Traffic prediction device, traffic prediction method, and program

Abstract

A traffic prediction device (1) according to the present invention comprises: a data acquisition unit (11) that acquires line requirements and traffic data on the line; a prediction data generation unit (12) that generates a first feature amount corresponding to a traffic fluctuation due to changes in the line requirements, on the basis of the acquired requirements, and generates a second feature amount based on a traffic statistic representing a feature of traffic fluctuation over time, on the basis of the acquired traffic data; and a prediction function unit (13) that predicts future traffic flow from the first feature amount and the second feature amount.

Description

Technical Field

[0001] The present disclosure relates to a traffic prediction device, a traffic prediction method, and a program in a network in which a plurality of communication lines are accommodated in a link between communication devices.

Background Art

[0002] Conventionally, in traffic prediction, prediction by an RNN (Recurrent Neural Network) or prediction from the required conditions of a line by a machine learning model has been performed. The required conditions of a line refer to the bandwidth upper limit value (hereinafter also referred to as the contracted bandwidth), etc.

[0003] FIG. 10 is a diagram showing a configuration example of a network system including a conventional traffic prediction device. A plurality of user communication lines (hereinafter referred to as lines) are accommodated in a link arranged between two communication devices. Each line has its own required conditions, and a contracted bandwidth is set according to the required conditions. The physical interfaces provided in each communication device measure traffic data such as traffic flow rate. The measured traffic data is stored in a traffic database. The traffic prediction device includes a bandwidth prediction function unit, and the bandwidth prediction function unit predicts the future traffic flow rate flowing through each link based on the information stored in the traffic database and the line database.

[0004] FIG. 11 is a diagram showing traffic prediction by a conventional RNN. The bandwidth prediction function unit predicts the future traffic flow rate from past traffic data using a time series prediction model (such as LSTM (Long Short Term Memory)) by an RNN.

[0005] FIG. 12 is a diagram showing traffic prediction from required conditions by a conventional machine learning model. The bandwidth prediction function unit predicts future traffic flow rates from future line data (data such as contract bandwidth based on the required conditions of the line) using a deep learning model (such as SVAE (Supervised Variational AutoEncoder)).

[0006] Non-Patent Document 1 describes an evaluation of the effectiveness of various RNN architectures for traffic prediction. Non-Patent Document 2 describes a bandwidth design method for calculating the required bandwidth based on traffic prediction by machine learning. Also, Non-Patent Document 3 describes a method for predicting the statistical upper limit of traffic.

Prior Art Documents

Non-Patent Documents

[0007]

Non-Patent Document 1

Non-Patent Document 2

Non-Patent Document 3

Summary of the Invention

Problems to be Solved by the Invention

[0008] However, the number of lines to be accommodated or the traffic volume may change over time due to new addition, deletion of lines, change of required conditions, etc. In such a case, in the prediction by RNN, there is a problem that the predicted value cannot follow the traffic fluctuation accompanying the change of the required conditions of the line and the error is large. Also, in the prediction from the required conditions by the machine learning model, since the traffic volume is predicted for each required condition of the line, there is a problem that it cannot follow the traffic fluctuation over time, or there is a problem that the prediction accuracy becomes low because the correlation between the required conditions and the traffic volume is low.

[0009] [[ID=1,4]]In view of such circumstances, an object of the present invention is to realize a more accurate prediction of the traffic volume based on the required conditions of the line and the traffic statistics.

Means for Solving the Problems

[0010] [[ID=,20]] In order to solve the above problems, a traffic prediction device according to the present disclosure is a traffic prediction device that predicts future traffic flow rates in a link accommodating a plurality of lines, and includes a data acquisition unit that acquires the required conditions of the lines and traffic data in the lines, a prediction data generation unit that generates a first feature amount corresponding to traffic fluctuations due to changes in the required conditions of the lines based on the acquired required conditions, and generates a second feature amount based on traffic statistics representing the characteristics of traffic fluctuations for each time based on the acquired traffic data, and a prediction function unit that predicts future traffic flow rates from the first feature amount and the second feature amount.

[0011] In order to solve the above problems, a traffic prediction method according to the present disclosure is a traffic prediction method that predicts future traffic flow rates in a link accommodating a plurality of lines, and includes steps of acquiring, by a traffic prediction device, the required conditions of the lines and traffic data in the lines, generating a first feature amount corresponding to traffic fluctuations due to changes in the required conditions of the lines based on the acquired required conditions, generating a second feature amount based on traffic statistics representing the characteristics of traffic fluctuations for each time based on the acquired traffic data, and predicting future traffic flow rates from the first feature amount and the second feature amount.

[0012] In order to solve the above problems, a program according to the present disclosure causes a computer to function as the above traffic prediction device.

Advantages of the Invention

[0013] According to the present disclosure, it becomes possible to realize a more accurate prediction of traffic flow rates based on required conditions and traffic statistics.

Brief Description of the Drawings

[0014]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Embodiments for Carrying Out the Invention

[0015] Hereinafter, embodiments according to the present disclosure will be described in detail with reference to the drawings.

[0016] (First Embodiment) FIG. 1 is a block diagram showing a configuration example of a traffic prediction device 1 according to a first embodiment. The traffic prediction device 1 according to the first embodiment will be described below. As shown in FIG. 1, the traffic prediction device 1 includes a data acquisition unit 11, a prediction data generation unit 12, and a prediction function unit 13. The traffic prediction device 1 predicts a future traffic flow in a link that accommodates a plurality of lines.

[0017] The data acquisition unit 11, the prediction data generation unit 12, and the prediction function unit 13 constitute a control unit 10 (controller 10). The control unit 10 (controller 10) may be configured by dedicated hardware such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field-Programmable Gate Array), or may be configured by a processor, or may be configured to include both.

[0018] The data acquisition unit 11 acquires a line request condition 21 that changes every arbitrary period and traffic data 22 in the line. The data acquisition unit 11 transmits the acquired request condition 21 and traffic data 22 to the prediction data generation unit 12.

[0019] The prediction data generation unit 12 generates a feature amount 23a (hereinafter referred to as a first feature amount 23a) corresponding to traffic fluctuations due to changes in the line request condition based on the line request condition acquired by the data acquisition unit 11. Further, the prediction data generation unit 12 generates a feature amount 23b (hereinafter referred to as a second feature amount 23b) based on a traffic statistic representing the feature of traffic fluctuations over time based on the traffic data 22 acquired by the data acquisition unit 11. The prediction data generation unit 12 transmits prediction data 23 composed of the first feature amount 23a and the second feature amount 23b to the prediction function unit 13.

[0020] FIG. 2 is a diagram showing an example of the structure of prediction data 23 generated by the prediction data generation unit 12. As shown in FIG. 2, the prediction data generation unit 12 generates prediction data 23 including a first feature quantity 23a and a second feature quantity 23b. The first feature quantity 23a is a feature quantity based on the required conditions of a line such as the total contract bandwidth and the average contract bandwidth. The first feature quantity 23a changes every arbitrary period (for example, one day), and the traffic flow fluctuates greatly along with the change. The second feature quantity 23b is a feature quantity based on the traffic statistics for each time. The second feature quantity 23b is, for example, the traffic flow several steps before (1, 2 , ··· Y steps before), the traffic moving average during the past Z steps, etc. The traffic one step before, ···, the traffic Y steps before, and the traffic moving average during the past Z steps are an example of traffic statistics representing the characteristics of traffic fluctuations for each time. Here, a step indicates the time interval of measurement. For example, assuming measurement is performed at 5-minute intervals, 1 step represents 00:05 (HH:MM), 2 steps represent 00:10, and 3 steps represent 00:15. The traffic flow several steps before (1, 2, ···, Y steps before) represents the traffic flow at each measurement time. For example, assuming the current time is t steps, it indicates the traffic flow at step t - 1, the traffic flow at step t - 2, ···, the traffic flow at step t - Y. The traffic moving average during the past Z steps represents the average value of the traffic flow during the past measurement period. For example, assuming the current time is t steps, it is (the traffic flow at step t - 1 + the traffic flow at step t - 2 + ··· + the traffic flow at step t - Z) / Z.

[0021] For Y and Z shown in FIG. 2, any numerical values are applied. The method for determining the numerical values of Y and Z includes a method of experimentally exploring and determining using prediction accuracy as an evaluation index, a method of determining according to the second embodiment described below, and the like. In addition to the traffic flow rates several steps before (1, 2, ···, Y steps before) and the traffic moving average over the past Z steps, various statistics such as the change amount between the current traffic flow rate and the traffic flow rates at each past time point and the weighted moving average can be adopted for the traffic statistic quantity.

[0022] Referring to FIG. 1 again, the prediction function unit 13 predicts the future traffic flow rate 24 (hereinafter, also referred to as the prediction result 24) from the first feature quantity 23a and the second feature quantity 23b. The prediction function unit 13 receives the prediction data 23 composed of the first feature quantity 23a and the second feature quantity 23b transmitted from the prediction data generation unit 12. The received prediction data 23 is input to the prediction model 13a that predicts the future traffic flow rate based on the first feature quantity 23a and the second feature quantity 23b, and the future traffic flow rate is predicted. As the prediction model 13a, it is possible to adopt the above-mentioned deep learning model SVAE, machine learning models LightGBM, TabNet, etc.

[0023] FIG. 3 is a diagram showing an example of learning data for the prediction function unit 13 to learn a model used for predicting traffic flow. The prediction model 13a can be created by learning learning data including a first feature quantity 23a, a second feature quantity 23b, and traffic at predetermined time intervals as shown in FIG. 3. As shown in FIG. 3, the learning data includes data indicating the total contract bandwidth, the average contract bandwidth, the traffic one step before, ···, the traffic Y steps before, the moving average of the traffic between the past Z steps, and the traffic at predetermined time intervals (in FIG. 3, 5-minute intervals). The total contract bandwidth and the average contract bandwidth are an example of the requirement condition 21. Assuming that the total contract bandwidth and the average contract bandwidth change daily, the rows on the same day have the same value. The traffic statistics shown in FIG. 3 include traffic data 22 (traffic flow rate) measured at each time and their moving average. The traffic statistics for each time are generated from the traffic data in the rightmost column. The requirement condition 21 changes for any period, but within the period of the requirement condition 21, a traffic flow rate equal to or less than the contract bandwidth is generated. For example, when the requirement condition 21 changes daily, the traffic flow rate may vary greatly with the change in the contract bandwidth of the requirement condition 21.

[0024] The prediction function unit 13 predicts the traffic flow rate n (n is an arbitrary natural number) steps ahead from time t. However, since the accuracy decreases as n increases, n is preferably within 10.

[0025] FIG. 4 is a flowchart showing an example of a traffic prediction method executed by the traffic prediction device 1 according to the first embodiment.

[0026] In step S101, the data acquisition unit 11 acquires the requirement condition 21 of the line that changes for any period.

[0027] In step S102, the data acquisition unit 11 acquires traffic data 22 for generating a traffic statistic representing the characteristics of traffic fluctuations over time. The processes of step S101 and step S102 may be performed in parallel as shown in FIG. 4, or either one of the processes may be performed first and the other process may be performed later.

[0028] In step S103, the prediction data generation unit 12 generates a first feature quantity 23a and a second feature quantity 23b.

[0029] In step S104, the prediction function unit 13 predicts the future traffic flow rate 24 from the first feature quantity 23a and the second feature quantity 23b.

[0030] In the traffic prediction using SVAE, which is a conventional technique, since the input data is a feature quantity based on the required conditions, the prediction content is the traffic average value for each required condition. Therefore, in the case of data with a low correlation between the required conditions and the traffic flow rate, there is a problem that the error between the prediction result and the actual traffic flow rate is large. Also, according to SVAE, prediction for each time interval cannot be performed. Further, in the traffic prediction using LSTM, which is a conventional technique, for example, in order to predict the future (e.g., 10 steps ahead) traffic flow rate from past traffic data, there is a problem that it cannot cope with large traffic fluctuations due to changes in required conditions (such as line addition, change of bandwidth upper limit value, etc.).

[0031] However, according to the traffic prediction device 1 according to the present embodiment, since the traffic flow rate is predicted based on both the required conditions and the traffic statistic, the problems of the conventional techniques described above can be solved as follows, and a highly accurate traffic flow rate prediction can be realized. The highly accurate traffic flow rate prediction is a highly accurate traffic flow rate prediction corresponding to traffic fluctuations due to changes in required conditions and traffic fluctuations over time due to traffic statistics.

[0032] First, according to the traffic prediction device 1 according to the present embodiment, since the prediction is performed based on the required conditions, even when the traffic fluctuation is large due to the change of the contract bandwidth (line addition, change of the bandwidth upper limit value, etc.), the traffic fluctuation can be dealt with by the prediction using the feature amount (the first feature amount 23a) based on the required conditions, and the highly accurate prediction of the traffic flow can be realized.

[0033] Second, according to the traffic prediction device 1 according to the present embodiment, even when the correlation between the required conditions and the traffic statistics is low, the traffic fluctuation due to the change of the required conditions can be dealt with by the prediction using the traffic statistics (the second feature amount 23b), and the highly accurate prediction of the traffic flow can be realized.

[0034] Third, according to the traffic prediction device 1 according to the present embodiment, since the prediction is performed based on the traffic statistics, the traffic fluctuation for each time interval can be dealt with. For example, in the above-described conventional technique SVAE, in order to calculate the predicted value for the required conditions, only one value can be set every day, but according to the traffic prediction device 1, since it is possible to calculate the predicted value for each time interval, it is also possible to deal with the small-scale traffic fluctuation for each time interval.

[0035] FIG. 5 is a table showing the quantitative effect of the traffic prediction device 1 according to the first embodiment. In FIG. 5, the prediction error by the conventional method and the prediction error by the traffic prediction device 1 are compared. Here, the prediction error is the RMSE (Root Mean Squared Error) between the measured value and the predicted value. As shown in FIG. 5, the prediction error by the traffic prediction device 1 according to the present embodiment is 12.791, and good data was obtained as compared with the prediction error of 57.594 using LSTM by the prior art and the prediction error of 34.168 using SVAE.

[0036] (Second Embodiment) FIG. 6 is a block diagram showing a configuration example of the traffic prediction device 2 according to the second embodiment. As shown in FIG. 6, the traffic prediction device 2 includes a data acquisition unit 11, a prediction data generation unit 12, a prediction function unit 13, and a data recording unit 14. The traffic prediction device 2 predicts the future traffic flow rate in a link that accommodates a plurality of lines. The traffic prediction device 2 according to the present embodiment is different from the traffic prediction device 1 according to the first embodiment in that it further includes a data recording unit 14. For the same configuration as that of the first embodiment, the same reference numerals as those in the first embodiment are given, and the description thereof is omitted as appropriate.

[0037] The data acquisition unit 11, the prediction data generation unit 12, and the prediction function unit 13 constitute a control unit 10 (controller 10). The control unit 10 (controller 10) may be configured by dedicated hardware such as an ASIC or an FPGA, may be configured by a processor, or may be configured to include both.

[0038] The prediction data generation unit 12 may predict the period of traffic fluctuation based on the past traffic statistics recorded in the data recording unit 14, and generate a second feature amount 23b based on the traffic statistics in a period corresponding to the predicted period. The prediction data generation unit 12 sequentially transmits the generated traffic statistics to the data recording unit 14 for recording. The prediction data generation unit 12 can extract useful traffic statistics from various past traffic statistics recorded in the data recording unit 14 and generate traffic statistics for predicting the period of traffic fluctuation.

[0039] The data recording unit 14 sequentially stores the traffic statistics transmitted from the prediction data generation unit 12. The data recording unit 14 transmits the requested past traffic statistics to the prediction data generation unit 12 in response to a request from the prediction data generation unit 12.

[0040] FIG. 7 is a flowchart showing an example of a traffic prediction method executed by the traffic prediction device according to the second embodiment.

[0041] In step S201, the data acquisition unit 11 acquires the line request condition 21 that changes for each arbitrary period.

[0042] In step S202, traffic data 22 for generating a traffic statistic representing the characteristics of traffic fluctuations over time is acquired. The processing of step S201 and the processing of step S202 may be performed in parallel as shown in FIG. 7, or either one of the processes may be performed first and the other process may be performed later.

[0043] In step S203, the prediction data generation unit 12 predicts the period of traffic fluctuations from past traffic statistics.

[0044] In step S204, the prediction data generation unit 12 extracts past traffic feature amounts recorded in the data recording unit 14.

[0045] In step S205, the prediction data generation unit 12 generates a first feature amount 23a and a second feature amount 23b.

[0046] In step S206, the prediction data generation unit 12 causes the generated traffic statistic to be recorded in the data recording unit 14.

[0047] In step S207, the prediction function unit 13 predicts the future traffic flow rate 24 from the first feature amount 23a and the second feature amount 23b.

[0048] The traffic statistic according to the present embodiment can be various traffic statistics such as the traffic statistic at each past time point, the amount of change between the current traffic and the traffic at each past time point, and the weighted moving average. It is determined which feature amounts are to be extracted from such various statistics. In the following example, examples of setting the number of steps of past traffic and the number of days of moving average are given. Similarly, the traffic statistic according to the present embodiment can also be created using a determination based on whether the request conditions are the same or a determination by general statistical processing.

[0049] According to the traffic prediction device 2 according to this embodiment, as shown in the following four examples, it is possible to realize a highly accurate prediction of traffic flow. FIG. 8 is a diagram showing an example of input data used to generate prediction data.

[0050] <Example of setting the number of days of moving average - 1> For example, at the prediction time shown in the input data 1 of FIG. 8, a moving average for a period shorter than the application period of the same request condition (1 day in this example) is adopted. By adopting such a short-period moving average, the prediction accuracy is improved by predicting from a moving average that does not include traffic flows under different request conditions that have different behaviors from the prediction time.

[0051] <Example of setting the number of days of moving average - 2> Calculate the moving average in multiple periods, and by judging whether each moving average is an outlier through statistical processing, adopt the moving average value after the moving average of the most recent outlier. General methods such as the Smrinov-Grubbs test are used for the statistical test used to determine outliers. By predicting the period of traffic fluctuations other than those due to changes in request conditions and predicting from a moving average that does not include traffic flows with different behaviors from the prediction time, the prediction accuracy is improved.

[0052] <Example of setting the number of steps of past traffic - 1> For example, at the prediction time shown in the input data 2 of FIG. 8, adopt the traffic flow within the application period of the same request condition. By adopting such a traffic flow, the prediction accuracy is improved by predicting from past traffic flows that do not include traffic flows under different request conditions that have different behaviors from the prediction time.

[0053] <Example of setting the number of steps of past traffic - 2> By determining whether each traffic flow is an outlier through statistical processing of the past traffic flows during any period, the traffic flows after the most recent outlier are adopted as prediction input data with the traffic flows several steps before. By predicting the period of traffic fluctuations other than those due to changes in the requirement conditions and making a prediction from the past traffic flows that do not include traffic flows with behaviors different from the prediction time point, the prediction accuracy is improved.

[0054] In order to make the above-described traffic prediction apparatuses 1 and 2 function, it is also possible to use a computer capable of executing program instructions. FIG. 9 is a block diagram showing a schematic configuration of a computer that functions as a traffic prediction apparatus. Here, the computer that functions as the traffic prediction apparatuses 1 and 2 may be a general-purpose computer, a dedicated computer, a workstation, a PC (Personal Computer), an electronic notebook pad, or the like. The program instructions may be a program code, a code segment, etc. for executing necessary tasks.

[0055] As shown in FIG. 9, the computer 100 includes a processor 110, a ROM (Read Only Memory) 120, a RAM (Random Access Memory) 130, and a storage 140 as storage units, an input unit 150, an output unit 160, and a communication interface (I / F) 170. Each component is communicably connected to each other via a bus 180.

[0056] The ROM 120 stores various programs and various data. The RAM 130 temporarily stores a program or data as a work area. The storage 140 is composed of an HDD (Hard Disk Drive) or an SSD (Solid State Drive) and stores various programs and various data including an operating system. In the present disclosure, the program according to the present disclosure is stored in the ROM 120 or the storage 140.

[0057] The processor 110 is specifically a CPU (Central Processing Unit), MPU (Micro Processing Unit), GPU (Graphics Processing Unit), DSP (Digital Signal Processor), SoC (System on a Chip), etc., and may be composed of a plurality of processors of the same type or different types. The processor 110 reads a program from the ROM 120 or the storage 140, and executes the program using the RAM 130 as a working area, thereby controlling each of the above components and performing various arithmetic processes. Note that at least a part of these processing contents may be realized by hardware.

[0058] The program may be recorded on a recording medium readable by the traffic prediction devices 1 and 2. By using such a recording medium, it is possible to install it on the traffic prediction devices 1 and 2. Here, the recording medium on which the program is recorded may be a non-transitory recording medium. The non-transitory recording medium is not particularly limited, and may be, for example, a CD-ROM, a DVD-ROM, a USB (Universal Serial Bus) memory, etc. Also, this program may be in a form downloaded from an external device via a network.

[0059] Regarding the above embodiments, the following additional remarks are further disclosed.

[0060] (Additional item 1) A traffic prediction device that predicts future traffic flow rates in a link accommodating a plurality of lines, A traffic prediction device comprising: a controller that acquires the required conditions of the line and traffic data in the line, generates a first feature quantity corresponding to traffic fluctuations due to changes in the required conditions of the line based on the acquired required conditions, generates a second feature quantity based on a traffic statistic representing the characteristics of traffic fluctuations over time based on the acquired traffic data, and predicts a future traffic flow rate from the first feature quantity and the second feature quantity. (Additional item 2) The traffic prediction device further comprises a memory that sequentially records the traffic statistics. The controller predicts a traffic fluctuation cycle based on past traffic statistics recorded in the memory, and generates the second feature quantity based on traffic statistics in a period corresponding to the predicted cycle. The traffic prediction device according to claim 1. (Additional item 3) A traffic prediction method for predicting a future traffic flow rate in a link accommodating a plurality of lines, comprising: a step of acquiring, by a traffic prediction device, the required conditions of the line and traffic data in the line; a step of generating a first feature quantity corresponding to traffic fluctuations due to changes in the required conditions of the line based on the acquired required conditions; a step of generating a second feature quantity based on a traffic statistic representing the characteristics of traffic fluctuations over time based on the acquired traffic data; and a step of predicting a future traffic flow rate from the first feature quantity and the second feature quantity. (Additional item 4) A non-transitory storage medium storing a program executable by a computer, the non-transitory storage medium storing a program that causes the computer to function as the traffic prediction device according to claim 1 or 2.

[0061] Although the above-described embodiments have been described as representative examples, it is apparent to those skilled in the art that many changes and substitutions can be made within the spirit and scope of the present disclosure. Therefore, the present invention should not be construed as being limited by the above-described embodiments, and various modifications or changes are possible without departing from the scope of the claims. For example, it is possible to combine a plurality of constituent blocks described in the configuration diagrams of the embodiments into one, or to divide one constituent block.

Explanation of Reference Numerals

[0062] 1,2 Traffic prediction device 10 Control unit (controller) 11 Data acquisition unit 12 Prediction data generation unit 13 Prediction function unit 13a Prediction model 14 Data recording unit (memory) 21 Line requirement conditions 22 Traffic data 23 Prediction data 23a First feature quantity (feature quantity based on line requirement conditions) 23b Second feature quantity (feature quantity based on traffic statistics) 24 Future traffic flow rate (prediction result) 100 Computer 110 Processor 120 ROM 130 RAM 140 Storage 150 Input unit 160 Output unit 170 Communication interface (I / F) 180 Bus

Claims

1. A traffic prediction device for predicting future traffic flow rates in a link that accommodates a plurality of lines, comprising: a data acquisition unit that acquires the required conditions of the lines and traffic data in the lines; a prediction data generation unit that generates a first feature amount corresponding to traffic fluctuations due to changes in the required conditions of the lines based on the acquired required conditions, and generates a second feature amount based on a traffic statistic representing the characteristics of traffic fluctuations over time based on the acquired traffic data; a data recording unit that sequentially records the traffic statistics; a prediction function unit that predicts future traffic flow rates from the first feature amount and the second feature amount; and the prediction data generation unit predicts the period of traffic fluctuations based on past traffic statistics recorded in the data recording unit, and generates the second feature amount based on traffic statistics for a period corresponding to the predicted period, the traffic prediction device.

2. A traffic prediction method executed by a traffic prediction device for predicting future traffic flow rates in a link that accommodates a plurality of lines, comprising: acquiring the required conditions of the lines and traffic data in the lines; generating a first feature amount corresponding to traffic fluctuations due to changes in the required conditions of the lines based on the acquired required conditions; a second feature amount generation step of generating a second feature amount based on a traffic statistic representing the characteristics of traffic fluctuations over time based on the acquired traffic data; a recording step of sequentially recording the traffic statistics; predicting future traffic flow rates from the first feature amount and the second feature amount; and including the second feature amount generation step includes predicting the period of traffic fluctuations based on past traffic statistics recorded in the recording step; generating the second feature amount based on traffic statistics for a period corresponding to the predicted period; and including the traffic prediction method.

3. A program for causing a computer to function as the traffic prediction device according to Claim 1.

Images