JP2022109135A - Information processing device, information processing method and information processing program - Google Patents

Abstract

To improve prediction accuracy about a prediction object.SOLUTION: An information processing device according to the present application includes a determination part and a calculation part. The determination part determines a ratio of combining a first prediction value about a prediction object predicted by prediction processing and an actual value on the basis of the reliability of a past actual value about the prediction object. The calculation part calculates a second prediction value about the prediction object from the first prediction value and the actual value on the basis of the ratio determined by the determination part.SELECTED DRAWING: Figure 4

Description

The present invention relates to an information processing device, an information processing method, and an information processing program.

Conventionally, techniques for calculating predicted values for various prediction targets are known. For example, when receiving a request from a vehicle to predict the travel time of a specific route, the travel time of that route is predicted based on the actual values and average values of the travel time of each section included in the route, and the prediction result is obtained. to the vehicle.

JP-A-2000-259977

A technique for improving the prediction accuracy of a prediction target is desired.

The information processing apparatus according to the embodiment determines a ratio of combining a first predicted value for the prediction target predicted by prediction processing and the actual value based on the reliability of past actual values for the prediction target. and a calculating unit that calculates a second predicted value related to the prediction target from the first predicted value and the actual value based on the ratio determined by the determining unit.

FIG. 1 is a diagram illustrating a configuration example of an information processing system according to an embodiment. FIG. 2 is a diagram illustrating a configuration example of an information processing apparatus according to the embodiment; FIG. 3 is an explanatory diagram showing an overview of information processing according to the embodiment. FIG. 4 is an explanatory diagram showing an overview of information processing according to the embodiment. FIG. 5 is a flow chart showing an information processing procedure according to the embodiment. FIG. 6 is a flow chart showing an information processing procedure according to the embodiment. FIG. 7 is a hardware configuration diagram showing an example of a computer that implements the functions of the information processing apparatus.

Embodiments for implementing an information processing apparatus, an information processing method, and an information processing program according to the present application (hereinafter referred to as "embodiments") will be described in detail below with reference to the drawings. The information processing apparatus, information processing method, and information processing program according to the present application are not limited to this embodiment. Also, in each of the following embodiments, the same parts are denoted by the same reference numerals, and overlapping descriptions are omitted.

(embodiment)
[1. Introduction]
In recent years, there has been growing interest in MaaS (Mobility as a Service), which is a concept of providing movement from a departure point to a destination as a single service. MaaS-based services allow users to use dedicated applications such as smartphones to search for routes from their departure point to their destination, make reservations, and make payments all at once.

Among MaaS services, services related to route search may be provided using a route search API (Application Programming Interface). For example, when the route search API receives from the user a prediction request regarding a search for a travel route from a starting point to a destination designated by the user, the route search API searches for a travel route. Then, the route search API returns to the user, together with the search result of the travel route, a predicted value of travel time required to travel the travel route corresponding to the search result.

Here, in order to improve the prediction accuracy of the travel time, it is conceivable to use the actual value of the travel time required to actually move the travel route for prediction. For example, newer performance values are considered more reliable than older performance values. Therefore, for example, it is conceivable to utilize the latest actual value of travel time for prediction. However, the latest actual value is not necessarily a reliable value. For example, depending on the weather, road conditions, etc., the latest actual value may deviate greatly from the expected travel time value. In this way, individual performance values contain uncertainties and are not necessarily reliable values. Therefore, in order to improve the travel time prediction accuracy, it is necessary to consider the reliability of the actual values and use the actual values for prediction.

On the other hand, the information processing apparatus 100 according to the present embodiment uses an existing route search API based on the reliability of the past performance value regarding the travel time required for the mobile object to move from the departure point to the arrival point. A ratio is determined that combines the first predicted value for the predicted travel time with the actual value. Then, the information processing apparatus 100 calculates a second predicted value regarding travel time from the first predicted value and the actual value based on the determined ratio.

Thereby, for example, when the reliability of the actual value is low, the information processing apparatus 100 can calculate the second predicted value by increasing the ratio of combining the first predicted value by the route search API with the actual value. can. On the other hand, for example, when the reliability of the actual value is high, the information processing apparatus 100 can calculate the second predicted value by increasing the ratio of combining the actual value with respect to the first predicted value by the route search API. . In this way, the information processing apparatus 100 can use the actual value for prediction according to the reliability of the actual value, after considering the reliability of the actual value. That is, the information processing apparatus 100 can utilize the actual value for prediction while covering the uncertainty of the actual value. Therefore, the information processing apparatus 100 can improve the prediction accuracy of the travel time.

Also, one of MaaS services is an on-demand bus (a type of mobile) dispatch service. Among such on-demand bus dispatch services, there are services provided by various servicers according to the type of service that the user receives through the dispatch service, such as nursing care service, mobile sales service, and nursing service. . Since the operation mode of the on-demand bus is usually different for each servicer, the operation time (also referred to as travel time) required for the on-demand bus to move from the departure point to the destination differs for each servicer. Therefore, in order to improve the prediction accuracy of the operation time, it is necessary to optimize the operation time prediction process for each servicer.

On the other hand, the information processing apparatus 100 according to the present embodiment provides the type of metadata (for example, the type of servicer, etc.) associated with the actual value of the travel time required for the mobile object to travel from the departure point to the arrival point. Classify the actual values into clusters based on . Then, the information processing apparatus 100 calculates a ratio of combining the first predicted value of the travel time predicted by the existing route search API and the actual value based on the reliability of the past actual value belonging to the classified cluster. decide. Further, the information processing apparatus 100 calculates a second predicted value regarding travel time from the first predicted value and the actual value based on the determined ratio.

In this way, the information processing apparatus 100 predicts travel times for each type of servicer, based on actual values of travel times classified according to the type of servicer, for example. As a result, the information processing apparatus 100 can, for example, predict the travel time according to different operation modes for each type of servicer, so that the prediction accuracy of the travel time can be further improved.

[2. Configuration of information processing system]
Next, the configuration of the information processing system according to the embodiment will be described using FIG. FIG. 1 is a diagram illustrating a configuration example of an information processing system according to an embodiment. The information processing system 1 may include a user device 10 , a mobile device 20 and an information processing device 100 . The user device 10, the mobile device 20, and the information processing device 100 may be communicatively connected via a predetermined network N by wire or wirelessly. The information processing system 1 may include any number of user devices 10 , any number of mobile devices 20 , and any number of information processing devices 100 .

The user device 10 is an information processing device used by a user who uses the on-demand bus dispatch service. The user device 10 is realized by, for example, a smart phone, a tablet terminal, a notebook PC (Personal Computer), a desktop PC, a mobile phone, a PDA (Personal Digital Assistant), or the like. In addition, the user device 10 displays information on the vehicle dispatch service distributed by the information processing device 100 (for example, information displayed on the screen of the user device 10 when making a vehicle dispatch request) using a web browser or an application. In the present embodiment, the user device 10 displays information about the dispatch service distributed by the information processing device 100 by means of an application (hereinafter also referred to as a dispatch application).

Also, the user device 10 transmits a prediction request regarding the on-demand bus to the information processing device 100 . For example, the user device 10, according to the operation of the user, information (hereinafter sometimes referred to as metadata) about the departure place, arrival place, departure time (or arrival time) of the on-demand bus, and servicer type. on the dispatch app screen. Subsequently, the user device 10 transmits to the information processing device 100 a prediction request relating to prediction of travel time and travel route together with metadata in accordance with the user's operation of pressing the route search button. For example, if the on-demand bus departs from the departure point at a designated departure time, the user device 10 may provide information about the estimated travel time required for the on-demand bus to travel from the departure point to the arrival point. A prediction request is transmitted to the information processing apparatus 100 . In addition, for example, if the on-demand bus arrives at the destination at the designated arrival time, the user device 10 may calculate the amount of travel predicted to be required for the on-demand bus to travel from the departure point to the destination. A prediction request regarding time is transmitted to the information processing apparatus 100 .

Also, the user device 10 receives a prediction response to the prediction request from the information processing device 100 . For example, the user device 10 receives, as an example of the prediction response, information about the predicted travel time. When user device 10 receives the information about the predicted value of travel time, user device 10 displays the received information about the predicted value on the screen of user device 10 .

The mobile device 20 is an information processing device that manages operation information of the on-demand bus. When the operation of the on-demand bus is completed, the mobile device 20 transmits operation information regarding the operation record of the on-demand bus to the information processing device 100 . For example, the mobile device 20 includes, as an example of operation information, information on the departure place, arrival place, departure time (or arrival time) of the on-demand bus, servicer type (hereinafter sometimes referred to as metadata), and information on the actual travel time required for the on-demand bus to travel from the departure point to the arrival point to the information processing apparatus 100 .

For example, when the on-demand bus departs from the point of departure at a designated departure time, the mobile device 20 stores information about the travel time from the point of departure to the point of arrival. to the information processing apparatus 100 . Further, for example, when the on-demand bus arrives at the destination at the designated arrival time, the mobile device 20 may store the travel time from the departure place to the destination as an example of the information related to the actual value of the travel time. Information on the actual value of the travel time required to do so is transmitted to the information processing apparatus 100 .

The information processing apparatus 100 classifies the actual values into clusters based on the types of metadata associated with the actual values of the travel time required for the on-demand bus to move from the departure point to the destination. Subsequently, the information processing apparatus 100 generates a first predicted value of travel time predicted by an existing route search API (Application Programming Interface) based on the reliability of past actual values belonging to the classified clusters, and the actual travel time. Determines the ratio in which to combine values. Subsequently, the information processing apparatus 100 calculates a second predicted value regarding travel time from the first predicted value and the actual value based on the determined ratio.

[3. Configuration of Information Processing Device]
Next, the configuration of the information processing apparatus 100 according to the embodiment will be described using FIG. FIG. 2 is a diagram illustrating a configuration example of the information processing apparatus 100 according to the embodiment. The information processing device 100 may have a communication section 110 , a storage section 120 and a control section 130 . The information processing apparatus 100 includes an input unit (for example, a keyboard, a mouse, etc.) for receiving various operations from an administrator of the information processing apparatus 100, and a display unit (for example, a liquid crystal display) for displaying various information. may have.

(Communication unit 110)
The communication unit 110 is realized by, for example, a NIC (Network Interface Card) or the like. The communication unit 110 is connected to a network by wire or wirelessly, and transmits and receives information to and from the user device 10 and the mobile device 20, for example.

(storage unit 120)
The storage unit 120 is realized by, for example, a semiconductor memory device such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disk. For example, the storage unit 120 stores operation information acquired from the mobile device 20 .

(control unit 130)
The control unit 130 is a controller, and for example, various programs (information processing programs) stored in a storage device inside the information processing apparatus 100 are controlled by a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like. (equivalent to one example) is implemented by executing the RAM as a work area. Also, the control unit 130 is a controller, and is implemented by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

The control unit 130 has an acquisition unit 131, a classification unit 132, a reception unit 133, a determination unit 134, a determination unit 135, a calculation unit 136, and a transmission unit 137 as functional units, which will be described below. Realize or perform information processing operations. Note that the internal configuration of the control unit 130 is not limited to the configuration shown in FIG. 2, and may be another configuration as long as it performs the information processing described later. Moreover, each functional unit indicates the function of the control unit 130 and may not necessarily be physically distinguished.

(Acquisition unit 131)
The acquisition unit 131 acquires operation information related to the operation record of the on-demand bus from the mobile device 20 . For example, the acquisition unit 131 obtains, as an example of operation information, information on the departure place, arrival place, departure time (or arrival time) of the on-demand bus, servicer type (hereinafter sometimes referred to as metadata), and To obtain information about the actual travel time required for an on-demand bus to travel from a departure point to an arrival point. Upon acquiring the operation information, the acquisition unit 131 stores the acquired operation information in the storage unit 120 .

For example, when the on-demand bus departs from the departure point at a specified departure time, the acquisition unit 131 obtains an example of information related to the actual value of travel time. Get information about the actual travel time taken. In addition, for example, when the on-demand bus arrives at the destination at the specified arrival time, the acquisition unit 131 obtains, as an example of the information related to the travel time actual value, the travel time of the on-demand bus from the departure point to the destination. Get information about the actual value of the travel time required to After acquiring the information on the actual value of the travel time, the acquiring unit 131 stores the acquired information on the actual value of the travel time in the storage unit 120 as an example of operation information in association with the metadata.

After acquiring the actual value, the acquiring unit 131 determines whether the actual value is an outlier. For example, the acquisition unit 131 uses the Smirnov-Grubbs test or the interquartile range (IQR) to determine whether the actual value is an outlier. When determining that the acquired performance value is an outlier, the acquisition unit 131 removes the acquired performance value as an outlier. After removing the outliers, the acquisition unit 131 updates the past performance values stored in the storage unit 120 . On the other hand, when the acquisition unit 131 determines that the acquired performance value is not an outlier, the acquisition unit 131 adds the acquired performance value to the past performance values. Acquisition part 131 will update the past performance value memorized by storage part 120, if a performance value is added.

(Classification unit 132)
From here, the process of classifying the actual values into clusters by the classifying unit 132 will be described with reference to FIG. 3 . FIG. 3 is an explanatory diagram showing an overview of information processing according to the embodiment. As shown in FIG. 3, the classification unit 132 classifies the actual values into clusters based on the types of metadata associated with the actual values. For example, the classification unit 132 refers to the storage unit 120 and acquires past performance values. Subsequently, when the past performance values are acquired, the classification unit 132 classifies the past performance values based on the type of metadata associated with the past performance values. For example, the classification unit 132 classifies past performance values for each condition based on metadata such as travel route, service type (servicer type), day of the week, and boarding status. In this way, the classification unit 132 classifies the past performance values for each condition in advance based on the metadata. For example, the classification unit 132 classifies past performance values into classification condition clusters that satisfy classification conditions based on metadata associated with past performance values.

Subsequently, the classification unit 132 classifies the past performance values according to the conditions, and classifies the past performance values according to the time zone indicating the departure time (or arrival time) associated with the past performance value. time clusters. For example, the classification unit 132 classifies past performance values into time zone clusters according to time zones using a known method such as k-means clustering. After classifying the past performance values into time zone clusters, the classification unit 132 updates the information of the time zone clusters into which the past performance values have been classified. In this way, the classification unit 132 classifies the past performance values classified for each condition into time zone clusters in advance.

Further, when it is determined that the actual value acquired by the acquiring unit 131 is not an outlier, the classifying unit 132 determines the metadata associated with the actual value acquired by the acquiring unit 131 based on the metadata associated with the actual value. It is determined whether or not there is a classification condition cluster that satisfies the classification condition based on the data. When the classification unit 132 determines that there is no classification condition cluster that satisfies the classification condition, the classification unit 132 newly generates a classification condition cluster corresponding to the classification condition. After generating the classification condition cluster, the classification unit 132 classifies the performance values acquired by the acquisition unit 131 into the generated classification condition cluster. After classifying the performance values acquired by the acquisition unit 131 into the generated classification condition clusters, the classification unit 132 updates the past performance values belonging to the classification condition clusters.

On the other hand, when determining that there is a classification condition cluster that satisfies the classification condition, the classification unit 132 classifies the actual values acquired by the acquisition unit 131 into the classification condition cluster corresponding to the classification condition. Subsequently, after classifying the performance values acquired by the acquisition unit 131 into classification condition clusters corresponding to the classification conditions, the classification unit 132 updates the past performance values belonging to the classification condition clusters.

Subsequently, when the past performance values belonging to the classification condition cluster are updated, the classification unit 132 updates the classification condition according to the time zone indicating the departure time (or arrival time) associated with the past performance value belonging to the classification condition cluster. The past performance values belonging to the cluster are classified into time zone clusters according to the time zone. For example, the classification unit 132 classifies past performance values into time zone clusters according to time zones using a known clustering method such as k-means clustering. After classifying the past performance values belonging to the classification condition clusters into time zone clusters according to the departure time (or arrival time), the classification unit 132 updates the past performance values belonging to the time zone clusters.

(Reception unit 133)
From here, the travel time prediction processing by the information processing apparatus 100 will be described with reference to FIG. 4 . FIG. 4 is an explanatory diagram showing an overview of information processing according to the embodiment. In the example shown in FIG. 4 , first, the accepting unit 133 accepts a prediction request from the user device 10 . For example, the reception unit 133 receives the information required for the on-demand bus to travel from the specified departure place to the specified arrival place, along with the departure place, arrival place, and departure time (or arrival time) specified by the user. It then accepts a prediction request about the predicted travel time. In addition to the prediction request, the reception unit 133 also receives a search request regarding a travel route from the departure point to the arrival point specified by the user. The reception unit 133 also receives metadata such as the service type (servicer type), day of the week, and boarding status specified by the user, together with the prediction request.

For example, if the reception unit 133 receives the designation of the departure time from the user, and if the on-demand bus leaves the designated departure place at the designated departure time, the reception unit 133 Receive prediction requests regarding the estimated travel time required to travel from to a specified destination. Further, for example, if the reception unit 133 receives a specification of the arrival time from the user, and if the on-demand bus arrives at the arrival place at the specified arrival time, the on-demand bus will travel from the departure place to the arrival place. Accepts prediction requests for estimated travel times required to travel. Subsequently, upon receiving the prediction request, the receiving unit 133 acquires a first predicted value related to travel time related to the prediction request, which is predicted using the existing route search API.

(Determination unit 134)
When the reception unit 133 receives the prediction request, the determination unit 134 identifies the classification condition corresponding to the second predicted value regarding the travel time related to the prediction request based on the metadata associated with the prediction request. For example, the determination unit 134 responds to the second predicted value based on metadata received together with the prediction request, such as the travel route to the departure point and the arrival point, the type of service (type of servicer), the day of the week, and the boarding status. Identify classification criteria. After identifying the classification conditions, the determination unit 134 then determines whether or not there is a classification condition cluster that satisfies the classification conditions. In the example illustrated in FIG. 4, the determination unit 134 determines that there is a classification condition cluster that satisfies the classification condition for the second predicted value.

(Determination unit 135)
When the determination unit 134 determines that there is a classification condition cluster that satisfies the classification condition of the second predicted value, the determination unit 135 determines the time zone indicating the departure time (or arrival time) associated with the second predicted value. , to identify the time clusters corresponding to the second predicted values. Subsequently, when identifying the time period cluster corresponding to the second predicted value, the determination unit 135 refers to the storage unit 120 to acquire past performance values belonging to the identified time period cluster.

Subsequently, when acquiring the past performance values belonging to the time period cluster into which the second predicted values have been classified, the determination unit 135 evaluates the reliability of the performance values. Specifically, the determination unit 135 performs statistical processing on the past performance values belonging to the time period cluster into which the second predicted values are classified, and calculates a reliability value indicating the reliability of the performance values.

More specifically, it is assumed that the probability distribution of actual values belonging to the same time cluster follows a normal distribution. This can be expressed as a formula, where X is the random variable corresponding to the actual values belonging to the same time zone cluster, μ is the average, and σ is the standard deviation.

Also, the sample mean of the normal distribution follows the normal distribution with the following population variance as a parameter. When this is represented by a formula, it is shown by the following Formula (2).

Here, using the sample number ⁿ of the actual value and the sample variance s2, the standard error of the actual value is given by the following equation (3).

The determining unit 135 calculates the reliability value by dividing the standard error of the actual values shown in the above equation (3) by the sample average of the actual values. Here, the confidence value t is represented by the following formula (4).

The determining unit 135 calculates a reliability value indicating that the smaller the variance of the actual value is, the higher the reliability of the actual value is, as shown in the above equation (4). The determination unit 135 also calculates a reliability value indicating that the reliability of the actual value increases as the number of samples of the actual value increases.

Here, since the actual value is time-series data, it is considered that the latest actual value (the date and time when the actual value was obtained is the latest) is the most reliable as the actual value. Therefore, the determination unit 135 calculates a reliability value indicating the reliability of the actual value by increasing the weight as the actual value is newer. Specifically, the determination unit 135 uses the actual value exponential moving average _mt and the actual value exponential moving variance _vt to replace the above equation (4).

Here, the exponential moving average refers to a weighted average of actual values calculated by assigning greater weight to individual actual values as the actual value becomes newer. The exponential moving average refers to the average value of actual values calculated by assigning exponentially decreasing weights to individual actual values. Also, the exponential movement variance refers to the variance of actual values calculated by assigning greater weight to individual actual values as the newer the actual value is.

More specifically, the determining unit 135 replaces the sample average of the actual values shown in the above equation (4) with the exponential moving average _mt of the actual values. The exponential moving average mt of actual values is _expressed by the following equation ( ₅ ) using the latest actual value _gt and weight β1.

Further, the determining unit 135 replaces the sample variance s2 of the actual values shown in the above equation ( ⁴ ) with the exponential movement variance _vt of the actual values. The actual value exponential movement variance v _t is expressed by the following equation (6) using the latest actual value g _t and weight β ₂ .

Further, the determining unit 135 calculates a corrected exponential moving average obtained by correcting the exponential moving average _mt of the actual values by the following formula (7). The corrected exponential moving average is represented by the following formula (7).

Further, the determining unit 135 calculates a corrected exponential moving variance obtained by correcting the actual value exponential moving variance _vt using the following formula (8). The corrected exponential moving variance is given by Equation (8) below.

As described above, the determining unit 135 calculates the corrected exponential moving average and the corrected exponential moving variance of the actual values. After calculating the corrected exponential moving average and the corrected exponential moving variance, the determining unit 135 then calculates a confidence value represented by the following formula (9) based on the calculated corrected exponential moving average and corrected exponential moving variance. .

As described above, the determining unit 135 calculates the reliability value indicating the reliability of the actual value based on the exponential moving average and the exponential moving variance of the actual value. The determination unit 135 also calculates a reliability value indicating that the smaller the exponential movement variance of the actual value, the higher the reliability of the actual value. In this way, the determination unit 135 uses the exponential moving average and the exponential moving variance of the actual values to calculate the reliability value indicating that the newer the actual value, the higher the reliability of the actual value.

Subsequently, after calculating the reliability value, the determination unit 135 determines a ratio of combining the first predicted value regarding the travel time predicted by the existing route search API and the actual value based on the calculated reliability value. Specifically, the determining unit 135 determines the ratio of combining the first predicted value and the representative value representing the actual value. For example, an example of the representative value is the exponential moving average of the actual values, and an example of the exponential moving average of the actual values is the corrected exponential moving average of the actual values shown in the above equation (7).

More specifically, the determining unit 135 determines a function of the confidence value t that determines the ratio of combining the first predicted value and the corrected exponential moving average of the actual value. For example, as a function of the confidence value t, the determination unit 135 obtains a polynomial of the confidence value t that satisfies the continuity condition shown in Equation (10) below.

For example, as a polynomial that satisfies the continuity condition (when k=2) shown in the above equation (10), the polynomial α(t) of the confidence value t shown in the following equation (11) is obtained.

The determining unit 135 substitutes the calculated reliability value t for the function α(t) represented by the above equation (11), and determines the ratio of combining the first predicted value and the corrected exponential moving average of the actual values. do. Specifically, the determining unit 135 calculates the first predicted value and the corrected exponential moving average of actual values. For example, the determining unit 135 sets the ratio of combining the first predicted value and the corrected exponential moving average of the actual values to the weight α of the corrected exponential moving average of the actual values as the weight of the first predicted value (1−α). decide to be The determination unit 135 determines the ratio of combining the first predicted value and the corrected exponential moving average of the actual values, as shown in the following equation (12).

As shown in the above formula (12), the determination unit 135 determines that the higher the reliability of the corrected exponential moving average of the actual values (that is, the smaller the confidence value), the more predicted by the existing route search API. The ratio of combining the corrected exponential moving average of the actual value with respect to the first predicted value of the travel time is increased. In addition, the determining unit 135 increases the ratio of combining the first predicted values to the corrected exponential moving average of the actual values as the reliability of the corrected exponential moving average of the actual values is lower (that is, the larger the value of the reliability value). do.

On the other hand, when the determination unit 134 determines that there is no classification condition cluster that satisfies the classification condition for the second predicted value, the determination unit 135 determines that the actual value does not exist because there is no actual value for the classification condition for the second predicted value. is determined to be "1" (maximum confidence value) indicating the lowest confidence value. Subsequently, when the determination unit 135 determines the confidence value to be “1”, the determination unit 135 substitutes the determined confidence value “1” into the function represented by the above equation (11) to set the value of α to “0”. calculate. Next, based on the calculated value of α “0”, the determining unit 135 sets the ratio of combining the first predicted value and the actual value to the weight “0” of the actual value and the weight of the first predicted value “ 1”.

In this way, when the determination unit 134 determines that there is no classification condition cluster that satisfies the classification condition for the second predicted value, the determination unit 135 does not have a performance value that satisfies the classification condition for the second predicted value. The reliability of the actual value is evaluated as the lowest. Moreover, since the reliability of the actual value is the lowest, the determination unit 135 determines not to utilize the actual value for prediction. That is, the determining unit 135 determines to calculate the second predicted value based only on the first predicted value of travel time predicted by the existing route search API.

(Calculation unit 136)
Based on the ratio determined by the determination unit 135, the calculation unit 136 calculates a second predicted value regarding travel time from the first predicted value and the corrected exponential moving average of the actual values. Specifically, the calculating unit 136 calculates the corrected exponential moving average of the actual values based on the weight α of the corrected exponential moving average of the actual values determined by the determining unit 135 and the weight (1−α) of the first predicted value. A second predicted value is calculated from the average and the first predicted value. For example, the calculation unit 136 calculates the value obtained by multiplying the corrected exponential moving average of the actual values by the weight α of the corrected exponential moving average of the actual values determined by the determining unit 135, and the first predicted value determined by the determining unit 135. A value obtained by multiplying the first predicted value by the weight (1-α) is calculated. Subsequently, the calculation unit 136 multiplies the weight α of the corrected exponential moving average of the actual values by the corrected exponential moving average of the actual values and the weight of the first predicted value (1−α) by multiplying the first predicted value. The second predicted value is calculated by adding the calculated value and the calculated value.

(Sending unit 137)
The transmission unit 137 transmits the predicted response to the user device 10. FIG. Specifically, when the calculating unit 136 calculates the second predicted value related to the travel time, the transmitting unit 137 sends the second predicted value, the travel route retrieved by the existing route search API, as an example of the prediction response. Along with the search result, a second predicted value of travel time required to travel the travel route corresponding to the search result is transmitted to the user device 10 .

[4. Information processing procedure]
Next, the procedure of information processing according to the embodiment will be described with reference to FIG. FIG. 5 is a flow chart showing an information processing procedure according to the embodiment.

In the example shown in FIG. 5, the on-demand service ends (step S101). When the on-demand operation ends, the mobile device 20 transmits operation information regarding the on-demand operation to the information processing device 100 .

The information processing device 100 acquires operation information from the mobile device 20 (step S102). When acquiring the operation information, the information processing device 100 acquires the performance value included in the operation information. Subsequently, when acquiring the actual value, the information processing apparatus 100 determines whether or not the actual value is an outlier (step S103).

When the information processing apparatus 100 determines that the acquired performance value is an outlier (step S103; Yes), the information processing apparatus 100 removes the acquired performance value as an outlier (step S105). After removing the outliers, the information processing apparatus 100 updates the past performance values (step S106).

On the other hand, when the information processing apparatus 100 determines that the acquired performance value is not an outlier (step S103; No), it adds the acquired performance value to the past performance values (step S104). After adding the actual value, the information processing apparatus 100 updates the past actual value (step S106).

After updating the past performance values, the information processing apparatus 100 classifies the acquired performance values according to the classification conditions based on the metadata associated with the performance values (step S107). Specifically, the information processing apparatus 100 determines whether or not there is a classification condition cluster that satisfies the classification condition based on the metadata associated with the actual value (step S108).

When the information processing apparatus 100 determines that there is no classification condition cluster that satisfies the classification condition (step S108; No), the information processing apparatus 100 newly generates a classification condition cluster corresponding to the classification condition (step S110). After generating the classification condition clusters, the information processing apparatus 100 classifies the performance values into the generated classification condition clusters. After classifying the performance values into the generated classification condition cluster, the information processing apparatus 100 updates the past performance values belonging to the classification condition cluster (step S112).

On the other hand, when the information processing apparatus 100 determines that there is a classification condition cluster that satisfies the classification condition (step S108; Yes), the information processing apparatus 100 classifies the actual value into the determined classification condition cluster. Subsequently, the information processing apparatus 100 acquires past performance values belonging to the classification condition cluster (step S109).

When the information processing apparatus 100 acquires the past performance values, the information processing apparatus 100 stores the past performance values in time zones according to the time zone indicating the departure time (or arrival time) associated with the past performance values. Classify into corresponding time zone clusters (step S111). After classifying the past performance values into time clusters according to the departure time (or arrival time), the information processing apparatus 100 may update the information of the time cluster into which the past performance values are classified (step S112). .

Next, the procedure of information processing according to the embodiment will be described with reference to FIG. FIG. 6 is a flow chart showing an information processing procedure according to the embodiment.

In the example shown in FIG. 6, the information processing device 100 receives a prediction request from the user device 10 (step S201). Upon receiving the prediction request, the information processing apparatus 100 acquires a first predicted value regarding travel time predicted by the existing route search API (step S210).

Further, when receiving the prediction request, the information processing apparatus 100 responds to the second predicted value based on the metadata associated with the predicted value of the travel time to be predicted (hereinafter also referred to as the second predicted value). A classification condition is specified (step S202). After identifying the classification condition, the information processing apparatus 100 determines whether or not there is a classification condition cluster that satisfies the classification condition (step S203).

When the information processing apparatus 100 determines that the classification condition cluster does not exist (step S203; No), the information processing apparatus 100 sets the confidence value to "1" (step S208). Subsequently, when the information processing apparatus 100 determines the value of the reliability value to be "1", the information processing apparatus 100 calculates the first predicted value of the travel time predicted by the existing route search API based on the determined value of the reliability value "1". and the actual value are calculated (step S209). Subsequently, after calculating the ratio, the information processing apparatus 100 calculates a second predicted value from the first predicted value based on the calculated ratio (step S211). Subsequently, after calculating the second predicted value, the information processing device 100 transmits a predicted response to the user device 10 (step S212).

On the other hand, when the information processing apparatus 100 determines that a classification condition cluster exists (step S203; Yes), the second predicted value (step S204).

After identifying the time period cluster corresponding to the second predicted value, the information processing apparatus 100 acquires past performance values belonging to the identified time period cluster (step S205). The information processing apparatus 100 acquires the past performance values belonging to the time zone cluster into which the second predicted values are classified, and then calculates the corrected exponential moving average and the corrected exponential moving variance of the past performance values (step S206). After calculating the corrected exponential moving average and the corrected exponential moving variance, the information processing apparatus 100 then calculates a confidence value based on the calculated corrected exponential moving average and corrected exponential moving variance (step S207).

After calculating the confidence value, the information processing apparatus 100 determines a ratio of combining the first predicted value regarding the travel time predicted by the existing route search API and the corrected exponential moving average based on the confidence value (step S209). ). Subsequently, after calculating the ratio, the information processing apparatus 100 calculates a second predicted value from the first predicted value and the corrected exponential moving average based on the calculated ratio (step S211). Subsequently, after calculating the second predicted value, the information processing device 100 transmits a predicted response to the user device 10 (step S212).

[5. Modification]
In the above-described embodiment, an example in which the prediction target is travel time has been described, but the prediction target is not limited to travel time. Specifically, the determination unit 135 predicts the travel distance predicted by the existing route search API based on the reliability of the past performance value related to the travel distance traveled by the moving object from the departure point to the destination as a prediction target. Determine a ratio that combines the first predicted value for and the actual value for . The calculation unit 136 calculates a second predicted value regarding the movement distance from the first predicted value and the actual value based on the ratio determined by the determination unit 135 .

[6. effect〕
As described above, the information processing apparatus 100 according to the embodiment includes the determination unit 135 and the calculation unit 136. FIG. The determining unit 135 determines a ratio of combining the first predicted value regarding the prediction target predicted by the prediction process and the actual value based on the reliability of the past actual value regarding the prediction target. The calculation unit 136 calculates a second prediction value for the prediction target from the first prediction value and the actual value based on the ratio determined by the determination unit 135 . Specifically, the determination unit 135 increases the ratio of combining the actual value with respect to the first predicted value as the reliability of the actual value is higher.

Thereby, for example, when the reliability of the actual value is low, the information processing apparatus 100 can calculate the second predicted value by increasing the ratio of combining the first predicted value with the actual value. On the other hand, for example, when the reliability of the performance value is high, the information processing apparatus 100 can calculate the second prediction value by increasing the ratio of combining the performance value with respect to the first prediction value. In this way, the information processing apparatus 100 can calculate the second predicted value by combining the actual value with the first predicted value according to the reliability of the actual value, after considering the reliability of the actual value. . That is, the information processing apparatus 100 can utilize the actual value for prediction while covering the uncertainty of the actual value. Therefore, the information processing device 100 can improve the prediction accuracy of the prediction target.

Further, the determining unit 135 determines the ratio based on the reliability value indicating that the reliability of the actual value is higher as the actual value is newer.

Thereby, the information processing apparatus 100 can evaluate the reliability of the actual value, taking into account the newness of the actual value. In this way, the information processing apparatus 100 can evaluate the reliability of the actual values in consideration of the newness of the actual values. be able to. Therefore, the information processing device 100 can improve the prediction accuracy of the prediction target.

In addition, the determination unit 135 determines the ratio based on the reliability value indicating that the smaller the variance of the actual value, the higher the reliability of the actual value.

Accordingly, the information processing apparatus 100 can statistically evaluate the reliability of the actual values, taking into consideration the dispersion of the actual values. In this way, the information processing apparatus 100 can statistically evaluate the reliability of the actual values, so that the actual values can be used for prediction while covering the uncertainty of the actual values. Therefore, the information processing device 100 can improve the prediction accuracy of the prediction target.

In addition, the determining unit 135 determines the ratio based on the reliability value indicating that the smaller the exponential movement variance of the actual value, the higher the reliability of the actual value.

Thereby, the information processing apparatus 100 can statistically evaluate the reliability of the actual value, taking into account the newness of the actual value. In this way, the information processing apparatus 100 can statistically evaluate the reliability of the actual values in consideration of the newness of the actual values. can be used for Therefore, the information processing device 100 can improve the prediction accuracy of the prediction target.

In addition, the determining unit 135 determines the ratio based on the reliability value indicating that the reliability of the actual value increases as the number of samples of the actual value increases.

Thereby, the information processing apparatus 100 can statistically evaluate the reliability of the actual value after considering the number of samples of the actual value. In this way, the information processing apparatus 100 can statistically evaluate the reliability of the actual values, so that the actual values can be used for prediction while covering the uncertainty of the actual values. Therefore, the information processing device 100 can improve the prediction accuracy of the prediction target.

Further, the determination unit 135 determines a ratio of combining the first predicted value and the representative value representing the actual value. Calculation unit 136 calculates a second prediction value from the first prediction value and the representative value based on the ratio determined by determination unit 135 .

In this way, the information processing apparatus 100 can utilize the actual values for prediction while covering the uncertainty of the actual values by using the representative values that represent the actual values. Therefore, the information processing device 100 can improve the prediction accuracy of the prediction target.

Also, the representative value is the sample average of the actual values.

In this way, the information processing apparatus 100 can utilize the actual values for prediction while covering the uncertainty of the actual values by using the sample average of the actual values. Therefore, the information processing device 100 can improve the prediction accuracy of the prediction target.

Also, the representative value is a weighted average of the actual values calculated by increasing the weight as the actual value becomes newer.

As a result, the information processing apparatus 100 uses the weighted average of the actual values calculated by increasing the weight for newer actual values, thereby covering the uncertainty of the actual values and utilizing the actual values for prediction. can. Therefore, the information processing device 100 can improve the prediction accuracy of the prediction target.

Also, the representative value is the exponential moving average of the actual values.

As a result, the information processing apparatus 100 can utilize the actual values for prediction while covering the uncertainty of the actual values by using the exponential moving average of the actual values. Therefore, the information processing device 100 can improve the prediction accuracy of the prediction target.

Also, the target of prediction is the travel time required for the moving object to travel from the departure point to the arrival point. Specifically, the prediction target is the travel time when the moving object leaves the departure point at the designated departure time. Also, the target of prediction is the travel time required for the moving object to arrive at the destination at the specified arrival time.

Thereby, for example, when the reliability of the actual value is low, the information processing apparatus 100 can calculate the second predicted value by increasing the ratio of combining the first predicted value with the actual value. On the other hand, for example, when the reliability of the performance value is high, the information processing apparatus 100 can calculate the second prediction value by increasing the ratio of combining the performance value with respect to the first prediction value. In this way, the information processing apparatus 100 can calculate the second predicted value by combining the actual value with the first predicted value according to the reliability of the actual value, after considering the reliability of the actual value. . That is, the information processing apparatus 100 can utilize the actual value for prediction while covering the uncertainty of the actual value. Therefore, the information processing apparatus 100 can improve the prediction accuracy of the travel time.

Also, the prediction target is the travel time predicted by an existing route search API (Application Programming Interface).

Thereby, for example, when the reliability of the actual value is low, the information processing apparatus 100 can calculate the second predicted value by increasing the ratio of combining the first predicted value by the route search API with the actual value. can. On the other hand, for example, when the reliability of the actual value is high, the information processing apparatus 100 can calculate the second predicted value by increasing the ratio of combining the actual value with respect to the first predicted value by the route search API. . In this way, the information processing apparatus 100 can calculate the second predicted value by combining the actual value with the first predicted value according to the reliability of the actual value, after considering the reliability of the actual value. . That is, the information processing apparatus 100 can utilize the actual value for prediction while covering the uncertainty of the actual value. Therefore, the information processing apparatus 100 can improve the prediction accuracy of the travel time.

The information processing apparatus 100 further includes a classification section 132 . The classification unit 132 classifies the actual values into clusters based on the types of metadata associated with the actual values. The determination unit 135 determines the ratio based on the reliability of the actual values belonging to the clusters classified by the classification unit 132 . The calculating unit 136 calculates a second predicted value from the first predicted value and the actual value belonging to the cluster based on the ratio determined by the determining unit 135 .

In this way, the information processing apparatus 100 predicts a prediction value related to a prediction target for each type of metadata, for example, based on actual values classified according to the type of metadata. Thereby, the information processing apparatus 100 can further improve the prediction accuracy of the prediction target.

Also, the metadata is a movement route along which the moving object travels from the departure point to the arrival point.

In this way, the information processing apparatus 100 predicts the predicted value of the travel time for each travel route based on the actual values classified according to the travel route. Thereby, the information processing apparatus 100 can further improve the prediction accuracy of the travel time.

Metadata is also the type of mobile service provided by the mobile.

In this way, the information processing apparatus 100 predicts the predicted value of travel time for each type of mobile service based on the actual values classified according to the type of mobile service. Thereby, the information processing apparatus 100 can further improve the prediction accuracy of the travel time.

Also, the metadata is the departure time when the mobile body departs from the departure place or the arrival time when the mobile body arrives at the destination.

In this way, the information processing apparatus 100 predicts the travel time for each departure time or arrival time of the mobile object based on the performance values classified according to the departure time or arrival time of the mobile object. Predict value. Thereby, the information processing apparatus 100 can further improve the prediction accuracy of the travel time.

[7. Hardware configuration]
Also, the information processing apparatus 100 according to the above-described embodiments is implemented by a computer 1000 configured as shown in FIG. 7, for example. FIG. 7 is a hardware configuration diagram showing an example of a computer that implements the functions of the information processing apparatus 100. As shown in FIG. Computer 1000 includes CPU 1100 , RAM 1200 , ROM 1300 , HDD 1400 , communication interface (I/F) 1500 , input/output interface (I/F) 1600 and media interface (I/F) 1700 .

The CPU 1100 operates based on programs stored in the ROM 1300 or HDD 1400 and controls each section. The ROM 1300 stores a boot program executed by the CPU 1100 when the computer 1000 is started up, a program depending on the hardware of the computer 1000, and the like.

The HDD 1400 stores programs executed by the CPU 1100, data used by these programs, and the like. Communication interface 1500 receives data from another device via a predetermined communication network, sends the data to CPU 1100, and transmits data generated by CPU 1100 to another device via a predetermined communication network.

The CPU 1100 controls output devices such as displays and printers, and input devices such as keyboards and mice, through an input/output interface 1600 . CPU 1100 acquires data from an input device via input/output interface 1600 . CPU 1100 also outputs the generated data to an output device via input/output interface 1600 . Note that instead of the CPU 1100, an MPU (Micro Processing Unit) or a GPU (Graphics Processing Unit) may be used since a large amount of calculation power is required.

Media interface 1700 reads programs or data stored in recording medium 1800 and provides them to CPU 1100 via RAM 1200 . CPU 1100 loads such a program from recording medium 1800 onto RAM 1200 via media interface 1700, and executes the loaded program. The recording medium 1800 is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) or a PD (Phase change rewritable disc), a magneto-optical recording medium such as an MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. etc.

For example, when the computer 1000 functions as the information processing apparatus 100 , the CPU 1100 of the computer 1000 implements the functions of the control unit 130 by executing programs loaded on the RAM 1200 . CPU 1100 of computer 1000 reads these programs from recording medium 1800 and executes them, but as another example, these programs may be obtained from another device via a predetermined communication network.

As described above, some of the embodiments of the present application have been described in detail based on the drawings. It is possible to carry out the invention in other forms with modifications.

[8. others〕
Further, among the processes described in the above embodiments and modifications, all or part of the processes described as being performed automatically can be performed manually, or described as being performed manually. All or part of the processing can also be performed automatically by known methods. In addition, information including processing procedures, specific names, various data and parameters shown in the above documents and drawings can be arbitrarily changed unless otherwise specified. For example, the various information shown in each drawing is not limited to the illustrated information.

Also, each component of each device illustrated is functionally conceptual, and does not necessarily need to be physically configured as illustrated. In other words, the specific form of distribution and integration of each device is not limited to the one shown in the figure, and all or part of them can be functionally or physically distributed and integrated in arbitrary units according to various loads and usage conditions. Can be integrated and configured.

Also, the above-described embodiments and modifications can be appropriately combined within a range that does not contradict the processing content.

Also, the above-mentioned "section, module, unit" can be read as "means" or "circuit". For example, the determination unit can be read as determination means or a determination circuit.

1 information processing system 100 information processing device 110 communication unit 120 storage unit 130 control unit 131 acquisition unit 132 classification unit 133 reception unit 134 determination unit 135 determination unit 136 calculation unit 137 transmission unit

Claims (20)

a determination unit that determines a ratio of combining a first predicted value for the prediction target predicted by prediction processing and the actual value based on the reliability of past actual values for the prediction target;
a calculation unit that calculates a second predicted value related to the prediction target from the first predicted value and the actual value based on the ratio determined by the determination unit;
Information processing device. The decision unit
The higher the reliability of the actual value, the higher the ratio of combining the actual value than the first predicted value,
The information processing device according to claim 1 . The decision unit
determining the ratio based on a confidence value indicating that the newer the actual value, the higher the reliability of the actual value;
The information processing apparatus according to claim 1 or 2. The decision unit
determining the ratio based on a confidence value indicating that the smaller the variance of the actual value, the higher the reliability of the actual value;
The information processing apparatus according to any one of claims 1 to 3. The decision unit
determining the ratio based on a confidence value indicating that the smaller the exponential moving variance of the actual value, the more reliable the actual value;
The information processing apparatus according to any one of claims 1 to 4. The decision unit
Determining the ratio based on a confidence value indicating that the larger the number of samples of the actual value, the higher the reliability of the actual value.
The information processing apparatus according to any one of claims 1 to 5. The decision unit
Determining a ratio of combining the first predicted value and a representative value representing the actual value,
The calculation unit
calculating the second predicted value from the first predicted value and the representative value based on the ratio determined by the determination unit;
The information processing apparatus according to any one of claims 1 to 6. The representative value is a sample average of the actual values,
The information processing apparatus according to claim 7. The representative value is a weighted average of the actual values calculated by increasing the weight as the actual value is newer,
The information processing apparatus according to claim 7 or 8. The representative value is an exponential moving average of the actual values,
The information processing apparatus according to any one of claims 7 to 9. The prediction target is the travel time required for the moving object to travel from the departure point to the arrival point.
The information processing apparatus according to any one of claims 1 to 10. The prediction target is the travel time when the moving body departs from the departure point at a designated departure time.
The information processing device according to claim 11 . The prediction target is the travel time when the mobile body arrives at the arrival point at a specified arrival time.
The information processing device according to claim 11 . The prediction target is the travel time predicted by an existing route search API (Application Programming Interface),
The information processing apparatus according to any one of claims 11 to 13. further comprising a classification unit that classifies the actual values into clusters based on the type of metadata associated with the actual values;
The decision unit
Determining the ratio based on the reliability of the actual values belonging to the cluster classified by the classifying unit;
The calculation unit
calculating the second predicted value from the first predicted value and the actual value belonging to the cluster, based on the ratio determined by the determination unit;
The information processing apparatus according to any one of claims 1 to 14. The metadata is a movement route along which the mobile object moves from the departure point to the arrival point,
The information processing device according to claim 15 . the metadata is the type of mobile service provided by the mobile;
The information processing apparatus according to claim 15 or 16. The metadata is the departure time when the mobile body departs from the departure point or the arrival time when the mobile body arrives at the destination,
The information processing apparatus according to any one of claims 15-17. A computer-executed information processing method comprising:
a determination step of determining a ratio of combining a first predicted value for the prediction target predicted by prediction processing and the actual value based on the reliability of past actual values for the prediction target;
a calculating step of calculating a second predicted value related to the prediction target from the first predicted value and the actual value based on the ratio determined by the determining step;
Information processing method including. A determination procedure for determining a ratio of combining a first predicted value for the prediction target predicted by prediction processing and the actual value based on the reliability of past actual values for the prediction target;
a calculation procedure for calculating a second predicted value related to the prediction target from the first predicted value and the actual value based on the ratio determined by the determination procedure;
An information processing program that causes a computer to execute

Images