Training method, passenger flow prediction method, device, electronic device and storage medium

Abstract

The disclosed embodiments provide a training method, a passenger flow prediction method, a device, an electronic device, and a storage medium. The training method includes: according to an iterative incremental strategy for a basic training sample set, increasing the sample size in an iterative manner on the basis of the basic training sample set to obtain a target training sample set, wherein the basic training sample set and the target training sample set include feature data related to the interest surface, and the feature data is data related to the passenger flow used to predict the interest surface; and using the target training sample set to train a preset model corresponding to the target training sample set to obtain a passenger flow prediction model.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and more specifically, to a training method for a passenger flow prediction model, a passenger flow prediction method, a device, an electronic device, a computer-readable storage medium, and a computer program product. Background Art

[0002] At present, offline passenger flow scenarios pay more attention to the statistics and prediction of passenger flow. Offline passenger flow scenarios include office buildings, bus stations, subway stations, shopping malls or scenic spots. In related technologies, resources can be reasonably allocated according to passenger flow. For example, appropriate advertising points can be determined according to passenger flow; advertising delivery strategies for advertising points can be adjusted according to passenger flow, etc.

[0003] In the process of implementing the concept of the present disclosure, the inventors found that there are at least the following problems in the related art: the prediction accuracy of the passenger flow prediction model is not high. Summary of the invention

[0004] In view of this, the present disclosure provides a training method for a passenger flow prediction model, a passenger flow prediction method, an apparatus, an electronic device, a computer-readable storage medium, and a computer program product.

[0005] One aspect of an embodiment of the present disclosure provides a method for training a passenger flow prediction model, comprising: according to an iterative incremental strategy for a basic training sample set, increasing the sample size in an iterative manner on the basis of the basic training sample set to obtain a target training sample set, wherein the basic training sample set and the target training sample set include feature data related to an interest surface, and the feature data is data related to the passenger flow used to predict the interest surface; and, using the target training sample set to train a preset model corresponding to the target training sample set to obtain the passenger flow prediction model.

[0006] Another aspect of an embodiment of the present disclosure provides a passenger flow prediction method, comprising: obtaining feature data corresponding to each of a plurality of target interest surfaces; and inputting the feature data corresponding to each of the plurality of target interest surfaces into a passenger flow prediction model, and outputting a prediction result corresponding to each of the target interest surfaces, wherein the passenger flow prediction model is trained using the method described above.

[0007] Another aspect of the disclosed embodiments provides a training device for a passenger flow prediction model, comprising: an acquisition module, for iteratively increasing the sample size on the basis of a basic training sample set according to an iterative incremental strategy for the basic training sample set, to obtain a target training sample set, wherein the basic training sample set and the target training sample set include feature data related to an interest surface, and the feature data is data related to the passenger flow used to predict the interest surface; and a training module, for training a preset model corresponding to the target training sample set using the target training sample set, to obtain the passenger flow prediction model.

[0008] Another aspect of an embodiment of the present disclosure provides a passenger flow prediction device, comprising: an acquisition module for acquiring feature data corresponding to each of a plurality of target interest surfaces; and an output module for inputting the feature data corresponding to each of the plurality of target interest surfaces into a passenger flow prediction model, and outputting a prediction result corresponding to each of the target interest surfaces, wherein the passenger flow prediction model is trained using the above-mentioned device.

[0009] Another aspect of an embodiment of the present disclosure provides an electronic device, including: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method described above.

[0010] Another aspect of an embodiment of the present disclosure provides a computer-readable storage medium storing computer-executable instructions, which are used to implement the method described above when executed.

[0011] Another aspect of an embodiment of the present disclosure provides a computer program product, which includes computer executable instructions, and the instructions are used to implement the above method when executed.

[0012] According to an embodiment of the present disclosure, the target training sample set is obtained by iteratively increasing the sample size based on the basic training sample set according to an iterative incremental strategy for the basic training sample set, and the preset model corresponding to the target training sample set is trained using the target training sample set to obtain a passenger flow prediction model. Since the target training sample set is obtained by continuous iteration, the sample size of the target training sample set is large and the quality is high, thus improving the prediction accuracy of the passenger flow prediction model obtained based on this training, thereby at least partially overcoming the technical problem of low prediction accuracy of the passenger flow prediction model. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The above and other objects, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0014] Figure 1 The exemplary system architecture of the training method of the passenger flow prediction model, the passenger flow prediction method and the device according to the embodiment of the present disclosure is schematically shown;

[0015] Figure 2 A flowchart schematically shows a method for training a passenger flow prediction model according to an embodiment of the present disclosure;

[0016] Figure 3 A schematic diagram schematically shows a process of generating a target training sample set according to an embodiment of the present disclosure;

[0017] Figure 4 A flowchart of a passenger flow prediction method according to an embodiment of the present disclosure is schematically shown;

[0018] Figure 5 A block diagram of a training device for a passenger flow prediction model according to an embodiment of the present disclosure is schematically shown;

[0019] Figure 6 A block diagram schematically shows a training device for a passenger flow prediction model according to an embodiment of the present disclosure; and

[0020] Figure 7 A block diagram of an electronic device suitable for implementing the method described above according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0021] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0022] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0023] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0024] In the case of using expressions such as "at least one of A, B, and C, etc.", it should generally be interpreted in accordance with the meaning of the expression generally understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.). In the case of using expressions such as "at least one of A, B, or C, etc.", it should generally be interpreted in accordance with the meaning of the expression generally understood by those skilled in the art (for example, "a system having at least one of A, B, or C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0025] The problem of passenger flow prediction can be solved in the following three ways.

[0026] Method 1: Determine the passenger flow based on data collected using hardware equipment.

[0027] Method 2: Determine passenger flow based on positioning data. For example, social software or travel software can be used to obtain positioning data.

[0028] Method three: Use the passenger flow prediction model obtained through machine learning training to predict passenger flow.

[0029] In the process of implementing the concept of the present disclosure, it was found that for method 1, due to the low coverage of hardware equipment, the data collected by the hardware equipment is not accurate enough, that is, the data is relatively sparse, which makes the prediction accuracy of passenger flow prediction low. For method 2, since the positioning data needs to be purchased, and the purchase cost is high, the prediction cost is increased. For method 3, since the data of the training sample is relatively sparse, the prediction accuracy of the passenger flow prediction model is not high.

[0030] To this end, it is found that the prediction accuracy of the passenger flow prediction model can be improved by increasing the number and quality of training samples. Therefore, the embodiment of the present disclosure proposes a passenger flow prediction scheme based on an ensemble learning method, that is, a target training sample set with higher quality is screened out from the initial training sample set in a continuous iterative manner, and the preset model is trained using the target training sample set to obtain a passenger flow prediction model that can be used to predict passenger flow. Since the target training sample set is obtained by continuous iterative screening, the target training sample set has a larger sample size and higher quality, thereby improving the prediction accuracy of the passenger flow prediction model obtained based on this training.

[0031] Based on the above analysis, the embodiments of the present disclosure provide a training method for a passenger flow prediction model, a passenger flow prediction method, an apparatus, an electronic device, a computer-readable storage medium, and a computer program product. The training method for the passenger flow prediction model includes: according to an iterative incremental strategy for a basic training sample set, the sample size is iteratively increased on the basis of the basic training sample set to obtain a target training sample set, wherein the basic training sample set and the target training sample set include feature data related to the interest surface, the feature data is data related to the passenger flow used to predict the interest surface, and the preset model corresponding to the target training sample set is trained using the target training sample set to obtain a passenger flow prediction model.

[0032] Figure 1 The exemplary system architecture of the training method, passenger flow prediction method and device that can be applied to the passenger flow prediction model according to the embodiment of the present disclosure is schematically shown. It should be noted that: Figure 1 What is shown is merely an example of a system architecture to which the embodiments of the present disclosure can be applied, in order to help those skilled in the art understand the technical content of the present disclosure, but it does not mean that the embodiments of the present disclosure cannot be used in other devices, systems, environments or scenarios.

[0033] like Figure 1 As shown, the system architecture 100 according to this embodiment may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired and / or wireless communication links, etc.

[0034] The user can use the terminal devices 101, 102, 103 to interact with the server 105 through the network 104 to receive or send data, etc. Various passenger flow prediction applications and data transmission applications can be installed on the terminal devices 101, 102, 103.

[0035] The terminal devices 101 , 102 , and 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers.

[0036] The server 105 may be a server that provides various services, such as a background management server (only an example) that provides support for websites browsed by users using the terminal devices 101, 102, and 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0037] It should be noted that the training method of the passenger flow prediction model and the passenger flow prediction method provided in the embodiments of the present disclosure can generally be executed by the terminal device 101, 102, or 103, or can also be executed by other terminal devices different from the terminal devices 101, 102, or 103. Correspondingly, the training device of the passenger flow prediction model and the passenger flow prediction device provided in the embodiments of the present disclosure can also be set in the terminal device 101, 102, or 103, or in other terminal devices different from the terminal devices 101, 102, or 103.

[0038] Alternatively, the training method of the passenger flow prediction model and the passenger flow prediction method provided in the embodiment of the present disclosure may also be executed by the server 105. Accordingly, the training device of the passenger flow prediction model and the passenger flow prediction device provided in the embodiment of the present disclosure may generally be arranged in the server 105. The training method of the passenger flow prediction model and the passenger flow prediction method provided in the embodiment of the present disclosure may also be executed by a server or a server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105. Accordingly, the training device of the passenger flow prediction model and the passenger flow prediction device provided in the embodiment of the present disclosure may also be arranged in a server or a server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105.

[0039] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to the implementation requirements.

[0040] Figure 2 A flowchart of a method for training a passenger flow prediction model according to an embodiment of the present disclosure is schematically shown.

[0041] like Figure 2 As shown, the method may include operations S201 to S202.

[0042] In operation S201, according to the iterative incremental strategy for the basic training sample set, the sample size is iteratively increased on the basis of the basic training sample set to obtain a target training sample set, wherein the basic training sample set and the target training sample set include feature data related to the interest surface, and the feature data is data related to the passenger flow used to predict the interest surface.

[0043] In operation S202, a preset model corresponding to the target training sample set is trained using the target training sample set to obtain a passenger flow prediction model.

[0044] According to an embodiment of the present disclosure, the iterative increment strategy may include a strategy for how to gradually increase the sample size and an iterative termination strategy. The iterative termination strategy may be used to set the conditions that need to be met for the iterative termination. The iterative process may be implemented using a clustering algorithm or an active learning algorithm.

[0045] According to an embodiment of the present disclosure, an Area of ​​Interest (AOI) may refer to a target object. The target object may include, but is not limited to, an office building, an airport, a bus stop, a subway station, or a shopping mall. The number of interest surfaces may include one or more. Each interest surface has a basic training sample and a target training sample corresponding to the interest surface. The dimension of the feature data may include multiple, that is, the feature data may include multiple feature dimension data.

[0046] According to an embodiment of the present disclosure, a basic training sample set may refer to a training sample set with a higher sample quality, that is, the prediction accuracy of passenger flow prediction using the basic training sample set is higher. The basic training sample set may be a training sample set obtained by processing an initial training sample set. For example, the initial training sample set may be sorted according to a sorting rule to obtain a sorting result, and the basic training sample set may be determined according to the sorting result. The sorting rule may include a sample decile rule. Determining the basic training sample set according to the sorting result may include determining the initial training sample set corresponding to the highest decile as the basic training sample set.

[0047] According to an embodiment of the present disclosure, the preset model may include an extreme gradient boosting (eXtreme Gradient Boosting, XGBoost) model, a random forest (Random Forest, RF) model or a deep learning (Deep Learning) model. The preset model can be configured according to actual business needs and is not limited here.

[0048] According to an embodiment of the present disclosure, the sample size can be increased iteratively on the basis of the basic training sample set to obtain a target training sample set, that is, for the i-th iteration, the training samples to be added corresponding to the iteration are determined. The training samples to be added are added on the basis of the basic training sample set to obtain a new basic training sample set. When it is determined that the iteration termination condition is not met, the i+1-th iteration is performed. For the i+1-th iteration, the training samples to be added are added on the basis of the new basic training sample set obtained in the i-th iteration to obtain a new basic training sample set. Repeat the above operation until the iteration termination condition is met. The new basic training sample set finally obtained is determined as the target training sample set. i≥1. The training samples to be added may include one or more.

[0049] According to an embodiment of the present disclosure, in the process of updating the basic training sample set, the preset model is also updated accordingly. After obtaining the target training sample set, the preset model corresponding to the target training sample set is trained using the target training sample set to obtain a passenger flow prediction model that can predict the passenger flow of the interest surface.

[0050] According to an embodiment of the present disclosure, the target training sample set is obtained by iteratively increasing the sample size based on the basic training sample set according to an iterative incremental strategy for the basic training sample set, and the preset model corresponding to the target training sample set is trained using the target training sample set to obtain a passenger flow prediction model. Since the target training sample set is obtained by continuous iteration, the sample size of the target training sample set is large and the quality is high, thus improving the prediction accuracy of the passenger flow prediction model obtained based on this training, thereby at least partially overcoming the technical problem of low prediction accuracy of the passenger flow prediction model.

[0051] According to an embodiment of the present disclosure, according to an iterative incremental strategy for a basic training sample set, the sample size is iteratively increased one by one on the basis of the basic training sample set to obtain a target training sample set, which may include the following operations.

[0052] According to the iterative incremental strategy for the basic training sample set, the second training sample set is determined from the first training sample set. The second training sample set is added to the basic training sample set to obtain a new basic training sample set. According to the first training sample set and the second training sample set, a new first training sample set is obtained. When it is determined that the new basic training sample set meets the preset conditions, the operations of obtaining the new basic training sample set and determining whether the new basic training sample set meets the preset conditions are repeatedly performed until the new basic training sample set does not meet the preset conditions. The new basic training sample set is determined as the target training sample set.

[0053] According to an embodiment of the present disclosure, the first training sample set can be used to provide training samples added to the basic training sample set. The preset condition can be used as a condition that needs to be met to terminate the iteration. The preset condition may include a condition that needs to be met by an evaluation indicator corresponding to the new basic training sample set. The evaluation indicator may include at least one of a mean value, a standard deviation, an absolute percentage error, or other indicators.

[0054] According to an embodiment of the present disclosure, a second training sample set is added to the basic training sample set to obtain a new basic training sample set. Accordingly, the second training sample set is deleted from the first training sample set to obtain a new first training sample set, so that the second training sample added to the new basic training sample set is selected from the new first training sample set.

[0055] According to an embodiment of the present disclosure, the operations of obtaining a new basic training sample set and determining whether the new basic training sample set meets a preset condition are repeatedly performed, that is, according to an iterative incremental strategy for the basic training sample set, a new second training sample set is determined from a new first training sample set. A new second training sample set is added to the new basic training sample set to obtain a new basic training sample set. A new first training sample set is obtained based on the new first training sample set and the new second training sample set, and it is determined whether the new basic training sample set meets the preset condition.

[0056] According to the embodiments of the present disclosure, since the target training sample set is obtained by screening in a continuous iterative manner, the target training sample set has a large sample size and high quality, thereby improving the prediction accuracy of the passenger flow prediction model obtained based on this training.

[0057] Reference below Figure 3 , the generation process of the target training sample set according to the embodiment of the present disclosure is further explained in combination with specific embodiments.

[0058] Figure 3 The following is a schematic diagram schematically showing a process of generating a target training sample set according to an embodiment of the present disclosure.

[0059] like Figure 3 As shown, the initial training sample set can be divided into a basic training sample set and a first training sample set according to a preset sample set division rule. Based on an iterative incremental strategy for the basic training sample set, the first training sample to be added to the basic training sample set can be determined from the first training sample set in an iterative manner, thereby obtaining a target training sample set.

[0060] According to an embodiment of the present disclosure, the first training sample set includes a plurality of first training samples.

[0061] According to the iterative increment strategy for the basic training sample set, determining the second training sample set from the first training sample set may include the following operations.

[0062] The initial preset model is trained using the basic training sample set to obtain a preset model corresponding to the basic training sample set. Each first training sample in the first training sample set is predicted using the preset model corresponding to the basic training sample set to obtain a prediction result corresponding to each first training sample. The prediction result corresponding to each first training sample is evaluated to obtain an evaluation result corresponding to each first training sample. A second training sample set is determined from the first training sample set according to the evaluation result corresponding to each first training sample.

[0063] According to an embodiment of the present disclosure, the initial preset model may refer to an untrained model. The evaluation result may include at least one of the mean value, standard deviation, mean absolute percentage error and other indicators.

[0064] According to an embodiment of the present disclosure, the initial preset model can be trained using the basic training sample set to obtain a preset model corresponding to the basic training sample set. After obtaining the preset model corresponding to the basic training sample set, the preset model corresponding to the basic training sample set can be used to determine the second training sample set from the first training sample set.

[0065] According to the embodiment of the present disclosure, the second training sample set is determined by evaluating the prediction result obtained by predicting the first training sample in the first training sample set using the preset model corresponding to the basic training sample set, so as to obtain the second training sample set based on the iterative increment strategy. On this basis, the target training sample set obtained based on the second training sample set has a larger sample size and higher quality, thereby improving the prediction accuracy of the passenger flow prediction model obtained based on this training.

[0066] According to an embodiment of the present disclosure, evaluating the prediction result corresponding to each first training sample to obtain the evaluation result corresponding to each first training sample may include the following operations.

[0067] Determine the absolute value of the difference between the predicted result corresponding to each first training sample and the true result corresponding to each first training sample. Determine the ratio between the absolute value of the difference corresponding to each first training sample and the true result corresponding to each first training sample. Determine the ratio corresponding to each first training sample as the evaluation result corresponding to each first training sample.

[0068] Determining the second training sample set from the first training sample set according to the evaluation result corresponding to each first training sample may include the following operations.

[0069] The first training sample corresponding to the target ratio is determined as the second training sample, wherein the target ratio is a ratio whose ratio is less than or equal to the ratio threshold. A second training sample set is obtained according to the second training sample.

[0070] According to an embodiment of the present disclosure, a ratio threshold value can be used as one of the bases for determining whether a first training sample is a second training sample. The value of the ratio threshold value can be configured according to actual business needs and is not limited here. For example, the ratio threshold value can be 0.3. According to an embodiment of the present disclosure, the smaller the ratio value, the better the quality of the first training sample corresponding to the ratio value. On this basis, the second training sample set can be determined from the first training sample set according to the ratio value. The second training sample set may include one or more second training samples. The ratio corresponding to the second training sample is the target ratio value.

[0071] According to an embodiment of the present disclosure, by using the ratio between the absolute value of the difference corresponding to each first training sample and the true result corresponding to each first training sample as the evaluation result corresponding to each first training sample, and using the evaluation result to determine the second training sample set, it is possible to more accurately determine the second training sample set, thereby improving the sample size and quality of the target training sample set obtained based on the second training sample set.

[0072] According to an embodiment of the present disclosure, the training method of the above passenger flow prediction model may further include the following operations.

[0073] The preset model corresponding to the basic training sample set is trained using the new basic training sample set to obtain the preset model corresponding to the new basic training sample set. Each new first training sample in the new first training sample set is predicted using the preset model corresponding to the new basic training sample set to obtain a prediction result corresponding to each new first training sample. The prediction result corresponding to each new first training sample is evaluated to obtain an evaluation result corresponding to each new first training sample. Based on the evaluation result corresponding to each new first training sample, an evaluation result corresponding to the new first training sample set is determined. Based on the evaluation result corresponding to each first training sample, an evaluation result corresponding to the first training sample set is determined. When it is determined that the evaluation result corresponding to the new first training sample set is better than the evaluation result corresponding to the first training sample set, it is determined that the new basic training sample set meets the preset condition.

[0074] According to an embodiment of the present disclosure, in order to determine whether the new basic training sample set meets the preset conditions, the evaluation results corresponding to the new first training sample set can be compared with the evaluation results corresponding to the first training sample set, and the new first training sample set and the first training sample set are the first training sample sets in two adjacent iterative processes.

[0075] According to an embodiment of the present disclosure, if it is determined that the evaluation result corresponding to the new basic training sample set is better than the evaluation result corresponding to the basic training sample set, it can be said that the new basic training sample set meets the preset condition. If it is determined that the evaluation result corresponding to the basic training sample set is better than the evaluation result corresponding to the new basic training sample set, it can be said that the new basic training sample set does not meet the preset condition.

[0076] According to an embodiment of the present disclosure, by comparing the evaluation results corresponding to the new first training sample set with the evaluation results corresponding to the first training sample set, it is determined whether the new basic training sample set meets the preset conditions, thereby effectively ensuring the quality of the determined new basic training sample set, thereby improving the quality of the target training sample set obtained based on the new basic training sample set.

[0077] According to an embodiment of the present disclosure, evaluating the prediction result corresponding to each new first training sample to obtain the evaluation result corresponding to each new first training sample may include the following operations.

[0078] Determine the absolute value of the difference between the predicted result corresponding to each new first training sample and the true result corresponding to each new first training sample. Determine the ratio between the absolute value of the difference corresponding to each new first training sample and the true result corresponding to each new first training sample. Determine the ratio corresponding to each new first training sample as the evaluation result corresponding to each new first training sample.

[0079] According to an embodiment of the present disclosure, for each new first training sample, the absolute value of the difference between the predicted result and the true result corresponding to the new first training sample can be determined, and the ratio between the absolute value of the difference between the predicted result and the true result corresponding to the new first training sample and the true result corresponding to the new first training sample can be determined, and the ratio can be determined as the evaluation result corresponding to the new first training sample.

[0080] According to an embodiment of the present disclosure, by using the ratio between the absolute value of the difference corresponding to each new first training sample and the true result corresponding to each new first training sample as the evaluation result corresponding to each new first training sample, it is possible to more accurately determine the evaluation result, thereby improving the quality of the target training sample set.

[0081] According to an embodiment of the present disclosure, determining the evaluation result corresponding to the new first training sample set according to the evaluation result corresponding to each new first training sample may include the following operations.

[0082] Determine a first target number, wherein the first target number represents the number of new second training samples included in the new first training sample set, wherein the new second training samples represent new first training samples whose ratio is greater than or equal to the ratio threshold. Determine a first mean absolute percentage error, wherein the first mean absolute percentage error represents the mean absolute percentage error corresponding to the new first training sample set. Determine the first target number and / or the first mean absolute percentage error as an evaluation result corresponding to the new first training sample set.

[0083] According to an embodiment of the present disclosure, a new second training sample can be determined from a new first training sample based on a ratio threshold and a ratio, that is, if the ratio corresponding to the new first training sample is greater than or equal to the ratio threshold, the new first training sample can be determined as a new second training sample.

[0084] According to an embodiment of the present disclosure, the number of new second training samples can be determined, and the number of new second training samples can be determined as the first target number. The mean absolute percentage error can be determined according to the following formula (1).

[0085]

[0086] Where t represents the tth training sample, N represents the number of samples in the training sample set, t∈{1, 2, ..., N-1, N}, N≥1. t Represents the prediction result corresponding to the training sample t, Groundtruth t Represents the true result corresponding to the training sample t.

[0087] According to an embodiment of the present disclosure, the mean absolute percentage error corresponding to the new first training sample set may be determined according to formula (1), that is, the first mean absolute percentage error may be determined.

[0088] According to an embodiment of the present disclosure, the evaluation result corresponding to the new first training sample set may include the first target quantity, the first mean absolute percentage error, or the first target quantity and the first mean absolute percentage error.

[0089] Determining the evaluation result corresponding to the first training sample set according to the evaluation result corresponding to each first training sample may include the following operations.

[0090] A second target number is determined, wherein the second target number represents the number of second training samples included in the first training sample set. A second mean absolute percentage error is determined, wherein the second mean absolute percentage error represents the mean absolute percentage error corresponding to the first training sample set. The second target number and / or the second mean absolute percentage error are determined as an evaluation result corresponding to the first training sample set.

[0091] According to an embodiment of the present disclosure, the number of second training samples can be determined and the number of second training samples can be determined as the second target number. The mean absolute percentage error corresponding to the first training sample can be determined according to the above formula (1), that is, the second mean absolute percentage error can be determined.

[0092] According to an embodiment of the present disclosure, the evaluation result corresponding to the first training sample set may include the second target quantity, the second mean absolute percentage error, or the second target quantity and the second mean absolute percentage error.

[0093] According to the embodiment of the present disclosure, the evaluation results corresponding to the new first training sample set and the evaluation results corresponding to the first training sample set are determined more accurately, thereby improving the quality of the target training sample set. According to the embodiment of the present disclosure, the training method of the passenger flow prediction model can also include the following operations.

[0094] When it is determined that at least one of the following exists, it is determined that the evaluation result corresponding to the new first training sample set is better than the evaluation result corresponding to the first training sample set, wherein the following at least one of the following includes: the first target number is greater than the second target number, the first mean absolute percentage error is less than the second mean absolute percentage error, and the first target number is greater than a preset number threshold.

[0095] According to the embodiments of the present disclosure, the preset quantity threshold can be configured according to actual business needs and is not limited here. For example, the preset quantity threshold can be 0.

[0096] According to an embodiment of the present disclosure, by providing at least one method for determining whether an evaluation result corresponding to a new first training sample set is better than an evaluation result corresponding to the first training sample set, the diversity of methods is improved.

[0097] According to an embodiment of the present disclosure, the target training sample set may include multiple target training samples, and the preset model may include a classification model and a regression model. Using the target training sample set to train the preset model corresponding to the target training sample set to obtain a passenger flow prediction model may include the following operations.

[0098] Input each of the multiple target training samples into a classification model corresponding to the target training sample set, and output a classification result corresponding to each target training sample. Train the classification model according to the classification result corresponding to each target training sample. For each of the multiple target training samples, when it is determined that the classification result corresponding to the target training sample is a preset result, input the target training sample into a regression model corresponding to the target training sample set, and output a prediction result corresponding to the target training sample. Train the regression model according to the prediction result corresponding to the target training sample. Determine the trained classification model and regression model as a passenger flow prediction model.

[0099] According to an embodiment of the present disclosure, the classification result may include "can be predicted" and "cannot be predicted". The preset result may include "can be predicted". The preset model may include an XGboost model, the classification model may include an XGBClassifier model, and the regression model may include an XGBRegressor model.

[0100] According to an embodiment of the present disclosure, each target training sample among multiple target training samples can be input into a classification model corresponding to the target training sample set, and a classification result corresponding to each target training sample can be output. The classification model can be trained based on the classification result corresponding to each target training sample and the true classification result.

[0101] According to an embodiment of the present disclosure, if the classification result corresponding to the target training sample is a preset result, that is, the classification result corresponding to the target training sample is a "predictable" classification result, it can be said that the target training sample can be predicted. On this basis, the target training sample can be input into a regression model corresponding to the target training sample set, and the prediction result corresponding to the target training sample is output. The regression model is trained according to the prediction result corresponding to the target training sample and the true prediction result.

[0102] According to an embodiment of the present disclosure, the regression model and the classification model may be trained jointly or separately. The joint training may refer to adjusting the model parameters of the regression model and the classification model at the same time, and the separate training may refer to adjusting the model parameters of the regression model and the classification model separately.

[0103] According to the embodiments of the present disclosure, since the target training sample set is obtained by screening in a continuous iterative manner, the target training sample set has a large sample size and high quality, thereby improving the prediction accuracy of the passenger flow prediction model obtained based on this training.

[0104] According to an embodiment of the present disclosure, the feature data related to the interest surface may include at least one of the following: basic feature dimension data related to the interest surface, auxiliary feature dimension data related to the interest surface, and static feature dimension data related to the interest surface.

[0105] According to an embodiment of the present disclosure, the basic feature dimension data related to the interest surface may include but is not limited to the area of ​​the interest surface, gender information, age information, the number information and distance information of the interest points included in the interest surface. The interest points may include but are not limited to office buildings, airports, railway stations, bus stations, hospitals and schools.

[0106] According to an embodiment of the present disclosure, auxiliary feature dimension data related to the interest surface can be used to assist in predicting passenger flow. Static feature dimension data related to the interest surface can be feature data related to the area where the interest surface is located. This type of feature data has the following characteristics: static feature dimension data of different interest surface areas related to the same area are the same.

[0107] According to the embodiments of the present disclosure, the types of feature data related to the interest surface may include multiple types, which improves the comprehensiveness of the feature dimension data related to the interest surface, thereby improving the quality of the target training sample set.

[0108] According to an embodiment of the present disclosure, the auxiliary feature dimension data related to the interest surface may include at least one of the following: the maximum passenger flow, the minimum passenger flow, the average passenger flow and the median passenger flow of the relevant interest surface related to the interest surface.

[0109] Among them, the static feature dimension data related to the interest surface may include at least one of the following: the maximum passenger flow, the minimum passenger flow, the average passenger flow, the standard deviation of the passenger flow, and the variance of the passenger flow of the city corresponding to the interest surface.

[0110] According to an embodiment of the present disclosure, the auxiliary feature dimension data may include relevant passenger flow data of public places related to the interest surface. The relevant interest surface related to the interest surface may refer to the interest surface whose distance from the interest surface is less than or equal to the distance threshold, and the relevant passenger flow data may include at least one of the maximum passenger flow, minimum passenger flow, average passenger flow and median passenger flow of the relevant interest surface. For example, the interest surface may be an office building, and the relevant interest surfaces may include residential areas and supermarkets.

[0111] According to an embodiment of the present disclosure, the maximum passenger flow of a city corresponding to an interest facet may refer to the maximum passenger flow among the passenger flows of different interest faces included in the city. The minimum passenger flow may refer to the minimum passenger flow among the passenger flows of different interest faces included in the city. The average passenger flow may refer to the average of the passenger flows of various interest faces included in the city. The median passenger flow may refer to the median of the passenger flows of various interest faces included in the city. The standard deviation of the passenger flow and the variance of the passenger flow may be determined based on the average passenger flow and the passenger flow of each interest facet.

[0112] According to an embodiment of the present disclosure, the static feature dimension data related to the interest surface may also include a city code and / or a city name.

[0113] According to an embodiment of the present disclosure, by introducing static feature dimension data and auxiliary feature dimension data related to the interest surface into the feature data of the interest surface, the quality of the feature data is improved, thereby improving the prediction accuracy of the passenger flow prediction model.

[0114] According to an embodiment of the present disclosure, the training method of the above passenger flow prediction model may further include the following operations.

[0115] The basic training sample set and the first training sample set are obtained by data preprocessing, wherein the data preprocessing includes at least one of the following: data elimination, data missing value supplementation, data aggregation and data discretization.

[0116] According to the embodiments of the present disclosure, data preprocessing can enable the model to better learn data rules and improve the prediction accuracy of the model. The example of interest surface is office buildings.

[0117] According to an embodiment of the present disclosure, feature dimensions in which the number of null values ​​included in the original feature data is greater than or equal to the null value number threshold can be deleted. For example, the null value number threshold is set to 20%. Since the number of null values ​​of feature dimensions such as office building age, office building greening rate, and office building area is greater than the null value number threshold, the above feature dimensions can be deleted.

[0118] According to an embodiment of the present disclosure, non-numeric feature dimensions may also be deleted from the feature dimensions. For example, since feature dimensions such as building age, geographic fence, scene classification, and longitude and latitude are non-numeric feature dimensions, the above feature dimensions may be deleted.

[0119] According to the embodiments of the present disclosure, the standard deviation of each feature dimension can also be determined. If the standard deviation is 0, it can be explained that the feature dimension data of the feature dimension has no difference in each interest facet and is useless data for the model. Therefore, the feature dimension with a standard deviation of 0 is deleted.

[0120] According to the embodiments of the present disclosure, continuous feature dimension data may be discretized, and since the discretization process can increase the robustness of the feature data, the passenger flow prediction model is made more stable. The continuous feature dimension data may include feature dimension data that characterizes quantity.

[0121] According to an embodiment of the present disclosure, the feature dimension data representing time can be divided according to a preset division rule. The preset division rule may include a division rule according to the rules of full-week average, mid-week average, and weekend average.

[0122] According to an embodiment of the present disclosure, a binary process is performed on the city, that is, if the city corresponding to the interest face is the target city, the characteristic dimension data representing the city dimension can be set as the first identifier. If the city corresponding to the interest face is not the target city, the characteristic dimension data representing the city dimension can be set as the second identifier. The target city can refer to a city whose characteristic data is relatively accurate. For example, the target cities may include Beijing, Shanghai, Guangzhou, Shenzhen, etc.

[0123] Figure 4 The flowchart of the passenger flow prediction method according to the embodiment of the present disclosure is schematically shown.

[0124] like Figure 4 As shown, the method may include operations S401-S402.

[0125] In operation S401, feature data corresponding to each interest surface of a plurality of target interest surfaces is acquired.

[0126] In operation S402, feature data corresponding to each of the multiple target interest surfaces is input into a passenger flow prediction model, and a prediction result corresponding to each target interest surface is output, wherein the passenger flow prediction model is trained using the passenger flow prediction model training method described in the embodiment of the present disclosure.

[0127] According to an embodiment of the present disclosure, by inputting feature data corresponding to each of a plurality of target interest surfaces into a passenger flow prediction model, a prediction result corresponding to each target interest surface is output. The passenger flow prediction model is based on an iterative incremental strategy for a basic training sample set, and the sample size is iteratively increased on the basis of the basic training sample set to obtain a target training sample set, and the target training sample set is used to train a preset model corresponding to the target training sample set. Since the target training sample set is obtained by screening in a continuous iterative manner, the sample size of the target training sample set is large and the quality is high, thereby improving the prediction accuracy of the passenger flow prediction model obtained based on this training.

[0128] According to an embodiment of the present disclosure, the passenger flow prediction model may include a classification model and a regression model.

[0129] Inputting feature data corresponding to each target interest surface among multiple target interest surfaces into a passenger flow prediction model and outputting a prediction result corresponding to each target interest surface may include the following operations.

[0130] The feature data corresponding to each of the multiple target interest surfaces are respectively input into the classification model, and the classification result corresponding to each target interest surface is output. For each of the multiple target interest surfaces, when it is determined that the classification result corresponding to the target interest surface is a preset result, the feature data corresponding to the target interest surface is input into the regression model, and the prediction result corresponding to the target interest surface is output.

[0131] According to an embodiment of the present disclosure, the classification result may include “can be predicted” and “cannot be predicted.” The preset result may include “can be predicted.”

[0132] According to the embodiments of the present disclosure, the classification result is obtained by first processing the feature data using the classification model, and then processing the feature data using the regression model to obtain the prediction result when the classification result is determined to be the preset result, thereby reducing the amount of data processing and improving the prediction efficiency.

[0133] Figure 5 A block diagram of a training device for a passenger flow prediction model according to an embodiment of the present disclosure is schematically shown.

[0134] like Figure 5 As shown, the training device 500 for the passenger flow prediction model may include an acquisition module 501 and a first training module 502 .

[0135] The acquisition module 501 is used to iteratively increase the sample size on the basis of the basic training sample set according to the iterative incremental strategy for the basic training sample set, so as to obtain the target training sample set, wherein the basic training sample set and the target training sample set include feature data related to the interest surface, and the feature data is data related to the passenger flow used to predict the interest surface.

[0136] The first training module 502 is used to train a preset model corresponding to the target training sample set using the target training sample set to obtain a passenger flow prediction model.

[0137] According to an embodiment of the present disclosure, the target training sample set is obtained by iteratively increasing the sample size based on the basic training sample set according to an iterative incremental strategy for the basic training sample set, and the preset model corresponding to the target training sample set is trained using the target training sample set to obtain a passenger flow prediction model. Since the target training sample set is obtained by continuous iteration, the sample size of the target training sample set is large and the quality is high, thus improving the prediction accuracy of the passenger flow prediction model obtained based on this training, thereby at least partially overcoming the technical problem of low prediction accuracy of the passenger flow prediction model.

[0138] According to an embodiment of the present disclosure, the obtaining module 501 may include a first determining submodule, an adding submodule, an obtaining submodule, an executing submodule and a second determining submodule.

[0139] The first determination submodule is used to determine the second training sample set from the first training sample set according to an iterative increment strategy for the basic training sample set.

[0140] The adding submodule is used to add a second training sample set on the basis of the basic training sample set to obtain a new basic training sample set.

[0141] The obtaining submodule is used to obtain a new first training sample set according to the first training sample set and the second training sample set.

[0142] The execution submodule is used to repeatedly execute the operations of obtaining a new basic training sample set and determining whether the new basic training sample set meets the preset conditions when it is determined that the new basic training sample set meets the preset conditions, until the new basic training sample set does not meet the preset conditions.

[0143] The second determination submodule is used to determine the new basic training sample set as the target training sample set.

[0144] According to an embodiment of the present disclosure, the first training sample set includes a plurality of first training samples.

[0145] According to an embodiment of the present disclosure, the first determining submodule may include a training unit, a prediction unit, an evaluation unit, and a determination unit.

[0146] The training unit is used to train the initial preset model using the basic training sample set to obtain the preset model corresponding to the basic training sample set.

[0147] The prediction unit is used to predict each first training sample in the first training sample set by using a preset model corresponding to the basic training sample set to obtain a prediction result corresponding to each first training sample.

[0148] The evaluation unit is used to evaluate the prediction result corresponding to each first training sample to obtain the evaluation result corresponding to each first training sample.

[0149] The determining unit is used to determine the second training sample set from the first training sample set according to the evaluation result corresponding to each first training sample.

[0150] According to an embodiment of the present disclosure, the evaluation unit may include a first determining subunit, a second determining subunit, and a third determining subunit.

[0151] The first determining subunit is used to determine the absolute value of the difference between the predicted result corresponding to each first training sample and the true result corresponding to each first training sample.

[0152] The second determining subunit is used to determine the ratio between the absolute value of the difference corresponding to each first training sample and the true result corresponding to each first training sample.

[0153] The third determining subunit is used to determine the ratio corresponding to each first training sample as the evaluation result corresponding to each first training sample.

[0154] The determining unit may include a fourth determining subunit and an obtaining subunit.

[0155] The fourth determining subunit is configured to determine the first training sample corresponding to the target ratio as the second training sample, wherein the target ratio is a ratio greater than or equal to a ratio threshold.

[0156] The obtaining subunit is used to obtain a second training sample set according to the second training sample.

[0157] According to an embodiment of the present disclosure, the training device 500 of the above-mentioned passenger flow prediction model may further include a second training module, a prediction module, an evaluation module, a first determination module, a second determination module and a third determination module.

[0158] The second training module is used to train the preset model corresponding to the basic training sample set using the new basic training sample set to obtain the preset model corresponding to the new basic training sample set.

[0159] The prediction module is used to predict each new first training sample in the new first training sample set by using a preset model corresponding to the new basic training sample set to obtain a prediction result corresponding to each new first training sample.

[0160] The evaluation module is used to evaluate the prediction result corresponding to each new first training sample to obtain the evaluation result corresponding to each new first training sample.

[0161] The first determination module is used to determine the evaluation result corresponding to the new first training sample set according to the evaluation result corresponding to each new first training sample.

[0162] The second determination module is used to determine the evaluation result corresponding to the first training sample set according to the evaluation result corresponding to each first training sample.

[0163] The third determination module is used to determine that the new basic training sample set meets the preset condition when it is determined that the evaluation result corresponding to the new first training sample set is better than the evaluation result corresponding to the first training sample set.

[0164] According to an embodiment of the present disclosure, the evaluation module may include a third determination submodule, a fourth determination submodule, and a fifth determination submodule.

[0165] The third determination submodule is used to determine the absolute value of the difference between the predicted result corresponding to each new first training sample and the true result corresponding to each new first training sample.

[0166] The fourth determination submodule is used to determine the ratio between the absolute value of the difference corresponding to each new first training sample and the true result corresponding to each new first training sample.

[0167] The fifth determining submodule is used to determine the ratio corresponding to each new first training sample as the evaluation result corresponding to each new first training sample.

[0168] According to an embodiment of the present disclosure, the first determining module may include a sixth determining submodule, a seventh determining submodule and an eighth determining submodule.

[0169] The sixth determination submodule is used to determine a first target quantity, wherein the first target quantity represents the number of new second training samples included in the new first training sample set, wherein the new second training samples represent new first training samples whose ratios are greater than or equal to a ratio threshold.

[0170] The seventh determination submodule is configured to determine a first mean absolute percentage error, wherein the first mean absolute percentage error represents a mean absolute percentage error corresponding to the new first training sample set.

[0171] An eighth determination submodule is used to determine the first target quantity and / or the first mean absolute percentage error as an evaluation result corresponding to the new first training sample set.

[0172] According to an embodiment of the present disclosure, the second determining module may include a ninth determining submodule, a tenth determining submodule and an eleventh determining submodule.

[0173] The ninth determination submodule is configured to determine a second target quantity, wherein the second target quantity represents the quantity of second training samples included in the second training sample set.

[0174] The tenth determination submodule is configured to determine a second mean absolute percentage error, wherein the second mean absolute percentage error represents a mean absolute percentage error corresponding to the first training sample set.

[0175] An eleventh determination submodule is used to determine a second target quantity and / or a second mean absolute percentage error as an evaluation result corresponding to the first training sample set.

[0176] According to an embodiment of the present disclosure, the training device 500 of the passenger flow prediction model may further include a fourth determination module.

[0177] A fourth determination module is used to determine that an evaluation result corresponding to the new first training sample set is better than an evaluation result corresponding to the first training sample set when it is determined that at least one of the following exists, wherein the following at least one of the following includes: the first target number is greater than the second target number, the first mean absolute percentage error is less than the second mean absolute percentage error, and the first target number is greater than a preset number threshold.

[0178] According to an embodiment of the present disclosure, the target training sample set includes multiple target training samples, and the preset model includes a classification model and a regression model.

[0179] According to an embodiment of the present disclosure, the first training module 502 may include a first classification submodule, a first training submodule, a first prediction submodule, a second training submodule, and a twelfth determination submodule.

[0180] The first classification submodule is used to input each target training sample among the multiple target training samples into the classification model corresponding to the target training sample set, and output the classification result corresponding to each target training sample.

[0181] The first training submodule is used to train the classification model according to the classification result corresponding to each target training sample.

[0182] The first prediction submodule is used to input the target training sample into a preset model corresponding to the target training sample set for each target training sample in the multiple target training samples, and output the prediction result corresponding to the target training sample when it is determined that the classification result corresponding to the target training sample is a preset result.

[0183] The second training submodule is used to train the regression model according to the prediction results corresponding to the target training samples.

[0184] The twelfth determination submodule is used to determine the trained classification model and regression model as a passenger flow prediction model.

[0185] According to an embodiment of the present disclosure, the feature data related to the interest surface includes at least one of the following: basic feature dimension data related to the interest surface, auxiliary feature dimension data related to the interest surface, and static feature dimension data related to the interest surface.

[0186] According to an embodiment of the present disclosure, the auxiliary feature dimension data related to the interest surface includes at least one of the following: the maximum passenger flow, the minimum passenger flow, the average passenger flow and the median passenger flow of the relevant interest surface related to the interest surface.

[0187] Among them, the static feature dimension data related to the interest surface includes at least one of the following: the maximum passenger flow, the minimum passenger flow, the average passenger flow, the standard deviation of the passenger flow, and the variance of the passenger flow of the city corresponding to the interest surface.

[0188] Figure 6 A block diagram of a training device for a passenger flow prediction model according to an embodiment of the present disclosure is schematically shown.

[0189] like Figure 6 As shown, the passenger flow prediction device 600 may include an acquisition module 601 and an output module 602 .

[0190] The acquisition module 601 is used to acquire feature data corresponding to each interest surface of multiple target interest surfaces.

[0191] Output module 602 is used to input feature data corresponding to each of multiple target interest surfaces into a passenger flow prediction model, and output prediction results corresponding to each target interest surface, wherein the passenger flow prediction model is trained using a training device for a passenger flow prediction model according to an embodiment of the present disclosure.

[0192] According to an embodiment of the present disclosure, the passenger flow prediction model includes a classification model and a regression model.

[0193] According to an embodiment of the present disclosure, the output module 602 may include a second classification submodule and a second prediction submodule.

[0194] The second classification submodule is used to input feature data corresponding to each target interest surface among the multiple target interest surfaces into the classification model respectively, and output a classification result corresponding to each target interest surface.

[0195] The second prediction submodule is used to input the feature data corresponding to the target interest surface into the regression model for each target interest surface among the multiple target interest surfaces, and output the prediction result corresponding to the target interest surface when it is determined that the classification result corresponding to the target interest surface is a preset result.

[0196] According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits, or at least part of the functions of any one of them can be implemented in one module. According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits can be split into multiple modules for implementation. According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits can be at least partially implemented as hardware circuits, such as field programmable gate arrays (Field Programmable Gate Array, FPGA), programmable logic arrays (Programmable Logic Arrays, PLA), systems on chips, systems on substrates, systems on packages, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), or can be implemented by hardware or firmware of any other reasonable way of integrating or packaging the circuit, or by any one of the three implementation methods of software, hardware, and firmware or by a suitable combination of any of them. Alternatively, according to the embodiments of the present invention, one or more of the modules, submodules, units, and subunits can be at least partially implemented as computer program modules, and when the computer program modules are run, the corresponding functions can be performed.

[0197] For example, any of the acquisition module 501 and the first training module 502, or the acquisition module 601 and the output module 602 can be combined in one module / submodule / unit / subunit for implementation, or any of the modules / submodules / units / subunits can be split into multiple modules / units / subunits. Alternatively, at least part of the functions of one or more of these modules / submodules / units / subunits can be combined with at least part of the functions of other modules / submodules / units / subunits and implemented in one module / submodule / unit / subunit. According to an embodiment of the present disclosure, the acquisition module 501 and the first training module 502, or at least one of the acquisition module 601 and the output module 602 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or by any one of the three implementation methods of software, hardware and firmware or by a suitable combination of any of them. Alternatively, at least one of the acquisition module 501 and the first training module 502, or the acquisition module 601 and the output module 602 may be at least partially implemented as a computer program module, and when the computer program module is executed, the corresponding function may be executed.

[0198] It should be noted that the training device part of the passenger flow prediction model in the embodiment of the present disclosure corresponds to the training method part of the passenger flow prediction model in the embodiment of the present disclosure. The description of the training device part of the passenger flow prediction model specifically refers to the training method part of the passenger flow prediction model, which will not be repeated here. Similarly, the passenger flow prediction device part in the embodiment of the present disclosure corresponds to the passenger flow prediction method part in the embodiment of the present disclosure. The description of the passenger flow prediction device part specifically refers to the passenger flow prediction method part, which will not be repeated here.

[0199] Figure 7 A block diagram of an electronic device suitable for implementing the method described above according to an embodiment of the present disclosure is schematically shown. Figure 7 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0200] like Figure 7 As shown, the electronic device 700 according to the embodiment of the present disclosure includes a processor 701, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 702 or the program loaded from the storage part 708 to the random access memory (RAM) 703. The processor 701 may include, for example, a general-purpose microprocessor (such as a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (for example, an application-specific integrated circuit (ASIC)), etc. The processor 701 may also include an onboard memory for caching purposes. The processor 701 may include a single processing unit or multiple processing units for performing different actions of the method flow according to the embodiment of the present disclosure.

[0201] In RAM 703, various programs and data required for the operation of electronic device 700 are stored. Processor 701, ROM 702 and RAM 703 are connected to each other via bus 704. Processor 701 performs various operations of the method flow according to the embodiment of the present disclosure by executing the program in ROM 702 and / or RAM 703. It should be noted that the program can also be stored in one or more memories other than ROM 702 and RAM 703. Processor 701 can also perform various operations of the method flow according to the embodiment of the present disclosure by executing the program stored in the one or more memories.

[0202] According to an embodiment of the present disclosure, the electronic device 700 may further include an input / output (I / O) interface 705, which is also connected to the bus 704. The electronic device 700 may further include one or more of the following components connected to the I / O interface 705: an input portion 706 including a keyboard, a mouse, etc.; an output portion 707 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 708 including a hard disk, etc.; and a communication portion 709 including a network interface card such as a LAN card, a modem, etc. The communication portion 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. A removable medium 711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 710 as needed, so that a computer program read therefrom is installed into the storage portion 708 as needed.

[0203] According to an embodiment of the present disclosure, the method flow according to an embodiment of the present disclosure can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program contains a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 709, and / or installed from the removable medium 711. When the computer program is executed by the processor 701, the above-mentioned functions defined in the system of the embodiment of the present disclosure are executed. According to an embodiment of the present disclosure, the system, equipment, device, module, unit, etc. described above can be implemented by a computer program module.

[0204] The present disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present disclosure is implemented.

[0205] According to an embodiment of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium. For example, it may include, but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, an apparatus, or a device.

[0206] For example, according to an embodiment of the present disclosure, the computer-readable storage medium may include the ROM 702 and / or the RAM 703 described above and / or one or more memories other than the ROM 702 and the RAM 703 .

[0207] An embodiment of the present disclosure also includes a computer program product, which includes a computer program, which contains program code for executing the method provided by the embodiment of the present disclosure. When the computer program product is run on an electronic device, the program code is used to enable the electronic device to implement the training method of the passenger flow prediction model or the passenger flow prediction method provided by the embodiment of the present disclosure.

[0208] When the computer program is executed by the processor 701, the above functions defined in the system / device of the embodiment of the present disclosure are executed. According to the embodiment of the present disclosure, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0209] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 709, and / or installed from the removable medium 711. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0210] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiment of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, Python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any variety of networks, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (e.g., using an Internet service provider to connect through the Internet).

[0211] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions. It can be understood by those skilled in the art that the features recorded in the various embodiments and / or claims of the present disclosure can be combined and / or combined in a variety of ways, even if such a combination or combination is not explicitly recorded in the present disclosure. In particular, without departing from the spirit and teaching of the present disclosure, the features described in the various embodiments and / or claims of the present disclosure may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present disclosure.

[0212] The embodiments of the present disclosure are described above. However, these embodiments are only for illustrative purposes and are not intended to limit the scope of the present disclosure. Although the embodiments are described above separately, this does not mean that the measures in the various embodiments cannot be used in combination to advantage. The scope of the present disclosure is defined by the attached claims and their equivalents. Without departing from the scope of the present disclosure, those skilled in the art may make a variety of substitutions and modifications, which should all fall within the scope of the present disclosure.

Claims

1. A training method for a passenger flow prediction model, comprising: According to the iterative increment strategy for the basic training sample set, the sample size is iteratively increased on the basis of the basic training sample set to obtain the target training sample set, including: According to an iterative incremental strategy for the basic training sample set, a second training sample set is determined from the first training sample set, wherein the first training sample set includes a plurality of first training samples; an initial preset model is trained using the basic training sample set to obtain a preset model corresponding to the basic training sample set; each first training sample in the first training sample set is predicted using the preset model corresponding to the basic training sample set to obtain a prediction result corresponding to each first training sample; the prediction result corresponding to each first training sample is evaluated to obtain an evaluation result corresponding to each first training sample; and the second training sample set is determined from the first training sample set according to the evaluation result corresponding to each first training sample; Adding the second training sample set to the basic training sample set to obtain a new basic training sample set; Deleting the second training sample set from the first training sample set to obtain a new first training sample set; In the case where it is determined that the new basic training sample set meets the preset condition, repeatedly performing the operations of obtaining the new basic training sample set and determining whether the new basic training sample set meets the preset condition until the new basic training sample set does not meet the preset condition; and Determining the new basic training sample set as the target training sample set, wherein the basic training sample set and the target training sample set include feature data related to the interest surface, and the feature data is data related to the passenger flow used to predict the interest surface; and The target training sample set is used to train a preset model corresponding to the target training sample set to obtain the passenger flow prediction model.

2. The method according to claim 1, wherein: The step of evaluating the prediction result corresponding to each first training sample to obtain the evaluation result corresponding to each first training sample includes: Determine an absolute value of a difference between a predicted result corresponding to each first training sample and a true result corresponding to each first training sample; determining a ratio between an absolute value of a difference corresponding to each of the first training samples and a true result corresponding to each of the first training samples; and Determine the ratio corresponding to each of the first training samples as the evaluation result corresponding to each of the first training samples; The step of determining the second training sample set from the first training sample set according to the evaluation result corresponding to each first training sample comprises: Determining a first training sample corresponding to a target ratio as a second training sample, wherein the target ratio is a ratio where the ratio is less than or equal to a ratio threshold; and The second training sample set is obtained according to the second training sample.

3. The method according to claim 2, further comprising: Using the new basic training sample set to train a preset model corresponding to the basic training sample set, to obtain a preset model corresponding to the new basic training sample set; Using a preset model corresponding to the new basic training sample set, predicting each new first training sample in the new first training sample set, to obtain a prediction result corresponding to each new first training sample; Evaluate the prediction result corresponding to each new first training sample to obtain an evaluation result corresponding to each new first training sample; Determine, according to the evaluation result corresponding to each of the new first training samples, an evaluation result corresponding to the new first training sample set; Determine, according to the evaluation result corresponding to each of the first training samples, an evaluation result corresponding to the first training sample set; as well as In the case where it is determined that the evaluation result corresponding to the new first training sample set is better than the evaluation result corresponding to the first training sample set, it is determined that the new basic training sample set meets a preset condition.

4. The method according to claim 3, wherein: The step of evaluating the prediction result corresponding to each new first training sample to obtain the evaluation result corresponding to each new first training sample includes: Determine an absolute value of a difference between a predicted result corresponding to each of the new first training samples and a true result corresponding to each of the new first training samples; determining a ratio between an absolute value of a difference corresponding to each of the new first training samples and a true result corresponding to each of the new first training samples; and The ratio corresponding to each new first training sample is determined as an evaluation result corresponding to each new first training sample.

5. The method according to claim 4, wherein: The step of determining the evaluation result corresponding to the new first training sample set according to the evaluation result corresponding to each new first training sample comprises: Determine a first target number, wherein the first target number represents the number of new second training samples included in the new first training sample set, wherein the new second training samples represent new first training samples whose ratios are greater than or equal to the ratio threshold; Determining a first mean absolute percentage error, wherein the first mean absolute percentage error represents a mean absolute percentage error corresponding to the new first training sample set; and Determine the first target quantity and / or the first mean absolute percentage error as an evaluation result corresponding to the new first training sample set; The step of determining the evaluation result corresponding to the first training sample set according to the evaluation result corresponding to each first training sample includes: Determine a second target number, wherein the second target number represents the number of second training samples included in the first training sample set; determining a second mean absolute percentage error, wherein the second mean absolute percentage error represents a mean absolute percentage error corresponding to the first training sample set; and The second target quantity and / or the second mean absolute percentage error are determined as evaluation results corresponding to the first training sample set.

6. The method according to claim 5, further comprising: In the case where it is determined that at least one of the following exists, it is determined that the evaluation result corresponding to the new first training sample set is better than the evaluation result corresponding to the first training sample set, wherein the at least one of the following includes: the first target number is greater than the second target number, the first mean absolute percentage error is less than the second mean absolute percentage error, and the first target number is greater than a preset number threshold.

7. The method according to any one of claims 1 to 6, wherein The target training sample set includes multiple target training samples, and the preset model includes a classification model and a regression model; The step of training a preset model corresponding to the target training sample set by using the target training sample set to obtain the passenger flow prediction model includes: Input each target training sample of the multiple target training samples into a classification model corresponding to the target training sample set, and output a classification result corresponding to each target training sample; Training the classification model according to the classification results corresponding to each target training sample; For each target training sample among the multiple target training samples, when it is determined that the classification result corresponding to the target training sample is a preset result, input the target training sample into a regression model corresponding to the target training sample set, and output a prediction result corresponding to the target training sample; Training the regression model according to the prediction results corresponding to the target training samples; and The trained classification model and regression model are determined as the passenger flow prediction model.

8. The method according to any one of claims 1 to 6, wherein The feature data related to the interest surface includes at least one of the following: basic feature dimension data related to the interest surface, auxiliary feature dimension data related to the interest surface, and static feature dimension data related to the interest surface.

9. The method according to claim 8, wherein: The auxiliary feature dimension data related to the interest surface includes at least one of the following: the maximum passenger flow, the minimum passenger flow, the average passenger flow and the median passenger flow of the relevant interest surface related to the interest surface; Among them, the static feature dimension data related to the interest surface includes at least one of the following: the maximum passenger flow, the minimum passenger flow, the average passenger flow, the standard deviation of the passenger flow, and the variance of the passenger flow of the city corresponding to the interest surface.

10. A passenger flow prediction method, comprising: Acquire feature data corresponding to each interest surface of the plurality of target interest surfaces; as well as Input feature data corresponding to each of the multiple target interest surfaces into a passenger flow prediction model, and output a prediction result corresponding to each target interest surface, wherein the passenger flow prediction model is trained using the method according to any one of claims 1 to 9.

11. The method according to claim 10, wherein: The passenger flow prediction model includes a classification model and a regression model; The step of inputting the feature data corresponding to each target interest facet among the plurality of target interest faces into the passenger flow prediction model and outputting the prediction result corresponding to each target interest facet comprises: inputting feature data corresponding to each target interest facet of the plurality of target interest faces into the classification model respectively, and outputting a classification result corresponding to each target interest facet; and For each target interest surface among the multiple target interest surfaces, when it is determined that the classification result corresponding to the target interest surface is a preset result, the feature data corresponding to the target interest surface is input into the regression model, and the prediction result corresponding to the target interest surface is output.

12. A training device for a passenger flow prediction model, comprising: The obtaining module is used to increase the sample size in an iterative manner based on the basic training sample set according to the iterative increment strategy for the basic training sample set to obtain the target training sample set, including: A first determination submodule is used to determine a second training sample set from a first training sample set according to an iterative incremental strategy for the basic training sample set, wherein the first training sample set includes a plurality of first training samples; the first determination submodule includes: a training unit, used to train an initial preset model using the basic training sample set to obtain a preset model corresponding to the basic training sample set; a prediction unit, used to predict each first training sample in the first training sample set using the preset model corresponding to the basic training sample set to obtain a prediction result corresponding to each first training sample; an evaluation unit, used to evaluate the prediction result corresponding to each first training sample to obtain an evaluation result corresponding to each first training sample; and a determination unit, used to determine the second training sample set from the first training sample set according to the evaluation result corresponding to each first training sample; An adding submodule, used for adding the second training sample set to the basic training sample set to obtain a new basic training sample set; An obtaining submodule, used for deleting the second training sample set from the first training sample set to obtain a new first training sample set; an execution submodule, configured to, when it is determined that the new basic training sample set satisfies the preset condition, repeatedly execute the operations of obtaining the new basic training sample set and determining whether the new basic training sample set satisfies the preset condition, until the new basic training sample set does not satisfy the preset condition; and a second determination submodule, configured to determine the new basic training sample set as the target training sample set, wherein the basic training sample set and the target training sample set include feature data related to the interest surface, and the feature data is data related to the passenger flow used to predict the interest surface; and The training module is used to train a preset model corresponding to the target training sample set using the target training sample set to obtain the passenger flow prediction model.

13. A passenger flow prediction device, comprising: An acquisition module, used to acquire feature data corresponding to each interest surface of the plurality of target interest surfaces; as well as An output module is used to input feature data corresponding to each of the multiple target interest surfaces into a passenger flow prediction model, and output a prediction result corresponding to each target interest surface, wherein the passenger flow prediction model is trained using the device according to claim 12.

14. An electronic device comprising: one or more processors; a memory for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method of any one of claims 1 to 9 or the method of any one of claims 10 to 11.

15. A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, enables the processor to implement the method of any one of claims 1 to 9 or any one of claims 10 to 11.

16. A computer program product, comprising a computer program, wherein the computer program is used to implement the method of any one of claims 1 to 9 or any one of claims 10 to 11 when executed by a processor.

Images