Recommended method, system and medium for slow-moving space connection network in urban mature recreational areas

By analyzing the 'crowdsourcing' data to identify the stay characteristics and clusters, a slow-moving space connection network for mature urban recreational areas is generated, which solves the problems of too short tour time and excessive gathering of attractions, and achieves a more balanced and personalized recreational path Recommended, improving travel experience.

CN117575117BActive Publication Date: 2024-11-29HARBIN INSTITUTE OF TECHNOLOGY (SHENZHEN) (INSTITUTE OF SCIENCE AND TECHNOLOGY INNOVATION HARBIN INSTITUTE OF TECHNOLOGY SHENZHEN) +1
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202311372653.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-10-23
Publication Date
2024-11-29
Estimated Expiration
2043-10-23

AI Technical Summary

Technical Problem

The existing technology ignores the specific status and information sharing of scenic spots in the recommendation of mature urban recreation areas, resulting in too short tour time and excessive gathering of attractions, which cannot effectively extend the tour path and reduce the risk of attractions gathering.

Method used

By acquiring and analyzing the 'crowdsourcing' slow recreation GPS trajectory data, identifying the stay feature trajectory, performing spatial clustering and connection heat calculations, generating a topological network based on the maximum three connections of the stay area, and balancing recommendations are performed in combination with the location recommendation algorithm. Avoid excessive gathering of attractions.

Benefits of technology

Extend the tour time, reduce the risk of attractions gathering, provide differentiated tourism services, improve individual tourism experience, and adapt to urban life and citizens' recreational needs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117575117B_ABST
    Figure CN117575117B_ABST
Patent Text Reader

Abstract

The present invention discloses a method, system and medium for recommending a connection network of slow-moving spaces in a mature urban recreation area. The method comprises: obtaining a "crowd-sourced" slow-moving recreation GPS track within a given geographical range to form a GPS track data set; identifying tracks with stop features by screening the tracks in the GPS track data set by time, distance and track points, and integrating the tracks into stop points; spatially clustering the stop points, and converting the clustered point elements into surface elements to obtain stop areas; using the stop areas and the track lines intersecting the stop areas as a two-mode network, using the length of the track in the stop areas as a weight, and calculating the connection heat between the stop areas; regenerating a connection network based on the maximum three connection stop areas of each stop area, calculating the structural parameters of the topological network and visualizing them, and combining the balance of the recommendation to perform algorithm recommendation. The present invention can effectively provide a reference basis for tourism planning and policy formulation from the bottom up.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to a method, system and medium for recommending a slow-moving space connection network in a mature urban recreation area, and belongs to the technical field of urban planning. Background Art

[0002] At present, the current scenic spot recommendation algorithm mainly considers the individual characteristics of tourists and emphasizes personalized recommendation algorithms, but ignores the specific status of the scenic spot (Xu Guoliang, 2022); Li Xiaofei et al. (2022) began to consider the crowding level of tourists and the time arrangement of tourists, but their patent was limited to the data of smart cameras in scenic spots, which was limited by the general information collection infrastructure of scenic spots and the information sharing between scenic spots. With the continuous development of smart tourism, how to use mature data mining and analysis technology to analyze crowdsourcing data to provide tourists with scenic spot recommendations has become a research focus. In the field of tourism, based on crowdsourcing data, Zheng et al. (2020) combined urban transportation mode selection with recreational destinations to optimize recommended routes; Liang Jingyu et al. clustered the spatial units within the research scope through the recreational network system and divided different types of hiking tourism communities, but did not recommend recreational routes. Most of the existing algorithms in the computer field focus on improving the recommendation accuracy (He, 2022) or combining recommended routes with traveler characteristics (Bin et al., 2019). However, in practice, the connection network of recreational spaces needs to extend the tour time and tour paths of the recommended routes, while minimizing the risk of excessive concentration of scenic spots. Moreover, tourists have higher and higher requirements for tourism quality, and many tourists tend to choose slow-moving spaces with a more comfortable experience. Therefore, for mature urban recreational areas, providing a recommended method for slow-moving space connection networks has become a key topic that needs to be paid attention to in research. Summary of the invention

[0003] In view of this, the present invention provides a method, system, computer device and storage medium for recommending a slow-moving space connection network in a mature urban recreation area. The method, system and computer device and storage medium are in line with the actual situation and can avoid excessive concentration of popular attractions, prolong the tour time and tour route when recommending attractions in mature urban recreation areas. Based on crowdsourcing data, the method can effectively provide a reference basis for tourism planning and policy formulation from the bottom up.

[0004] The first object of the present invention is to provide a method for recommending a slow-moving space connection network in a mature urban recreation area.

[0005] The second object of the present invention is to provide a slow-moving space connection network recommendation system for mature urban recreation areas.

[0006] A third object of the present invention is to provide a computer device.

[0007] A fourth object of the present invention is to provide a storage medium.

[0008] The first object of the present invention can be achieved by adopting the following technical solutions:

[0009] A method for recommending a slow-moving space connection network in a mature urban recreation area, the method comprising:

[0010] Obtain the “crowd-sourced” slow-travel recreation GPS trajectories within a given geographical area to form a GPS trajectory dataset;

[0011] By screening the trajectories in the GPS trajectory data set by time, distance and number of trajectory points, the trajectories with stay features are identified and integrated into stay points.

[0012] Spatial clustering of the stay points is performed, and the clustered point features are converted into surface features to obtain the stay area;

[0013] The stay area and the trajectory lines intersecting the stay area are regarded as a two-mode network, the length of the trajectory in the stay area is used as the weight, and the connection heat between the stay areas is calculated;

[0014] The connection network is regenerated based on the maximum three connected stay areas of each stay area, the structural parameters of the topological network are calculated and visualized, and the algorithm recommendation is made in combination with the balance of the recommendation.

[0015] Furthermore, the "crowd-sourced" slow-travel recreation GPS tracks within a given geographical area are obtained to form a GPS track data set, which specifically includes:

[0016] Select keywords to obtain "crowd-sourced" slow-moving recreation GPS trajectories, clean the trajectory data, delete all trajectory data with missing attribute information, duplication, and high speed, filter out the slow-moving recreation trajectory data, and obtain a trajectory data set consisting of several trajectories. Each trajectory includes several trajectory points with trajectory ID, longitude, latitude, timestamp information, and driving speed.

[0017] Furthermore, the method of screening the trajectories in the GPS trajectory data set by time, distance and number of trajectory points, identifying trajectories with stop features, and integrating the trajectories with stop features into stop points specifically includes:

[0018] De-duplicate the trajectory data, filter out the first trajectory point of each trajectory, and use this trajectory point as the trajectory label;

[0019] Calculate each trajectory in turn, and for each trajectory point of a trajectory, calculate the time difference between the trajectory point and the first trajectory point, and unify the time difference into units to form the time difference values ​​between several trajectory points of each trajectory and the first trajectory point;

[0020] The first screening function is defined as calculating the time difference between each track point in each track and other track points, and calculating the distance difference between each track point in each track and other track points according to the longitude and latitude of the existing track points;

[0021] The second screening function is defined to screen the distance of each trajectory. When the distance difference between two trajectory points is less than the first distance threshold, the screening is qualified.

[0022] The third filtering function is defined as filtering the time and number of trajectory points of each trajectory based on the filtering distance. When the time difference between two trajectory points is greater than the time threshold and the number of trajectory points is greater than or equal to the point threshold, the screening meets the criteria, and the trajectory segment that meets the screening criteria is assigned an ID;

[0023] Define a first distance threshold, a time threshold, and a point threshold;

[0024] According to the first distance threshold, time threshold and point number threshold, all trajectories are calculated and screened to obtain trajectories with stay characteristics;

[0025] The track points are grouped according to the track segment ID, and the average values ​​of the longitude, latitude and time of the grouped track points are calculated and integrated to obtain the stop points.

[0026] Furthermore, the spatial clustering of the stay points specifically includes:

[0027] Add the range area map and stop points to the current map, and generate track lines based on the track ID;

[0028] predefine a second distance threshold between the stay points and a minimum cluster point;

[0029] Generate a cluster of stay points according to the second distance threshold and the minimum cluster point, each cluster is defined by a cluster ID;

[0030] The cluster points are screened against the map, clusters that do not belong to the recreation type are deleted, and clusters are merged and decomposed.

[0031] Furthermore, the step of converting the clustered point elements into surface elements to obtain the stay area specifically includes:

[0032] Input clustered point features, generate TIN, and convert TIN to line features;

[0033] Convert TIN line features to surface features, spatially connect TIN lines to cluster points, and spatially connect TIN surface features to TIN lines that are already spatially connected to cluster points;

[0034] According to the cluster ID, the segmented TIN surfaces are merged to generate the stay area.

[0035] Furthermore, the staying area and the trajectory line intersecting the staying area are used as a two-mode network, the length of the trajectory in the staying area is used as a weight, and the connection heat between the staying areas is calculated, which specifically includes:

[0036] Cut off the part of the trajectory line outside the stay area;

[0037] Explode the track line in the stay area, so that one track line is decomposed into multiple track lines located in multiple stay areas;

[0038] Connect the trajectory line after explosion with the stay area in space to obtain the trajectory line layer data;

[0039] Generate a stay area-trajectory data table based on the trajectory line layer data;

[0040] According to the stay area-trajectory data table, a matrix of the "stay area-trajectory" two-mode network is formed, which is defined as V. i and j are defined as the number of rows of V, representing the stay area, and k is defined as the number of columns of V, representing the trajectory. V i,k and V j,k Respectively represent the trajectory length of trajectory k within the range of stay area i and j, as weights;

[0041] The matrix of the two-mode network is integrated into a stay area-stay area connection network, and the connection between stay areas i and j is calculated by the co-occurrence network algorithm;

[0042] According to the location recommendation algorithm, the sum of the lengths of the trajectory k passing through the stop area and the sum of the lengths of all trajectory lines within the stop area i are calculated, thereby calculating the connection between the stop areas i and j;

[0043] According to the interest community discovery algorithm, the cosine similarity is calculated, the similarity between the stay areas i and j is analyzed, and the number of trajectories inside the stay areas i and j is calculated, so as to calculate the connection between the stay areas i and j.

[0044] Furthermore, the connection network is regenerated based on the maximum three connection stay areas of each stay area, the structural parameters of the topological network are calculated and visualized, and the algorithm recommendation is performed in combination with the balance of the recommendation, specifically including:

[0045] Calculate the maximum three connected stay areas of each stay area obtained by the co-occurrence network algorithm, location recommendation algorithm and interest community discovery algorithm respectively, and generate a matrix;

[0046] Calculate the average distance, diameter and reciprocity of the topological network. The average distance refers to the average value of the number of edges of the path with the least number of edges between any two nodes in the network. The diameter refers to the maximum value of the number of edges of the shortest path between any two nodes in the network. The reciprocity includes Arc-reciprocity and Dyad-reciprocity. Arc-reciprocity refers to the ratio of directed edges with reciprocating relationships to all edges. Dyad-reciprocity refers to the ratio of the number of tuples with reciprocating relationships to all tuples connected by edges.

[0047] For the co-occurrence network algorithm and the community of interest discovery algorithm, the generated matrix is ​​copied and XY transposed, and the data before and after the transposition are added to obtain a symmetrical data table. Half of the data along the diagonal of the data table is deleted to obtain a triangular matrix.

[0048] Generate and visualize the stay area connection network diagram of the co-occurrence network algorithm, location recommendation algorithm and interest community discovery algorithm based on the average distance, diameter and reciprocity of the topological network;

[0049] For mature urban recreation areas, a partial chain structure of recommendation pairs is formed based on the location recommendation algorithm.

[0050] The second object of the present invention can be achieved by adopting the following technical solutions:

[0051] A network recommendation system for slow-moving space connections in mature urban recreational areas, the system comprising:

[0052] The acquisition module is used to obtain the “crowd-sourced” slow-travel recreation GPS trajectories within a given geographical area to form a GPS trajectory dataset;

[0053] The integration module is used to identify the trajectories with the stay feature by screening the trajectories in the GPS trajectory data set by time, distance and number of trajectory points, and integrate the trajectories with the stay feature into stay points;

[0054] The clustering module is used to spatially cluster the stay points and convert the clustered point features into surface features to obtain the stay area;

[0055] A calculation module is used to use the stay area and the trajectory lines intersecting the stay area as a two-mode network, use the length of the trajectory in the stay area as a weight, and calculate the connection heat between the stay areas;

[0056] The recommendation module is used to regenerate the connection network based on the maximum three connected stay areas of each stay area, calculate and visualize the structural parameters of the topological network, and make algorithm recommendations based on the balance of the recommendations.

[0057] The third object of the present invention can be achieved by adopting the following technical solutions:

[0058] A computer device includes a processor and a memory for storing a program executable by the processor. When the processor executes the program stored in the memory, the above-mentioned method for recommending a connection network of slow-moving spaces in mature urban recreation areas is implemented.

[0059] The fourth object of the present invention can be achieved by adopting the following technical solutions:

[0060] A storage medium stores a program, and when the program is executed by a processor, the above-mentioned method for recommending a connection network of slow-moving spaces in a mature urban recreation area is implemented.

[0061] The present invention has the following beneficial effects compared with the prior art:

[0062] 1. Based on a "bottom-up" perspective, the present invention proposes a relatively complete and novel analysis framework to identify the stay areas of slow-moving recreational spaces and construct a connection network. In addition, in line with urban life and citizens' recreational needs, the invention explores the connection network between stay areas and conducts algorithm comparison for route recommendation, providing a basis for extending the tour time and tour routes, avoiding excessive concentration of scenic spots, differentiating tourism service management, and improving individual tourism experience.

[0063] 2. The present invention adopts a location recommendation algorithm to regenerate a connection network based on the maximum three connection points of each stay area, calculate the structural parameters of the topological network, and make algorithm recommendations based on the balance of the recommendations to reduce the risk of clustering of scenic spots and avoid overheating. While separating out generally well-known scenic spots, it reduces the risk of clustering in local areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0065] Figure 1 This is a simplified flowchart of a method for recommending a network connection for slow-moving spaces in mature urban recreation areas according to Example 1 of the present invention.

[0066] Figure 2 This is a detailed flow chart of the method for recommending a network connection for slow-moving spaces in mature urban recreation areas according to Example 1 of the present invention.

[0067] Figure 3 This is a spatial distribution diagram of the stay point clusters of Example 1 of the present invention.

[0068] Figure 4This is a spatial distribution diagram of the stay area of ​​Example 1 of the present invention.

[0069] Figure 5 This is a network topology diagram of the co-occurrence network algorithm of Example 1 of the present invention.

[0070] Figure 6 This is a network topology diagram of a recommended perspective of the location recommendation algorithm of Example 1 of the present invention.

[0071] Figure 7 This is a network topology diagram of the recommended perspective of the location recommendation algorithm of Example 1 of the present invention.

[0072] Figure 8 This is a network topology diagram of the interest community discovery algorithm according to Example 1 of the present invention.

[0073] Fig. 9 This is a structural block diagram of a slow-moving space connection network recommendation system for urban mature recreation areas according to Example 2 of the present invention.

[0074] Fig.10 This is a structural block diagram of a computer device according to Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0075] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0076] Embodiment 1:

[0077] like Figure 1 and Figure 2 As shown, this embodiment provides a method for recommending a slow-moving space connection network for mature urban recreational areas. The method takes the old urban area of ​​Guangzhou as the research object, and uses "bottom-up" crowdsourcing data to construct a slow-moving space connection network to recommend mature recreational areas in the old urban area of ​​Guangzhou. The specific area of ​​the old urban area is based on the "Guangzhou Urban Renewal Special Master Plan (2018-2035)", with a total area of ​​about 50.24 square kilometers, covering three administrative districts of Haizhu District, Liwan District, and Yuexiu District. The area has rich historical and cultural relics and natural scenery, and has a large number of popular scenic spots. It is the main area for Guangzhou to carry out daily recreational activities for citizens and sightseeing activities for foreign tourists. The method specifically includes the following steps:

[0078] S201. Obtain “crowd-sourced” slow-moving recreation GPS trajectories within a given geographical area to form a GPS trajectory dataset.

[0079] Specifically, appropriate keywords are selected to obtain "crowd-sourced" slow-moving recreation GPS trajectories, and the trajectory data is cleaned to delete all trajectory data that lack attribute information, are repeated, and have a high speed, and to filter out the slow-moving recreation trajectory data. Finally, a trajectory data set consisting of several trajectories is obtained, each trajectory including several trajectory points with trajectory ID, longitude, latitude, timestamp information, and driving speed.

[0080] In this embodiment, the crowdsourced trajectory data comes from the "Two-Bu Road" APP, which can realize functions such as planning recreational routes, recording recreational trajectories, and sharing travel notes and experiences; the recreational GPS trajectories (the time range is from 2018 to 2022) are searched and captured on the "Two-Bu Road" software platform using the Guangzhou urban administrative units Haizhu District, Liwan District and Yuexiu District as keywords. The GPS trajectory data consists of several trajectory points of several trajectory data, including the trajectory ID, geographic coordinates (longitude, latitude), timestamp information, driving speed and other related information of each trajectory point; the GPS trajectory data is cleaned, and all trajectory data with missing attribute information and duplicates are deleted. In addition, combined with the provisions of the "Safety Technical Specifications for Electric Bicycles" (GB17761-2018) that the maximum driving speed of electric vehicles shall not exceed 15km / h, the trajectory data with a speed greater than 15km / h is deleted. After multiple cleanings, a total of 16,143 GPS trajectories (a total of 211,489 GPS trajectory points) were obtained as the research data set.

[0081] S202, identifying trajectories with stay features by screening the trajectories in the GPS trajectory data set by time, distance and number of trajectory points, and integrating the trajectories with stay features into stay points.

[0082] Furthermore, step S202 specifically includes:

[0083] S2021. De-duplicate the trajectory data, and filter out the first trajectory point of each trajectory, and use this trajectory point as the label of the trajectory.

[0084] Specifically, run the Python tool, import the pandas, numpy, os, tqdm and datetime libraries, and read the trajectory point data csv file, that is, the GPS trajectory data set obtained previously; use the drop_duplicates() function of the pandas library to deduplicate the trajectory data while filtering out the first trajectory point of each trajectory, and use this trajectory point as the label of the trajectory.

[0085] S2022, calculating each trajectory in turn, for each trajectory point of a trajectory, calculating the time difference between the trajectory point and the first trajectory point, and unifying the time difference into units to form a time difference value between a plurality of trajectory points of each trajectory and the first trajectory point.

[0086] Specifically, a for loop is used to calculate each trajectory in turn. For each trajectory point of a trajectory, the time difference between it and the first trajectory point is calculated, and the time difference is unified into units. Finally, the time difference values ​​between several trajectory points of each trajectory and the first trajectory point of the trajectory are formed, and it is added as a column of content "delta" to the original csv file.

[0087] The Python tool code is as follows:

[0088]

[0089] S2023. Define a first screening function to calculate the time difference between each trajectory point and other trajectory points in each trajectory, and calculate the distance difference between each trajectory point and other trajectory points in each trajectory according to the longitude and latitude of the existing trajectory points.

[0090] Theoretically, a stop point has dual connotations in time and space, and can be regarded as an individual imprint of activities that last for more than a certain period of time within a given spatial range. Therefore, when screening stop points from a trajectory, it is generally necessary to consider the thresholds of both time and space factors (time_threshld and distance_threshold), and control the threshold of the number of trajectory points (point_threshold) according to actual conditions; therefore, the first screening function is defined, specifically as follows: first, the time difference and distance difference between the trajectory points within the trajectory are calculated: using a while loop, the previously calculated "delta" is used to calculate the time difference "real_delta" between each trajectory point and other trajectory points, and the distance difference "real_distance" between each trajectory point and other trajectory points is calculated based on the latitude and longitude of the existing trajectory data points.

[0091] S2024. Define a second screening function to screen the distance of each trajectory. When the distance difference between two trajectory points is less than the first distance threshold, the screening is qualified.

[0092] Define the second screening function, specifically: use a while loop and an if else statement to perform screening. When the screening condition is that the distance difference "real_distance" is less than the first distance threshold (distance_threshold), the screening is qualified.

[0093] S2025. Define the third filtering function to filter the time and number of trajectory points of each trajectory based on the filtering distance. When the time difference between two trajectory points is greater than the time threshold and the number of trajectory points is greater than or equal to the point threshold, the filtering meets the criteria, and an ID is assigned to the trajectory segment that meets the filtering criteria.

[0094] Define the third filtering function, specifically: based on the filtering distance, filter the time and number of trajectory points of each trajectory, use the if statement for filtering, when the filtering condition is that the time difference "real_delta" is greater than the time threshold (time_threshold) and the number of trajectory points is greater than or equal to the point threshold (point_threshlod), the filtering meets the criteria, and the trajectory segment that meets the filtering criteria is assigned an ID of "Staypoint_id_in".

[0095] After the screening is completed, a table consisting of the column data information of the trajectory points such as "trajectory ID", "trajectory segment ID", "longitude", "latitude", "time", "real_delta", and "real_distance" is returned. At this point, the entire function completes the capture of a trajectory segment that meets the time, distance, and point count thresholds.

[0096] The Python tool code is as follows:

[0097]

[0098]

[0099] S2026. Define a first distance threshold, a time threshold, and a point threshold.

[0100] Specifically, predefined thresholds are set in combination with specific urban scenes and general tourist behaviors such as walking and other generalized stop behaviors to obtain a first distance threshold (distance_threshold), a time threshold (time_threshold) and a point threshold (point_threshlod). In this embodiment, the first distance threshold of the recreational trajectory data of the old city of Guangzhou is defined as 250 meters, the time threshold is 20 minutes, and the point threshold is 4.

[0101] S2027. Calculate and filter all trajectories according to the first distance threshold, the time threshold, and the point threshold to obtain trajectories with stay features.

[0102] Specifically, the defined filtering function is calculated according to the previously defined first distance threshold, time threshold and point number threshold; since the defined filtering function is to calculate and filter a trajectory, a for loop is used to calculate and filter all trajectories to obtain trajectories with stay features, and the final result is exported as a csv file, which contains data information such as "trajectory ID", "trajectory segment ID", "longitude", "latitude", "real_delta", "real_distance", "trajectory points" and so on of the trajectory points.

[0103] The Python tool code is as follows:

[0104]

[0105] S2028. Group the track points according to the track segment ID, calculate the average values ​​of the longitude, latitude and time of the grouped track points, and integrate them to obtain the stop points.

[0106] Specifically, the groupby function is used to group the trajectory points according to the label "trajectory segment ID", and the mean method is used to calculate the average values ​​of the longitude, latitude and time of the grouped trajectory points. The stop points are obtained by integration and the data information of the stop points is exported to a csv file, including the "trajectory segment ID", "longitude", "latitude" and "real_delta" data information of the stop points.

[0107] S203, spatially clustering the stay points, and converting the clustered point elements into surface elements to obtain the stay areas.

[0108] Furthermore, step S203 specifically includes:

[0109] S2031. Add the range area map and the stop points to the current map, and generate a track line according to the track ID.

[0110] Specifically, the scope area of ​​this embodiment is the old city of Guangzhou. Open ArcGIS Pro, create a new project page, click "Directory", right-click "Folder", click "Add Folder Connection", add the default folder where the data is located to the page, right-click the default folder, click "New", and create a new "File Geodatabase" to store the subsequently generated files; select the map shp file of the old city of Guangzhou from the directory-default folder, right-click and select "Add to Current Map", add the base map of the old city of Guangzhou to the map page, click "Add Data", select the stop point csv file, add the independent table to the map, use the "Create XY Event Layer" tool, select "Longitude" for the X field, and select "Latitude" for the Y field, convert the table into a point feature and add it to the map, use the "Point Set to Line" tool, select "Track ID" as the line field, and generate a track line.

[0111] S2032: predefine a second distance threshold between the stay points and a minimum cluster point.

[0112] Specifically, the second distance threshold (Distance Threshold) and the minimum clustering points (Minpts) are defined in consideration of the real space scale, and in this embodiment, they are set to 150 meters and 20 respectively.

[0113] S2033: Generate a cluster of stay points according to the second distance threshold and the minimum cluster point.

[0114] Specifically, use the "Density-based Clustering" tool, select the stop point feature layer, select "Defined Distance (DBSCAN)" as the clustering method, enter the "Minimum number of features per cluster" and "Search distance", that is, the previously defined minimum number of cluster points (Minpts) and distance threshold (Distance Threshold), generate clusters of trajectory points, click on the cluster layer, right-click and select "Attribute Table", and each cluster is limited by a cluster ID.

[0115] S2034. Filter the cluster points according to the map, delete the clusters that do not belong to the recreation type, and merge and decompose the clusters.

[0116] Theoretically, the point cloud identified by the DBSCAN clustering method is essentially the spatial aggregation of tourist trajectory points and cannot be directly equated with recreational areas. Therefore, the cluster points should be screened according to the map, clusters that do not belong to the recreational type should be deleted, and clusters should be fused and decomposed based on daily experience and actual conditions.

[0117] In this embodiment, cluster points are screened by comparing with the map, non-recreational points such as industrial parks and residential areas are deleted, and scattered clusters with the same recreational content and adjacent recreational spaces are merged. For example, two relatively small cluster areas generated near Chen Clan Ancestral Hall are merged, and finally 37 trajectory point clusters are formed. Figure 3 shown.

[0118] S2035. Input cluster point features, generate TIN, and convert TIN into line features.

[0119] Specifically, use the "Create TIN" tool, input the cluster point features, select "Cluster ID" in the height field, click Run, and generate TIN. Use the "TIN Edge" tool to convert TIN into line features. There are many lines in the TIN line feature layer that do not belong to the cluster. Manually select these lines and delete them.

[0120] S2036. Convert the TIN line features into surface features, spatially connect the TIN lines with the cluster points, and spatially connect the TIN surface features and the TIN lines that are already spatially connected to the cluster points.

[0121] Specifically, use the "Feature to Polygon" tool to convert TIN line features into polygon features; use the "Spatial Join" tool, select the TIN line as the target feature, select the cluster point as the connecting feature, and spatially connect the TIN line with the cluster point; use the "Spatial Join" tool again, select the TIN line that has been spatially connected to the cluster point as the connecting feature, and spatially connect the TIN polygon feature and the TIN line that has been spatially connected to the cluster point; click on the polygon feature layer after the spatial connection, right-click and select "Attribute Table", and check whether the "Cluster ID" field already exists.

[0122] S2037. According to the cluster ID, the segmented TIN surfaces are merged to generate a stay area.

[0123] Specifically, observe the attribute table of the TIN surface feature layer after spatial connection. The TIN surface in each cluster is divided by TIN lines. Therefore, it is necessary to merge the divided TIN surfaces according to the cluster ID. Use the "Merge" tool, select "Cluster ID" as the fusion field, "Cluster ID" as the statistical data field, select "Sum" as the statistical type, and cancel the line segmentation. Fusion generates stay areas. In this embodiment, 37 stay areas are finally generated by fusion, as shown in the figure below. Figure 4 shown.

[0124] S204: The stay area and the trajectory line intersecting the stay area are used as a two-mode network, the length of the trajectory in the stay area is used as a weight, and the connection heat between the stay areas is calculated.

[0125] Furthermore, step S204 specifically includes:

[0126] S2041, cutting off the portion of the trajectory line outside the stay area.

[0127] Specifically, use the "Crop" tool, select the stay area as the cropping element, and crop the portion of the trajectory line generated in step S302 that is outside the stay area.

[0128] S2042: Explode the trajectory line in the stay area, so that one trajectory line is decomposed into multiple trajectory lines located in multiple stay areas.

[0129] Specifically, observing the trajectory line at this time, one trajectory line includes multiple parts located in multiple stop areas, so the "Multiple Parts to Single Part" tool is used to explode the trajectory line in the stop area. At this time, one trajectory line is decomposed into multiple trajectory lines located in multiple stop areas.

[0130] S2043, spatially connecting the exploded trajectory line with the stay area to obtain trajectory line layer data containing stay area information.

[0131] Specifically, use the "Spatial Join" tool, select the exploded trajectory line as the target element, select the stay area as the connecting element, spatially connect the exploded trajectory line with the stay area, and obtain the trajectory line layer data containing the stay area information; click the trajectory line layer after spatial connection, right-click and select "Attribute Table", and there are already two fields "Cluster ID" and "Trajectory ID".

[0132] S2044. Generate a stop area-trajectory data table based on the trajectory line layer data.

[0133] Specifically, observe the attribute table of the track line layer and check that the "cluster ID", "track ID" and "track length" of any track line are complete. Use the "table to excel" tool to export the attribute table to an excel file; open the excel file, select all data areas, click Insert-Pivot Table, select "cluster ID" for rows, "track ID" for columns, and "track length" for values ​​to generate a stop area-track data table.

[0134] S2045. According to the stay area-trajectory data table, a matrix of the "stay area-trajectory" two-mode network is formed, which is defined as V. i and j are defined as the number of rows of V, representing the stay area, and k is defined as the number of columns of V, representing the trajectory. V i,k and V j,k They represent the trajectory lengths of trajectory k within the stay areas i and j, respectively, as weights.

[0135] Specifically, open Mathcad, click Insert-Data-Table, insert a table, copy the stay area-trajectory data table in Excel, and paste it into the table in Mathcad to form a matrix of the "stay area-trajectory" two-mode network, defined as V, i and j are defined as the number of rows of V, representing the stay area, k is defined as the number of columns of V, representing the trajectory, and finally a matrix of 37 (0-36) rows and 9575 (0-9574) columns is formed; V i,k and V j,k They represent the trajectory lengths of trajectory k within the stay areas i and j, respectively, and are used as weights for subsequent calculations.

[0136] S2046. Calculate the stay area-stay area connection network based on the matrix of the two-mode network, and obtain the connection between stay areas i and j through the co-occurrence network algorithm.

[0137] Specifically, in the field of bibliometrics, based on the "co-occurrence" network formed by word frequency, the data matrix of the stay area-trajectory two-mode network is integrated into a stay area-stay area connection network to measure spatial connections, and the co-occurrence network algorithm (CNM) is calculated in Mathcad. Define and input the function equation as follows:

[0138]

[0139] Among them, C i,j Represents the connection between stay areas i and j calculated by the co-occurrence network algorithm.

[0140] S2047. According to the location recommendation model, the sum of the lengths of the trajectory k passing through the stop area and the sum of the lengths of all trajectory lines within the stop area i are calculated, thereby calculating the connection between the stop areas i and j.

[0141] In order to reduce the impact of the inherent defects of the co-occurrence network algorithm, such as ignoring the number of stopover areas along the trajectory and the scale of trajectories within the stopover areas, the two parameters, the length of the stopover area along the trajectory and the total length of all trajectories in the stopover area, are incorporated into the network recommendation process. That is, the slow-moving trajectory distance is used as a parameter to allocate resources successively in each trajectory and each stopover area, thereby showing the regional differences (i.e., the asymmetry of the connection) and truly depicting the slow-moving space network of the old city. This algorithm is called the location recommendation model (LRM). First, the sum of the lengths of the stopover areas along the trajectory k is calculated as follows:

[0142]

[0143] Among them, FX i,0 Represents the sum of the lengths of all trajectory lines within the stop zone i.

[0144] Then calculate the connection between any stay areas i and j, and input the function equation as follows:

[0145]

[0146] Among them, DI i,j It indicates that the connection between stay areas i and j is calculated by the location recommendation model. In theory, the LRM algorithm takes into account the asymmetry and recommendation effect between the connection between recreation areas i and j.

[0147] S2048. Calculate the cosine similarity according to the interest community discovery algorithm, analyze the similarity between the stay areas i and j, calculate the number of trajectories inside the stay areas i and j, and thus calculate the connection between the stay areas i and j.

[0148] Specifically, the third interest community discovery algorithm is based on the similarity of the stay area. For the stay areas i and k involved in the same trajectory, the similarity between the stay areas i and j is first analyzed by calculating the cosine similarity, as shown in the following formula:

[0149]

[0150] Among them, SI i,j Represents the cosine similarity between stay areas i and j. The closer the value is to 1, the higher the similarity between stay areas i and j.

[0151] The community of interest discovery algorithm also takes into account the size of the stay area and calculates the number of trajectories inside stay areas i and j as follows:

[0152]

[0153] Among them, ST i,j It indicates that the connection between stay areas i and j is calculated by the interest community discovery algorithm.

[0154] This embodiment tests the effects of three network connection algorithms, as follows:

[0155] 1) Delete 5% of the original trajectory data

[0156] 1.1) Randomly divide the original data into twenty equal parts: Create a new Excel file, paste the trajectory-dwelling area data table, use the randbetween(101, 120) function and the int function to randomly generate integers between 101 and 120 in the same area as the trajectory-dwelling area data table, and divide the generated random integers into twenty equal parts on average; click Insert-Data-Table in Mathcad to create a new table, copy the randomly generated integers between 101 and 120 into it to form a matrix RA, use this as a reference to the original real data, and divide the original real data into twenty equal parts.

[0157] 1.2) Using a set of twenty equally divided random integers, delete 5% of the original trajectory data, as shown below:

[0158]

[0159]

[0160] Among them, S j,k represents the remaining 95% of the original data, which is used as the subsequent training set, and NT represents the number of missing trajectories.

[0161] 2) Based on the training set, 5% of the missing data is predicted as follows:

[0162]

[0163]

[0164] Among them, UR j,0 represents the average length of the trajectory in the stay zone j in the training set, CS j,k It represents the difference between the remaining 95% of the trajectory length data in the stop zone j in the training set and the average trajectory length. This difference will be used to predict the missing trajectory later.

[0165] 3) Predict the missing trajectory based on the data set obtained by the collinear network algorithm and examine its difference with the original real data

[0166] 3.1) The missing trajectory is predicted based on the data set calculated by the co-occurrence network algorithm as follows:

[0167]

[0168] Assume that there are several “neighboring” regions i in the stop zone j (i.e., the stop zone i has trajectories belonging to the stop zone j), C j,i Represents the connection between stay area i and stay area j obtained by the co-occurrence network algorithm, thus obtaining the predicted value VPINM of the censored trajectory j,k .

[0169] 3.2) The root mean square error and mean absolute error between the predicted value of the censored trajectory and the original true data are calculated as follows:

[0170]

[0171]

[0172] Through this method, the root mean square error (RMSEINM) and mean absolute error (MAEINM) between the original true data of the censored data and the predicted set calculated according to the co-occurrence network algorithm were obtained.

[0173] 4) Predict the missing trajectory based on the data set obtained by the location recommendation algorithm and examine its difference with the original real data

[0174] 4.1) The missing trajectory is predicted based on the data set calculated by the location recommendation algorithm, as follows:

[0175]

[0176] Among them, DI j,i Represents the connection between stay area i and stay area j obtained by the location recommendation algorithm, thus obtaining the predicted value VPLR of the censored trajectory j,k .

[0177] 4.2) Calculate the root mean square error and mean absolute error between the predicted value of the censored trajectory and the original true data as follows:

[0178]

[0179]

[0180] Through this method, the root mean square error (RMSELRM) and mean absolute error (MAELRM) between the original real data of the censored data and the predicted set obtained according to the calculation results of the location recommendation algorithm are obtained.

[0181] 5) Predict the missing trajectory based on the data set obtained by the community of interest discovery algorithm and examine its difference with the original real data

[0182] 5.1) The missing trajectory is predicted based on the data set calculated by the interest community discovery algorithm, as follows:

[0183]

[0184] Among them, ST j,i It represents the connection between the stay area i and the stay area j obtained by the community of interest discovery algorithm, thus obtaining the predicted value VPDICM of the censored trajectory j,k .

[0185] 5.2) Calculate the root mean square error and mean absolute error between the predicted value of the deleted trajectory and the original true data as follows:

[0186]

[0187]

[0188] Through this method, the root mean square error and mean absolute error between the original real data of the deleted data calculated according to the three algorithms and the predicted set are obtained, as shown in Table 1 below; Generally speaking, the larger the calculation results of RMSE and AME, the greater the difference from the original real data results, and the lower the effectiveness and accuracy of the stay area prediction.

[0189] Table 1 Root mean square error and mean absolute error calculated according to three algorithms

[0190]

[0191] S205. Regenerate a connection network based on the three largest connection stay areas of each stay area, calculate and visualize the structural parameters of the topological network, and make an algorithm recommendation based on the balance of the recommendation.

[0192] Furthermore, this step specifically includes:

[0193] S2051. Calculate the maximum three connected stay areas of each stay area obtained by the co-occurrence network algorithm, the location recommendation algorithm and the interest community discovery algorithm respectively, and generate a matrix.

[0194] S20511. Export the stay area-stay area two-mode network data table: Select the matrix obtained by the co-occurrence network algorithm, left-click and select export, select Microsoft Excel as the file format, and export the Excel data table of the connection network between stay areas calculated by the algorithm. The rows and columns of the data table are a number of stay areas, and the values ​​represent the connection values ​​between the stay areas. According to the above method, export the Excel data table of the connection network between stay areas calculated by the location recommendation algorithm and the interest community discovery algorithm, as shown in Table 2-4 below.

[0195] Table 2 Matrix obtained by co-occurrence network algorithm

[0196]

[0197]

[0198] Table 3 Matrix obtained by location recommendation algorithm

[0199]

[0200]

[0201] Table 4 Matrix obtained by interest community discovery algorithm

[0202]

[0203]

[0204] S20512. Global analysis, calculate the three stay areas with the strongest connectivity with other stay areas: open the Excel data table obtained by the co-occurrence network algorithm, calculate the average value of the connection value between each column of the stay area and other stay areas, compare the average values ​​to obtain the three stay areas with the strongest connectivity with other stay areas, and then pass this on to tourists as a reminder.

[0205] S20513. Open the Excel data table obtained by the co-occurrence network algorithm, first use the large function to filter out the value of the maximum fourth connected stay area of ​​each stay point, and then use the if function. If the connection value between the two stay areas is greater than the maximum fourth value obtained by the previous screening, then retain the value; if it is less than the maximum fourth value obtained by the previous screening, then change the value to 0, so as to obtain the maximum three connected stay areas of each stay area.

[0206] S20514. Due to the asymmetry of the location recommendation algorithm, it is necessary to analyze the connection network of the stay area from the perspective of recommendation (→) and recommendation (←). From the definition of the location recommendation model, DI i,j Refers to the recommendation from stay area i to stay area j, with the first row header of the Excel data table as the object, which is called the "recommended (←) perspective"; on the contrary, the Excel data table obtained by the location recommendation algorithm is XY transposed, which is called the "recommended (→) perspective", so that an asymmetric location recommendation algorithm data table in two directions is obtained, and the above S20513 operation is repeated, that is, the Excel data table obtained by the location recommendation algorithm is opened, and the large function is first used to filter out the value of the maximum fourth connected stay area of ​​each stay point, and then the if function is used. If the connection value between the two stay areas is greater than the maximum fourth value obtained by the previous screening, the value is retained; if it is less than the maximum fourth value obtained by the previous screening, the value is changed to 0, so that the maximum three connected stay areas of each stay area are obtained.

[0207] S20515. Repeat the operation of S20512, that is, open the Excel data table obtained by the interest community discovery algorithm, first use the large function to filter out the value of the maximum fourth connected stay area of ​​each stay point, and then use the if function. If the connection value between the two stay areas is greater than the maximum fourth value obtained by the previous screening, then retain the value; if it is less than the maximum fourth value obtained by the previous screening, then change the value to 0, so as to obtain the maximum three connected stay areas of each stay area.

[0208] S20516. Copy the above four data tables to four sheets of the same Excel file respectively, so as to calculate the subsequent topological connection network structure parameters.

[0209] S2052. Calculate the average distance, diameter and reciprocity of the topological network.

[0210] S20521. Open Ucinet software, modify the default save location at the bottom of the software, click Data-Import Excel-Full matrix w / multiple sheets (obsolete), select the Excel file containing four sheets, click "Ok", generate a ##H file containing four sheet data tables, and save it in the system default folder.

[0211] S20522. Considering the balance of the recommended stay areas, the connection network topology parameters of the three algorithms are calculated: the average distance, diameter and reciprocity of the topological network are calculated, where the average distance refers to the average number of edges of the path with the least number of edges between any two nodes in the network, and the diameter refers to the maximum number of edges of the shortest path between any two nodes in the network. The purpose of calculating reciprocity is to calculate whether the relationship between members in the network is mutual. Two methods are used to calculate it. Arc-reciprocity refers to the proportion of directed edges with reciprocating relationships to all edges, and Dyad-reciprocity refers to the proportion of the number of tuples with reciprocating relationships to all tuples connected by edges. Click Network-cohesion-multiple cohesion measures of Ucinet, enter the previously generated ##H file, click "OK", view Avg distance, Diameter, Arc Reciprocity and Dyad Reciprocity in the generated results, and obtain the four topological network parameters of average distance, diameter, Arc-reciprocity and Dyad-reciprocity of the topological network, as shown in Table 5 below.

[0212] Table 5. Some topological parameters of networks with different algorithms

[0213]

[0214] Analyzing the above table, considering the balance of recommended stay areas, examining the connection network topology parameters of different algorithms shows that in the connection network that takes the top three recommended hotspots, the location recommendation algorithm of the recommendation perspective (→) is outstanding in parameters such as average distance, diameter, and reciprocity, indicating that the network concentration obtained by this algorithm is low and the interaction balance is good. On the basis of ensuring the co-occurrence network, it promotes the extension of the tour time and tour path, which is conducive to expanding the stay area and reducing the recommendation ratio of overheated areas.

[0215] S2053. For the co-occurrence network algorithm and the interest community discovery algorithm, the generated matrix is ​​copied and XY transposed, and the data before and after the transposition are added to obtain a symmetrical data table, and half of the data is deleted along the diagonal of the data table to obtain a triangular matrix.

[0216] Specifically, the sheet matrix generated by the previous co-occurrence network algorithm and interest community discovery algorithm is copied and XY transposed, and the data before and after the transposition are added to obtain a symmetrical data table. Half of the data is deleted along the diagonal of the data table to obtain a triangular matrix sheet for subsequent connection network construction.

[0217] S2054. Generate a connection network diagram of the stay area of ​​the co-occurrence network algorithm, the location recommendation algorithm and the interest community discovery algorithm according to the average distance, diameter and reciprocity of the topological network, and visualize it;

[0218] Specifically, open the Ucinet software, click Data-Import Excel-Full matrix w / multiple sheets (obsolete), generate a ##H file containing four sheets of triangular matrices of the co-occurrence network algorithm and the interest community discovery algorithm, and save them in the default folder; click Visualize-NetDraw of Ucinet, click File, open the previously stored ##H file, generate a total of four stay area connection network diagrams for three algorithms, and adjust the topological structure according to the real geographical location of Guangzhou Old City, such as Figure 5-Figure 8 shown.

[0219] observe Figure 5-Figure 8 , the co-occurrence network, interest community discovery and location recommendation algorithms from the perspective of recommendation (←) all have obvious agglomeration centers, namely Haizhu Square, Beijing Road, Zhongshan Memorial Hall, etc., and partially present a star-shaped topological network structure, among which the core agglomeration of the interest community discovery algorithm is slightly weaker; in addition, there is no obvious star-shaped structure in the stay point connection network system of the location recommendation algorithm from the perspective of recommendation (→), and the overall balance of the topological network is relatively obvious. In the northern part of the old city, Liwan Lake Park - Longjin West Road - Chen Clan Academy - Xihua Road - Liuhua Lake Park - Orchid Garden - Zhongshan Memorial Hall, Shamian - Cultural Park - Yide Road - Haizhu Square and other areas present a chain structure under the constraint of proximity. At the same time, Haizhu Square, Beijing Road and other stay points do not show significant agglomeration characteristics.

[0220] S2055. For mature urban recreation areas, based on the location recommendation algorithm, a partial chain structure of recommendation pairs is formed.

[0221] The connection network structures and structural parameters generated by the four recommendation methods of the three algorithms are observed; specifically, in the co-occurrence network algorithm, the trajectory is not allocated, which is relatively beneficial to relatively well-known recreational areas such as traditional large parks and scenic spots (because the larger the area, the easier it is to cover more trajectory lengths in theory), thus forming a more obvious star-shaped clustering feature; in the interest community discovery algorithm, due to the dilution effect of the resource allocation process, some relatively weak connections will be blurred to a certain extent, so the overall structure is also relatively clustered; in contrast, from the recommendation (→) perspective of the location recommendation algorithm, since the recreational connection is affected by the number of trajectories covered by the area itself, there will be obvious directional recommendation features, forming a partial chain structure recommendation pair; for mature urban recreational areas, the connection network of this algorithm is conducive to avoiding excessive agglomeration of scenic spots in the area and avoiding overheating, while reducing the agglomeration risk of local areas while extracting generally well-known scenic spots, which is relatively in line with the actual situation and has more recommendation significance.

[0222] It should be noted that although the method operations of the above embodiments are described in a specific order, this does not require or imply that these operations must be performed in this specific order, or that all the operations shown must be performed to achieve the desired results. On the contrary, the steps depicted can be performed in a different order. Additionally or alternatively, some steps can be omitted, multiple steps can be combined into one step, and / or one step can be decomposed into multiple steps.

[0223] Embodiment 2:

[0224] like Fig. 9 As shown, this embodiment provides a slow-moving space connection network recommendation system for mature urban recreation areas, which includes an acquisition module 901, an integration module 902, a clustering module 903, a calculation module 904 and a recommendation module 905. The specific functions of each module are as follows:

[0225] The acquisition module 901 is used to acquire the “crowd-sourced” slow-moving recreational GPS tracks within a given geographical area to form a GPS track data set.

[0226] The integration module 902 is used to identify the trajectories with the stop feature by screening the trajectories in the GPS trajectory data set by time, distance and number of trajectory points, and integrate the trajectories with the stop feature into stop points.

[0227] The clustering module 903 is used to perform spatial clustering on the stay points and convert the clustered point elements into surface elements to obtain the stay areas.

[0228] The calculation module 904 is used to use the stay area and the trajectory line intersecting the stay area as a two-mode network, use the length of the trajectory in the stay area as a weight, and calculate the connection heat between the stay areas.

[0229] The recommendation module 905 is used to regenerate the connection network based on the maximum three connected stay areas of each stay area, calculate the structural parameters of the topological network and visualize it, and make algorithm recommendations based on the balance of the recommendations.

[0230] It should be noted that the system provided in this embodiment is only illustrated by the division of the above-mentioned functional modules. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure can be divided into different functional modules to complete all or part of the functions described above.

[0231] Embodiment 3:

[0232] This embodiment provides a computer device, such as Fig.10 As shown, it includes a processor 1002, a memory, an input device 1003, a display device 1004 and a network interface 1005 connected through a system bus 1001, the processor is used to provide computing and control capabilities, the memory includes a non-volatile storage medium 1006 and an internal memory 1007, the non-volatile storage medium 1006 stores an operating system, a computer program and a database, the internal memory 1007 provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium, when the processor 1002 executes the computer program stored in the memory, the method for recommending a connection network for slow-moving spaces in urban mature recreation areas of the above-mentioned embodiment 1 is implemented.

[0233] Embodiment 4:

[0234] This embodiment provides a storage medium, which is a computer-readable storage medium and stores a computer program. When the computer program is executed by a processor, the method for recommending a connection network of slow-moving spaces in mature urban recreation areas of the above-mentioned embodiment 1 is implemented.

[0235] It should be noted that the computer-readable storage medium of the present embodiment may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0236] In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device, or device. In this embodiment, the computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium may also be any computer-readable storage medium other than a computer-readable storage medium, which may send, propagate, or transmit a program used by or in combination with an instruction execution system, device, or device. The computer program contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0237] The computer readable storage medium can be written in one or more programming languages ​​or a combination thereof to execute the computer program of the present embodiment, and the programming language includes an object-oriented programming language, such as Java, Python, C++, and also includes a conventional procedural programming language, such as C language or a similar programming language. The program can be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on a remote computer, or completely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., using an Internet service provider to connect through the Internet).

[0238] In summary, the present invention provides a "bottom-up" recommendation method for slow-moving space connection networks in mature urban recreational areas, which can avoid overcrowding of tourists in scenic spots, provide a reference for scenic spot safety management, and improve individual tourism experience.

[0239] The above description is only a preferred embodiment of the present invention. It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention.

Claims

1. A method for recommending a connection network of slow-moving spaces in mature urban recreational areas, characterized in that: The method comprises: Obtain the "crowd-sourced" slow-travel recreation GPS trajectories within a given geographical area to form a GPS trajectory dataset; By screening the trajectories in the GPS trajectory data set by time, distance and number of trajectory points, the trajectories with stay features are identified and integrated into stay points. Spatial clustering of the stay points is performed, and the clustered point features are converted into surface features to obtain the stay area; The stay area and the trajectory lines intersecting the stay area are regarded as a two-mode network, the length of the trajectory in the stay area is used as the weight, and the connection heat between the stay areas is calculated; Regenerate the connection network based on the three largest connection stay areas of each stay area, calculate and visualize the structural parameters of the topological network, and make algorithm recommendations based on the balance of recommendations; The method of obtaining the "crowd-sourced" slow-moving recreational GPS tracks within a given geographical area to form a GPS track data set specifically includes: Select keywords to obtain the "crowd-sourced" slow-moving recreation GPS tracks, clean the track data, delete all track data with missing attribute information, duplication, and high speed, filter out the slow-moving recreation track data, and obtain a track data set consisting of several tracks, each track includes several track points with track ID, longitude, latitude, timestamp information, and driving speed; The spatial clustering of the stay points specifically includes: Add the range area map and stop points to the current map, and generate track lines based on the track ID; predefine a second distance threshold between the stay points and a minimum cluster point; Generate a cluster of stay points according to the second distance threshold and the minimum cluster point, each cluster is defined by a cluster ID; The cluster points are screened by comparing with the map, the clusters that do not belong to the recreation type are deleted, and the clusters are merged and decomposed; The step of converting the clustered point elements into surface elements to obtain the stay area specifically includes: Input clustered point features, generate TIN, and convert TIN to line features; Convert TIN line features to surface features, spatially connect TIN lines to cluster points, and spatially connect TIN surface features to TIN lines that are already spatially connected to cluster points; According to the cluster ID, the segmented TIN surfaces are merged to generate the stay area; The method uses the stay area and the trajectory line intersecting the stay area as a two-mode network, uses the length of the trajectory in the stay area as a weight, and calculates the connection heat between the stay areas, specifically including: Cut off the part of the trajectory line outside the stay area; Explode the track line in the stay area, so that one track line is decomposed into multiple track lines located in multiple stay areas; The trajectory line after explosion is spatially connected with the stay area to obtain the trajectory line layer data containing the stay area information; Generate a stop area-trajectory data table based on the trajectory line layer data; According to the stay area-trajectory data table, a matrix of the "stay area-trajectory" two-mode network is formed, which is defined as V. i and j are defined as the number of rows of V, representing the stay area, and k is defined as the number of columns of V, representing the trajectory. V i,k and V j,k Respectively represent the trajectory length of trajectory k within the range of stay area i and j, as weights; The stay area-stay area connection network is calculated based on the matrix of the two-mode network, and the connection between stay areas i and j is calculated by the co-occurrence network algorithm; According to the location recommendation algorithm, the sum of the lengths of the trajectory k passing through the stop area and the sum of the lengths of all trajectory lines within the stop area i are calculated, thereby calculating the connection between the stop areas i and j; According to the interest community discovery algorithm, the cosine similarity is calculated, the similarity between the stay areas i and j is analyzed, and the number of trajectories inside the stay areas i and j is calculated, so as to calculate the connection between the stay areas i and j; The method regenerates the connection network based on the maximum three connected stay areas of each stay area, calculates and visualizes the structural parameters of the topological network, and performs algorithm recommendation in combination with the balance of the recommendation, specifically including: Calculate the maximum three connected stay areas of each stay area obtained by the co-occurrence network algorithm, location recommendation algorithm and interest community discovery algorithm respectively, and generate a matrix; Calculate the average distance, diameter and reciprocity of the topological network. The average distance refers to the average value of the number of edges of the path with the least number of edges between any two nodes in the network. The diameter refers to the maximum value of the number of edges of the shortest path between any two nodes in the network. The reciprocity includes Arc-reciprocity and Dyad-reciprocity. Arc-reciprocity refers to the ratio of directed edges with reciprocating relationships to all edges. Dyad-reciprocity refers to the ratio of the number of tuples with reciprocating relationships to all tuples connected by edges. For the co-occurrence network algorithm and the community of interest discovery algorithm, the generated matrix is ​​copied and XY transposed, and the data before and after the transposition are added to obtain a symmetrical data table. Half of the data along the diagonal of the data table is deleted to obtain a triangular matrix. Generate and visualize the stay area connection network diagrams of the co-occurrence network algorithm, location recommendation algorithm, and interest community discovery algorithm based on the average distance, diameter, and reciprocity of the topological network; For mature urban recreation areas, based on the location recommendation algorithm, recommendation pairs with partial chain structures are formed.

2. The method for recommending a connection network of slow-moving spaces in mature urban recreation areas according to claim 1 is characterized in that: The method of screening the trajectories in the GPS trajectory data set by time, distance and number of trajectory points, identifying the trajectories with the stop feature, and integrating the trajectories with the stop feature into the stop points specifically includes: De-duplicate the trajectory data, filter out the first trajectory point of each trajectory, and use this trajectory point as the trajectory label; Calculate each trajectory in turn, and for each trajectory point of a trajectory, calculate the time difference between the trajectory point and the first trajectory point, and unify the time difference into units to form the time difference values ​​between several trajectory points of each trajectory and the first trajectory point; The first screening function is defined as calculating the time difference between each track point in each track and other track points, and calculating the distance difference between each track point in each track and other track points according to the longitude and latitude of the existing track points; The second screening function is defined to screen the distance of each trajectory. When the distance difference between two trajectory points is less than the first distance threshold, the screening is qualified. The third filtering function is defined as filtering the time and number of trajectory points of each trajectory based on the filtering distance. When the time difference between two trajectory points is greater than the time threshold and the number of trajectory points is greater than or equal to the point threshold, the screening meets the criteria, and the trajectory segment that meets the screening criteria is assigned an ID; Define a first distance threshold, a time threshold, and a point threshold; According to the first distance threshold, time threshold and point number threshold, all trajectories are calculated and screened to obtain trajectories with stay characteristics; The track points are grouped according to the track segment ID, and the average values ​​of the longitude, latitude and time of the grouped track points are calculated and integrated to obtain the stop points.

3. A network recommendation system for slow-moving space connections in mature urban recreational areas, characterized in that: The system comprises: The acquisition module is used to obtain the "crowd-sourced" slow-moving recreational GPS trajectories within a given geographical area to form a GPS trajectory dataset; The integration module is used to identify the trajectories with the stay feature by screening the trajectories in the GPS trajectory data set by time, distance and number of trajectory points, and integrate the trajectories with the stay feature into stay points; The clustering module is used to spatially cluster the stay points and convert the clustered point features into surface features to obtain the stay areas; A calculation module is used to use the stay area and the trajectory lines intersecting the stay area as a two-mode network, use the length of the trajectory in the stay area as a weight, and calculate the connection heat between the stay areas; The recommendation module is used to regenerate the connection network based on the maximum three connection stay areas of each stay area, calculate the structural parameters of the topological network and visualize it, and make algorithm recommendations based on the balance of the recommendations; The method of obtaining the "crowd-sourced" slow-moving recreational GPS tracks within a given geographical area to form a GPS track data set specifically includes: Select keywords to obtain the "crowd-sourced" slow-moving recreation GPS tracks, clean the track data, delete all track data with missing attribute information, duplication, and high speed, filter out the slow-moving recreation track data, and obtain a track data set consisting of several tracks, each track includes several track points with track ID, longitude, latitude, timestamp information, and driving speed; The spatial clustering of the stay points specifically includes: Add the range area map and stop points to the current map, and generate track lines based on the track ID; predefine a second distance threshold between the stay points and a minimum cluster point; Generate a cluster of stay points according to the second distance threshold and the minimum cluster point, each cluster is defined by a cluster ID; The cluster points are screened by comparing with the map, the clusters that do not belong to the recreation type are deleted, and the clusters are merged and decomposed; The step of converting the clustered point elements into surface elements to obtain the stay area specifically includes: Input clustered point features, generate TIN, and convert TIN to line features; Convert TIN line features to surface features, spatially connect TIN lines to cluster points, and spatially connect TIN surface features to TIN lines that are already spatially connected to cluster points; According to the cluster ID, the segmented TIN surfaces are merged to generate the stay area; The method uses the stay area and the trajectory line intersecting the stay area as a two-mode network, uses the length of the trajectory in the stay area as a weight, and calculates the connection heat between the stay areas, specifically including: Cut off the part of the trajectory line outside the stay area; Explode the track line in the stay area, so that one track line is decomposed into multiple track lines located in multiple stay areas; The trajectory line after explosion is spatially connected with the stay area to obtain the trajectory line layer data containing the stay area information; Generate a stop area-trajectory data table based on the trajectory line layer data; According to the stay area-trajectory data table, a matrix of the "stay area-trajectory" two-mode network is formed, which is defined as V. i and j are defined as the number of rows of V, representing the stay area, and k is defined as the number of columns of V, representing the trajectory. V i,k and V j,k Respectively represent the trajectory length of trajectory k within the range of stay area i and j, as weights; The stay area-stay area connection network is calculated based on the matrix of the two-mode network, and the connection between stay areas i and j is calculated by the co-occurrence network algorithm; According to the location recommendation algorithm, the sum of the lengths of the trajectory k passing through the stop area and the sum of the lengths of all trajectory lines within the stop area i are calculated, thereby calculating the connection between the stop areas i and j; According to the interest community discovery algorithm, the cosine similarity is calculated, the similarity between the stay areas i and j is analyzed, and the number of trajectories inside the stay areas i and j is calculated, so as to calculate the connection between the stay areas i and j; The method regenerates the connection network based on the maximum three connected stay areas of each stay area, calculates and visualizes the structural parameters of the topological network, and performs algorithm recommendation in combination with the balance of the recommendation, specifically including: Calculate the maximum three connected stay areas of each stay area obtained by the co-occurrence network algorithm, location recommendation algorithm and interest community discovery algorithm respectively, and generate a matrix; Calculate the average distance, diameter and reciprocity of the topological network. The average distance refers to the average value of the number of edges of the path with the least number of edges between any two nodes in the network. The diameter refers to the maximum value of the number of edges of the shortest path between any two nodes in the network. The reciprocity includes Arc-reciprocity and Dyad-reciprocity. Arc-reciprocity refers to the ratio of directed edges with reciprocating relationships to all edges. Dyad-reciprocity refers to the ratio of the number of tuples with reciprocating relationships to all tuples connected by edges. For the co-occurrence network algorithm and the community of interest discovery algorithm, the generated matrix is ​​copied and XY transposed, and the data before and after the transposition are added to obtain a symmetrical data table. Half of the data along the diagonal of the data table is deleted to obtain a triangular matrix. Generate and visualize the stay area connection network diagrams of the co-occurrence network algorithm, location recommendation algorithm, and interest community discovery algorithm based on the average distance, diameter, and reciprocity of the topological network; For mature urban recreation areas, based on the location recommendation algorithm, recommendation pairs with partial chain structures are formed.

4. A computer device comprising a processor and a memory for storing a program executable by the processor, characterized in that: When the processor executes the program stored in the memory, the method for recommending a slow-moving space connection network in a mature urban recreation area as described in any one of claims 1-2 is implemented.

5. A storage medium storing a program, characterized in that: When the program is executed by a processor, the method for recommending a slow-moving space connection network in a mature urban recreation area as described in any one of claims 1-2 is implemented.

Citation Information

Patent Citations

  • Combined recommendation method of traveling scenic spots based on network data collection

    CN104881472A

  • Scenic spot recommendation method and device and storage medium

    CN110750737A