Metro station connection area shared bicycle flow mode identification method and device

Abstract

The application provides a subway station connection area shared bicycle flow mode recognition method and equipment. The method comprises steps 1 to 9. The application fully adopts the shared bicycle actual riding order data to generate the connection area, saves the complicated process of constructing the road network in the traditional method, improves the authenticity and effectiveness of the connection area, adopts the DTW distance capable of adaptive distortion matching on the time axis, reduces the influence of time lag on mode division, and improves the recognition efficiency of the shared bicycle flow mode.

Description

TECHNICAL FIELD

[0001] The embodiment of the present application relates to the technical field of geographic information, and particularly relates to a shared bicycle flow mode identification method and equipment in a subway station connection area. BACKGROUND

[0002] At present, the coverage of the subway is rapidly expanded, the service level is continuously improved, and more and more residents choose the subway as a travel mode, which drives the land development and economic growth along the subway. In addition to improving the travel service of the subway station, improving the connection service of the subway system is another important way to increase the attraction of rail transit and improve the operation efficiency of rail transit. The travel mode of the shared bicycle and the rail transit connection well solves the first kilometer / last kilometer problem. The problem caused by the disordered use of the shared bicycle is increasingly acute. Due to the commuting behavior of the residents and the separation of the urban function of residence and work, the imbalance between supply and demand of the shared bicycle frequently occurs in the peak commuting period of the day and the specific area. How to identify the space-time mode of the shared bicycle flow around the subway station becomes the key to solve the related problems. In the definition of the connection range, the existing method is mostly based on a certain distance to construct a circular buffer area with the subway station as the center or a buffer area based on the road network, but the connection area with the fixed radius does not match the shared bicycle riders with the bidirectional attribute (arrival / leave) in space, and the buffer area can contain places that cannot be reached. In the mode clustering, the existing method mostly uses the Euclidean distance measurement, but the Euclidean distance requires one-to-one correspondence between elements in a long sequence, and is very sensitive to the shift on the time length axis of the time length sequence data, and the morning and evening peak of the shared bicycle has the time period, and due to the "time lag" of the shared bicycle flow in different areas, the similarity measurement method of the "one-to-one correspondence" of the Euclidean distance is poor. Therefore, developing a shared bicycle flow mode identification method and equipment in the subway station connection area can effectively overcome the defects in the related art, and has become a technical problem to be solved in the industry. SUMMARY

[0003] In view of the above problems existing in the prior art, the embodiment of the present application provides a shared bicycle flow mode identification method and equipment in a subway station connection area.

[0004] In a first aspect, embodiments of the present application provide a subway station connection area shared bicycle flow pattern recognition method, comprising: step 1: collecting and processing the research area data set, including subway station entrance and exit data and shared bicycle order data; step 2: determining the connection area of the subway station and the shared bicycle; step 3: based on the connection area generated in step 2, screening the time and space information of the shared bicycle trip order in each subway station connection area; step 4: constructing two time length sequence indexes of standardized total flow and standardized net flow as the measurement indexes of the shared bicycle flow pattern in the subway station connection area; step 5: based on the contour coefficient and the sum of squared errors SSE, the optimal cluster number k of the K-medoids algorithm is determined, that is, k shared bicycle flow patterns; step 6: initializing the initial cluster center of the k shared bicycle flow patterns, calculating the DTW distance between the sample data and the k initial cluster centers, dividing all sample points to the nearest cluster center, and marking the sample after the first clustering; step 7: calculating the sum of the DTW distances between the sample points in the cluster and other sample points in the cluster, selecting the point with the smallest distance sum as the new cluster center, dividing all sample points to the nearest new cluster center and marking; step 8: if the iteration termination condition is met, step 9 is executed; if not, step 7 is executed until the iteration termination condition is met; wherein the iteration termination condition is that the cluster center of the class cluster does not change, or no sample point is re-assigned to a different class; step 9: output the shared bicycle clustering result and show its spatial distribution; wherein the DTW distance is a dynamic dynamic warping distance.

[0005] On the basis of the above method embodiment content, the subway station connection area shared bicycle flow pattern recognition method provided in the embodiments of the present application, the processing of the research area data set in step 1 includes: uniformly projecting and converting the spatial vector data to form a vector data set with consistent spatial coordinate system; cleaning the shared bicycle order data, considering the actual riding situation, taking into account the reasons for coordinate drift, not locking, GPS signal anomaly, deleting orders with abnormal spatial position, deleting orders with riding distance and time length not within the predetermined time length range, and deleting orders with riding speed not within the predetermined speed range; the determination of the connection area of the subway station and the shared bicycle in step 2 includes: assigning all riding orders starting or ending within 150m of the subway station entrance to the subway station, then regarding the orders within the first 85% of the riding distance of the subway station riding orders as potential connection orders and retaining them, and deleting the riding orders greater than the distance as leisure or commuting orders; using the DBSCAN algorithm to perform density clustering on the connection orders around the subway station, eliminating low-density noise values to avoid shape abnormalities in the connection area due to individual non-frequent orders; using the aggregation point tool in the ArcGIS software, setting the aggregation distance, and creating a polygon of the order points around the subway station to obtain the connection area of the subway station.

[0006] On the basis of the above method embodiment content, the subway station connection area in the shared bicycle flow mode recognition method provided in the embodiment of the application, the screening of the shared bicycle travel order space-time information in each subway station connection area in step 3 comprises: for the shared bicycle order whose starting point or terminal point falls in the connection area, extracting the unlocking longitude and latitude coordinates, the locking longitude and latitude coordinates, the unlocking time length and the locking time length.

[0007] On the basis of the above method embodiment content, the subway station connection area in the shared bicycle flow mode recognition method provided in the embodiment of the application, the calculation method of the two time length sequence indexes of the standardized total flow and the standardized net flow in step 4 comprises: dividing each day into 24 intervals with 1 hour as a step, as the length of the two time length sequence indexes of the standardized total flow and the standardized net flow, defining the inside flow of each subway station as the shared bicycle flow inside the subway station connection area, the outflow as the shared bicycle flow leaving the subway station connection area, and the inflow as the shared bicycle flow flowing into the subway station connection area, and the total flow TF and the net flow NF of any subway station i in the time length interval t are calculated by the following formula:

[0008] TF(i,t)=inf low(i,t)+outflow(i,t)+in side f low(i,t)

[0009] NF(i,t)=inf low(i,t)-outflow(i,t)

[0010] The total flow and the net flow are standardized by using the z-score standardization method, and the calculation formula comprises:

[0011]

[0012] Wherein, x is the total flow or the net flow, x * is the standardized total flow or net flow, is the mean value of the total flow or the net flow in a day, and sigma is the standard deviation of the total flow or the net flow in a day.

[0013] Based on the two-dimensional time length sequence vector composed of the standardized total flow and the standardized net flow, the two-dimensional time sequence vector with a length of 24 composed of the standardized total flow tf i and the standardized net flow nf i of any subway station k i is defined as follows, to represent the shared bicycle flow characteristics:

[0014]

[0015] On the basis of the above method embodiment content, the subway station connection area shared bicycle flow mode recognition method provided in the embodiment of the application, the calculation formula of the contour coefficient in step 5 includes:

[0016]

[0017] For the contour coefficient s of vector i (i) , a (i) is cohesion, the average distance from point i to other points in the cluster; b (i) is separation, the average distance from point i to all points outside the cluster, s (i) is closer to 1, the better the clustering performance, and the global contour coefficient is obtained by averaging the contour coefficients of all vectors; the calculation of the sum of squared errors includes:

[0018]

[0019] Where p is a sample vector, m i is the cluster center of the class cluster i, and k is the number of clusters; the elbow rule is used to determine the optimal number of clusters k, and the optimal number of clusters k will be the point at which the contour coefficient stops increasing rapidly and the sum of squared errors SSE stops decreasing rapidly, at which time increasing the number of clusters will not effectively improve the clustering result.

[0020] On the basis of the above method embodiment content, the subway station connection area shared bicycle flow mode recognition method provided in the embodiment of the application, the calculation of the DTW distance between the sample data and the k initial cluster centers in step 6 includes: for sample data a and cluster center b, corresponding to two time length sequences S1 and S2 respectively, a distance matrix D of a x b is created by calculating the squared Euclidean distance between any two elements of S1 and S2, then any twisted path passing through the distance matrix D is a set of matrix elements between S1 and S2, which satisfies three constraints: boundary condition: the first point and the last point of S1 and S2 are aligned, that is, the starting point and the ending point of the twisted path are located in the diagonal cells; continuity: the twisted path is aligned with adjacent points and cannot be cross-aligned; monotonicity: each point of S1 is matched with one or more points of S2, and vice versa, then the DTW distance is the minimum length of all twisted paths:

[0021]

[0022] Where P is the path set between S1 and S2, p k is a path in the path set, D[i,j] is the Euclidean distance between the i-th point S1 i of S1 and the j-th point S2 j of S2, including: D[i,j] = d(S1 i , S2 j ).

[0023] On the basis of the above method embodiment content, the subway station connection area shared bicycle flow mode recognition method provided in the embodiment of the application, the new cluster center calculation formula in step 7 includes:

[0024]

[0025] Wherein, x i is the sample i in the original cluster, x j is the sample j in the new cluster, and min is the minimum value.

[0026] In a second aspect, the embodiment of the application provides a subway station connection area shared bicycle flow mode recognition device, which comprises: a first main module, which is used for realizing steps 1: collecting and processing the research area data set, including subway station entrance and exit data and shared bicycle order data; step 2: determining the connection area of the subway station and the shared bicycle; and step 3: based on the connection area generated in step 2, screening the shared bicycle trip order space-time information in each subway station connection area; a second main module, which is used for realizing steps 4: constructing two time length sequence indexes of standardized total flow and standardized net flow, which are used as the measurement indexes of the shared bicycle flow mode in the subway station connection area; and step 5: based on the contour coefficient and the sum of squared errors SSE, determining the optimal cluster number k of the K-medoids algorithm, that is, k shared bicycle flow modes; a third main module, which is used for realizing steps 6: initializing the initial cluster centers of the k shared bicycle flow modes, calculating the DTW distance between the sample data and the k initial cluster centers, dividing all sample points to the nearest cluster center, and marking the sample to complete the first clustering; and step 7: calculating the sum of the DTW distances between the sample points in the cluster and other sample points in the cluster, selecting the point with the minimum distance sum as the new cluster center, and dividing all sample points to the nearest new cluster center and marking; a fourth main module, which is used for realizing steps 8: if the iteration termination condition is met, executing step 9; if not, executing step 7 until the iteration termination condition is met; wherein the iteration termination condition is that the cluster center of the cluster does not change, or no sample point is re-assigned to a different category; and step 9: outputting the shared bicycle clustering result and showing the spatial distribution; wherein the DTW distance is the dynamic dynamic warping distance.

[0027] In a third aspect, the embodiment of the application provides an electronic device, comprising:

[0028] at least one processor; and

[0029] at least one memory connected with the processor, wherein:

[0030] The memory stores program instructions executable by the processor, and the processor calling the program instructions can execute the subway station connection area shared bicycle flow mode recognition method provided by any one of various implementation manners of the first aspect.

[0031] In a fourth aspect, embodiments of the present application provide a non-transitory computer-readable storage medium storing computer instructions, which cause a computer to execute the subway station connection area shared bicycle flow mode recognition method provided by any one of various implementation manners of the first aspect.

[0032] The subway station connection area shared bicycle flow mode recognition method and device provided by the embodiments of the present application fully generate the connection area by using the actual shared bicycle riding order data, saves the complicated process of constructing the road network in the traditional method, improves the authenticity and effectiveness of the connection area, uses the DTW distance that can be adaptively twisted and matched on the time axis, reduces the influence of time lag on mode division, and improves the recognition efficiency of the shared bicycle flow mode. BRIEF DESCRIPTION OF DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0034] Figure 1 The subway station connection area shared bicycle flow mode recognition method flowchart provided by the embodiments of the present application;

[0035] Figure 2 The subway station connection area shared bicycle flow mode recognition device structure schematic diagram provided by the embodiments of the present application;

[0036] Figure 3 The physical structure schematic diagram of the electronic device provided by the embodiments of the present application;

[0037] Figure 4 The three method principle schematic diagrams for generating the subway station and the shared bicycle connection area provided by the embodiments of the present application;

[0038] Figure 5 The shared bicycle flow schematic diagram provided by the embodiments of the present application;

[0039] Figure 6 The shared bicycle convergence and divergence schematic diagram provided by the embodiments of the present application;

[0040] Figure 7A principle schematic view of SSE and contour coefficient values corresponding to different clustering clusters provided by the embodiment of the present application;

[0041] Figure 8 A DTW distance schematic view provided by the embodiment of the present application;

[0042] Figure 9 A shared bicycle tide type flow mode spatial distribution and TF and NF time-sharing statistical view provided by the embodiment of the present application;

[0043] Figure 10 A shared bicycle non-tide type flow mode spatial distribution and TF and NF time-sharing statistical view provided by the embodiment of the present application. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical scheme and advantages of the embodiment of the present application clearer, the technical scheme in the embodiment of the present application will be described clearly and completely below in combination with the drawings in the embodiment of the present application. Obviously, the described embodiment is a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the present application. In addition, the technical features in each embodiment or single embodiment provided by the present application can be combined with each other arbitrarily to form a feasible technical scheme, and this combination is not restricted by the order of steps and / or structure mode, but must be based on the realization by those skilled in the art, when the combination of technical schemes appears contradictory or cannot be realized, it should be considered that the combination of technical schemes does not exist, and is not within the protection scope required by the present application.

[0045] The embodiment of the present application provides a shared bicycle flow mode recognition method in a subway station connection area, referring to Figure 1The method comprises the following steps: step 1, collecting and processing the research area data set, including the subway station entrance and exit data and the shared bicycle order data; step 2, determining the transfer area of the subway station and the shared bicycle; step 3, screening the shared bicycle travel order space-time information in each subway station transfer area based on the transfer area generated in step 2; step 4, constructing two time length sequence indexes of standardized total flow and standardized net flow as the measurement indexes of the shared bicycle flow mode in the subway station transfer area; step 5, determining the optimal clustering number k of the K-medoids algorithm based on the contour coefficient and the sum of squared errors SSE, that is, k shared bicycle flow modes; step 6, initializing the initial clustering center of the k shared bicycle flow modes, calculating the DTW distance between the sample data and the k initial clustering centers, dividing all sample points to the nearest clustering center, and marking the sample to complete the first clustering; step 7, calculating the sum of the DTW distances of the sample points in the cluster to other sample points in the cluster, selecting the point with the minimum distance sum as the new clustering center, dividing all sample points to the nearest new clustering center and marking; step 8, if the iteration termination condition is met, step 9 is executed; if not, step 7 is executed until the iteration termination condition is met; wherein the iteration termination condition is that the clustering center of the cluster is not changed or the sample point is not re-assigned to a different category; step 9, outputting the shared bicycle clustering result and showing the spatial distribution; wherein the DTW distance is a dynamic dynamic warping distance.

[0046] Based on the content of the above method embodiment, as an optional embodiment, the subway station transfer area shared bicycle flow mode recognition method provided in the embodiment of the application comprises the following steps: in step 1, the research area data set is processed, including: uniformly projecting and converting the spatial vector data to form a vector data set with consistent spatial coordinate system; cleaning the shared bicycle order data, considering the actual riding situation, the reasons for coordinate drift, not locking, and abnormal GPS signal, deleting the orders with abnormal spatial position, deleting the orders with riding distance and time length not within the predetermined time length range, and deleting the orders with riding speed not within the predetermined speed range; in step 2, the transfer area of the subway station and the shared bicycle is determined, including: all riding orders starting or ending within 150m of the subway station entrance and exit are assigned to the subway station, then the orders with riding distance of 85% of the riding distance of the subway station order are considered as potential transfer orders and are reserved, and the orders with riding distance greater than the distance are deleted as leisure or commuting orders; using the DBSCAN algorithm to perform density clustering on the transfer orders around the subway station, eliminating the noise values with low density to avoid abnormal shape of the transfer area due to individual non-frequent orders; using the aggregation point tool in the ArcGIS software, setting the aggregation distance, creating the order points around the subway station into a polygon, and the transfer area of the subway station.

[0047] Based on the content of the above method embodiment, as an optional embodiment, the subway station connection area shared bicycle flow mode recognition method provided in the embodiment of the application, the screening of the shared bicycle trip order space-time information in each subway station connection area in step 3 includes: for the shared bicycle order whose starting point or ending point falls in the connection area, extracting the unlocking longitude and latitude coordinates, the locking longitude and latitude coordinates, the unlocking time length and the locking time length.

[0048] Based on the content of the above method embodiment, as an optional embodiment, the subway station connection area shared bicycle flow mode recognition method provided in the embodiment of the application, the calculation method of the two time length sequence indexes of the standardized total flow and the standardized net flow in step 4 includes: dividing each day into 24 intervals with 1 hour as a step, as the length of the two time length sequence indexes of the standardized total flow and the standardized net flow, defining the inside flow of each subway station as the shared bicycle flow inside the subway station connection area, the outflow as the shared bicycle flow leaving the subway station connection area, and the inflow as the shared bicycle flow flowing into the subway station connection area, and the total flow TF and the net flow NF of any subway station i in the time length interval t are calculated by the following formulas:

[0049] TF(i,t)=inflow(i,t)+outflow(i,t)+insideflow(i,t) (1)

[0050] NF(i,t)=inflow(i,t)-outflow(i,t) (2)

[0051] The total flow and the net flow are standardized by using the z-score standardization method, and the calculation formula includes:

[0052]

[0053] Wherein, x * is the standardized total flow or net flow, is the mean value of the total flow or net flow in a day, and sigma is the standard deviation of the total flow or net flow in a day, and x is the total flow or net flow;

[0054] Based on the two-dimensional time length sequence vector composed of the standardized total flow and the standardized net flow, the two-dimensional time sequence vector with a length of 24 composed of the standardized total flow tf i and the standardized net flow nf i of any subway station k i is defined as follows, to represent the shared bicycle flow characteristics:

[0055]

[0056] Based on the content of the above method embodiment, as an optional embodiment, the subway station connection area shared bicycle flow mode recognition method provided in the embodiment of the application, the calculation formula of the contour coefficient in step 5 includes:

[0057]

[0058] For the contour coefficient s of the vector i (i) , a (i) is the cohesion, the average distance from point i to other points in the cluster; b (i) is the separation, the average distance from point i to all points outside the cluster, s (i) is closer to 1, the better the performance of the cluster, and the global contour coefficient is obtained by averaging the contour coefficients of all vectors; the calculation of the sum of squared errors includes:

[0059]

[0060] Where p is the sample vector, m i is the cluster center of the class cluster i, and k is the number of clusters; the elbow rule is used to determine the optimal number of clusters k, and the optimal number of clusters k will be the point at which the contour coefficient stops increasing rapidly and the sum of squared errors SSE stops decreasing rapidly, at which time increasing the number of clusters will not effectively improve the clustering result.

[0061] Based on the content of the above method embodiment, as an optional embodiment, the subway station connection area shared bicycle flow mode recognition method provided in the embodiment of the application, the calculation of the DTW distance between the sample data and the k initial cluster centers in step 6 includes: for sample data a and cluster center b, corresponding to two time length sequences S1 and S2 respectively, a distance matrix D of a x b is created by calculating the squared Euclidean distance between any two elements of S1 and S2, then any twisted path passing through the distance matrix D is a set of matrix elements mapping between S1 and S2, satisfying three constraints: boundary condition: the first and last points of S1 and S2 are aligned, that is, the starting point and the ending point of the twisted path are located in the diagonal cells; continuity: the twisted path is aligned with adjacent points and cannot be cross-aligned; monotonicity: each point of S1 is matched with one or more points of S2, and vice versa, then the DTW distance is the shortest path in all twisted paths:

[0062]

[0063] Where P is the path set between S1 and S2, p k is a path in the path set, and D[i,j] is the Euclidean distance between the i-th point S1 i of S1 and the j-th point S2 j of S2, including: D[i,j] = d(S1 i , S2j ).

[0064] Based on the content of the above method embodiments, as an optional embodiment, the subway station connection area shared bicycle flow mode recognition method provided in the embodiment of the application includes that the new cluster center calculation formula in step 7 comprises:

[0065]

[0066] Wherein, x i is a sample i in the original cluster, x j is a sample j in the new cluster, and min is the minimum value.

[0067] The subway station connection area shared bicycle flow mode recognition method provided in the embodiment of the application fully uses the shared bicycle actual riding order data to generate the connection area, saves the complicated process of constructing the road network in the traditional method, improves the authenticity and effectiveness of the connection area, uses the DTW distance that can adaptively distort and match on the time axis, reduces the influence of time lag on mode division, and improves the recognition efficiency of the shared bicycle flow mode.

[0068] In another embodiment, the subway station connection area shared bicycle flow mode recognition method comprises: step 1: taking the main urban area of a city as a research area, collecting order data of shared bicycles of users in the city from November 1 to 7, 2019 and subway station entrance and exit data of the city at the end of 2019. Among them, the shared bicycle order data is collected in a time period, the weather condition is good, there is no special event affecting the use of shared bicycles, and the data contains six types of attributes such as user ID, shared bicycle ID, order starting point and ending point coordinates, shared bicycle unlocking and locking time, and the data format example is shown in Table 1. The subway station entrance and exit data includes the names of all subway stations, entrance and exit names, corresponding lines, station longitude and latitude, entrance and exit longitude and latitude of the city at the end of 2019, and the data format example is shown in Table 2.

[0069] Table 1: Shared bicycle order data format example

[0070]

[0071] Table 2: Subway station data format example

[0072]

[0073] The processing of data includes: (1) uniform projection conversion processing of spatial vector data to form a vector data set with consistent spatial coordinate system; (2) cleaning of shared bicycle order data, in combination with actual riding conditions, deleting the riding orders not in the coordinate range of the main urban area of the city, setting the riding distance threshold value between 0.5km-6km, and the riding time between 1min-2h. The setting of the riding speed refers to the “2018 Wuhan Shared Bicycle Report”, in combination with the operation conditions of the traffic system in the main urban area, and the speed threshold value is set between 5-20km / h.

[0074] Step 2: Based on the actual order data of shared bicycles, the transfer area of shared bicycles and subway stations is generated, and the specific process is as follows:

[0075] 1) Riding order distribution: all riding orders starting or ending within 150m of the subway station entrance in working days are distributed to the subway station, and then the orders with a riding distance less than 1.3km in the riding orders of the subway station are reserved as potential transfer orders, and the orders greater than the distance are deleted.

[0076] 2) Riding order noise point elimination: the DBSCAN algorithm is used to cluster the transfer orders around the subway station, and the noise values with low density are eliminated to avoid abnormal shape of the transfer area due to individual non-frequent orders, and the algorithm parameters are set as: within the range of 45m of the order OD point, less than 8 other order points are considered as noise values.

[0077] 3) Transfer area aggregation: the aggregation point tool in the ArcGIS software is used to aggregate the order points around the subway station into a polygon, the transfer area of the subway station, and the aggregation distance is set to 800m, which is equivalent to a 10-minute walking distance.

[0078] As shown in Figure 4 , the range of the transfer area obtained by three different methods is shown, and the method used by the present application can more accurately capture the transfer riding orders around the subway station and save the complicated process of constructing the road network.

[0079] Step 3: Screening the space-time information of shared bicycle travel orders in each transfer area of the subway station. According to the range of the transfer area and the spatial position of the travel order data, the shared bicycle orders with the starting point or the terminal point falling within the transfer area are screened, and the unlocking longitude and latitude coordinates, the locking longitude and latitude coordinates, the unlocking time and the locking time are extracted.

[0080] Step 4: Calculate two time sequence indexes of standardized total flow and standardized net flow, and the specific process is as follows:

[0081] Taking the data of Wednesday (November 6, 2019) as an example, the day is divided into 24 intervals with an interval of 1 hour. It should be explained that the date selected and the time interval set here are only for the purpose of illustration and are not limiting of the present application. In other embodiments, other dates or other suitable time intervals are selected. Since the morning and evening peak commuting periods of the city are generally 7:00-9:00 and 17:00-19:00, the demand and order generation of shared bicycles around the subway station are huge during this period. Therefore, 1 hour is selected as the interval for subsequent statistics of the total flow and net flow of shared bicycle flow.

[0082] In each 1-hour interval, define the insideflow of each subway station as the shared bicycle flow inside the subway station, the outflow as the shared bicycle flow leaving the subway station, and the inflow as the shared bicycle flow flowing into the subway station, Figure 5 The insideflow, outflow and inflow of the subway station are shown, and the total flow TF and net flow NF of the subway station in the time interval are shown as formulas (1) and (2).

[0083] The total flow and net flow respectively represent the riding vitality and supply-demand stability of shared bicycles. The larger the total flow, the more shared bicycles are used in the subway station connection area during the period. The larger the absolute value of the net flow, the more likely it is to cause imbalance between supply and demand of shared bicycles. Positive net flow indicates that more shared bicycles arrive, defined as a convergent state, and if it exceeds a threshold, shared bicycle accumulation can occur. Negative net flow indicates that more shared bicycles leave the area, defined as a divergent state, and if it exceeds a threshold, shared bicycle supply shortage can occur, as shown in Figure 6 .

[0084] The total flow and net flow are standardized by the z-score standardization method, and the calculation formula is shown as formula (3). Further, based on the two-dimensional vector composed of the standardized total flow and the standardized net flow, the shared bicycle flow characteristics in any subway station are defined as formula (4).

[0085] Step 5: Based on the silhouette coefficient and the sum of squared errors (SSE), determine the optimal number of clusters k of the K-medoids algorithm, and the k shared bicycle flow patterns. The calculation formula of the silhouette coefficient is shown as formula (5). The calculation formula of the sum of squared errors is shown as formula (6). The elbow rule is used to determine the best clustering number as 6, as shown in Figure 7 .

[0086] Step 6: Initialize 6 initial cluster centers of shared bicycle flow patterns, calculate the DTW distance between sample data and 6 initial cluster centers, divide all sample points to the nearest cluster center, and complete the first clustering after labeling the sample. The calculation method of DTW distance is as follows: for sample data a and cluster center b, corresponding to two time series S1 and S2 respectively, create a x b distance matrix D by calculating the square of the Euclidean distance between any two elements of S1 and S2. Any twisted path passing through the distance matrix D is a set of matrix elements mapping between S1 and S2, which satisfies three constraints: boundary condition: the first and last points of S1 and S2 are aligned, i.e. the start and end points of the twisted path are located in the diagonal cells; continuity: the twisted path aligns with adjacent points and cannot cross align; monotonicity: each point of S1 is matched with one or more points of S2, and vice versa. As shown in (7), the DTW distance is the shortest path among all twisted paths. Figure 8

[0087] Step 7: Calculate the sum of DTW distances from sample points in the cluster to other sample points in the cluster, select the point with the smallest distance sum as the new cluster center, and divide all sample points to the nearest new cluster center and label. The formula for calculating the new cluster center is shown in (8).

[0088] Step 8: If the iteration termination condition is met, complete the clustering of shared bicycle flow patterns and execute step 9; if not, execute step 7 until the iteration termination condition is met. The iteration termination condition is that the cluster center of the class cluster does not change, or no sample point is reassigned to a different class.

[0089] Step 9: Output the clustering results of shared bicycles and show their spatial distribution, and further analyze the corresponding shared bicycle flow patterns in the metro station connection area.

[0090] Use ArcGIS software to draw the spatial distribution of each metro station after clustering, and draw the spatial distribution of the clustering results and the time curve of total flow and net flow. The TF cluster center of each type of metro station is highlighted with a yellow line, and the NF cluster center is highlighted with a blue line, which facilitates further analysis of the use patterns of shared bicycles in the metro station connection area. According to the two characteristics of riding intensity and stability, the riding intensity is divided into high riding intensity (HF) and low riding intensity (LF), and the stability is divided into all-day stability (S), morning convergence and evening divergence (CD), morning divergence and evening convergence (DC), and all-day oscillation (O). As shown in (7) and (8). Figure 9 Figure 10 ​​As shown, three tidal flow patterns were found, including low-flow early-in late-out type (LF-CD), high-flow early-out late-in type (HF-DC), and high-flow early-in late-out type (HF-CD); three non-tidal flow patterns, including low-flow stable type (LF-S), high-flow stable type (HF-S), and high-flow oscillation type (HF-O). The characteristics of each type of shared bicycle flow pattern are as follows:

[0091] (1) Tidal flow pattern

[0092] The low-flow early-in late-out type is characterized by shared bicycles flowing into the subway station range during the morning peak period and flowing out of the subway station range during the evening peak period, with relatively low riding vitality. This type accounts for 22.79%, and the number is second only to LF-S type. LF-CD type is distributed in residential community-intensive areas near the second ring line. Although it has tidal flow characteristics, it is not obvious, and the total flow of shared bicycles is relatively small.

[0093] The high-flow early-out late-in type is characterized by shared bicycles flowing out of the subway station range during the morning peak period and flowing into the subway station range during the evening peak period, with relatively low riding vitality. This type accounts for 16.18%. HF-DC type is distributed in areas with high employment opportunities such as enterprises, companies, and office buildings near the second ring line. Its tidal characteristics are that the net outflow during the morning peak period is relatively large, but the net inflow during the evening peak period is small or tends to be stable, and the total flow of shared bicycles is slightly higher than that of LF-CD. This type of shared bicycle flow solves the last kilometer problem of commuting. A large number of residents use shared bicycles to travel from the subway station to the workplace during the morning peak period, resulting in a large number of shared bicycles flowing out of the subway station; during the evening peak period, commuting populations use shared bicycles to travel from the workplace to the subway station, but due to the inconsistency of the length of the evening peak, the inflow characteristics of shared bicycles are not obvious.

[0094] The high-flow early-in late-out type is characterized by shared bicycles flowing into the subway station range during the morning peak period and flowing out of the subway station range during the evening peak period. Compared with LF-CD type, the tidal flow characteristics of HF-CD type are more obvious, and the riding vitality is higher. This type accounts for 13.24%. Some areas are away from the city center, and residents have a strong willingness to use public transportation for commuting, and the residential density is high, with a large amount of riding.

[0095] (2) Non-tidal flow pattern

[0096] The low-flow stable type is characterized by low total flow and net flow of shared bicycles, and small fluctuations throughout the day, with strong randomness of shared bicycle use. This type accounts for 30.88%.

[0097] The high flow stability type is characterized by a large total flow of shared bicycles, but the net flow remains relatively stable throughout the day without obvious morning and evening tidal phenomena, and the proportion of this type is 9.56%. Most of the HF-S types are located in mixed functional areas, and the commercial, residential, medical, educational, and recreational land types have a high degree of integration, and the shared bicycle travel purposes are diversified, but the inflow and outflow of shared bicycles remain relatively balanced.

[0098] The high flow oscillation type is characterized by a large total flow of shared bicycles, and the net flow fluctuates significantly throughout the day, and the proportion of this type is 7.35%. The HF-O type subway stations are all located in the Jiangbei central area, which has the Jianghan Road commercial district and the Wuhan-Guangzhou commercial district large commercial agglomeration area. The total flow of shared bicycles is high throughout the day, and the intensive use characteristics of the morning and evening peaks are obvious, and the total flow peak is large. The areas with large net inflow values in the morning peak include Hanzhengjie subway station and Shuishuilou subway station, which exhibit the characteristics of the HF-CD type, and the fluctuations are more intense; the net flow curves of some areas are more like the HF-DC type; the net flow of the remaining subway stations presents disordered fluctuations throughout the day.

[0099] The implementation basis of each embodiment of the present application is realized by programmed processing of a device with processor function. Therefore, in engineering practice, the technical solutions and functions of each embodiment of the present application can be packaged into various modules. Based on this actual situation, on the basis of the above embodiments, an embodiment of the present application provides a shared bicycle flow mode recognition device in a subway station connection area, which is used to execute the shared bicycle flow mode recognition method in the above method embodiment. Referring to Figure 2The device comprises: a first main module for implementing steps 1: collecting and processing the research area data set, including the subway station entrance data and the shared bicycle order data; step 2: determining the connection area of the subway station and the shared bicycle; and step 3: screening the shared bicycle travel order space-time information in each subway station connection area based on the connection area generated in step 2; a second main module for implementing steps 4: constructing two time length sequence indexes of standardized total flow and standardized net flow as the measurement indexes of the shared bicycle flow mode in the subway station connection area; and step 5: determining the optimal clustering number k of the K-medoids algorithm based on the contour coefficient and the sum of squared errors (SSE) two standards, that is, k shared bicycle flow modes; a third main module for implementing steps 6: initializing the initial clustering center of the k shared bicycle flow modes, calculating the DTW distance between the sample data and the k initial clustering centers, dividing all sample points to the nearest clustering center, and marking the sample to complete the first clustering; and step 7: calculating the sum of the DTW distances of the sample points in the cluster to other sample points in the cluster, selecting the point with the minimum distance sum as the new clustering center, dividing all sample points to the nearest new clustering center and marking; a fourth main module for implementing steps 8: if the iteration termination condition is met, step 9 is executed; if not, step 7 is executed until the iteration termination condition is met; wherein the iteration termination condition is that the clustering center of the class cluster does not change, or no sample point is re-assigned to a different class; and step 9: outputting the shared bicycle clustering result and showing the spatial distribution; wherein the DTW distance is a dynamic dynamic warping distance.

[0100] The subway station connection area shared bicycle flow mode recognition device provided by the embodiment of the application adopts Figure 2 several modules in the method, fully uses the shared bicycle actual riding order data to generate the connection area, saves the complicated process of constructing the road network in the traditional method, improves the authenticity and effectiveness of the connection area, uses the DTW distance that can be adaptively twisted and matched on the time axis, reduces the influence of time lag on mode division, and improves the recognition efficiency of the shared bicycle flow mode.

[0101] It should be noted that the device in the device embodiment provided by the application can be used to implement the methods in the above method embodiments, and can also be used to implement the methods in other method embodiments provided by the application. The difference is only that the corresponding function modules are set, and the principle is basically the same as that of the above device embodiments provided by the application. As long as the person skilled in the art can obtain the corresponding technical means by combining technical features on the basis of the above device embodiments, and the technical solutions formed by these technical means, on the premise of ensuring the practicability of the technical solutions, the device in the above device embodiment can be improved to obtain the corresponding device class embodiment for implementing the methods in other method class embodiments. For example:

[0102] Based on the content of the above device embodiment, as an optional embodiment, the subway station connection area shared bicycle flow mode recognition device provided in the embodiment of the application further comprises: a first sub-module for realizing the processing of the research area data set in step 1, including: uniformly projecting and converting the spatial vector data to form a vector data set with consistent spatial coordinate system; cleaning the shared bicycle order data, considering the actual riding situation, the reasons for coordinate drift, not locking and abnormal GPS signal, deleting the orders with abnormal spatial position, deleting the orders with riding distance and time length not within the predetermined time length range, and deleting the orders with riding speed not within the predetermined speed range; the determination of the connection area of the subway station and the shared bicycle in step 2 includes: distributing all the riding orders starting or ending within 150m of the entrance of the subway station to the subway station, then regarding the orders within the first 85% of the riding distance in the subway station riding orders as potential connection orders for reservation, and deleting the orders with riding distance greater than the distance as leisure or commuting orders; using the DBSCAN algorithm to perform density clustering on the connection orders around the subway station, eliminating the noise values with low density to avoid abnormal shape of the connection area due to individual non-frequent orders; using the aggregation point tool in the ArcGIS software, setting the aggregation distance, creating the order points around the subway station into a polygon, and the connection area of the subway station.

[0103] Based on the content of the above device embodiment, as an optional embodiment, the subway station connection area shared bicycle flow mode recognition device provided in the embodiment of the application further comprises: a second sub-module for realizing the screening of the shared bicycle trip order space-time information in each subway station connection area in step 3, including: for the shared bicycle orders with starting point or ending point falling within the connection area, extracting the unlocking longitude and latitude coordinates, locking longitude and latitude coordinates, unlocking time length and locking time length.

[0104] Based on the content of the above device embodiment, as an optional embodiment, the subway station connection area shared bicycle flow mode recognition device provided in the embodiment of the application further comprises: a third sub-module for realizing the calculation method of the two time length sequence indexes of the standardized total flow and the standardized net flow in step 4, including: dividing each day into 24 intervals with 1 hour as a step, as the length of the two time length sequence indexes of the standardized total flow and the standardized net flow, defining the inside flow of each subway station as the shared bicycle flow within the subway station connection area, the outflow as the shared bicycle flow leaving the subway station connection area, and the inflow as the shared bicycle flow flowing into the subway station connection area, and calculating the total flow TF and the net flow NF of any subway station i in the time length interval t by the following formula:

[0105] TF(i, t) = inflow(i, t) + outflow(i, t) + insideflow(i, t)

[0106] NF(i, t) = inflow(i, t) - outflow(i, t)

[0107] The total flow and the net flow are standardized by using the z-score standardization method, and the calculation formula includes:

[0108]

[0109] wherein, x * is the standardized total flow or net flow, is the mean of the total flow or net flow in a day, and sigma is the standard deviation of the total flow or net flow in a day, and x is the total flow or net flow;

[0110] Based on the two-dimensional time sequence vector composed of the standardized total flow and the standardized net flow, the two-dimensional time sequence vector with a length of 24 composed of the standardized total flow tf i and the standardized net flow nf i in any subway station k i is defined as follows, to represent the shared bicycle flow characteristics:

[0111]

[0112] Based on the content of the above device embodiment, as an optional embodiment, the shared bicycle flow pattern recognition device in the subway station connection area provided in the embodiment of the application further includes: a fourth sub-module, configured to realize the calculation formula of the silhouette coefficient in step 5, which includes:

[0113]

[0114] For the silhouette coefficient s (i) of the vector i, a (i) is the cohesion, the average distance from point i to other points in the cluster; b (i) is the separation, the average distance from point i to all points outside the cluster, and s (i) is closer to 1, the better the clustering performance is, the global silhouette coefficient is obtained by averaging the silhouette coefficients of all vectors, and the calculation of the sum of squared errors includes:

[0115]

[0116] wherein p is the sample vector, m iThe cluster center is a cluster of the class cluster i, and k is the number of clusters; the elbow rule is used to determine the optimal number of clusters k, and the optimal number of clusters k will be the point at which the silhouette coefficient stops increasing rapidly and the sum of squared errors SSE stops decreasing rapidly, at which time the cluster result cannot be effectively improved by continuously increasing the number of clusters.

[0117] Based on the content of the above device embodiment, as an optional embodiment, the subway station connection area shared bicycle flow mode recognition device provided in the embodiment of the application further includes: a fifth sub-module for implementing the calculation of the DTW distance between the sample data and the k initial cluster centers in step 6, including: for sample data a and cluster center b, corresponding to two time length sequences S1 and S2 respectively, a distance matrix D of a x b is created by calculating the Euclidean distance square of any two elements between S1 and S2, then any twisted path passing through the distance matrix D is a matrix element set mapping between S1 and S2, satisfying three constraint conditions: boundary condition: the first point and the last point of S1 and S2 are aligned, that is, the starting point and the ending point of the twisted path are located in the diagonal cells; continuity: the twisted path is aligned with adjacent points and cannot be cross-aligned; monotonicity: each point of S1 is matched with one or more points of S2, and vice versa, and the DTW distance is the minimum length of one of all twisted paths:

[0118]

[0119] Wherein, P is the path set between S1 and S2, p k is one of the path set, D[i,j] is the Euclidean distance from the i th point S1 i of S1 to the j th point S2 j of S2. i j

[0120] Based on the content of the above device embodiment, as an optional embodiment, the subway station connection area shared bicycle flow mode recognition device provided in the embodiment of the application further includes: a sixth sub-module for implementing the new cluster center calculation formula in step 7, including:

[0121]

[0122] Wherein, x i is a sample i in the original cluster, x j is a sample j in the new cluster, and min is the minimum value.

[0123] The method of the embodiment of the application is realized by relying on an electronic device, so it is necessary to introduce the related electronic device. For this purpose, the embodiment of the application provides an electronic device, such as Figure 3 ​​As shown, the electronic device includes at least one processor, a communications interface, at least one memory, and a communications bus, wherein the at least one processor, the communications interface, and the at least one memory complete mutual communication through the communications bus. The at least one processor can invoke a logical instruction in the at least one memory to execute all or part of the steps of the method provided by the foregoing various method embodiments.

[0124] In addition, the logical instruction in the at least one memory described above can be implemented in the form of a software functional unit and sold or used as an independent product, and can be stored in a computer-readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes several instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the methods described in the various method embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0125] The device embodiments described above are only schematic, wherein the units shown as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, i.e., can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment. Those skilled in the art can understand and implement without creative labor.

[0126] From the above description of the embodiments, those skilled in the art can clearly understand that the embodiments can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware. Based on such understanding, the technical solutions described above essentially or the part that contributes to the prior art can be embodied in the form of a software product. The computer software product can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes several instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in the various embodiments or some parts of the embodiments.

[0127] The computer program product of the present application can be a computer program implemented on one or more various apparatuses, e.g., general purpose computers or dedicated computers. When implemented on a general purpose computer, the computer program can be stored in a tangible computer readable storage medium, e.g., a floppy disk, a hard disk, an optical storage, a memory, etc., and implemented as one or more computer programs running on the general purpose computer. When implemented on a dedicated computer, the computer program can be stored in a tangible computer readable storage medium, e.g., a memory, etc., and implemented as one or more computer programs running on the dedicated computer.

[0128] It should be noted that the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a" does not, without more constraints, exclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0129] Finally, it should be noted that the above-mentioned embodiments are merely used to illustrate the technical solutions of the present application, rather than limit the present application; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features thereof can be replaced equivalently; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for recognizing the flow pattern of shared bicycles in a subway station connection area, characterized in that, include: Step 1: Collect and process the dataset for the study area, including subway station entrance and exit data and shared bicycle order data; Step 2: Determine the connection area between the subway station and the shared bicycles; Step 3: Based on the connection areas generated in Step 2, filter the spatiotemporal information of shared bicycle travel orders within each subway station connection area; Step 4: Construct two time-series indicators, standardized total flow and standardized net flow, to serve as a measure of the flow pattern of shared bicycles within the subway station connection area; Step 5: Based on the two criteria of silhouette coefficient and squared error and SSE, determine the optimal number of clusters k for the K-medoids algorithm, i.e. k shared bicycle flow patterns; The formula for calculating the contour coefficient in step 5 include: For the silhouette coefficient of vector i , For cohesion, the average distance from point i to other points within the cluster; For separability, point The average distance to all points outside the cluster. Clusters closer to 1 perform better. The global silhouette coefficient is obtained by averaging the silhouette coefficients of all vectors. The calculation of the sum of squared errors includes: Where p is the sample vector Let k be the cluster center of cluster i, and k be the number of clusters. The elbow rule is used to determine the optimal number of clusters k. The optimal number of clusters k will be the point where the silhouette coefficient stops increasing rapidly and the squared error and SSE stop decreasing rapidly. At this point, continuing to increase the number of clusters will not effectively improve the clustering results. Step 6: Initialize k initial cluster centers for shared bicycle flow patterns, calculate the DTW distance between the sample data and the k initial cluster centers, assign all sample points to the nearest cluster center, and complete the first clustering after labeling the samples; Step 6, calculating the DTW distance between the sample data and the k initial cluster centers, includes: For sample data a and cluster center b, corresponding to two time sequences S1 and S2 respectively, a distance matrix D of a×b is created by calculating the squared Euclidean distance between any two elements of S1 and S2. Any twisted path through this distance matrix D is a set of matrix elements mapping between S1 and S2, satisfying three constraints: Boundary condition: The first and last points of S1 and S2 are aligned, i.e., the start and end points of the twisted path are located in diagonal cells; Continuity: The twisted path is aligned with adjacent points and cannot intersect; Monotonicity: Each point in S1 matches one or more points in S2, and vice versa. Therefore, the DTW distance is the shortest among all twisted paths. Where P is the path set between S1 and S2. Let D[i,j] be one of the paths in the path set, and let D[i,j] be the i-th point in S1. To the j-th point in S2 Euclidean distances include: ; Step 7: Calculate the sum of the DTW distances from a sample point in a cluster to other sample points in that cluster. Select the point with the smallest sum of distances as the new cluster center, and assign all sample points to the nearest new cluster center and label them. Step 8: If the iteration termination condition is met, proceed to Step 9; otherwise, proceed to Step 7 until the iteration termination condition is met. The iteration termination condition is that no cluster center changes or no sample points are reassigned to different categories. Step 9: Output the shared bicycle clustering results and display their spatial distribution. The DTW distance is the dynamic warp distance.

2. The method for recognizing the flow pattern of shared bicycles in the subway station connection area according to claim 1, characterized in that, Step 1, processing the study area dataset, includes: performing a unified projection transformation on the spatial vector data to form a vector dataset with a consistent spatial coordinate system; cleaning the shared bicycle order data, considering actual riding conditions, coordinate drift, unlocked bicycles, and abnormal GPS signals, and deleting orders with abnormal spatial locations, riding distances and durations outside the predetermined time range, and riding speeds outside the predetermined speed range. Step 2, determining the connection zone between the subway station and the shared bicycles, includes: assigning all riding orders starting or ending within 150m of the subway station entrance / exit to that subway station; then retaining the first 85% of the riding orders within that subway station as potential connection orders, while deleting riding orders beyond that distance as leisure or commuting orders; using the DBSCAN algorithm to perform density clustering on the connection orders around the subway station, removing low-density noise values ​​to avoid abnormal shapes in the connection zone due to individual infrequent orders; and using the aggregation point tool in ArcGIS software, setting the aggregation distance, and creating polygons from the order points around the subway station to define the subway station's connection zone.

3. The method for recognizing the flow pattern of shared bicycles in the subway station connection area according to claim 2, characterized in that, Step 3 involves filtering the spatiotemporal information of shared bicycle trip orders within each subway station transfer area, including: for shared bicycle orders whose origin or destination falls within the transfer area, extracting the unlock latitude and longitude coordinates, lock latitude and longitude coordinates, unlock duration, and lock duration.

4. The method for recognizing the flow pattern of shared bicycles in the subway station connection area according to claim 3, characterized in that, The calculation method for the two time-series indicators, standardized total flow and standardized net flow, in step 4 includes: dividing each day into 24 intervals with a step size of 1 hour, which serve as the lengths of the two time-series indicators, standardized total flow and standardized net flow; defining the insideflow of each subway station as the flow of shared bicycles within the subway station's connecting area, the outflow as the flow of shared bicycles leaving the subway station's connecting area, and the inflow as the flow of shared bicycles flowing into the subway station's connecting area; and for any subway station... The total flow TF and net flow NF over a time interval t are calculated using the following formulas: The total flow and net flow are standardized using the z-score standardization method. The calculation formula includes: Where x represents the total flow or net flow. The standardized total flow or net flow. This represents the average of total or net flow over a day. The standard deviation of total or net flow over one day; Based on a two-dimensional duration sequence vector composed of standardized total flow and standardized net flow, any subway station Internal Standardized Total Flow and standardized net flow The resulting two-dimensional time series vector of length 24 is defined as follows, used to characterize the flow characteristics of shared bicycles: .

5. The method for recognizing the flow pattern of shared bicycles in the subway station connection area according to claim 1, characterized in that, The new formula for calculating cluster centers in step 7 includes: in, For sample i in the original cluster, Let j be a sample in the new cluster, and min be the minimum value.

6. A device for recognizing the movement patterns of shared bicycles in a subway station connection area, characterized in that, The method for recognizing the flow pattern of shared bicycles in the subway station connection area as described in any one of claims 1-5 includes: a first main module for implementing step 1: collecting and processing the dataset of the study area, including subway station entrance and exit data and shared bicycle order data; step 2: determining the connection area between the subway station and the shared bicycles; step 3: based on the connection area generated in step 2, filtering the spatiotemporal information of shared bicycle travel orders in each subway station connection area; a second main module for implementing step 4: constructing two time-series indicators, standardized total flow and standardized net flow, as a measure of the flow pattern of shared bicycles in the subway station connection area; step 5: determining the optimal number of clusters k, i.e., k shared bicycle flow patterns, for the K-medoids algorithm based on two criteria: silhouette coefficient and squared error and SSE; a third main module for implementing step 6: initialization... Step 7: Calculate the sum of the DTW distances between the sample data and the k initial cluster centers for the shared bicycle flow patterns. Assign all sample points to the nearest cluster center and label the samples to complete the first clustering. Step 8: If the iteration termination condition is met, execute step 9; otherwise, execute step 7 until the iteration termination condition is met. The iteration termination condition is that no cluster center changes or no sample points are reassigned to different categories. Step 9: Output the shared bicycle clustering results and display their spatial distribution. The DTW distance is the dynamic warping distance.

7. An electronic device, characterized in that, include: At least one processor, at least one memory, and a communication interface; wherein, The processor, memory, and communication interface communicate with each other; The memory stores program instructions that can be executed by the processor, which invokes the program instructions to perform the method described in any one of claims 1 to 5.

8. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium stores computer instructions that cause the computer to perform the method described in any one of claims 1 to 5.

Images