimension: perform time series analysis on the extracted key factors, calculate the importance weights of key factors in different time windows, and calculate the time-weighted impact value W through the sliding window method. t )x i ); S23, analyze the impact of key factors in the spatial dimension: Based on the geographical division of railway hubs and surrounding areas, aggregate the impact of key factors by region and calculate the weighted impact value W of different regions s (x i ); S24, construction of factor feature matrix: Based on the extracted key factors, a standardized factor feature matrix F is generated by weighted analysis of the time dimension and space dimension. key .

4. The method for predicting passenger demand of railway hubs based on multi-dimensional analysis according to claim 3 is characterized in that: The regional linkage prediction in S3 includes: S31, Dynamic Weighted Graph Construction: Construct a dynamic weighted graph based on the geographical scope and transportation network characteristics of the railway hub and its surrounding areas; S32, Spatial Dependency Modeling: Based on the dynamic weighted graph, the spatial dependency relationship of nodes in the graph is modeled using the graph neural network model; S33, Regional Interconnection Passenger Transport Demand Forecast: Utilize the node embedding features generated by the graph neural network and combine the results of key factor analysis to predict the regional interconnection passenger transport demand.

5. The method for predicting passenger demand of railway hubs based on multi-dimensional analysis according to claim 4 is characterized in that: The dynamic weighted graph construction in S31 includes: S311, determine the basic elements of the dynamic weighted graph: take the railway hub and its surrounding areas as the node set V = {v1,v2,...,v n }, with the traffic network and passenger flow relationship as the edge set E = {e ij }, each edge e ij Represents the connection between two regions, for each edge e ij Assign dynamic weight w ij , indicating the strength of the relationship between regions; S312, calculation of dynamic weight: dynamic weight w ij Calculation is based on the intensity of inter-regional passenger flow and the characteristics of the transportation network; S313, the complete definition of dynamic weighted graph: construct a dynamic weighted graph G = (V, E, W), where V is the node set, E is the edge set, and W = {w ij } is the dynamic weight matrix of the edge.

6. The method for predicting passenger demand at a railway hub based on multi-dimensional analysis according to claim 5, characterized in that: The spatial dependency modeling in S32 includes: S321, node feature update: Based on the dynamic weighted graph, the graph convolutional network model is used to model spatial dependencies. The node features of each layer are updated by aggregating the information of neighboring nodes. S322, multi-layer feature propagation and embedding generation: By stacking multi-layer graph convolutional networks, the features of each node are aggregated for multiple rounds to capture spatial dependencies and finally generate node embedding features H = h1,h2,...,h l }, used for regional linkage passenger transport demand forecasting.

7. The method for predicting passenger demand at a railway hub based on multi-dimensional analysis according to claim 6, characterized in that: The regional linkage passenger transport demand forecast in S33 includes: S331, fusion of key factor features: the node embedding feature H and the factor feature matrix F generated by key factor analysis key Fusion is performed to form the input feature matrix Z; S332, regional passenger demand prediction: input the fused input feature matrix Z to the multi-layer perceptron model to perform regional linkage passenger demand prediction.

8. The method for predicting passenger demand at a railway hub based on multi-dimensional analysis according to claim 7, characterized in that: The prediction result verification and resource optimization suggestions in S4 include: S41, compare the regional linkage prediction results with the actual operation data: compare the passenger demand results predicted by regional linkage with the actual operation data by region, and calculate the residual e for each region i ; S42, evaluate the prediction accuracy: use residual analysis and mean square error to quantitatively evaluate the prediction accuracy, including: Residual analysis: Count the residual distribution of all regions and calculate the mean of the residuals and standard deviation σ e , to assess the bias and volatility of forecast results; Mean square error: Calculates the overall error of the prediction results, which is used to measure the prediction accuracy; S43, generate resource optimization suggestions: Generate resource optimization suggestions for railway hubs based on the residuals of each area and the evaluation results of prediction accuracy.

9. A railway hub passenger transport demand forecasting method based on multi-dimensional analysis according to claim 8, characterized in that: The generated resource optimization suggestion in S43 includes: S431, Train adjustment: When the residual e i The demand in the region is higher than the forecast value, that is, e i >0, it is recommended to increase the number of trains during peak hours. i <0, the train interval is optimized to reduce the vacancy rate; S432, platform scheduling optimization: adjust the order and frequency of train platform use based on actual operation data and regional linkage predicted passenger demand results; S433, Peak Resource Allocation Strategy: Increase resources during peak hours and in high-demand areas.

10. A railway hub passenger demand forecasting system based on multi-dimensional analysis, used to implement a railway hub passenger demand forecasting method based on multi-dimensional analysis as claimed in any one of claims 1 to 9, characterized in that: Includes the following modules: Data collection module: collects multi-dimensional data related to railway hub passenger demand, including historical passenger flow data, weather data, holiday information, surrounding traffic flow and socio-economic indicators; Data preprocessing module: cleans and normalizes the collected multi-dimensional data; Key factor analysis module: Analyze the key factors that affect the passenger demand of railway hubs based on the extracted multi-dimensional data, and analyze the impact of key factors in combination with time and space scope; Regional linkage prediction module: By constructing a dynamic weighted graph with railway hubs and surrounding areas as nodes and transportation networks and passenger flow relationships as edges, the graph neural network model is used to capture passenger flow dependencies in the spatial dimension, and combined with the results of key factor analysis, the passenger demand of regional linkage is predicted; Result verification and optimization module: Compare the results of regional linkage prediction with actual operating data, evaluate the prediction accuracy through residual analysis and mean square error, and generate resource optimization suggestions, including train adjustment, platform scheduling optimization and peak resource allocation strategy.

Images