WO2021171591A1

WO2021171591A1 - Information processing method, information processing device, and program

Info

Publication number: WO2021171591A1
Application number: PCT/JP2020/008445
Authority: WO
Inventors: 裕造園生
Original assignee: 裕造園生
Priority date: 2020-02-28
Filing date: 2020-02-28
Publication date: 2021-09-02
Also published as: JPWO2021171591A1; JP7448992B2

Abstract

An information processing method for predicting an animal's potential athletic performance, the method comprising: acquiring parameters of an object animal; identifying, on the basis of the acquired parameters, animals that are ancestors of said object animal; and predicting, on the basis of evaluation values of the identified animals that are ancestors of said object animal, potential athletic performance of the object animal, wherein the evaluation values are an average and a variance, determined, for each of the animals that are ancestors of said object animal, from an athletic performance of the ancestral animal of said object animal and athletic performances of animals that are descendants of said ancestral animal.

Description

Information processing methods, information processing devices, and programs

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

So far, methods have been proposed for predicting the potential athletic performance of a subject, such as a thoroughbred racehorse, based on the biological parameters of the subject. Patent Document 1 discloses a method for predicting competition performance based on the presence of one or more single nucleotide polymorphisms.

Special Table 2012-50167A

According to one of the embodiments of the invention, the information processing method for predicting the potential competitive performance of an animal is based on an acquisition step of acquiring the parameters of the target animal and the parameters acquired in the acquisition step. Prediction to predict the potential competitive performance of the target animal based on the specific step of identifying the animal corresponding to the ancestor of the target animal and the competitive performance of the animal corresponding to the ancestor of the target animal specified in the specific step. The competition performance, including the steps, is the average and dispersion of at least one of the winnings and the number of appearances in the competition between the animal that is the ancestor of the animal of interest and its descendants.

The accompanying drawings are included in the specification and are used to form a part thereof, show embodiments of the present invention, and explain the principles of the present invention together with the description thereof.
The hardware block diagram of the information processing apparatus which concerns on 1st Embodiment. The software block diagram of the information processing apparatus which concerns on 1st Embodiment. The figure which shows an example of the data structure of the racehorse data which the information processing apparatus which concerns on 1st Embodiment has. The figure which shows an example of the data structure of the breeding horse data which the information processing apparatus which concerns on 1st Embodiment has. The flowchart which shows an example of the learning process executed by the information processing apparatus which concerns on 1st Embodiment. The flowchart which shows an example of the learning process executed by the information processing apparatus which concerns on 1st Embodiment. The flowchart which shows an example of the prediction processing executed by the information processing apparatus which concerns on 1st Embodiment. The flowchart which shows an example of the complementary processing executed by the information processing apparatus which concerns on 2nd Embodiment. The figure which shows the blood relation of a racehorse and an example of the competition performance of a racehorse. The figure which shows an example of the explanatory variable of a machine learning model. The figure which shows an example of the objective variable of a machine learning model. The figure which shows an example of the machine learning model.

Hereinafter, embodiments will be described in detail with reference to the attached drawings. The following embodiments do not limit the invention according to the claims, and not all combinations of features described in the embodiments are essential to the invention. Two or more of the plurality of features described in the embodiments may be arbitrarily combined. In addition, the same or similar configuration will be given the same reference number, and duplicated explanations will be omitted.

<First Embodiment>
Hereinafter, the information processing device according to the present embodiment will be described. The information processing device implements an information processing method that predicts the potential competitive performance of the animal. In the present embodiment, the case where the animal to be predicted is a horse, particularly a young piece, will be described. Specifically, the potential competitive performance of the young piece is predicted based on the competitive performance of horses having a pedigree relationship with the young piece (referred to as related horses in the present specification) such as the father horse and the mother horse of the young piece. In the present embodiment, the potential competition performance will be described as the number of appearances in the horse racing race and the prize money won. In this embodiment, a horse is used as an example for predicting the potential competition performance, but in order to predict the potential competition performance of any animal such as a cow, a dog, or a chicken. The present embodiment can also be applied to. In this case, the competition performance can be selected according to the competition in which the animal participates.

FIG. 1 is a hardware configuration diagram of the information processing device 1 according to the present embodiment. The information processing device 1 has a processor 101, a memory 102, a storage 103, and a network I / F (interface) 104. These components can communicate with each other via the bus 105.

The processor 101 is a system control unit and controls the entire information processing device 1. The processor 101 includes at least one of a CPU, a GPU (Graphics Processing Unit), and a microcontroller that executes a program (instruction group) or an operating system (OS) stored in the storage 103.

The memory 102 is composed of a DRAM (Dynamic Random Access Memory) used as a work memory of the processor 101, a SRAM (Static Random Access Memory), and the like.

The storage 103 includes an HDD (Hard Disk Drive) and an SSD (Solid State Drive) that store at least one of a program executed by the processor 101, an OS, and a data table and setting values used by the program. Further, the storage 103 may include a PROM (Programmable Read Only Memory), an EPROM (Erasable ROM), and an EEPROM (Electrically EPROM).

The network I / F 104 is a communication unit for connecting to an external device (not shown) and executing data communication, and includes at least one of a wired network I / F and a wireless network I / F. The wireless network I / F can perform wireless communication compliant with Wi-Fi (Wireless Fidelity) (registered trademark), Bluetooth (registered trademark), and the like.

In one example, the information processing device 1 may include a user interface (UI) including at least one of an input unit such as a keyboard and a touch panel that accepts user operations, and an output unit such as a display or a speaker.

The information processing device 1 can be realized as an arbitrary information processing terminal such as a personal computer, a mobile phone, a smartphone, a PDA, a tablet terminal, a wearable terminal, or a performance prediction device.

FIG. 2 is a diagram showing a software configuration of the information processing device 1 according to the present embodiment. The software module of the information processing apparatus 1 shown in FIG. 2 is realized by the processor 101 executing a program stored in the storage 103. The software module of the information processing device 1 includes a racehorse data management unit 201, a breeding horse data management unit 202, a young piece parameter acquisition unit 203, a relative horse identification unit 204, a learning unit 205, a prediction unit 206, and a providing unit 207. ..

The racehorse data management unit 201 manages data such as past racehorse results and pedigree in the storage 103. In one example, the racehorse data management unit 201 may update the data stored in the storage 103 when the racehorse data is received via the network I / F 104. In addition, the racehorse data management unit 201 may periodically access a database on the Internet to acquire new racehorse data.

The breeding horse data management unit 202 manages the data of horses bred on the ranch to produce foals in the storage 103. In one example, the breeding horse data management unit 202 may update the data stored in the storage 103 when the breeding horse data is newly received via the network I / F 104. In addition, the breeding horse data management unit 202 may periodically access a database on the Internet to acquire new breeding horse data.

The young piece parameter acquisition unit 203 acquires the parameters of the young piece for which the potential competition performance should be predicted. In one example, the young piece parameter acquisition unit 203 may acquire the young piece parameter from an external device via the network I / F 104. Further, in one example, the information processing device 1 may include a user interface (UI) (not shown), and may acquire parameters of a young piece whose potential competition performance should be predicted via the UI. The parameters of the young piece include the sex of the young piece, information that can identify the parents, and the like. The details of the parameters of the young piece will be described later.

The blood-related horse identification unit 204 identifies a horse that has a blood relationship with the young horse whose potential competition performance should be predicted, based on the parameters acquired by the young piece parameter acquisition unit 203. In one example, the blood-related horse identification part 204 is the father horse, mother horse, father horse (father of the father horse, that is, the paternal grandfather of the child piece), parent horse (father of the mother horse, that is, the mother of the child piece). Identify the grandfather), the mother-father horse (the mother of the father horse, that is, the paternal grandmother of the young piece), and the mother-mother horse (the mother of the mother horse, that is, the maternal grandmother of the young piece). Further, in one example, the blood-related horse identification unit 204 identifies horses corresponding to younger brothers, cousins, and cousins. Further, in one example, the blood relative horse identification unit 204 identifies a father horse, a mother horse, a father horse, a parent horse, a mother father horse, and a mother mother horse, and horses corresponding to these descendants.

The learning unit 205 adjusts the parameters of the machine learning model held by the prediction unit 206 based on the racehorse data and the breeding horse data managed by the racehorse data management unit 201 and the breeding horse data management unit 202. For example, the learning unit 205 can execute a learning process for adjusting parameters of a machine learning model by using a known machine learning algorithm such as a random forest, a neural network, or a perceptron. A detailed description of the learning process executed by the learning unit 205 will be described later with reference to FIGS. 5A and 5B.

The prediction unit 206 generates input data to be input to the machine learning model based on the parameters of the young piece acquired by the young piece parameter acquisition unit 203. Further, the prediction unit 206 inputs the generated input data to the machine learning model, and obtains a predicted value of the competition performance as an output. A detailed description of the prediction unit 206 will be described later with reference to FIG.

The providing unit 207 provides the user or an external device with the potential competition performance of the young piece predicted by the predicting unit 206. For example, the providing unit 207 may transmit a predicted value of the competition performance of the young piece to an external device via the network I / F 104. Alternatively, when the information processing device 1 is provided with an output device such as a display, a speaker, or a printer, at least one of these output devices may present the predicted value of the competition performance of the young piece to the user of the information processing device 1. good.

Next, an example of racehorse data managed by the racehorse data management unit 201 will be described with reference to FIG.

The data managed by the racehorse data management unit 201 includes identification information 301, horse name 302, date of birth 303, gender 304, breed 305, coat color 306, performance information 307, and relative horse identification information 308.

The identification information 301 is an identifier of the racehorse. Horse name 302 is the name given to the racehorse. The date of birth 303 is information corresponding to the date of birth of the racehorse. Gender 304 is information regarding the gender of the racehorse. The breed 305 is information indicating a racehorse breed such as "Thoroughbred", "Thoroughbred", "Semi-Thoroughbred", and "Anglo-Arab". The coat color 306 is information indicating the coat color of a race horse such as "brown hair", "liver brown hair", "bay", "black bay", and "blue bay".

The performance information 307 is information indicating the past performance information of the racehorse, and in the present embodiment, the performance information 307 includes the number of races information 309 and the prize information 310. The number of races information 309 is information indicating how many times the racehorse has participated in the race so far. In one example, the run count information 309 may include information about the run count for each type of race. The prize information 310 is information indicating the amount of prize money (earned prize money) of the racehorse so far.

In one example, the performance information 307 includes the order of arrival of the racehorse, the type of race in which the racehorse participated (turf, dirt, obstacle, clockwise, counterclockwise), the order of arrival of the race in which the racehorse participated, or It is possible to further include the number of times each arrival order is reached. That is, the performance information 307 may include any information as long as it is an index indicating the competition performance of the racehorse.

Blood-related horse identification information 308 indicates identification information of horses that are related to the racehorse. The blood-related horse identification information 308 according to the present embodiment includes father-horse identification information 311 and mother-horse identification information 312, father-father horse identification information 313, parent-mother horse identification information 314, mother-father horse identification information 315, and mother-mother horse identification information. 316 is included.

The father horse identification information 311 is identification information of the father horse, for example, the father horse identification information 311 included in the racehorse data of a certain racehorse and the identification information 301 included in the racehorse data of the father horse of the racehorse. May be associated with each other. Further, the father horse identification information 311 included in the racehorse data of a certain racehorse may be associated with the identification information included in the breeding horse data of the father horse of the racehorse. The mare identification information 312 indicates the identification information of the mother horse, and the father-father horse identification information 313 indicates the identification information of the father-father horse (father horse of the father horse). Similarly, the parent-mother horse identification information 314 indicates the parent-mother horse, the mother-father horse identification information 315 indicates the mother-father horse, and the mother-mother horse identification information 316 indicates the identification information of the mother-mother horse. In the present embodiment, the information regarding the blood relationship of the race horse includes information that can identify the ancestral horses of the father horse, the mother horse, the father horse, the parent horse, the mother father horse, and the mother mother horse up to two generations ago. Of course, more than three generations (father-father horse, father-parent horse, parent-mother-father horse, parent-mother-mother horse, mother-father-father horse, mother-parent-mother horse, mother-mother-father horse, and mother-mother-mother horse) Horses of ancestors may be considered.

In the present embodiment, in one example, the blood relative horse identification information 308 may include information that can identify offspring such as a foal and a grandchild. Even in this case, the offspring to be specified may be limited by the perspective of the blood relationship. For example, it may be limited to foal horses, grandchild horses, great-grandson horses, etc. Horses may be identified.

Further, in one example, the racehorse data management unit 201 has producer information indicating information on producers such as the ranch where the racehorse was born and the owner of the ranch, and trainer information indicating information on the trainer in charge of training the racehorse. , Data including owner information indicating information on racehorse owners may be managed.

In this embodiment, the potential competition performance of the young piece will be described as predicting the number of races and prizes, but if the purpose is to predict other parameters, the other parameters will race. It may be included in the horse data management unit 201. For example, when the purpose is to predict the number of winning races of a racehorse, the racehorse data management unit 201 may manage the racehorse data including the total number of first-place finishes, or the number of winning races by distance. If the purpose is to predict, the number of first arrivals for each distance may be managed.

Next, the breeding horse data managed by the breeding horse data management unit 202 will be described with reference to FIG.

Breeding horse data includes identification information 401, date of birth 402, gender 403, breed 404, coat color 405, father horse breeding registration information 406, and mare breeding registration information 407.

The identification information 401 is the identification information of the breeding horse, and in one example, the identification information 401 of a certain breeding horse is the identification information 301 included in the racehorse data when the breeding horse is active as a racehorse. It may be associated. Since the date of birth 402, the sex 403, the breed 404, and the coat color 405 are the same as the date of birth 303, the sex 304, the breed 305, and the coat color 306 of the racehorse, the description thereof will be omitted.

The father horse breeding registration information 406 is the identification information of the mother horse of the breeding horse, and the mother horse breeding registration information 407 is the identification information of the mother horse of the breeding horse. The paternal horse breeding registration information 406 and the mare breeding registration information 407 may be associated with the identification information 401 of the paternal horse and mare breeding horse data.

<Learning process>
Next, the process of the learning unit 205 that learns the machine learning model of the prediction unit 206 will be described with reference to FIGS. 5A and 5B. The learning process of FIGS. 5A and 5B may be executed when an execution request for the learning process is received via the UI or the network I / F 104, or may be executed by the information processing apparatus 1 at predetermined time intervals. good. The learning process of FIGS. 5A and 5B is realized by the processor 101 executing a program stored in the storage 103 using the memory 102 as a workspace.

First, in S501, the processor 101 acquires the racehorse data managed by the racehorse data management unit 201. Here, if the race horse data managed by the race horse data management unit 201 includes duplicate records or data included in the race horse data that is not used for learning, these records and data are deleted. Data shaping processing such as Further, in S501, the racehorse may be learned for a limited period of time, for example, by limiting the date of birth 402 of the racehorse data and acquiring the racehorse data. As a result, it is possible to prevent the competition performance of the racehorse at a time when the number of races held and the prize amount are significantly different from excessively affecting the learning process.

Subsequently, in S502, the processor 101 uses the racehorse data management unit 201 to manage the racehorse data management unit 201 for the racehorse identification information 308, and the breeding horse data management unit 202 to manage the breeding horse data. It is specified based on at least one of the paternal horse breeding registration information 406 and the mother horse breeding registration information 407. For example, if the racehorse identification information 308 of the racehorse data corresponding to a certain racehorse is acquired, the ancestor horse of the racehorse can be identified. Further, by searching whether the identification information 301 of the racehorse data corresponding to a certain racehorse is included in the blood relative horse identification information 308 of the other racehorse data, the horse of the descendant of the racehorse is specified. be able to.

Subsequently, in S503, the processor 101 calculates the average value and dispersion of the racehorse run count information 309 and the prize information 310 of the racehorses of the respective descendants of the blood relative horses for each of the blood relative horses of each race horse, and the blood relative horses. Get the processed competition performance of. The processed competition performance of a relative horse is treated as an evaluation value of the relative horse based on the competition performance of the relative horse and its descendants.

Subsequently, in S503, the competition performance of the pedigree horse is processed. Here, the details of S503 will be described with reference to FIG. 5B. First, in S5031, the processor 101 identifies the offspring of one of the relatives. Subsequently, in S5032, the processor acquires the competitive performance of any of the identified relatives and their offspring. Next, in S5033, the processor 101 calculates the average and variance of the competitive performance of the kin and the offspring of the kin. Subsequently, in S5034, the processor 101 stores the average and variance of the competition performance of the relative horse and the descendants of the relative horse as the processed competition performance of the relative horse. This is repeated for all related horses (S5035). In this embodiment, the average and variance of the competition performance between the kinship horse and the offspring of the kinship horse are calculated, but weighting may be performed according to the perspective of the kinship relationship with the kinship horse. good.

Here, the processed competition performance will be described with reference to the figure showing the pedigree of the race horses in FIG. 8 and the information on the number of races and the prize money for each. FIG. 8 is a pedigree diagram showing racehorses descendants of horses A801, B802, and C803. In FIG. 8, when horse A801 is described as an example, the upper row 8011 shows the identification information, the lower row left 8012 shows the gender, the lower row center 8013 shows the prize money, and the lower row right 8014 shows the number of runs. The other horses are also illustrated in the same way. For example, the horse D804 has run 24 times and the prize money is 55M (55000000) yen.

In FIG. 8, horse A801 and horse B802 give birth to horse D804 and horse E805. It can also be seen that horse E805 gave birth to a foal containing horse I809 with horse F806. It can also be seen that horse B802 and horse C803 gave birth to horse G807 and horse H808.

For example, regarding horse D804, in S502, a horse having a pedigree relationship of horse B802 as a father horse and horse A801 as a mare is specified. In another example, horse A801, horse B802, horse E805, horse G807, horse H808, and horse I809 may be identified as pedigree horses in S502.

Here, a case of calculating the processed competition performance of horse A801, which is a relative horse of horse D804, will be described. The descendants of horse A801 are horse D804, horse E805, and horse I809. Therefore, the average prize amount of the horse A801 and the descendant horse of the horse A801 is 48.75, and the variance is 279.69. Further, the average number of run times of the horses A801 and the descendants of the horse A801 is 17.75, and the variance is 35.19. These values are stored in memory as the competition performance of horse A801 (referred to as processed competition performance). Similarly, for each pedigree horse, the processed competition performance is calculated. By calculating in this way, it is possible to calculate the competition performance of a blood-related horse in consideration of the blood-related relationship.

Subsequently, in S504, the processor 101 sets the age of the mare at the time of giving birth to the racehorse birth date 303, the birth date 402 of the mare breeding horse data, or the racehorse data of the mare. It is calculated and obtained from the date of birth 303 of. If the racehorse database stores the age of the mare at the time of delivery of the racehorse, the age of the mare at the time of delivery may be obtained without calculation.

Subsequently, in S505, the processor 101 generates a set of explanatory variables and objective variables. The objective variable points to the competition performance that you want to estimate as the potential competition performance of the young piece. For example, in S505, the processor 101 extracts the run number information 309 and the prize information 310 as objective variables for each racehorse. Further, as an explanatory variable that affects the objective variable, the processor 101 includes the racehorse data managed by the racehorse data management unit 201 for each racehorse, such as gender 304, breed 305, hair color 306, and the time when the racehorse is delivered. Extract the age of the mother horse and the processed race performance of the relative horse in.

Here, the explanatory variables and the objective variables will be described with reference to FIGS. 9A and 9B.

FIG. 9A is a diagram showing explanatory variables. The identification information 901 is the identification information 901 of the racehorse, and corresponds to the identification information 301 in FIG. 3 in one example. Gender 902 and coat color 903 are data indicating the gender and route of the racehorse, respectively, and correspond to gender 304 and coat color 306 in FIG. 3 in one example. The mare birth age 904 is the age of the mare at the time of giving birth to the racehorse, and is the age of the mare at the time of delivery acquired in S504.

The average and variance 905 of the prizes of the father horse and the average and variance 906 of the number of runs of the father horse are the processed competition performances of the father horse calculated in S503. In the example of FIG. 8, when the racehorse is horse D804, it is the processed competition performance of horse B802. The average and variance (907-916) of mare-to-mare prizes and run counts is the processed competition performance of mare-to-mare.

FIG. 9B shows the prize money 951 and the number of runs 952 corresponding to the explanatory variables. The prize money 951 corresponds to the prize money information 310 and the number of runs information 309 in FIG. In the example of FIG. 8, when the racehorse is horse D804, it is the prize money and the number of runs (55M, 24) of horse D804.

When the racehorse data of the racehorse and the horse related to the racehorse is stored in the racehorse data management unit 201, the processor 101 can acquire both the explanatory variable and the objective variable. .. Therefore, the parameters of the machine learning model can be tuned using both the explanatory variables and the objective variables.

In S505, the processor 101 may extract at least a part of the set of explanatory variables and objective variables and standardize the objective variables and explanatory variables of the extracted set. For example, when standardizing the number of runs in a processed competition performance, the scale may be changed so that the average number of runs of all racehorses is 0 and the variance is 1. For items for which a value such as coat color 405 cannot be obtained, standardization processing may not be performed.

Subsequently, in S506, the processor 101 adjusts the machine learning model so that when the explanatory variable is input to the machine learning model of the prediction unit 206, the objective variable corresponding to the explanatory variable is output. For example, in S506, the branch condition of the node of the random forest is set. In another example, a known technique such as grid search can be used for parameter tuning. Further, in S507, the processor 101 stores a set of objective variables and explanatory variables used for learning, and also stores a random forest model after learning. When learning S506, a known technique for learning a machine learning model may be applied, such as randomly selecting explanatory variables and adjusting a decision tree.

As described above, in the processing of FIGS. 5A and 5B, for a racehorse that already has a competition performance, the machine learning model of the prediction unit 206 is learned based on the competition performance of the horse that is related to the racehorse. I do.

<Prediction processing>
Next, with reference to FIG. 6, the prediction processing of the potential competition performance of the young piece using the machine learning model, which is executed by the prediction unit 206, will be described. The process of FIG. 6 is realized by the processor 101 executing a program stored in the storage 103 using the memory 102 as a workspace.

First, in S601, the processor 101 acquires the parameters of the young piece including the identification information of the father horse, the identification information of the mother horse, the sex, the coat color, and the birth age of the mother horse. In one example, the identification information of the father horse acquired in S601 may be associated with the identification information 401 managed by the breeding horse data management unit 202. In one example, the information processing apparatus 1 includes a user I / F (not shown) and may accept at least one of the parameters specified via the user I / F. In another example, the information processing device 1 may operate as a web server and acquire at least one of the parameters from another communication device via the network I / F 104. Alternatively, for example, the processor 101 that has received the identifier of the young piece or the horse name may access an external database and search for the identification information of the father horse and the mother horse.

Subsequently, the processor 101 advances the process to S602 and identifies a horse that has a blood relationship with the young piece for which the parameter has been acquired. In the present embodiment, in S602, the father horse and the mother horse are identified based on the identification information of the father horse of the young piece and the identification information of the mother horse.

Subsequently, the processor 101 advances the processing to S603 and calculates the processed competition performance of the father horse and the mother horse. In addition, the machine learning model of the prediction unit 206 and the explanatory variables used during the learning of the father horse and the mother horse explained with reference to FIG. 9A are acquired, and a part of the objective variables is extracted. For example, the average and variance of the father's horse's prize included in the explanatory variable containing the identification information 901 corresponding to the father's horse's identification information 905 can be used as the average and variance of the father's horse's prize for the young piece. Similarly, the average and variance of the mare prizes contained in the explanatory variables containing the identification information 901 corresponding to the identification information of the father horse can be used as the average and variance of the prize money of the mother horse for the young piece. In this way, at least some of the explanatory variables corresponding to the paternal horse identifier are used in predicting the potential competitive performance of the paternal horse's child, reducing the amount of computation in performing the prediction process. be able to. Similarly, the explanatory variables corresponding to the mare identifiers are used to predict the potential athletic performance of the mare's offspring. In S603, the processed competition performance of the father horse and the mother horse was explained as being calculated, but in one example, the processed competition of the father horse and the mother horse is associated with the explanatory variables shown in FIG. 9A during the learning process. By calculating and storing the performance, the prediction process can be executed even faster.

Subsequently, the processor 101 advances the processing to S604, and inputs the acquired parameters of the young piece and the processed competition performance of the blood relative horse of the young piece into the machine learning model. Then, as an output, the expected value of the competition performance is acquired as the potential competition performance of the young piece.

Here, the relationship between input and output will be described with reference to FIG.

FIG. 10 is a diagram showing a machine learning model 1000 that takes the explanatory variable shown in FIG. 9A as an input and outputs the objective variable shown in FIG. 9B. In FIG. 10, the machine learning model 1000 is described as being a random forest, but other machine learning techniques can be applied.

The machine learning model 1000 includes a plurality of decision trees 1 to n. At least a part of the explanatory variables shown in FIG. 9A is input to each of the decision trees 1 to n. Explanatory variables for different items may be input to different decision trees. Next, each of the decision trees 1 to n is subjected to regression analysis based on the input explanatory variables. Then, the average module 1001 averages the outputs from the respective decision trees 1 to n, and outputs the predicted values of the prize money 951 of the young piece and the number of runs 952. In this way, in S604, the prediction unit 206 inputs the explanatory variables to the model adjusted by machine learning and outputs the objective variables.

In the example of FIG. 10, the predicted values of the prize money 951 and the number of runs 952 have been described as one value each, but in one example, the predicted value of the prize money 951 is in the range of values (50M to 65M). It may be the average value and the variance of the prize money predicted by each decision tree 1 to n.

Subsequently, the processor 101 advances the processing to S605 and provides the potential competition performance of the acquired young piece. In S605, for example, when the information processing device 1 operates as a web server, the information processing device 1 may transmit a signal indicating an objective variable to another communication device via the network I / F 104. When the information processing device 1 includes a display unit (not shown), the information processing device 1 may display a screen showing the objective variable on the display unit. Thereby, it is possible to provide the user of the other information processing device or the information processing device 1 with the potential competition performance of the young piece corresponding to the input data.

As described above, the information processing device according to the present embodiment predicts the potential competition performance of the young piece based on the competitive performance of the blood-related horse having a blood relationship with the young piece. As a result, it is not necessary to investigate the genetic information of the young piece, so that the potential competitive performance of the young piece can be predicted quickly and easily.

In addition, the information processing device according to the present embodiment predicts the potential competitive performance of the young piece based on the age of the mare when the young piece is born. This makes it possible to more accurately predict the potential competitive performance of young pieces.

Further, in the information processing apparatus according to the present embodiment, the competition performance of the racehorse having a blood relationship with the young piece is processed based on the competition performance of the descendants of the racehorse having the blood relationship, and the potential competition performance of the young piece is processed. Used to predict. As a result, the pedigree is expected to have high competitive performance, but the potential competitive performance of young pieces is excessively low due to horses that have exceptionally low competitive performance or horses that have not been able to perform sufficiently due to injuries. It can be prevented from being estimated.

Further, the information processing apparatus according to the present embodiment predicts the potential competitive performance of the young piece based on the competitive performance of the father horse, the mother horse, the father father horse, the parent horse, the mother father horse, and the mother mother horse as blood relative horses. .. This makes it possible to analyze in detail the influence of blood relatives on the competitive performance of young pieces, and to predict the potential competitive performance of young pieces more accurately.

<Second Embodiment>
In the first embodiment, the racehorse data management unit 201 and the breeding horse data management that adjust the parameters of the machine learning model based on the competition performance of the racehorse that already has the competition performance and the horse that is related to the racehorse. The information processing device that predicts the potential competition performance of the young piece based on the data possessed by the part 202 has been described. In one example, the lack of racehorse or breeding horse data may make it impossible to obtain at least some of the competitive performance of horses that are related to the young piece. In the second embodiment, when the competition performance of one related horse cannot be obtained, the process of estimating and complementing the competition performance of the one related horse will be described. The description of the configuration, function, or processing similar to that of the first embodiment will be omitted.

FIG. 7 is a flowchart showing an example of the prediction process according to the second embodiment. S5031 and S5032 are the same as those in the first embodiment.

In S701, the processor 101 determines whether or not there is a horse that cannot acquire the competition performance among the horses that are related to the competition horse used for learning. If there is no horse whose competition performance cannot be obtained (No in S701), the processor 101 advances the process to S5033. Since the following is the same as that of the first embodiment, the description thereof will be omitted.

If there is a horse whose competition performance cannot be obtained in S701 (Yes in S701), the processor 101 advances the process to S702 and identifies a horse having a blood relationship with the horse whose competition performance cannot be obtained. The processing of S702 can be performed in the same manner as in S502 and S5031 by using the identification information of horses whose competition performance cannot be acquired. Subsequently, the processor 101 advances the process to S703 and acquires the competition performance of the horse having a blood relationship with the horse whose competition performance cannot be acquired. Subsequently, the processor 101 advances the process to S704, and estimates the competition performance of the horse whose competition performance cannot be acquired based on the competition performance of the horse which is related to the horse whose competition performance cannot be acquired. For example, the competition performance cannot be acquired by inputting the competition performance of a horse whose competition performance cannot be acquired as an objective variable and the competition performance of a horse having a blood relationship with a horse whose competition performance cannot be acquired as an explanatory variable into a machine learning model. You can get the objective variable of the horse. In this case, the machine learning model used in S704 may be a machine learning model adjusted based on a horse whose competitive performance can be acquired and a horse which is related to the horse. That is, the machine learning model may be adjusted based on the competition performance of a horse whose competition performance can be acquired once and the horse which is related to the horse, and then the competition performance of the horse whose competition performance cannot be acquired may be estimated. As a result, even when the competition performance does not exist in some horses, the competition performance can be estimated and used for the learning process, so that the training data of the machine learning model can be increased.

In the present embodiment, when there is a horse whose competition performance cannot be acquired at the time of learning, a case where the learning process is performed by complementing the competition performance of the horse whose competition performance cannot be acquired has been described. In one example, the same process can be performed when there is a horse whose competition performance cannot be obtained at the time of prediction. As a result, even if the competitive performance of a horse that is related to the young piece cannot be obtained, the competitive performance can be estimated, so that the potential competitive performance of the young piece is predicted by the horse that does not have the competitive performance. You can reduce the possibility that you will not be able to do it.

<Other Embodiments>
In the present embodiment, an application for predicting the potential competition performance of a young piece has been described as an example, but the present embodiment can also be applied to other applications.

For example, the configuration of the information processing apparatus according to the present embodiment may be changed for the purpose of identifying a father horse for producing a young piece having good potential competition performance for a specific mother horse. .. In this case, for a specific mother horse, a candidate father horse (and a father horse, a parent horse, a father father horse, a father parent horse, a parent parent horse, and a parent mother horse corresponding to the father horse) are selected. By inputting and assuming the birth age of the mother horse, it is possible to predict the potential competitive performance of the non-existent young piece. This makes it possible to determine which mare and father horse should produce a young piece with high potential competitive performance.

Further, the configuration of the information processing device according to the present embodiment may be changed for the purpose of specifying an appropriate age for the mare to produce a young piece. In this case, for a specific mare and a specific mare, the birth age of the mother horse is virtually set and predicted multiple times, so that the young piece born at which birth age has a high potential competitive performance. It is possible to determine the timing of giving birth to the mare.

Further, a program that realizes one or more functions of each of the above-described embodiments is supplied to the system or device via a network or storage medium, and one or more processors in the computer of the system or device reads and executes the program. It can also be realized by processing.

The invention is not limited to the above embodiment, and various modifications and changes can be made within the scope of the gist of the invention.

Claims

An information processing method that predicts the potential competitive performance of animals.
The acquisition process to acquire the parameters of the target animal,
A specific step of identifying an animal that is an ancestor of the target animal based on the parameters acquired in the acquisition step, and
Including a prediction step of predicting the potential athletic performance of the target animal based on the evaluation value of the animal corresponding to the ancestor of the target animal specified in the specific step.
The evaluation value is information obtained for each of the animals corresponding to the ancestors of the target animal, which is the average and variance of the competition performance between the animals corresponding to the ancestors of the target animal and the animals of its descendants. Processing method.
The information processing method according to claim 1, wherein the competition performance includes at least one of the prize money won in the competition and the number of appearances.
In the prediction step, the potential competition performance of the target animal is a model obtained by machine learning, and the potential competition of the target animal is input by inputting the evaluation value of the animal corresponding to the ancestor of the target animal. The information processing method according to claim 1 or 2, wherein the predicted value of performance is predicted based on the model.
The information processing method according to claim 3, further comprising a learning step of learning the model based on the evaluation values and the competition performance of a plurality of animals and animals corresponding to the ancestors of the plurality of animals.
The target animal is a horse,
The information processing method according to any one of claims 1 to 4, wherein the competition performance is at least one of a prize money won and a number of races in a horse racing race.
The animal, which is the ancestor of the target animal, is a mother horse, a father horse, a parent horse, a father horse, a mother mother horse, a mother father horse, a father father horse, a father parent horse, a parent mother horse, The information processing method according to claim 5, further comprising at least one of a parent-mother horse, a mother-father-father horse, a mother-parent-mother horse, a mother-mother-father horse, and a mother-mother-mother horse.
The information processing method according to claim 5 or 6, wherein the parameter includes identification information of a father horse and a mare of the target animal.
The parameters include at least one of the age of the mare that gave birth to the animal of interest, the coat color of the animal of interest, and the sex of the animal of interest.
In the prediction step, the potential competitive performance of the animal of interest is further based on at least one of the age of the mare that gave birth to the animal of interest, the coat color of the animal of interest, and the sex of the animal of interest. The information processing method according to any one of claims 5 to 7, wherein the information processing method is predicted.
Further provided, a providing step of providing the potential competition performance of the target animal predicted in the prediction step to an external device via a communication unit is provided.
The information processing method according to any one of claims 1 to 8, wherein in the acquisition step, the parameter is acquired from the external device via the communication unit.
The display unit further comprises a providing step of providing the user with the potential competitive performance of the target animal predicted in the prediction step.
The information processing method according to any one of claims 1 to 8, wherein in the acquisition step, the parameter is acquired via a user interface.
When the evaluation value of the animal corresponding to the ancestor of the target animal is missing, the missing evaluation value of the animal corresponding to the ancestor of the target animal is related to the animal corresponding to the ancestor of the target animal. The information processing method according to any one of claims 1 to 10, further comprising a complementing step of complementing based on the evaluation value of the animal in the above.
An information processing method that predicts the potential competitive performance of horses.
An acquisition process for acquiring parameters including the identification information of the target horse's father and mare, the sex and coat color of the target horse, and the age of the mother horse at the time of giving birth to the target horse.
An output process that outputs a potential competition performance predicted based on the parameters acquired in the acquisition process, including the number of times the target horse has run and the prize money won.
An information processing method characterized by including.
An information processing device that predicts the potential competitive performance of animals.
With one or more processors
Memory and
A group of instructions for causing the information processing apparatus to execute the information processing method according to any one of claims 1 to 12, when the program is stored in the memory and is executed by one or more processors. An information processing device including a program including.
An information processing device that predicts the potential competitive performance of animals.
An acquisition method for acquiring the parameters of the target animal,
A specific means for identifying an animal that is an ancestor of the target animal based on the parameters acquired by the acquisition means, and
A predictive means for predicting the potential athletic performance of the target animal based on the evaluation value of the animal which is the ancestor of the target animal specified by the specific means.
Including
The evaluation value is information obtained for each of the animals corresponding to the ancestors of the target animal, which is the average and variance of the competition performance between the animals corresponding to the ancestors of the target animal and the animals of its descendants. Processing equipment.
A program for causing a computer to execute the information processing method according to any one of claims 1 to 12.