Shared weight borrowing and returning method and system based on location-based service

Abstract

The invention discloses a shared weight borrowing and returning method and system based on position service, and the method comprises the steps: obtaining the position information of a leasing user and shared weight information, the position information of the leasing user comprises the position information of the leasing user and the position information of a returning user, and the shared weight information comprises the position information of a borrowing and returning point; according to the position information of the returning user, predicting to obtain a moving track of the returning user, and according to the moving track and the position information of the borrowing and returning point, obtaining recommended position information of the borrowing and returning point and predicting the number of shared weights of the borrowing and returning point information at different time nodes; and according to the number of shared weights of the borrowing and returning point information at different time nodes and the position information of the leasing user, obtaining recommended leasing point position information.

Description

Technical Field

[0001] The present invention belongs to the technical field of shared weight management, and particularly relates to a method and system for borrowing and returning shared weights based on location services. Background Art

[0002] There are various cheating scales on the market, and different types of scales have different types of cheating means; it is basically impossible to solve the problem of cheating scales by cracking the cheating means. Therefore, by setting shared metrological standard weights in the market, it is convenient for users to check whether various scales have cheating phenomena.

[0003] Regarding the borrowing and returning process of shared weights, when setting the borrowing and returning points of standard weights in the initial stage, in order to facilitate users to borrow and return weights, users can borrow and return standard weights at different borrowing and returning points. However, during the borrowing and returning process, users will judge the borrowing and returning points based on subjective experience. However, whether there are sufficient weights or insufficient remaining space at the borrowing and returning points will prevent effective borrowing and returning. At the same time, during the borrowing and returning process, the selection of the borrowing and returning points by users is generally judged by manual experience, and no relevant guiding content is given, resulting in a poor user experience. Summary of the Invention

[0004] To solve the above technical problems, the present invention proposes a method and system for borrowing and returning shared weights based on location services to solve the problems existing in the above prior art.

[0005] To achieve the above object, the present invention provides a method for borrowing and returning shared weights based on location services, including:

[0006] Obtaining the location information of the renting user and the shared weight information, where the location information of the renting user includes the location information of the renting user and the location information of the returning user, and the shared weight information includes the location information of the borrowing and returning points;

[0007] Predicting the moving trajectory of the returning user according to the location information of the returning user, and obtaining the recommended borrowing and returning point location information and the number of shared weights at different time nodes based on the moving trajectory and the location information of the borrowing and returning points;

[0008] Obtaining the recommended rental point location information based on the number of shared weights at different time nodes and the location information of the renting user.

[0009] Optionally, predicting the location information of the returning user through a prediction model to obtain the moving trajectory of the returning user, where the prediction model includes a cross-attention mechanism, an Add&Norm layer, a feed-forward neural network, an Add&Norm layer, a linear layer, and a softmax function layer connected in sequence, and an LSTM network and several subsequent fully connected layers are connected after the softmax function layer.

[0010] Optionally, the moving trajectory and the location information of the borrowing and returning points are processed by a deep learning model to obtain the recommended location information of the borrowing and returning points. The input of the deep learning model includes the location information at the current time node, the location information at the previous time node, and the location information at the next time node in the moving trajectory and the location information of the borrowing and returning points, where the location information of the borrowing and returning points is determined according to the number of weights in the borrowing and returning points.

[0011] Optionally, the process of obtaining the recommended rental point location information includes:

[0012] Based on the location information of the borrowing and returning points and the location information of the rental user, the moving distance and the moving time are obtained;

[0013] Based on the moving time, the shared weight number of the borrowing and returning point information at different time nodes is judged and calculated to obtain the waiting time. The moving distance, the moving time, and the waiting time are weighted and calculated to obtain a scoring result, and the scoring result is judged to obtain the recommended rental point location information.

[0014] Optionally, it further includes obtaining the recommended rental point location information corresponding to a number of rental users by an optimization algorithm based on the shared weight number of the borrowing and returning point information at different time nodes and the location information of a number of rental users.

[0015] On the other hand, the present invention also provides a shared weight borrowing and returning system based on location services, including

[0016] an acquisition module, a return module, and a rental module. The acquisition model is used to acquire the location information and the shared weight information of the rental user. The location information of the rental user includes the location information of the rental user and the location information of the return user, and the shared weight information includes the location information of the borrowing and returning points;

[0017] The return module is used to predict the moving trajectory of the return user based on the location information of the return user, and obtain the recommended location information of the borrowing and returning points and predict the shared weight number of the borrowing and returning point information at different time nodes according to the moving trajectory and the location information of the borrowing and returning points;

[0018] The rental module is used to obtain the recommended rental point location information based on the shared weight number of the borrowing and returning point information at different time nodes and the location information of the rental user.

[0019] Optionally, in the return module, the location information of the return user is predicted by a prediction model to obtain the moving trajectory of the return user. The prediction model includes a cross-attention mechanism, an Add&Norm layer, a feed-forward neural network, an Add&Norm layer, a linear layer, and a softmax function layer connected in sequence, and an LSTM network and a number of subsequent fully connected layers are connected after the softmax function layer.

[0020] Optionally, in the return module, the movement trajectory and the location information of the borrowing and returning points are processed by a deep learning model to obtain the recommended location information of the borrowing and returning points. The input of the deep learning model includes the location information at the current time node, the location information at the previous time node, and the location information at the next time node in the movement trajectory and the location information of the borrowing and returning points, where the location information of the borrowing and returning points is determined according to the number of weights in the borrowing and returning points.

[0021] Optionally, in the rental module, according to the location information of the borrowing and returning points and the location information of the rental user, the moving distance and the moving time are obtained;

[0022] According to the moving time, the shared weight number of the borrowing and returning point information at different time nodes is judged and calculated to obtain the waiting time. The moving distance, the moving time, and the waiting time are weighted and calculated to obtain a scoring result, and the scoring result is judged to obtain the recommended rental point location information.

[0023] Optionally, in the rental module, it further includes obtaining the recommended rental point location information corresponding to a number of rental users by an optimization algorithm according to the shared weight number of the borrowing and returning point information at different time nodes and the location information of a number of rental users.

[0024] Compared with the prior art, the present invention has the following advantages and technical effects:

[0025] Through the above technical solution, the present invention predicts the movement trajectory of the user to be returned, and predicts the time node of the return location based on the movement trajectory. Based on this time node and the location of the user in need of rental, the indicators existing between the user in need of rental and the weight sharing area are calculated, and based on the calculation result of the indicators, a better weight sharing area is recommended to the user in need of rental. In the above content, the present invention can effectively predict the behavior of the user to be returned, and at the same time recommend the return location when the user to be returned is on the way. At the same time, the weight sharing area is evaluated according to the user location, and the evaluation result is fed back to the user in need of rental to determine a better weight sharing area, so as to improve the rental user's experience by providing a shorter distance and moving time for the rental user while meeting the needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The drawings constituting a part of this application are used to provide a further understanding of this application. The schematic embodiments of this application and their descriptions are used to explain this application and do not constitute an improper limitation to this application. In the drawings:

[0027] Figure 1 It is a schematic flowchart of the method according to the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0028] It should be noted that, without conflict, the embodiments in the present application and the features in the embodiments may be combined with each other. The following will describe the present application in detail with reference to the accompanying drawings and in conjunction with the embodiments.

[0029] It should be noted that the steps shown in the flowchart of the accompanying drawings may be executed in a computer system such as a set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than here.

[0030] As Figure 1 shown, this embodiment provides a method for borrowing and returning shared weights based on location services, including:

[0031] Obtain the location information of the borrowing user and the shared weight information, where the shared weight information includes the location information of the shared weights and the real-time quantity information under different shared areas, and the location information of the borrowing user includes the location information of the user who needs to rent and the location information of the user who needs to return;

[0032] Obtain the historical borrowing user information, which includes the location information, movement trajectory, return time, and return location of the historical borrowing user;

[0033] Predict the return location of the user to be returned through a deep learning model, and count the number of weights in different shared areas. Select the rental location of the user who needs to rent according to the return location and the number of weights and the location information of the current borrowing user, inform the user according to the rental location, and explain when they can rent.

[0034] For the above technical solutions, the detailed technical content provided by the present invention is as follows:

[0035] First, obtain the location information of the borrowing user, where the location information of the borrowing user includes the location information of the user who needs to rent and the location information of the user who needs to return. The location information of the user who needs to rent is the location information that has been notified of the relevant content indicating that it needs to rent and is about to move towards the weight sharing area. The location information of the user who needs to return is the location information of the user who has rented and holds weights at the same time;

[0036] At the same time, obtain the location information and the number of weights in different shared areas, and the number of weights is recorded in real time.

[0037] Use a deep learning model to predict the possible movement trajectory, return location, and change information of the number of weights of the user to be returned;

[0038] Among them, the deep learning model uses the movement trajectory of the user to be returned in a time series, and on the basis of the movement trajectory, predicts the possible return location and the number of weights again.

[0039] The deep learning model includes two aspects, namely a prediction model for predicting the movement trajectory of the user to return the shared weights under the time series and a prediction model for predicting the possible return location and the number of weights.

[0040] The deep learning model learns the movement trajectory and return information of the user to return. After borrowing the shared weights and finishing using them, the user to return may not go directly to return. Most likely, during the movement process, after a certain distance from the shared area, the user will return on the way. Therefore, the present invention simulates the movement trajectory of the user by predicting the position of the user at different time nodes, so as to simulate the return process of the shared weights of the user to return. Based on the predicted return location, when the user to return is on the way, the present invention guides the user to return at a location as convenient as possible, improving the user experience.

[0041] First, the prediction model uses the position information at two adjacent time nodes for relevant fusion to predict the possible movement trajectory at the next time based on the positions at the two time nodes in the vector. At the same time, the prediction model is trained in real time with the historical movement trajectory of the user to return, and the parameters of the prediction model are adjusted in real time to ensure the prediction accuracy of the historical movement trajectory of the user to return.

[0042] The prediction model includes a cross-attention mechanism, an Add&Norm layer, a feed-forward neural network, an Add&Norm layer, a linear layer, and a softmax function layer connected in sequence. After the softmax function layer, there is an LSTM network and several subsequent fully connected layers. Among them,

[0043] The cross-attention is as follows:

[0044]

[0045] Among them, d represents the dimension of the Q, K, and V vectors. The dimensions of the Q, K, and V vectors are the same, all being d, which is used to scale the dot product result to prevent the softmax function from saturating due to excessive numerical values, thus affecting the accuracy of the attention distribution. The generation methods of the three matrices Q, K, and V are the same as those of the QKV matrix in the Transformer, and they are the key matrix, the query matrix, and the value matrix respectively. Among them, the subscript cur represents the current position information, and the subscript l represents the position information at the previous time node.

[0046] Under the above cross-attention mechanism, fusion and normalization are performed through the Add&Norm layer, and relevant feature extraction is carried out through the subsequent network structure. At the same time, LSTM is used to provide features under its time series, and the position information at the next time node is predicted through the fully connected layer. After the position information at the next time node is determined, the predicted next time node is used as the current time node, and the current time node is used as the previous time node to re-predict the next next time node. Repeat the above process until the maximum number of repetitions is reached or the position information at a certain time node coincides with the time node in the weight sharing area.

[0047] After predicting the position information of the next time node as above, based on the position information of the previous time, the current time node, and the next time node, the possible return positions are predicted through the neural network model:

[0048] The specific process in the neural network model is as follows:

[0049] The input data includes the position information of the previous time, the current time node, and the next time node, and the position information in the shared area. Calculate the output feature data of the first convolutional layer and output it to the i-th output feature data of the first convolutional layer

[0050]

[0051] where, W M1 is the convolutional filter matrix in the first convolutional layer, and the elements in this matrix are the weights in the weight set Θ. * represents the convolutional operation.

[0052] Take the output feature data result of the first convolutional layer as the i-th input feature map data of the first normalization layer in the model, and calculate the n-th output feature data of this layer according to the following formula

[0053]

[0054] In the formula, represents the output feature data of the first normalization layer of the k-th training sample feature, and f CBN (·) is the normalization operation.

[0055] Take the output feature data result of the first normalization layer as the i-th input feature map data of the first ReLU activation layer in the model, and calculate the i-th output feature data of this layer according to the following formula

[0056] where, f RELU (·) is the rectified linear unit (ReLU) activation function;

[0057] Use the first ReLU activation layer as the input feature data of the second convolutional layer, calculate the output feature data of the second convolutional layer, and output it to the i-th output feature data of the second convolutional layer

[0058]

[0059] where W M2 is the convolutional filter matrix in the second convolutional layer, and * represents the convolution operation.

[0060] Use the output feature data result of the second convolutional layer as the i-th input feature data of the second normalization layer in the model, and calculate the n-th output feature data of this layer according to the following formula

[0061]

[0062] In the formula, represents the output feature data of the second normalization layer of the k-th training sample feature, and f CBN (·) is the normalization operation;

[0063] Use the output feature data result of the second normalization layer as the i-th input feature map data of the first activation layer in the model, and calculate the i-th output feature data x of this layer according to the following formula i p2 .

[0064]

[0065] where f ReLU (·) is the rectified linear unit (ReLU) activation function;

[0066] Use the second ReLU activation layer as the input feature data of the second convolutional layer, calculate the output feature data of the third convolutional layer, and output it to the i-th output feature data of the third convolutional layer

[0067]

[0068] where W M3 is the convolutional filter matrix in the third convolutional layer, and * represents the convolution operation;

[0069] Use the output feature data result of the third convolutional layer as the i-th input feature data of the third normalization layer in the model, and calculate the i-th output feature data of this layer according to the following formula

[0070]

[0071] In the formula, f i b3 represents the output feature data of the third normalization layer of the feature of the i-th training sample, and f CBN (·) is a normalization operation;

[0072] Take the output feature data result of the third normalization layer as the i-th input feature data of the third activation layer in the AI model, and calculate the i-th output feature data f of this layer according to the following formula i p3 .

[0073] f i p3 = f ReLU (f i b2 )

[0074] Among them, f ReLU (·) is the rectified linear unit (ReLU) activation function;

[0075] Take the output feature data result of the third ReLU activation layer as the i-th input feature data of the global average pooling layer in the Softmax classifier, and calculate the i-th output feature data of this layer according to the following formula

[0076] Among them, f CGAP (·) is the global average pooling operation.

[0077] Take the output feature data result of the global average pooling layer as the i-th input feature data of the fully connected layer in the Softmax classifier, and calculate the i-th output feature map data of this layer according to the following formula

[0078] Among them, f CFC (·) is the fully connected layer operation.

[0079] Take the output feature data result of the fully connected layer as the i-th input feature data of the Softmax layer in the Softmax classifier, and calculate the i-th output feature data p of this layer according to the following formula i .

[0080]

[0081] Among them, p k represents the output prediction of the output layer of the Softmax classifier of the feature of the i-th training sample, that is, the possible return position, and f CSoftmax (·) is the Softmax function.

[0082] In the above content, the location information is displayed as an image of a satellite map or an image of a certain fixed - range area, where a certain pixel in the satellite - map image is marked as the relevant location information.

[0083] For the predicted trajectories of the above - mentioned different users to return (i.e., the location information at different time nodes), it is judged whether to return according to the location information, and the number of weights in different shared areas is predicted and statistically analyzed in real - time at different time nodes.

[0084] Under the guidance of the historical rental - user information, the above - mentioned deep - learning model is trained. Among them, the location information of different historical rental users at different time nodes is used to train the above - mentioned prediction model. At the same time, the neural - network model is trained according to the return location. The time nodes are intercepted at fixed time intervals, and the time - node information required for the training of historical rental - user information is divided according to the return time. It should be noted that when predicting the return of a certain user, the moving trajectory at the current time node is also used to train the prediction model to ensure that it conforms to the behavior characteristics of the current user.

[0085] After obtaining the location information of the user to return at different time nodes through the above - mentioned prediction model and neural - network model based on the location information at different time nodes, the moving trajectory is determined, and the moving trajectory is predicted in real - time. Based on the location information at different times, when the predicted location information coincides with the return - location information, the time node passed by it is calculated. Through the time node and the number, the possible return time is obtained;

[0086] At the same time, at the predicted return location, the corresponding number of weights at the corresponding time node in the corresponding shared area is determined, and it is judged whether the weights can be effectively returned. If effective return cannot be carried out, the corresponding weight - sharing area is removed, and the possible return location is predicted again through the neural - network model, and the possible return location is informed to the user to return, which is convenient for the user to return on the way and improves the user experience.

[0087] After predicting the location information and return time of the user to return at different future time nodes, at different time nodes, when the predicted location information of the user to return coincides with the shared - weight - area information, the predicted number of weights changes, and the number of weights that change and the location information of the rental user are used to judge the users who need to rent.

[0088] According to the location information of the users who need to rent, taking the moving distance, moving time, and waiting time as indicators, the shared area for renting is selected and provided to the users who need to rent. When there are many users who need to rent, the possible weight - sharing area locations and quantities at different times can be used to guide the decisions of different users.

[0089] After obtaining the number of weights in different shared areas at different time nodes, based on the location of the user in need of rental and obtaining the moving speed of the rental user, calculate the moving distance, moving time, whether waiting is required and the possible waiting time for the user in need of rental for different shared areas, and calculate the weighted sum of the above indicators as the final evaluation score. Among the evaluation scores, the moving distance and moving time are determined according to the location of the rental user and the shared area and the user's speed. After determination, count the number of weights in different shared areas at different time nodes, and judge whether there are relevant weights that can be lent in a certain shared area. If so, set the waiting time to 0. If not, judge when the weights can be provided. Based on the change in the number of weights in the shared area at future time nodes to meet the needs of the user in need of rental, if the moving time is less than the time for the change in the number of weights, then the waiting time is 0. Otherwise, calculate the difference between the time for the change in the number of weights and the time to move to the shared area, and use the difference as the waiting time. Based on the above different indicators, calculate the corresponding scoring scores for different shared areas, and use the shared area with the smallest scoring score as the recommended area.

[0090] In the above content, there may be a shared area with a short distance and a short moving time. However, if there are no weights in this shared area at present, then it is necessary to move to other shared areas, which may increase the distance and moving time. At this time, if you only need to wait for a short time, the user who has borrowed the weights will return the weights, and the rental user can rent the weights. For the rental user, it can improve the user experience to a certain extent. Based on the above content, provide relevant shared areas for the user in need of rental.

[0091] When multiple rental users need to rent weights in the same area, then optimize the above through relevant optimization algorithms to ensure the improvement of the comprehensive or overall user experience of multiple rental users.

[0092] Among them, the optimization algorithm, the content of the optimization algorithm includes:

[0093] Initialize the population. Among them, the population includes different individuals, and one individual includes the location of the shared area corresponding to the rental weights of different users in need of rental.

[0094] Set relevant objective functions, constraint conditions and corresponding fitness values. Among them, the objective function is:

[0095] f(x)=min∑B m A nm

[0096] Among them, A nm represents the value of the mth index of the nth user, and B represents the weight parameter corresponding to the index. The constraint conditions include the specific location information of the shared area location, and the fitness function is the reciprocal of the objective function.

[0097] Based on the above objective function, randomly assign the location of the shared area for each individual, and calculate the score by combining the location of the users in need of rental for each shared area location. Use the above score as the fitness, and for the fitness, guide the adjustment of the shared area location or the individual of each user in need of rental. By continuously and randomly adjusting the shared area location corresponding to the users in need of rental, after reaching the maximum number of iterations, output the final shared area location to obtain the final shared area location corresponding to each user in need of rental. And inform the users in need of rental of the shared area location to improve the user experience of the overall users in need of rental.

[0098] Through the above technical solution, the present invention predicts the movement trajectory of the user to be returned, and predicts the time node of the return location based on the movement trajectory. Based on this time node and the location of the users in need of rental, calculate the metrics existing between the users in need of rental and the weight sharing area, and based on the calculation results of the metrics, recommend a better weight sharing area to the users in need of rental. In the above content, the present invention can effectively predict the behavior of the user to be returned, and at the same time recommend the return location when the user to be returned is on the way. At the same time, evaluate the weight sharing area according to the user location, and feedback the evaluation result to the users in need of rental to determine a better weight sharing area, so as to improve the rental users' experience by providing a shorter distance and movement time for the rental users while meeting the needs.

[0099] On the other hand, the present invention also provides a weight sharing borrowing and returning system based on location services, including

[0100] an acquisition module, a return module and a rental module, wherein the acquisition model is used to acquire the location information of the renting users and the weight sharing information, wherein the location information of the renting users includes the location information of the rental users and the location information of the return users, and the weight sharing information includes the location information of the borrowing and returning points;

[0101] The return module is used to predict the movement trajectory of the return user according to the location information of the return user, and obtain the recommended borrowing and returning point location information and the number of weight sharing corresponding to the borrowing and returning points at different time nodes according to the movement trajectory and the location information of the borrowing and returning points;

[0102] The rental module is used to obtain the recommended rental point location information according to the number of weight sharing corresponding to the borrowing and returning points at different time nodes and the location information of the rental users.

[0103] The above system technical solution corresponds to the method technical solution and will not be elaborated here.

[0104] The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present application should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A shared weight borrowing and returning method based on location service, characterized in that: include: Obtaining the location information and shared weight information of the renting user, wherein the location information of the renting user includes the location information of the renting user and the location information of the returning user, and the shared weight information includes the location information of the borrowing and returning point; According to the location information of the returning user, the moving trajectory of the returning user is predicted, and according to the moving trajectory and the location information of the borrowing and returning point, the recommended location information of the borrowing and returning point and the number of shared weights of the predicted borrowing and returning point information at different time nodes are obtained; According to the number of shared weights of the borrowing and returning point information at different time nodes and the location information of the leasing user, the recommended leasing point location information is obtained.

2. The method according to claim 1, characterized in that The location information of the returning user is predicted through the prediction model to obtain the movement trajectory of the returning user. The prediction model includes a cross-attention mechanism, an Add&Norm layer, a feedforward neural network, an Add&Norm layer, a linear layer and a softmax function layer connected in sequence, and an LSTM network and several subsequent fully connected layers are connected after the softmax function layer.

3. The method according to claim 1, characterized in that The mobile trajectory and borrowing and returning point location information are processed by a deep learning model to obtain recommended borrowing and returning point location information, wherein the input of the deep learning model includes the location information at the current time node in the mobile trajectory, the location information at the previous time node, the location information at the next time node, and the borrowing and returning point location information, wherein the borrowing and returning point location information is determined according to the number of weights in the borrowing and returning point.

4. The method according to claim 1, characterized in that: The process of obtaining the recommended rental point location information includes: According to the location information of the borrowing and returning point and the location information of the renting user, the moving distance and moving time are obtained; According to the moving time, the number of shared weights of the borrowing and returning point information at different time nodes is judged and calculated to obtain the waiting time. The moving distance, moving time and waiting time are weighted and calculated to obtain the scoring result. The scoring result is judged to obtain the recommended rental point location information.

5. The method according to claim 1, characterized in that It also includes obtaining the recommended rental point location information corresponding to several rental users based on the number of shared weights of the borrowing and returning point information at different time nodes and the location information of several rental users through an optimization algorithm.

6. A shared weight borrowing and returning system based on location services, characterized in that: include An acquisition module, a return module and a rental module, wherein the acquisition module is used to acquire the location information of the renting user and the shared weight information, wherein the location information of the renting user includes the location information of the renting user and the location information of the returning user, and the shared weight information includes the location information of the borrowing and returning point; The return module is used to predict the movement trajectory of the returning user based on the location information of the returning user, and obtain the recommended borrowing and returning point location information and the number of shared weights of the borrowing and returning point information at different time nodes based on the movement trajectory and the borrowing and returning point location information; The leasing module is used to obtain the recommended leasing point location information based on the number of shared weights of the borrowing and returning point information at different time nodes and the location information of the leasing user.

7. The system according to claim 6, characterized in that In the return module, the location information of the returning user is predicted through the prediction model to obtain the movement trajectory of the returning user. The prediction model includes a cross-attention mechanism, an Add&Norm layer, a feedforward neural network, an Add&Norm layer, a linear layer and a softmax function layer connected in sequence, and an LSTM network and several subsequent fully connected layers are connected after the softmax function layer.

8. The system according to claim 6, characterized in that In the return module, the movement trajectory and the borrowing and returning point location information are processed by a deep learning model to obtain the recommended borrowing and returning point location information, wherein the input of the deep learning model includes the location information at the current time node in the movement trajectory, the location information at the previous time node, the location information at the next time node, and the borrowing and returning point location information, wherein the borrowing and returning point location information is determined according to the number of weights in the borrowing and returning point.

9. The system according to claim 6, characterized in that In the rental module, the moving distance and moving time are obtained according to the location information of the borrowing and returning point and the location information of the rental user; According to the moving time, the number of shared weights of the borrowing and returning point information at different time nodes is judged and calculated to obtain the waiting time. The moving distance, moving time and waiting time are weighted and calculated to obtain the scoring result. The scoring result is judged to obtain the recommended rental point location information.

10. The system according to claim 6, characterized in that In the leasing module, it also includes obtaining the recommended leasing point location information corresponding to several leasing users based on the number of shared weights of the borrowing and returning point information at different time nodes and the location information of several leasing users through an optimization algorithm.

Images