Consumer travel prediction method and device based on commercial link degree

Abstract

The invention provides a consumer travel prediction method based on a commercial link degree, and the method comprises the steps: obtaining consumption travel OD data through constructing a commercial district topology network model, calculating the attraction degree between a node and a main attraction point, and training a neural network model for travel prediction. The method can scientifically evaluate travel behaviors of consumers in a business district and improve prediction accuracy.

Description

Technical Field

[0001] The present invention belongs to the technical field of big data analysis, and in particular relates to a consumer travel prediction method based on commercial connectivity. Background Art

[0002] With the acceleration of urbanization, urban traffic problems have become increasingly prominent, especially traffic congestion and travel efficiency in commercial areas.

[0003] The existing business district format planning is mainly based on experience or reference to similar business districts. However, different business districts have different demographic characteristics, economic characteristics, infrastructure, etc., and simple copy-paste often results in business district formats not adapting to the local conditions.

[0004] Scientifically and accurately predicting the travel behavior of consumers in a business district can provide great help for business district business planning. In the field of business district consumer travel prediction, although some advanced technologies such as machine learning and artificial intelligence have begun to be applied, their popularity and application effects are still not ideal. Many prediction models still rely on traditional calculation methods, and the prediction accuracy is not high.

[0005] How to scientifically and accurately predict the travel behavior of consumers within a business district is a technical problem that needs to be solved. Summary of the invention

[0006] The technical problem to be solved by the present invention is to provide a consumer travel prediction method and device based on business connectivity in view of the deficiencies in the above-mentioned prior art, which can improve the accuracy of predicting the travel behavior of consumers in a business district.

[0007] The first aspect of the present invention discloses a consumer travel prediction method based on commercial link degree, comprising the following steps:

[0008] Step 1: Obtain the business district map data, mark the commercial facilities and transportation facilities in the business district as nodes, and mark the walking paths and bus paths between the nodes as edges to form a topological network model of the business district;

[0009] Step 2: Mark the main commercial facilities and main transportation facilities in the topological network model as the main attraction points, and assign weights to each main attraction point;

[0010] Step 3: Calculate the commercial connectivity of each node, commercial connectivity = J*spatial connectivity + K*business connectivity, where J and K are weights;

[0011] The spatial connection degree of the node is obtained by the following method: calculating the attraction degree between the node and each main attraction point, and summing the obtained multiple attraction degrees to obtain the spatial connection degree of the node; the attraction degree is obtained by the following method: calculating the walking time and distance between the node and the main attraction point, multiplying the time and distance by the weight of the main attraction point respectively, and then summing them to obtain the attraction degree between the node and the main attraction point;

[0012] The business connectivity is obtained by the following method:

[0013] Step 0, predefine the following:

[0014] Fi,j is the characteristic value of node i in business type j;

[0015] Fr,j is the characteristic value of the main attraction point r in the business format j;

[0016] n is the total number of business characteristics;

[0017] wj is the weight of business characteristic j;

[0018] Step 1. Standardize the business characteristic values ​​to make them have the same scale. Standardize Fi,j to Fnorm,i,j; standardize Fr,j to Fnorm,r,j;

[0019] Step 2. Calculate the cosine similarity of business characteristics: Use cosine similarity to evaluate the similarity between nodes and main attraction points in each business characteristic: CosSim i,j =F norm,i,j ×F norm,r,j ;

[0020] Step 3. Calculate the business format linkage degree: multiply the cosine similarity of all business format features by the weight, and then sum them up to get the business format linkage degree:

[0021] Step 4: Obtain the consumer travel OD between nodes in the business district, and mark the consumer travel OD into the topological network model;

[0022] Step 5: Take the consumption travel OD between nodes as the output variable, and the traditional attributes and commercial connection degree of the nodes as the input variables, train the neural network model, and establish a consumption travel prediction model;

[0023] Step 6: Calculate the commercial link degree of the corresponding node to be installed according to the location of the commercial facility to be installed or the transportation facility to be installed;

[0024] Step 7: Input the traditional attributes and commercial connectivity of the node to be set into the trained consumer travel prediction model to predict the consumer travel OD of the node to be set.

[0025] The above method also includes step 8, superimposing the consumer travel OD of the node to be set up into the topological network model of the current business district to obtain the topological network model after the node to be set up is built.

[0026] In the above method, the main commercial facility is a large shopping mall, and the main transportation facility is a rail transit station.

[0027] In the above method, when obtaining the consumer travel OD between nodes in the business district in step 4, it is obtained by the following method: through mobile phone signaling and LBS data, the traffic travel data of regular stops between nodes within a period of time are identified, marked as consumer travel OD, and summarized to form the consumer travel OD between each node. The total amount of each node O is the traffic generation amount, and the total amount of each node D is the traffic attraction amount.

[0028] The above method, before establishing the consumer travel prediction model, analyzes the correlation between the consumer travel OD between the nodes and the traditional attributes of the nodes through relationship clusters, and eliminates traditional attribute parameters whose correlation with the consumer travel OD is lower than a preset correlation value.

[0029] The second aspect of the present invention discloses a consumer travel prediction device based on business link degree, comprising:

[0030] The business district map data acquisition module is used to acquire business district map data, mark the commercial facilities and transportation facilities in the business district as nodes, and mark the walking paths and bus paths between the nodes as edges, so as to form a topological network model of the business district;

[0031] A main attraction point marking module is used to mark the main commercial facilities and main transportation facilities in the topological network model as main attraction points, and assign a weight to each main attraction point;

[0032] The commercial connectivity calculation module is used to calculate the commercial connectivity of each node. Commercial connectivity = J*spatial connectivity + K*business connectivity, where J and K are weights.

[0033] The spatial connection degree of the node is obtained by the following method: calculating the attraction degree between the node and each main attraction point, and summing the obtained multiple attraction degrees to obtain the spatial connection degree of the node; the attraction degree is obtained by the following method: calculating the walking time and distance between the node and the main attraction point, multiplying the time and distance by the weight of the main attraction point respectively, and then summing them to obtain the attraction degree between the node and the main attraction point;

[0034] The business connectivity is obtained by the following method:

[0035] Step 0, predefine the following:

[0036] Fi,j is the characteristic value of node i in business type j;

[0037] Fr,j is the characteristic value of the main attraction point r in the business format j;

[0038] n is the total number of business characteristics;

[0039] wj is the weight of business characteristic j;

[0040] Step 1. Standardize the business characteristic values ​​to make them have the same scale. Standardize Fi,j to Fnorm,i,j; standardize Fr,j to Fnorm,r,j;

[0041] Step 2. Calculate the cosine similarity of business characteristics: Use cosine similarity to evaluate the similarity between nodes and main attraction points in each business characteristic: CosSim i，j =F norm，i，j ×F norm，r，j ;

[0042] Step 3. Calculate the business format linkage degree: multiply the cosine similarity of all business format features by the weight, and then sum them up to get the business format linkage degree:

[0043] The consumption travel OD acquisition module is used to obtain the consumption travel OD between nodes in the business circle and mark the consumption travel OD into the topological network model;

[0044] The neural network model training module is used to train the neural network model and establish a consumer travel prediction model by taking the consumer travel OD between nodes as the output variable and the traditional attributes and commercial connection degree of the nodes as the input variables;

[0045] A commercial connectivity prediction module, used to calculate the commercial connectivity of the corresponding node to be built according to the location of the commercial facility to be built or the transportation facility to be built;

[0046] The consumer travel prediction module is used to input the traditional attributes and commercial connectivity of the node to be set into the trained consumer travel prediction model to predict the consumer travel OD of the node to be set.

[0047] The above device further comprises:

[0048] The topological network model updating module is used to superimpose the consumer travel OD of the node to be set up into the topological network model of the current business district to obtain the topological network model after the node to be set up is built.

[0049] In the above device, the main commercial facility is a large shopping mall, and the main transportation facility is a rail transit station.

[0050] In the above-mentioned device, in the consumer travel OD acquisition module, when obtaining the consumer travel OD between nodes in the business district, it is obtained by the following method: through mobile phone signaling and LBS data, the traffic travel data of regular stops between nodes within a period of time are identified, and they are marked as consumer travel OD, and summarized to form the consumer travel OD between each node. The total amount of each node O is the traffic generation amount, and the total amount of each node D is the traffic attraction amount.

[0051] Before establishing the consumer travel prediction model, the above-mentioned device analyzes the correlation between the consumer travel OD between the nodes and the traditional attributes of the nodes through the relationship cluster analysis module, and eliminates the traditional attribute parameters whose correlation with the consumer travel OD is lower than the preset correlation value.

[0052] Compared with the prior art, the present invention has the following advantages:

[0053] 1. Improve prediction accuracy: By marking the commercial facilities and transportation facilities in the business district as nodes and forming a topological network model, the actual traffic and business conditions in the business district can be more comprehensively reflected.

[0054] When calculating the commercial connectivity of a node, walking time and distance are taken into account. These factors directly affect consumers' travel choices, making the prediction model closer to actual travel behavior.

[0055] 2. Enhance the dynamic adaptability of the model: By acquiring business district map data in real time and updating the topological network model, changes within the business district, such as the opening of new commercial facilities or the adjustment of transportation facilities, can be reflected in a timely manner, thereby improving the dynamic adaptability of the model.

[0056] 3. Scientifically calculate the commercial connectivity: By combining walking time and distance with the weight of the main attraction points, the commercial connectivity of each node is scientifically calculated. This method is more objective and accurate than traditional subjective evaluation.

[0057] 4. Comprehensive consideration of multiple influencing factors: The prediction model not only takes into account the traditional attributes of the node, but also the commercial connection degree. This comprehensive consideration of multiple factors makes the prediction results more comprehensive and accurate.

[0058] 5. Optimize the traffic planning of the business district: By predicting the impact of the proposed commercial facilities or transportation facilities on the travel of consumers in the business district, it can provide a scientific basis for the traffic planning and commercial layout of the business district and optimize resource allocation.

[0059] 6. Improve decision-making support capabilities: This method can provide accurate consumer travel forecast data for government and business decision makers, helping them make more scientific and reasonable decisions, such as adjusting bus routes, optimizing commercial layout, etc.

[0060] 7. Promote business development: By predicting consumer travel OD, operators of commercial facilities can better understand consumers' travel habits and preferences, thereby adjusting their business strategies, attracting more consumers, and promoting business development.

[0061] 8. Technological advancement: The neural network model is used for training and prediction. This method has advantages in processing complex data and nonlinear relationships, and can improve the accuracy and reliability of predictions.

[0062] 9. Scalability: This method is not only applicable to a single business district, but can also be extended to multiple business districts or business districts in different cities. It has good versatility and scalability.

[0063] In summary, the present invention can significantly improve the accuracy and practicability of business district consumer travel predictions, and provide powerful technical support for urban planning, traffic management, commercial development and other fields.

[0064] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 This is a flow chart of the method of Example 1 of the present invention.

[0066] Figure 2 This is a block diagram of the device module of Example 2 of the present invention. DETAILED DESCRIPTION

[0067] Example 1

[0068] like Figure 1 As shown, a consumer travel prediction method based on business connectivity includes the following steps:

[0069] Step 1: Get the business district map data:

[0070] Use Geographic Information System (GIS) technology to obtain business district map data, including the location information of commercial facilities and transportation facilities.

[0071] GIS software is used to mark these facilities as nodes, and the walking paths and bus paths between them as edges, forming a topological network model of the business district.

[0072] Step 2: Mark the main attraction points and assign weights:

[0073] Identify the main commercial facilities (such as large shopping malls) and main transportation facilities (such as rail transit stations) within the business district and mark them as main attractions.

[0074] Assign a weight to each main attraction, which can be determined based on factors such as the size of the facility, historical passenger flow, and popularity.

[0075] Step 3: Calculate the commercial connectivity of each node: commercial connectivity = J*spatial connectivity + K*business connectivity, where J and K are weights;

[0076] The spatial connection degree of the node is obtained by the following method: calculating the attraction degree between the node and each main attraction point, and summing the obtained multiple attraction degrees to obtain the spatial connection degree of the node; the attraction degree is obtained by the following method: calculating the walking time and distance between the node and the main attraction point, multiplying the time and distance by the weight of the main attraction point respectively, and then summing them to obtain the attraction degree between the node and the main attraction point;

[0077] The business connectivity is obtained by the following method:

[0078] Step 0, predefine the following:

[0079] Fi,j is the characteristic value of node i in business type j;

[0080] Fr,j is the characteristic value of the main attraction point r in the business format j;

[0081] n is the total number of business characteristics;

[0082] wj is the weight of business format feature j, reflecting the importance of this feature in the business format connectivity;

[0083] Step 1. Standardize the business characteristic values ​​to make them have the same scale. Standardize Fi,j to Fnorm,i,j; standardize Fr,j to Fnorm,r,j;

[0084] Step 2. Calculate the cosine similarity of business characteristics: Use cosine similarity to evaluate the similarity between nodes and main attraction points in each business characteristic: CosSim i,j =F norm,i,j ×F norm,r,j ;

[0085] Step 3. Calculate the business format linkage degree: multiply the cosine similarity of all business format features by the weight, and then sum them up to get the business format linkage degree:

[0086] It should be noted that the standardization of business characteristics ensures that business characteristics of different magnitudes have the same scale in the calculation. The range of cosine similarity is between -1 and 1, with positive values ​​indicating similarity, negative values ​​indicating dissimilarity, and 0 indicating no correlation. The allocation of weights wj should be adjusted according to the actual impact of business characteristics on consumer attractiveness.

[0087] It should be noted that in actual implementation, first of all, the values ​​of weights J and K need to be determined. These weights reflect the importance of spatial connectivity and business connectivity in commercial connectivity. The determination of weights may need to be adjusted according to specific circumstances.

[0088] For each node, the walking time and distance between it and each main attraction point is calculated. This uses map data and an estimate of walking speed.

[0089] For each node and main attraction point, the attraction is calculated using the following formula: Attraction = (walking time × time weight) + (walking distance × distance weight), where time weight and distance weight are attributes of the main attraction point and are adjusted according to actual conditions.

[0090] For each node, the attraction between it and all the main attraction points is summed to obtain the spatial connection degree of the node: spatial connection degree = ∑ (attraction degree);

[0091] Business characteristics refer to the types and characteristics of different business entities in a business area. These characteristics can significantly affect consumer choices and the attractiveness of a business area. The following are some examples of business characteristics and how they affect the calculation of business connectivity:

[0092] 1. Retail formats:

[0093] Including supermarkets, department stores, specialty stores, etc.

[0094] Features: product variety, brand awareness, price range.

[0095] 2. Catering industry:

[0096] Including fast food restaurants, cafes, fine dining restaurants, etc.

[0097] Characteristics: type of cuisine, price level, dining environment.

[0098] 3. Entertainment industry:

[0099] Including cinemas, amusement parks, nightclubs, etc.

[0100] Features: types of entertainment programs, degree of modernization of facilities, and customer experience.

[0101] 4. Service formats:

[0102] Including banks, post offices, beauty salons, gyms, etc.

[0103] Features: professionalism, convenience, and quality of service.

[0104] 5. Cultural formats:

[0105] Including bookstores, art galleries, theaters, etc.

[0106] Characteristics: cultural atmosphere, educational value, and artistic level.

[0107] 6. Tourism industry:

[0108] Including hotels, B&Bs, tourist souvenir shops, etc.

[0109] Features: geographical location, service quality, and travel experience.

[0110] 7. Education industry:

[0111] Including schools, training centers, educational institutions, etc.

[0112] Characteristics: quality of education, diversity of courses, and strength of faculty.

[0113] 8. Medical and health industry:

[0114] Including hospitals, clinics, pharmacies, etc.

[0115] Characteristics: medical service level, advanced medical equipment, and professional ability of doctors.

[0116] When calculating the business connectivity, the business relevance between a node and a major attraction can be evaluated based on the similarity or difference in business characteristics. For example, if a node has a similar catering business to a major attraction, then the business connectivity between them may be high. On the contrary, if the business differences are large, the business connectivity may be low.

[0117] Business connectivity can also be calculated by the following steps:

[0118] s1. Collect and classify the business characteristics of each node and main attraction point;

[0119] s2. Determine the weights of business format characteristics based on the importance of the business format’s attractiveness to consumers.

[0120] s3. Calculate the deviation ratio of each business characteristic, which can be achieved by comparing the business characteristics of the node and the main attraction point.

[0121] s4. Take the weighted sum of the deviation ratios of all business characteristics to obtain the business connectivity.

[0122] The calculation method of business format connectivity can be adjusted according to specific circumstances to ensure that it accurately reflects the impact of business format characteristics on commercial attractiveness.

[0123] Finally, based on the calculation results, the distribution of business connectivity is analyzed, which can help business planners understand which areas or nodes have high business potential and thus make corresponding business layout and optimization.

[0124] The selection of weights J and K can be adjusted according to specific circumstances and objectives to ensure that the calculation results of commercial linkage degree can reflect the actual commercial value.

[0125] Step 4: Get consumer travel OD:

[0126] Using mobile phone signaling and LBS (location-based services) data, we can identify traffic data that regularly stops between nodes over a period of time (which can be a week, a month, or a year).

[0127] These data are identified as consumption travel OD and summarized to form consumption travel OD between each node. At the same time, the total amount of each node O (traffic generation) and the total amount of each node D (traffic attraction) are calculated.

[0128] Step 5: Train the neural network model:

[0129] The consumption travel OD between nodes is taken as the output variable, and the traditional attributes of the nodes (such as geographical location, population density, etc.) and commercial connectivity are taken as the input variables.

[0130] Use deep learning frameworks (such as TensorFlow or PyTorch) to train neural network models and build consumer travel prediction models. The model can use structures such as multi-layer perceptron (MLP) or convolutional neural network (CNN).

[0131] Step 6: Calculate the commercial link degree of the node to be set:

[0132] According to the location of the commercial facilities or transportation facilities to be built, the commercial connectivity of the corresponding nodes to be built is calculated in the same manner as in step 3.

[0133] Step 7: Predict the consumption travel OD of the node to be set:

[0134] The traditional attributes and commercial connectivity of the node to be set are input into the trained consumer travel prediction model to predict the consumer travel OD of the node to be set.

[0135] Step 8: Update the topology network model:

[0136] The consumer travel OD of the node to be built is superimposed on the topological network model of the current business district to obtain the topological network model after the node to be built is built. This step can be achieved through GIS software to ensure the real-time update and accuracy of the model.

[0137] In this embodiment, before establishing the consumption travel prediction model, a relationship cluster analysis is performed to identify the correlation between the consumption travel OD between each node and the traditional attributes of the node.

[0138] Traditional attribute parameters whose correlation with consumer travel OD is lower than the preset correlation value are eliminated to improve the prediction accuracy of the model.

[0139] Through the above implementation, the accuracy and reliability of consumer travel prediction in the business district can be significantly improved. Using GIS technology and deep learning models, the dynamic changes and complex factors in the business district can be more comprehensively considered.

[0140] By updating the business district map data and consumer travel OD in real time, the model can adapt to changes in the business district and provide more accurate prediction results.

[0141] Eliminating traditional attribute parameters with low correlation with consumer travel OD can reduce the noise of the model and improve the stability and accuracy of the prediction.

[0142] This method is not only applicable to a single business district, but can also be extended to multiple business districts or business districts in different cities. It has good versatility and scalability.

[0143] Example 2

[0144] like Figure 2 As shown, a consumer travel prediction device based on business connection degree includes:

[0145] The business district map data acquisition module is used to acquire business district map data, mark the commercial facilities and transportation facilities in the business district as nodes, and mark the walking paths and bus paths between the nodes as edges, so as to form a topological network model of the business district;

[0146] A main attraction point marking module is used to mark the main commercial facilities and main transportation facilities in the topological network model as main attraction points, and assign a weight to each main attraction point;

[0147] The commercial connectivity calculation module is used to calculate the commercial connectivity of each node: commercial connectivity = J*spatial connectivity + K*business connectivity, where J and K are weights;

[0148] The spatial connection degree of the node is obtained by the following method: calculating the attraction degree between the node and each main attraction point, and summing the obtained multiple attraction degrees to obtain the spatial connection degree of the node; the attraction degree is obtained by the following method: calculating the walking time and distance between the node and the main attraction point, multiplying the time and distance by the weight of the main attraction point respectively, and then summing them to obtain the attraction degree between the node and the main attraction point;

[0149] The business connectivity is obtained by the following method:

[0150] Step 0, predefine the following:

[0151] Fi,j is the characteristic value of node i in business type j;

[0152] Fr,j is the characteristic value of the main attraction point r in the business format j;

[0153] n is the total number of business characteristics;

[0154] wj is the weight of business format feature j, reflecting the importance of this feature in the business format connectivity;

[0155] Step 1. Standardize the business characteristic values ​​to make them have the same scale. Standardize Fi,j to Fnorm,i,j; standardize Fr,j to Fnorm,r,j;

[0156] Step 2. Calculate the cosine similarity of business characteristics: Use cosine similarity to evaluate the similarity between nodes and main attraction points in each business characteristic: CosSim i，j =F norm，i，j ×F norm，r，j ;

[0157] Step 3. Calculate the business format linkage degree: multiply the cosine similarity of all business format features by the weight, and then sum them up to get the business format linkage degree:

[0158] The consumption travel OD acquisition module is used to obtain the consumption travel OD between nodes in the business circle and mark the consumption travel OD into the topological network model;

[0159] The neural network model training module is used to train the neural network model and establish a consumer travel prediction model by taking the consumer travel OD between nodes as the output variable and the traditional attributes and commercial connection degree of the nodes as the input variables;

[0160] A commercial connectivity prediction module, used to calculate the commercial connectivity of the corresponding node to be built according to the location of the commercial facility to be built or the transportation facility to be built;

[0161] The consumer travel prediction module is used to input the traditional attributes and commercial connectivity of the node to be set into the trained consumer travel prediction model to predict the consumer travel OD of the node to be set.

[0162] In this embodiment, it also includes:

[0163] The topological network model updating module is used to superimpose the consumer travel OD of the node to be set up into the topological network model of the current business district to obtain the topological network model after the node to be set up is built.

[0164] In this embodiment, the main commercial facility is a large shopping mall, and the main transportation facility is a rail transit station.

[0165] In this embodiment, in the consumer travel OD acquisition module, when acquiring the consumer travel OD between nodes in the business district, it is obtained by the following method: through mobile phone signaling and LBS data, the traffic travel data of regular stops between nodes within a period of time (which can be a week, a month, or a year) is identified, and it is marked as consumer travel OD, and summarized to form the consumer travel OD between each node. The total amount of each node O is the traffic generation amount, and the total amount of each node D is the traffic attraction amount.

[0166] In this embodiment, before establishing the consumer travel prediction model, the correlation between the consumer travel OD between the nodes and the traditional attributes of the nodes is analyzed through the relationship cluster analysis module, and the traditional attribute parameters whose correlation with the consumer travel OD is lower than the preset correlation value are eliminated.

[0167] It should be noted that Example 2 shares the same technical basis and principle as Example 1. In order to keep the document concise and avoid repetition, we will not elaborate on the technical details. The reader can refer to the detailed description provided in Example 1 to obtain a comprehensive understanding of the technical details of Example 2.

[0168] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any way. Any simple modification, change and equivalent structural change made to the above embodiment based on the technical essence of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A consumer travel prediction method based on commercial connectivity, characterized in that: The following steps are involved: Step 1: Obtain the business district map data, mark the commercial facilities and transportation facilities in the business district as nodes, and mark the walking paths and bus paths between the nodes as edges to form a topological network model of the business district; Step 2: Mark the main commercial facilities and main transportation facilities in the topological network model as the main attraction points, and assign weights to each main attraction point; Step 3: Calculate the commercial connectivity of each node, commercial connectivity = J*spatial connectivity + K*business connectivity, where J and K are weights; The spatial connection degree of the node is obtained by the following method: calculating the attraction degree between the node and each main attraction point, and summing the obtained multiple attraction degrees to obtain the spatial connection degree of the node; the attraction degree is obtained by the following method: calculating the walking time and distance between the node and the main attraction point, multiplying the time and distance by the weight of the main attraction point respectively, and then summing them to obtain the attraction degree between the node and the main attraction point; The business connectivity is obtained by the following method: Step 0, predefine the following: Fi,j is the characteristic value of node i in business type j; Fr,j is the characteristic value of the main attraction point r in the business format j; n is the total number of business characteristics; wj is the weight of business characteristic j; Step 1. Standardize the business characteristic values ​​to make them have the same scale. Standardize Fi,j to Fnorm,i,j; standardize Fr,j to Fnorm,r,j; Step 2. Calculate the cosine similarity of business characteristics: CosSim i，j =F norm，i，j ×F norm，r，j ; Step 3. Calculate the business connection degree: Step 4: Obtain the consumer travel OD between nodes in the business district, and mark the consumer travel OD into the topological network model; Step 5: Take the consumption travel OD between nodes as the output variable, and the traditional attributes and commercial connection degree of the nodes as the input variables, train the neural network model, and establish a consumption travel prediction model; Step 6: Calculate the commercial link degree of the corresponding node to be installed according to the location of the commercial facility to be installed or the transportation facility to be installed; Step 7: Input the traditional attributes and commercial connectivity of the node to be set into the trained consumer travel prediction model to predict the consumer travel OD of the node to be set.

2. The method according to claim 1, characterized in that The method also includes step 8, superimposing the consumer travel OD of the node to be set up into the topological network model of the current business district to obtain the topological network model after the node to be set up is built.

3. The method according to claim 1, characterized in that The main commercial facilities are large shopping malls, and the main transportation facilities are rail transit stations.

4. The method according to claim 1, characterized in that: When obtaining the consumer travel OD between nodes in the business district in step 4, it is obtained through the following method: through mobile phone signaling and LBS data, the traffic travel data with regular stops between nodes within a period of time is identified, marked as consumer travel OD, and summarized to form the consumer travel OD between each node. The total amount of each node O is the traffic generation amount, and the total amount of each node D is the traffic attraction amount.

5. The method according to claim 1, characterized in that Before establishing the consumer travel prediction model, the correlation between the consumer travel OD between the nodes and the traditional attributes of the nodes is analyzed through relationship clusters, and the traditional attribute parameters whose correlation with the consumer travel OD is lower than the preset correlation value are eliminated.

6. A consumer travel prediction device based on business connectivity, characterized in that: include: The business district map data acquisition module is used to acquire business district map data, mark the commercial facilities and transportation facilities in the business district as nodes, and mark the walking paths and bus paths between the nodes as edges, so as to form a topological network model of the business district; A main attraction point marking module is used to mark the main commercial facilities and main transportation facilities in the topological network model as main attraction points, and assign a weight to each main attraction point; The commercial connectivity calculation module is used to calculate the commercial connectivity of each node. Commercial connectivity = J*spatial connectivity + K*business connectivity, where J and K are weights. The spatial connection degree of the node is obtained by the following method: calculating the attraction degree between the node and each main attraction point, and summing the obtained multiple attraction degrees to obtain the spatial connection degree of the node; the attraction degree is obtained by the following method: calculating the walking time and distance between the node and the main attraction point, multiplying the time and distance by the weight of the main attraction point respectively, and then summing them to obtain the attraction degree between the node and the main attraction point; The business connectivity is obtained by the following method: Step 0, predefine the following: Fi,j is the characteristic value of node i in business type j; Fr,j is the characteristic value of the main attraction point r in the business format j; n is the total number of business characteristics; wj is the weight of business characteristic j; Step 1. Standardize the business characteristic values ​​to make them have the same scale. Standardize Fi,j to Fnorm,i,j; standardize Fr,j to Fnorm,r,j; Step 2. Calculate the cosine similarity of business characteristics: CosSim i，j =F norm，i，j ×F norm，r，j ; Step 3. Calculate the business connection degree: The consumption travel OD acquisition module is used to obtain the consumption travel OD between nodes in the business circle and mark the consumption travel OD into the topological network model; The neural network model training module is used to train the neural network model and establish a consumer travel prediction model by taking the consumer travel OD between nodes as the output variable and the traditional attributes and commercial connection degree of the nodes as the input variables; A commercial connectivity prediction module, used to calculate the commercial connectivity of the corresponding node to be built according to the location of the commercial facility to be built or the transportation facility to be built; The consumer travel prediction module is used to input the traditional attributes and commercial connectivity of the node to be set into the trained consumer travel prediction model to predict the consumer travel OD of the node to be set.

7. The device according to claim 6, characterized in that Also includes: The topological network model updating module is used to superimpose the consumer travel OD of the node to be set up into the topological network model of the current business district to obtain the topological network model after the node to be set up is built.

8. The device according to claim 6, characterized in that The main commercial facilities are large shopping malls, and the main transportation facilities are rail transit stations.

9. The device according to claim 6, characterized in that In the consumer travel OD acquisition module, when obtaining the consumer travel OD between nodes in the business district, it is obtained by the following method: through mobile phone signaling and LBS data, the traffic travel data of regular stops between nodes within a period of time are identified, marked as consumer travel OD, and summarized to form the consumer travel OD between each node. The total amount of each node O is the traffic generation amount, and the total amount of each node D is the traffic attraction amount.

10. The device according to claim 6, characterized in that Before establishing the consumer travel prediction model, the correlation between the consumer travel OD between the nodes and the traditional attributes of the nodes is analyzed through the relationship cluster analysis module, and the traditional attribute parameters whose correlation with the consumer travel OD is lower than the preset correlation value are eliminated.

Images