WO2023175809A1 - Situation prediction system, situation prediction method, and recording medium - Google Patents

Situation prediction system, situation prediction method, and recording medium Download PDF

Info

Publication number
WO2023175809A1
WO2023175809A1 PCT/JP2022/012105 JP2022012105W WO2023175809A1 WO 2023175809 A1 WO2023175809 A1 WO 2023175809A1 JP 2022012105 W JP2022012105 W JP 2022012105W WO 2023175809 A1 WO2023175809 A1 WO 2023175809A1
Authority
WO
WIPO (PCT)
Prior art keywords
racehorse
prediction
information
state
future
Prior art date
Application number
PCT/JP2022/012105
Other languages
French (fr)
Japanese (ja)
Inventor
文秀 瀧本
宗裕 橋本
康彦 吉田
結佳 遠藤
Original Assignee
日本電気株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 日本電気株式会社 filed Critical 日本電気株式会社
Priority to PCT/JP2022/012105 priority Critical patent/WO2023175809A1/en
Publication of WO2023175809A1 publication Critical patent/WO2023175809A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/34—Betting or bookmaking, e.g. Internet betting

Definitions

  • the present invention relates to a state prediction system and the like.
  • racehorses In horse racing, which is a publicly managed competition, auctions of racehorses are held. Since racehorses are sold at auctions while they are young, buyers need to predict the racehorse's future condition before making a bid. However, there are cases where the racehorse you purchase does not grow as expected. There are also cases where a rider retires because he or she is unable to race enough due to injury. Therefore, it is desirable to have a system that can predict the future condition of racehorses.
  • the racehorse potential ability prediction system disclosed in Patent Document 1 uses a learning model to predict lifetime prize money.
  • Patent Document 1 When the potential ability prediction system of Patent Document 1 predicts the future state of a racehorse, it may be difficult to interpret the prediction results.
  • the condition prediction system of the present invention uses an acquisition means for acquiring information about racehorses and a prediction model that predicts the future condition of racehorses from the information about racehorses.
  • the present invention includes a prediction means for predicting the future state of a racehorse from information about the horse, and an output means for outputting the result of the prediction and the reason for the prediction.
  • the condition prediction method of the present invention acquires information about the racehorse, uses a prediction model that predicts the future condition of the racehorse from the information about the racehorse, and calculates the future condition of the racehorse from the acquired information about the racehorse.
  • the prediction result and the reason for the prediction are output.
  • the recording medium of the present invention uses a process of acquiring information about racehorses and a prediction model that predicts the future state of racehorses from the information about racehorses, and predicts the future state of racehorses from the acquired information about racehorses.
  • a prediction program that causes a computer to execute a process of predicting a state, a process of outputting a result of the prediction, and a reason for the prediction is recorded non-temporarily.
  • FIG. 1 is a diagram showing an example of a configuration of a first embodiment of the present invention.
  • FIG. 1 is a diagram showing an example of the configuration of a prediction system according to a first embodiment of the present invention. It is a figure showing an example of a display screen in a 1st embodiment of the present invention. It is a figure showing an example of a display screen in a 1st embodiment of the present invention. It is a figure showing an example of a display screen in a 1st embodiment of the present invention. It is a figure showing an example of a display screen in a 1st embodiment of the present invention. It is a figure showing an example of a display screen in a 1st embodiment of the present invention. It is a figure showing an example of a display screen in a 1st embodiment of the present invention.
  • FIG. 1st embodiment of the present invention It is a figure showing an example of a display screen in a 1st embodiment of the present invention. It is a figure showing an example of a display screen in a 1st embodiment of the present invention. It is a figure showing an example of a display screen in a 1st embodiment of the present invention. It is a figure showing an example of a display screen in a 1st embodiment of the present invention. It is a figure showing an example of a display screen in a 1st embodiment of the present invention. It is a figure showing an example of an operation flow of a prediction system of a 1st embodiment of the present invention. It is a figure showing an example of composition of a prediction system of a 2nd embodiment of the present invention. A diagram showing an example of an operation flow of a prediction system according to a second embodiment of the present invention. It is a figure showing the example of composition of other embodiments of the present invention.
  • FIG. 1 is a diagram showing an example of a racehorse prediction system.
  • the racehorse prediction system includes a state prediction system 10, a user terminal device 20, and an information management server 30.
  • the state prediction system 10 is connected to a user terminal device 20 via a network. Further, the state prediction system 10 is connected to the information management server 30 via a network.
  • the condition prediction system 10 is a system that predicts the future condition of racehorses. For example, the condition prediction system 10 predicts the future condition of a racehorse before an auction.
  • the racehorse before the auction is, for example, a racehorse owned by a producer.
  • the condition prediction system 10 may predict the future condition of racehorses that have not reached the age to run in races.
  • the future state of the racehorse is, for example, information regarding the evaluation of the racehorse in the future rather than at the time of prediction.
  • the future state of a racehorse is, for example, at least one of auction price, health condition, maintenance costs, race results, and earned prize money.
  • the auction price is the successful bid price of a racehorse in an auction.
  • the health condition is, for example, the presence or absence of an injury in the future.
  • the future condition of a racehorse is determined by changes in body weight, changes in muscle mass, training time, trainers suitable for racehorses, stables to outsource, jockeys suitable for racehorses, suitable race types, leg quality, etc. It may be at least one of the following: suitable race distance, preferred race development, first race start time, last race start time, and race entry period.
  • the future state of a racehorse is not limited to the above.
  • the condition prediction system 10 predicts the future condition of the racehorse, for example, using a prediction model that predicts the future condition of the racehorse from information about the racehorse.
  • the state prediction system 10 then outputs the prediction result and the reason for the prediction.
  • Information regarding racehorses is, for example, information that can affect the future condition of racehorses.
  • the information regarding the racehorse is, for example, at least one of parent horse information, biological information of the racehorse at the time of prediction, and breeding history.
  • Information regarding racehorses is not limited to the above.
  • the reason for prediction is, for example, information on which the prediction model is based when predicting the future state of the racehorse.
  • the reason for prediction is, for example, an item of information about a racehorse that has a greater influence on the prediction result of the future state of the racehorse than other items when the prediction model predicts the future state of the racehorse. .
  • the predictive model is, for example, a trained model generated using a machine learning algorithm.
  • the state prediction system 10 learns, for example, the relationship between information about racehorses that have run in races in the past and the future state of the racehorses. Then, the condition prediction system 10 generates a prediction model that predicts the future condition of the racehorse from the information regarding the racehorse to be predicted.
  • the prediction model may be a learning model generated outside the state prediction system 10. The prediction model will be explained later.
  • the user terminal device 20 is, for example, a terminal device owned by a person who uses the prediction results of the state prediction system 10.
  • a person who uses the prediction results is, for example, a person who makes a bid in a racehorse auction.
  • the person using the prediction result may be an auctioneer or an auction organizer.
  • the person who uses the prediction results may be a person who invests in racehorses owned by corporations or individuals.
  • a person who invests in racehorses owned by a corporation or an individual is also called a single owner.
  • the person using the prediction result may be a reporter or a commentator.
  • the person who uses the prediction result is not limited to the above example.
  • the information management server 30 is, for example, a server that holds information regarding racehorses.
  • the state prediction system 10 acquires information regarding racehorses from the information management server 30, for example. Then, the condition prediction system 10 receives the acquired information regarding the racehorse as input and predicts the future condition of the racehorse using the prediction model. After predicting the future state of the racehorse, the state prediction system 10 outputs the prediction result and the reason for the prediction to the user terminal device 20, for example.
  • the state prediction system 10 may acquire information regarding racehorses from a plurality of information management servers 30. Further, the state prediction system 10 may acquire information regarding the racehorse input by the user of the user terminal device 20 from the user terminal device 20.
  • the state prediction system 10 may output prediction results and prediction reasons to a plurality of user terminal devices 20.
  • the state prediction system 10 may output the prediction result and the reason for the prediction to the user terminal devices 20 used by a plurality of users.
  • the number of user terminal devices 20 and information management servers 30 may be set as appropriate.
  • FIG. 2 is a diagram showing an example of the configuration of the state prediction system 10. As shown in FIG.
  • the state prediction system 10 includes an acquisition section 11 , a prediction section 12 , an output section 13 , a generation section 14 , and a storage section 15 .
  • the acquisition unit 11 acquires information regarding racehorses.
  • Information regarding racehorses is, for example, information that can affect the future state of racehorses.
  • the acquisition unit 11 acquires information about the parent horse, biological information about the horse, and breeding history as information about the racehorse. When the racehorse to be predicted is running in a race, the acquisition unit 11 may acquire the race record of the racehorse to be predicted.
  • Information on the parent horse includes, for example, name, gender, breeder, trainer, stable, presence of injuries, history of training times, times for each race distance, race results, winnings, leg quality, stamina evaluation, and auction price. At least one or more.
  • the race results include, for example, at least one of the following: race course, race distance, weather, ranking, race development, field characteristics, number of runners, ranking, and earned prize money in races in which the racer ran in the past.
  • the race record may include the age at the time of first race and the age at retirement.
  • the parent horse information may include information on sibling horses of the racehorse to be predicted. Information on sibling horses includes, for example, name, gender, breeder, trainer, stable, presence or absence of injuries, training time by age, time by race distance, race record, winnings, leg quality, and auction price. One or more. Parent horse information and sibling horse information are not limited to the above.
  • the biological information includes at least one of sex, weight, weight change, blood data, and muscle mass. Biometric information is not limited to the above.
  • the training history is, for example, at least one of the history of the producer, trainer, and training time. The history of the trainer and training time is acquired when training has started. The training history is not limited to the above.
  • the acquisition unit 11 may acquire the selection results of the prediction models from the user terminal device 20.
  • the selection result of the prediction model is input to the user terminal device 20 by, for example, an operation by a person who uses the prediction result. Further, the acquisition unit 11 may acquire from the user terminal device 20 a selection of display items that is input to the user terminal device 20 through an operation by a person who uses the prediction results.
  • the acquisition unit 11 may acquire information regarding the racehorse and the future condition of the racehorse as training data for generating the prediction model.
  • the acquisition unit 11 stores, for example, the acquired information regarding the racehorse and the future state of the racehorse in the storage unit 15.
  • the prediction unit 12 predicts the future state of the racehorse from the acquired information about the racehorse using a prediction model that predicts the future state of the racehorse from the information about the racehorse. Furthermore, the prediction unit 12 extracts the reason why the prediction model predicted the future state of the racehorse as the prediction reason. For example, the prediction unit 12 acquires parameters used when the prediction model predicts the future state of the racehorse. The prediction unit 12 then extracts the reason for the prediction from the parameters that make a large contribution to predicting the future state of the racehorse.
  • the prediction unit 12 predicts at least one of the auction price of the racehorse, the maintenance cost of the purchased racehorse, and the prize money that the purchased racehorse will win as the future state of the racehorse.
  • the auction price of a racehorse is the winning bid price in an auction of a racehorse.
  • Racehorse maintenance costs are expenses paid to continue owning a racehorse.
  • the maintenance cost of a racehorse is, for example, the cost of feeding, maintaining health, and training the racehorse.
  • the cost of maintaining a racehorse may be the cost of entrusting a purchased racehorse to a stable.
  • the maintenance costs for racehorses are not limited to the above.
  • the amount of prize money that the purchased racehorse will receive is the amount of prize money that the purchased racehorse will receive when it runs in a race in the future.
  • the prediction unit 12 may predict, as the future state of the racehorse, the amount of prize money and maintenance costs that the purchased racehorse will receive for each age of the horse.
  • the prediction unit 12 may predict the amount of prize money and maintenance costs that the purchased racehorse will receive on a monthly basis.
  • the period for predicting the amount of prize money and maintenance costs that a purchased racehorse will receive may be set as appropriate.
  • the prediction unit 12 predicts the future income and expenditure as the future state of the racehorse based on the auction price of the racehorse, the maintenance cost of the purchased racehorse, and the prize money that the purchased racehorse will win. Good too.
  • the prediction unit 12 predicts, for example, the auction price of the racehorse, the maintenance cost of the purchased racehorse up to a certain point in the future, and the prize money that the purchased racehorse will win.
  • the prediction unit 12 then predicts the income and expenditure at a certain point in the future by subtracting the auction price of the racehorse and the accumulated amount of maintenance costs of the purchased racehorse from the total prize money that the purchased racehorse will win. do.
  • the prediction unit 12 predicts future income and expenditure, for example, on a monthly basis.
  • the interval for predicting future income and expenditure is not limited to monthly intervals, and may be determined as appropriate according to the preferences of the person using the prediction results of the state
  • the prediction unit 12 uses the purchase price of the racehorse, the maintenance cost of the purchased racehorse, and the prize money that the purchased racehorse will earn. You can also predict future income and expenditure.
  • the maintenance cost for the purchased racehorse may be a predetermined amount per month or year. Further, the future state of the racehorse predicted by the prediction model is not limited to the above example.
  • the prediction unit 12 may predict the future state of the racehorse using a prediction model depending on the purpose of prediction. For example, when the purpose of prediction is to predict an auction price, the prediction unit 12 predicts the auction price from information regarding the racehorse using a prediction model that predicts the auction price as the future state of the racehorse. Further, for example, when the purpose of prediction is to predict the winning prize money, the prediction unit 12 predicts the winning prize money from information about the racehorse using a prediction model that predicts the winning prize money as the future state of the racehorse. .
  • the prediction unit 12 may predict the recommendation level from information regarding the racehorse using a prediction model that predicts the recommendation level for purchasing the racehorse as the future state of the racehorse.
  • the degree of recommendation for purchase is calculated using, for example, an index set in advance by an expert who evaluates racehorses.
  • the prediction unit 12 may predict the future state of the racehorse using a prediction model that places weight on items that are important to the user.
  • the prediction unit 12 may determine the prediction result by weighting the prediction result of each prediction model.
  • the prediction unit 12 weights the prediction results of each prediction model, for example, depending on which of the plurality of prediction models the results of which prediction model is prioritized.
  • the prediction unit 12 may predict the recommendation level from information regarding the racehorse using a prediction model that predicts the recommendation level for purchasing a racehorse according to the items that are important to the person using the prediction results. For example, when a person using the prediction results places emphasis on pedigree, the prediction unit 12 predicts the recommendation level from information about the racehorse using a prediction model that places weight on data related to the pedigree at the time of prediction.
  • the prediction unit 12 may predict the future state of the racehorse using a prediction model according to the age of the racehorse to be predicted.
  • the prediction unit 12 predicts the future state of the racehorse using, for example, a prediction model for 0-year-olds, a prediction model for 1-year-olds, and a prediction model for 2-year-olds.
  • the age classification when making a prediction using a prediction model according to age is not limited to the above.
  • the prediction unit 12 may predict the future state of the racehorse using a prediction model according to the attributes of the person using the prediction results.
  • the attribute of the person who uses the prediction result is, for example, a beginner, an intermediate, or an expert.
  • the attributes of the person who uses the prediction results may be classified as an auction organizer, an auction participant, a horse owner, a trainer, or a reporter.
  • the classification of attributes of persons using prediction results is not limited to the above example.
  • the output unit 13 outputs the prediction result of the racehorse's future state and the reason for the prediction.
  • the output unit 13 outputs the prediction result of the future state of the racehorse and the reason for the prediction to the user terminal device 20, for example.
  • the output unit 13 may output the prediction result and the reason for prediction to a display device (not shown) connected to the state prediction system 10. Further, the output unit 13 may output the prediction result of the future state of the racehorse and the reason for the prediction to a server that distributes auction information, for example.
  • the output unit 13 may output the prediction results of the plurality of items.
  • the output unit 13 outputs, for example, the auction price and the amount of the won prize money as a result of predicting the future state of the racehorse.
  • the output unit 13 outputs, for example, the reason for predicting the auction price and the reason for predicting the amount of prize money to be won as the prediction reason. Outputs .
  • the output unit 13 may output prediction results of future states of items other than the auction price and reasons for the prediction for each price range of the auction price.
  • the output unit 13 may output the prediction result regarding the future income and expenditure as information for each lapse of time.
  • the output unit 13 outputs, for example, information regarding the income and expenditure for each age of the horse as a graph.
  • the output unit 13 outputs, for example, a graph showing the cumulative amount of expenditure and the cumulative amount of income for each age of the horse.
  • the output unit 13 may output a graph showing the difference between the cumulative amount of expenditure and the cumulative amount of income for each age of the horse as the income and expenditure.
  • the output unit 13 uses the investment amount as a single horse owner instead of the auction price, and outputs information on expenditures and income according to the investment amount. It's okay.
  • the output unit 13 may output information on items that are important in predicting the future state of the racehorse, which is added to the prediction result as reference information, among the information regarding the racehorse, along with the prediction result. Items to be emphasized in predicting the future state of a racehorse are, for example, items that have a high frequency of influencing the future state of the racehorse. Furthermore, when outputting information on items that are important in predicting the future state of a racehorse, the output unit 13 may output data that highlights items that correspond to the reason for the prediction. The output unit 13 may output, as reference information, items selected by a person who uses the prediction results from among the information regarding racehorses.
  • the output unit 13 may output information about the parent horse of the racehorse to be predicted.
  • the output unit 13 may output information about the parent horse of the racehorse to be predicted, when the age of the racehorse to be predicted is less than a standard.
  • the output unit 13 outputs information on the parent horse's race record and breeding history as reference information, for example, when the racehorse to be predicted is at an age where there is little information on the race record and breeding history.
  • the age standard and the items to be output are set in advance.
  • the output unit 13 may output performance data of racehorses having similar attributes to the racehorse to be predicted, as reference information.
  • a racehorse with similar attributes is, for example, a racehorse whose auction price and breeding history are similar to the racehorse to be predicted.
  • the racehorse with similar attributes may be a racehorse with a similar pedigree to the racehorse to be predicted. Racehorses with similar attributes are not limited to the above example.
  • the output unit 13 outputs, as reference information, information on races in which racehorses having attributes similar to those of the racehorse to be predicted have run, and the winning prize money.
  • the output unit 13 may output data that highlights items selected by the person using the prediction results among the reference information. Highlighting is performed, for example, by changing at least one of the color, font size, font thickness, and decoration around the font from other items.
  • the output unit 13 outputs, for example, at least one item among age, sex, weight, father horse, mother horse, producer, and training history as reference information.
  • the output unit 13 may output highlighted data corresponding to the prediction reason when there is an item corresponding to the prediction reason among the reference information items.
  • the output unit 13 may output an item that is a negative factor for the prediction result as the reason for the prediction.
  • Negative factors are items that have a large influence on the prediction that the racehorse's evaluation will be lowered, among the items that have a large influence on the future state of the racehorse.
  • a large influence means, for example, that when a certain item is changed, the fluctuation in the prediction result of the future state of the racehorse is larger than that of other items.
  • the publicly managed competition is horse racing, for example, if the parent horse has a history of injury or illness, items related to health status are extracted as negative factors.
  • the output unit 13 may output the prediction reason as a text.
  • the output unit 13 outputs a sentence indicating the reason for the prediction, for example, based on information regarding a racehorse that has a high degree of influence on the prediction of the future state of the racehorse. For example, the relationship between information about a racehorse that has a high degree of influence on the prediction of the future state of the racehorse and a sentence indicating the reason for the prediction is set in advance.
  • the output unit 13 may output a sentence such as "This horse is recommended because it is expected to win a lot of prize money" as the reason for prediction. good.
  • the output unit 13 may output an image of the racehorse along with the prediction result and the reason for the prediction. Further, the output unit 13 may output an image of the racehorse being trained together with the prediction result and the reason for the prediction. Further, the output unit 13 may output an image of the parent horse or sibling horse of the racehorse to be predicted. The image may be a still image or a moving image.
  • the output unit 13 may output the prediction result and the reason for the prediction superimposed on the image of the racehorse. Further, the output unit 13 may output either the prediction result or the prediction reason superimposed on the image of the racehorse.
  • FIG. 3 shows an example of a display screen of a prediction result of a racehorse's future state.
  • the example of the display screen in FIG. 3 is a display screen of a prediction result when an auction price is predicted as a prediction result of a racehorse's future state.
  • horse names, predicted prices, and reasons are displayed as a list for racehorses to be predicted.
  • the predicted price is the auction price predicted by the prediction model.
  • this is the prediction reason that the prediction model outputs together with the prediction result.
  • the prediction reason may include multiple items.
  • FIG. 4 shows an example of a display screen that further displays the amount of prize money won by a racehorse in the example of the display screen of FIG. 3.
  • the prediction result of the auction price is displayed as the expected price
  • the prediction result of the prize amount to be won is displayed as the prediction of the amount to be won.
  • FIG. 5 shows an example of a display screen that further displays information regarding the racehorse to be predicted in the example of the display screen of FIG. 3.
  • the names of the breeder, father horse, and mother horse of the racehorse to be predicted are further displayed.
  • the information regarding racehorses that is displayed together with the prediction results is not limited to the above example.
  • FIG. 6 shows an example of a display screen that displays changes in income and expenditure regarding racehorses as a graph.
  • a graph is displayed in which the horizontal axis represents the age of the racehorse to be predicted, and the vertical axis represents the amount of money representing expenditures and income.
  • the auction price, expenditures representing the cumulative amount of expenses necessary for maintenance, and income representing the cumulative amount of won prizes are displayed as a graph.
  • the graph indicated by a broken line indicates the cumulative amount of the auction price and the expenses necessary for maintenance.
  • the graph indicated by a solid line indicates the cumulative amount of earned prize money.
  • the reason for prediction may be displayed in a superimposed manner on the graph in the example display screen of FIG.
  • the graph in the example display screen of FIG. 6 may be displayed in combination with another example display screen.
  • FIG. 7 shows an example of a display screen that further displays negative factors as prediction reasons in the example of the display screen of FIG. 3.
  • the plus factor indicates a positive factor.
  • a positive factor is an item that has a large influence on the prediction that the price will be high among the items that have a large influence on the prediction result of the auction price.
  • negative factors are shown as negative factors. Negative factors are items that have a large influence on the prediction that the price will be low, among the items that have a large influence on the prediction result of the auction price.
  • FIG. 8 shows, in the example of the display screen in FIG. 3, the predicted auction price of a racehorse with a low auction price, the predicted reason why the auction price is low, and the high evaluation that is opposite to the predicted reason.
  • An example of a display screen showing related factors is shown.
  • the example of the display screen in FIG. 8 is a display screen that displays, for a racehorse with a low predicted price, the reason why the price is low and the factors that increase the evaluation.
  • an expected price indicating the prediction result of the auction price and a price reason indicating the reason for the prediction are displayed.
  • factors that increase the prediction result are displayed as evaluation reasons.
  • a factor that increases the evaluation is, for example, an item that causes the winning prize to be high while the auction price is low. For example, if the health condition is lowering the auction price, but the increased training time is expected to increase the prize money, the training time will be displayed as a factor that will increase the evaluation.
  • FIG. 9 shows an example of a display screen that further displays reference information in the example of the display screen of FIG. 3.
  • the prediction result and the reason for the prediction are displayed in the left frame.
  • information regarding racehorse B is shown as reference information in the right frame.
  • the horse's age, breeder, trainer, pedigree, weight, and training history are displayed as reference information. Items displayed as reference information are not limited to the above.
  • reference information about the selected runner is displayed in the reference information column. Good too.
  • FIG. 10 shows an example of a display screen that outputs an image of a racehorse in the example of the display screen of FIG. 9.
  • an image of a racehorse is displayed on the left side of the lower row.
  • the image of the selected racehorse may be displayed in the image display area.
  • an image of the parent horse of the racehorse to be predicted may be displayed.
  • the image of the parent horse of the racehorse to be predicted may be an image when the parent horse ran in a race in the past.
  • a video during training may be displayed as the image of the racehorse.
  • the image of the racehorse is acquired from the information management server 30, for example.
  • FIG. 11 shows an example of a display screen in which a prediction result and a reason for prediction are displayed superimposed on an image of a racehorse in the example of the display screen of FIG. 10.
  • a good pedigree is displayed as the reason for prediction.
  • the prediction reason displayed on the image of the racehorse may be a plurality of items. Further, either the prediction result or the prediction reason may be displayed superimposed on the image of the racehorse.
  • the generation unit 14 When generating a prediction model in the state prediction system 10, the generation unit 14 generates a prediction model that predicts the future state of the racehorse from information regarding the racehorse. For example, the generation unit 14 learns the relationship between information about racehorses and the future state of the racehorse, and generates a prediction model that predicts the future state of the racehorse from the information about the racehorse. The generation unit 14 may generate a prediction model that predicts a plurality of items regarding the future state of the racehorse. The generation unit 14 generates, for example, a prediction model that predicts an auction price and an acquired prize.
  • the generation unit 14 may generate each of the plurality of prediction models used by the prediction unit 12 for prediction.
  • the generation unit 14 When generating a prediction model for 0-year-olds, the generation unit 14, for example, learns the relationship between information about the racehorse, including the race record of the parent horse, and the future state of the racehorse, and generates the prediction model. .
  • the generation unit 14 When generating a prediction model for a one-year-old, the generation unit 14, for example, learns the relationship between information about the racehorse, including biological information of the racehorse, and the future state of the racehorse, and generates the prediction model. do.
  • the biological information of the racehorse includes, for example, one or more of sex, coat, gait, weight, body length, change in weight, change in physical condition, blood data, and personality.
  • the biological information of racehorses is not limited to the above.
  • the generation unit 14 When generating a prediction model for a 2-year-old horse, the generation unit 14 generates the prediction model by learning, for example, the relationship between information about the racehorse including its training history and its future state.
  • the generation unit 14 When generating a prediction model according to the items that the user of the prediction results places importance on, the generation unit 14 generates, for example, the relationship between information about the racehorse, including information on the items that the user places emphasis on, and the future state of the racehorse. Learn and generate predictive models. When generating a predictive model that emphasizes pedigree, the generation unit 14 learns, for example, the relationship between the biological information of the parent horse, information about the racehorse including the parent horse's race record, and the future state of the racehorse. and generate a predictive model. When generating a prediction model according to the items of the prediction results that the user considers important, the generation unit 14 may generate the prediction model by increasing the weight of the items that the user considers important. When increasing the weight of the item that the user places importance on, the generation unit 14 generates a prediction model such that, for example, the coefficient of the feature amount regarding the item that the user places importance on is larger than the coefficient of the feature amount regarding other items. generate
  • the generation unit 14 generates a predictive model using, for example, a learning algorithm based on factorized asymptotic Bayesian inference.
  • the generation unit 14 uses information about the racehorse as input data and the future state of the racehorse as correct answer data, and generates a case according to decision tree-style rules. Divide.
  • the generation unit 14 then generates a learning model that predicts the future state of the racehorse using a linear model that combines different explanatory variables in each case.
  • the generation unit 14 generates a learning model by sequentially optimizing data case classification conditions, generating a predictive model by optimizing combinations of explanatory variables, and deleting unnecessary predictive models.
  • This method of generating a learning model is also called heterogeneous mixture learning because it makes predictions by combining prediction models based on combinations of different explanatory variables.
  • By generating a prediction model using heterogeneous mixture learning it becomes possible to explain the prediction results of the future state of the racehorse using conditions that have a strong influence on the prediction results, which improves the explainability of the prediction results. will improve.
  • a method of heterogeneous mixture learning is disclosed in, for example, US Patent Application Publication No. 2014/0222741.
  • the learning algorithm used for machine learning to generate a predictive model is not limited to the above example.
  • the generation unit 14 may generate a learning model that predicts the future state of a racehorse from information about the racehorse by deep learning using a neural network.
  • the generation unit 14 for example, varies the data of each item and selects items that have a large impact on the future condition of the racehorse based on changes in the future condition of the racehorse.
  • the generation unit 14 changes the data of each item and extracts an item that has a large influence on the future state of the racehorse as the reason for prediction.
  • the storage unit 15 stores, for example, a prediction model. When a plurality of prediction models are used, the storage unit 15 stores the plurality of prediction models. Further, when the state prediction system 10 generates a prediction model, the storage unit 15 may store data that is used as teacher data and associates information regarding the racehorse with the future state of the racehorse. Further, when adding reference information to the prediction result, the storage unit 15 may store data used as the reference information. Note that the prediction model used by the prediction unit 12 may be stored in a storage means other than the storage unit 15.
  • the user terminal device 20 acquires the prediction result and the reason for the prediction from the state prediction system 10. Then, the user terminal device 20 outputs the prediction result and the reason for the prediction to, for example, a display device (not shown).
  • the user terminal device 20 When a prediction model is selected by the user, the user terminal device 20 obtains, for example, the name of the prediction model input by the user's operation as the selection result of the prediction model. Then, the user terminal device 20 outputs the name of the input prediction model to the state prediction system 10.
  • a smartphone, a tablet computer, a notebook computer, or a desktop computer is used as the user terminal device 20.
  • the terminal device used for the user terminal device 20 is not limited to the above example.
  • the information management server 30 is, for example, a server that stores or manages information regarding racehorses.
  • the information management server 30 may be a plurality of servers installed according to the content of information regarding racehorses.
  • Information regarding racehorses may be stored in a storage device managed by the information management server 30. Further, the information management server 30 may store images of racehorses.
  • FIG. 12 is a diagram showing an example of an operation flow when the state prediction system 10 predicts the future state of a racehorse.
  • the acquisition unit 11 acquires information regarding a racehorse whose future state is to be predicted (step S11).
  • the acquisition unit 11 acquires information regarding racehorses from the information management server 30, for example.
  • the prediction unit 12 uses the prediction model that predicts the future state of the racehorse from the information about the racehorse to predict the future state of the racehorse from the information about the racehorse acquired by the acquisition unit 11.
  • the future state is predicted (step S12).
  • the output unit 13 When the future state of the racehorse is predicted, the output unit 13 outputs the prediction result and the reason for the prediction (step S13). The output unit 13 outputs the prediction result and the reason for the prediction to the user terminal device 20, for example.
  • the user terminal device 20 that has received the prediction result and the prediction reason displays the prediction result and the prediction reason on the display device, for example.
  • FIG. 13 is a diagram illustrating an example of an operation flow when the state prediction system 10 generates a prediction model.
  • the acquisition unit 11 acquires information regarding the racehorse and the future state of the racehorse (step S21).
  • the generation unit 14 learns the relationship between the information about the racehorse and the future state of the racehorse, and calculates the future state of the racehorse from the information about the racehorse.
  • a prediction model that predicts the state of is generated (step S22). After generating the prediction model, the generation unit 14 stores the generated prediction model in the storage unit 15 (step S23).
  • the state prediction system 10 of the racehorse prediction system of this embodiment acquires information about racehorses and uses a prediction model to predict the future state of the racehorses. Then, the state prediction system 10 outputs the prediction result and the reason for the prediction to the user terminal device 20, for example. By outputting the reason for the prediction together with the prediction result of the future state of the racehorse, a person using the prediction result can easily interpret the prediction result of the future state of the racehorse. Therefore, by using the condition prediction system 10, it is possible to easily interpret the results of predicting the future condition of the racehorse.
  • the condition prediction system 10 can, for example, show the difference between the auction price and maintenance costs and the won prize money as the income and expenditure, so that a person using the prediction results can see the racehorse's profit. It is possible to output information for reference when considering. Further, when outputting information regarding the income and expenditure of a racehorse as a graph, the state prediction system 10 can output information regarding the income and expenditure, which is a prediction result, in a format that is easy for the user to visually understand.
  • the state prediction system 10 can, for example, use a prediction model that corresponds to the items that the person using the prediction results places importance on, so that the state prediction system 10 can use the prediction model that corresponds to the items that the person using the prediction results places importance on.
  • the reason for the prediction can be output together with the prediction result.
  • the state prediction system 10 can output an appropriate prediction result and the reason for the prediction according to the age of the preliminary racehorse, for example.
  • the state prediction system 10 When outputting reference information together with the prediction result, the state prediction system 10 adds and outputs the reference information to the prediction result and the reason for the prediction, so that a person using the prediction result can understand the reason for the prediction and the reference information. You can refer to the information to more easily interpret the reason for the prediction.
  • FIG. 14 is a diagram showing an example of the configuration of the state prediction system 100 of this embodiment.
  • the state prediction system 100 includes an acquisition section 101, a prediction section 102, and an output section 103.
  • the acquisition unit 101 acquires information regarding racehorses.
  • the prediction unit 102 predicts the future state of the racehorse from the acquired information about the racehorse using a prediction model that predicts the future state of the racehorse from the information about the racehorse.
  • the output unit 103 outputs the prediction result and the reason for the prediction.
  • the acquisition unit 11 of the first embodiment is an example of the acquisition unit 101. Further, the acquisition unit 101 is one aspect of an acquisition means.
  • the prediction unit 12 of the first embodiment is an example of the prediction unit 102. Furthermore, the prediction unit 102 is one aspect of prediction means.
  • the output unit 13 of the first embodiment is an example of the output unit 103. Furthermore, the output unit 103 is one aspect of output means.
  • FIG. 15 is a diagram illustrating an example of the operation flow of the state prediction system 100.
  • the acquisition unit 101 acquires information regarding racehorses (step S101).
  • the prediction unit 102 predicts the future condition of the racehorse from the acquired information regarding the racehorse using a prediction model that predicts the future condition of the racehorse from the information regarding the racehorse. Predict (step S102).
  • the output unit 103 outputs the prediction result and the reason for the prediction (step S103).
  • the state prediction system 100 of this embodiment predicts the future state of a racehorse using a prediction model. Then, the condition prediction system 10 outputs the prediction result of the future condition of the racehorse and the reason for the prediction. As a result, the condition prediction system 10 can facilitate the interpretation of prediction results of the racehorse's future condition.
  • FIG. 16 shows an example of the configuration of a computer 200 that executes a computer program that performs each process in the state prediction system 10 of the first embodiment and the state prediction system 100 of the second embodiment.
  • the computer 200 includes a CPU (Central Processing Unit) 201, a memory 202, a storage device 203, an input/output I/F (Interface) 204, and a communication I/F 205.
  • CPU Central Processing Unit
  • the CPU 201 reads computer programs for performing each process from the storage device 203 and executes them.
  • the CPU 201 may be configured by a combination of multiple CPUs. Further, the CPU 201 may be configured by a combination of a CPU and other types of processors. For example, the CPU 201 may be configured by a combination of a CPU and a GPU (Graphics Processing Unit).
  • the memory 202 is configured with a DRAM (Dynamic Random Access Memory) or the like, and temporarily stores computer programs executed by the CPU 201 and data being processed.
  • the storage device 203 stores computer programs executed by the CPU 201.
  • the storage device 203 is configured by, for example, a nonvolatile semiconductor storage device. Other storage devices such as a hard disk drive may be used as the storage device 203.
  • the input/output I/F 204 is an interface that receives input from a worker and outputs display data and the like.
  • the communication I/F 205 is an interface that transmits and receives data between the user terminal device 20 and the information management server 30. Further, the user terminal device 20 and the information management server 30 may also have similar configurations.
  • the computer program used to execute each process can also be stored and distributed in a computer-readable recording medium that non-temporarily records program data.
  • a computer-readable recording medium for example, a magnetic tape for data recording or a magnetic disk such as a hard disk can be used.
  • an optical disc such as a CD-ROM (Compact Disc Read Only Memory) can also be used.
  • a nonvolatile semiconductor memory device may be used as the recording medium.

Abstract

This situation prediction system comprises an acquisition unit, a prediction unit, and an output unit. The acquisition unit acquires information relating to racehorses. The prediction unit predicts the future situation of racehorses from the acquired information pertaining to the racehorses by using a prediction model that predicts the future situation of the racehorses from the information relating to the racehorses. The output unit outputs prediction results and the reason for predictions.

Description

状態予測システム、状態予測方法および記録媒体Condition prediction system, condition prediction method, and recording medium
　本発明は、状態予測システム等に関する。 The present invention relates to a state prediction system and the like.
　公営競技のうち競馬では、競走馬のオークションが行われる。オークションでは競走馬が若いうちに取引が行われるため、購入者は、競走馬の将来状態を予測して入札する必要がある。しかし、購入した競走馬が、期待通りには成長しない場合もある。また、怪我によって十分にレースに出走できずに引退する場合もある。そのため、競走馬の将来状態を予測できるシステムがあることが望ましい。 In horse racing, which is a publicly managed competition, auctions of racehorses are held. Since racehorses are sold at auctions while they are young, buyers need to predict the racehorse's future condition before making a bid. However, there are cases where the racehorse you purchase does not grow as expected. There are also cases where a rider retires because he or she is unable to race enough due to injury. Therefore, it is desirable to have a system that can predict the future condition of racehorses.
　特許文献１の競走馬の潜在的能力予測システムは、学習モデルを用いて、生涯獲得賞金を予測している。 The racehorse potential ability prediction system disclosed in Patent Document 1 uses a learning model to predict lifetime prize money.
特開２０２０－１４９８５３号公報Japanese Patent Application Publication No. 2020-149853
　特許文献１の潜在的能力予測システムは、競走馬の将来の状態の予測において、予測結果の解釈が難しい場合がある。 When the potential ability prediction system of Patent Document 1 predicts the future state of a racehorse, it may be difficult to interpret the prediction results.
　上記の課題を解決するため、競走馬の将来の状態の予測結果の解釈を容易にすることができる状態予測システム等を提供することを目的とする。 In order to solve the above-mentioned problems, it is an object of the present invention to provide a condition prediction system etc. that can facilitate the interpretation of prediction results of the future condition of racehorses.
　上記の課題を解決するため、本発明の状態予測システムは、競走馬に関する情報を取得する取得手段と、競走馬に関する情報から競走馬の将来の状態を予測する予測モデルを用いて、取得した競走馬に関する情報から、競走馬の将来の状態を予測する予測手段と、予測の結果と、予測の理由とを出力する出力手段とを備える。 In order to solve the above problems, the condition prediction system of the present invention uses an acquisition means for acquiring information about racehorses and a prediction model that predicts the future condition of racehorses from the information about racehorses. The present invention includes a prediction means for predicting the future state of a racehorse from information about the horse, and an output means for outputting the result of the prediction and the reason for the prediction.
　本発明の状態予測方法は、競走馬に関する情報を取得し、競走馬に関する情報から競走馬の将来の状態を予測する予測モデルを用いて、取得した競走馬に関する情報から、競走馬の将来の状態を予測し、予測の結果と、予測の理由とを出力する。 The condition prediction method of the present invention acquires information about the racehorse, uses a prediction model that predicts the future condition of the racehorse from the information about the racehorse, and calculates the future condition of the racehorse from the acquired information about the racehorse. The prediction result and the reason for the prediction are output.
　本発明の記録媒体は、競走馬に関する情報を取得する処理と、競走馬に関する情報から競走馬の将来の状態を予測する予測モデルを用いて、取得した競走馬に関する情報から、競走馬の将来の状態を予測する処理と、予測の結果と、予測の理由とを出力する処理とをコンピュータに実行させる予測プログラムを非一時的に記録する。 The recording medium of the present invention uses a process of acquiring information about racehorses and a prediction model that predicts the future state of racehorses from the information about racehorses, and predicts the future state of racehorses from the acquired information about racehorses. A prediction program that causes a computer to execute a process of predicting a state, a process of outputting a result of the prediction, and a reason for the prediction is recorded non-temporarily.
　本発明によると、競走馬の将来の状態の予測結果の解釈を容易にすることができる。 According to the present invention, it is possible to easily interpret the results of predicting the future state of a racehorse.
本発明の第１の実施形態の構成の一例を示す図である。1 is a diagram showing an example of a configuration of a first embodiment of the present invention. FIG. 本発明の第１の実施形態の予測システムの構成の例を示す図である。1 is a diagram showing an example of the configuration of a prediction system according to a first embodiment of the present invention. 本発明の第１の実施形態における表示画面の例を示す図である。It is a figure showing an example of a display screen in a 1st embodiment of the present invention. 本発明の第１の実施形態における表示画面の例を示す図である。It is a figure showing an example of a display screen in a 1st embodiment of the present invention. 本発明の第１の実施形態における表示画面の例を示す図である。It is a figure showing an example of a display screen in a 1st embodiment of the present invention. 本発明の第１の実施形態における表示画面の例を示す図である。It is a figure showing an example of a display screen in a 1st embodiment of the present invention. 本発明の第１の実施形態における表示画面の例を示す図である。It is a figure showing an example of a display screen in a 1st embodiment of the present invention. 本発明の第１の実施形態における表示画面の例を示す図である。It is a figure showing an example of a display screen in a 1st embodiment of the present invention. 本発明の第１の実施形態における表示画面の例を示す図である。It is a figure showing an example of a display screen in a 1st embodiment of the present invention. 本発明の第１の実施形態における表示画面の例を示す図である。It is a figure showing an example of a display screen in a 1st embodiment of the present invention. 本発明の第１の実施形態における表示画面の例を示す図である。It is a figure showing an example of a display screen in a 1st embodiment of the present invention. 本発明の第１の実施形態の予測システムの動作フローの例を示す図である。It is a figure showing an example of an operation flow of a prediction system of a 1st embodiment of the present invention. 本発明の第１の実施形態の予測システムの動作フローの例を示す図である。It is a figure showing an example of an operation flow of a prediction system of a 1st embodiment of the present invention. 本発明の第２の実施形態の予測システムの構成の例を示す図である。It is a figure showing an example of composition of a prediction system of a 2nd embodiment of the present invention. 本発明の第２の実施形態の予測システムの動作フローの例を示す図でA diagram showing an example of an operation flow of a prediction system according to a second embodiment of the present invention. 本発明の他の実施形態の構成の例を示す図である。It is a figure showing the example of composition of other embodiments of the present invention.
　本発明の第１の実施形態について、図を参照して詳細に説明する。図１は、競走馬予測システムの例を示す図である。一例として、競走馬予測システムは、状態予測システム１０と、利用者端末装置２０と、情報管理サーバ３０を備える。状態予測システム１０は、ネットワークを介して、利用者端末装置２０と接続する。また、状態予測システム１０は、ネットワークを介して、情報管理サーバ３０と接続する。 A first embodiment of the present invention will be described in detail with reference to the drawings. FIG. 1 is a diagram showing an example of a racehorse prediction system. As an example, the racehorse prediction system includes a state prediction system 10, a user terminal device 20, and an information management server 30. The state prediction system 10 is connected to a user terminal device 20 via a network. Further, the state prediction system 10 is connected to the information management server 30 via a network.
　状態予測システム１０は、競走馬の将来の状態を予測するシステムである。状態予測システム１０は、例えば、オークション前の競走馬の将来の状態を予測する。オークション前の競走馬は、例えば、生産者が保有している競走馬である。状態予測システム１０は、レースに出走する年齢に到達していない競走馬の将来の状態を予測してもよい。競走馬の将来の状態は、例えば、予測時よりも将来における競走馬の評価に関する情報である。競走馬の将来の状態は、例えば、オークション価格、健康状態、維持費用、レース成績および獲得賞金のうちの少なくとも１つ以上である。オークション価格は、オークションにおける競走馬の落札価格である。健康状態は、例えば、将来における怪我の有無である。また、競走馬の将来の状態は、体重の変化、筋肉量の変化、調教タイム、競走馬に適した調教師、委託する厩舎、競走馬に適した騎手、適性のあるレース種別、脚質、適性のあるレース距離、得意とするレース展開、初出走時期、最終出走時期、出走可能期間のうちの少なくとも１つ以上であってもよい。競走馬の将来の状態は、上記に限られない。 The condition prediction system 10 is a system that predicts the future condition of racehorses. For example, the condition prediction system 10 predicts the future condition of a racehorse before an auction. The racehorse before the auction is, for example, a racehorse owned by a producer. The condition prediction system 10 may predict the future condition of racehorses that have not reached the age to run in races. The future state of the racehorse is, for example, information regarding the evaluation of the racehorse in the future rather than at the time of prediction. The future state of a racehorse is, for example, at least one of auction price, health condition, maintenance costs, race results, and earned prize money. The auction price is the successful bid price of a racehorse in an auction. The health condition is, for example, the presence or absence of an injury in the future. In addition, the future condition of a racehorse is determined by changes in body weight, changes in muscle mass, training time, trainers suitable for racehorses, stables to outsource, jockeys suitable for racehorses, suitable race types, leg quality, etc. It may be at least one of the following: suitable race distance, preferred race development, first race start time, last race start time, and race entry period. The future state of a racehorse is not limited to the above.
　状態予測システム１０は、例えば、競走馬に関する情報から競走馬の将来の状態を予測する予測モデルを用いて、競走馬の将来の状態を予測する。そして、状態予測システム１０は、予測結果と、予測理由とを出力する。 The condition prediction system 10 predicts the future condition of the racehorse, for example, using a prediction model that predicts the future condition of the racehorse from information about the racehorse. The state prediction system 10 then outputs the prediction result and the reason for the prediction.
　競走馬に関する情報は、例えば、競走馬の将来の状態に影響を及ぼし得る情報である。競走馬に関する情報は、例えば、親馬の情報、予測時における競走馬の生体情報および育成履歴のうち、少なくとも１つである。競走馬に関する情報は、上記に限られない。予測理由は、例えば、予測モデルが競走馬の将来の状態を予測する際に根拠となった情報である。予測理由は、例えば、競走馬に関する情報のうち、予測モデルが競走馬の将来の状態を予測する際に、競走馬の将来の状態の予測結果への影響が他の項目よりも大きい項目である。 Information regarding racehorses is, for example, information that can affect the future condition of racehorses. The information regarding the racehorse is, for example, at least one of parent horse information, biological information of the racehorse at the time of prediction, and breeding history. Information regarding racehorses is not limited to the above. The reason for prediction is, for example, information on which the prediction model is based when predicting the future state of the racehorse. The reason for prediction is, for example, an item of information about a racehorse that has a greater influence on the prediction result of the future state of the racehorse than other items when the prediction model predicts the future state of the racehorse. .
　予測モデルは、例えば、機械学習アルゴリズムを用いて生成された学習済みモデルである。状態予測システム１０は、例えば、過去にレースに出走した競走馬に関する情報と、競走馬の将来の状態との関係を学習する。そして、状態予測システム１０は、予測対象の競走馬に関する情報から競走馬の将来の状態を予測する予測モデルを生成する。予測モデルは、状態予測システム１０の外部で生成された学習モデルであってもよい。予測モデルについては、後で説明する。 The predictive model is, for example, a trained model generated using a machine learning algorithm. The state prediction system 10 learns, for example, the relationship between information about racehorses that have run in races in the past and the future state of the racehorses. Then, the condition prediction system 10 generates a prediction model that predicts the future condition of the racehorse from the information regarding the racehorse to be predicted. The prediction model may be a learning model generated outside the state prediction system 10. The prediction model will be explained later.
　利用者端末装置２０は、例えば、状態予測システム１０の予測結果を利用する人物が所持している端末装置である。予測結果を利用する人物は、例えば、競走馬のオークションにおいて入札する人物である。予測結果を利用する人物は、オークションへの出品者またはオークションの主催者であってもよい。予測結果を利用する人物は、法人または個人によって保有されている競走馬に出資する人物であってもよい。法人または個人によって保有されている競走馬に出資する人物は、一口馬主とも呼ばれる。また、予測結果を利用する人物は、記者または解説者であってもよい。予測結果を利用する人物は、上記の例に限られない。情報管理サーバ３０は、例えば、競走馬に関する情報を保有しているサーバである。 The user terminal device 20 is, for example, a terminal device owned by a person who uses the prediction results of the state prediction system 10. A person who uses the prediction results is, for example, a person who makes a bid in a racehorse auction. The person using the prediction result may be an auctioneer or an auction organizer. The person who uses the prediction results may be a person who invests in racehorses owned by corporations or individuals. A person who invests in racehorses owned by a corporation or an individual is also called a single owner. Further, the person using the prediction result may be a reporter or a commentator. The person who uses the prediction result is not limited to the above example. The information management server 30 is, for example, a server that holds information regarding racehorses.
　状態予測システム１０は、例えば、情報管理サーバ３０から競走馬に関する情報を取得する。そして、状態予測システム１０は、取得した競走馬に関する情報を入力とし、予測モデルを用いて競走馬の将来の状態を予測する。競走馬の将来の状態を予測すると、状態予測システム１０は、例えば、利用者端末装置２０に、予測結果と、予測理由とを出力する。 The state prediction system 10 acquires information regarding racehorses from the information management server 30, for example. Then, the condition prediction system 10 receives the acquired information regarding the racehorse as input and predicts the future condition of the racehorse using the prediction model. After predicting the future state of the racehorse, the state prediction system 10 outputs the prediction result and the reason for the prediction to the user terminal device 20, for example.
　状態予測システム１０は、複数の情報管理サーバ３０から競走馬に関する情報を取得してもよい。また、状態予測システム１０は、利用者端末装置２０から、利用者端末装置２０の利用者が入力する競走馬に関する情報を取得してもよい。 The state prediction system 10 may acquire information regarding racehorses from a plurality of information management servers 30. Further, the state prediction system 10 may acquire information regarding the racehorse input by the user of the user terminal device 20 from the user terminal device 20.
　状態予測システム１０は、複数の利用者端末装置２０に、予測結果と、予測理由とを出力してもよい。状態予測システム１０は、例えば、複数の利用者がそれぞれ利用している利用者端末装置２０に、予測結果と、予測理由とを出力してもよい。利用者端末装置２０および情報管理サーバ３０の数は、適宜、設定され得る。 The state prediction system 10 may output prediction results and prediction reasons to a plurality of user terminal devices 20. For example, the state prediction system 10 may output the prediction result and the reason for the prediction to the user terminal devices 20 used by a plurality of users. The number of user terminal devices 20 and information management servers 30 may be set as appropriate.
　状態予測システム１０の構成について説明する。図２は、状態予測システム１０の構成の例を示す図である。状態予測システム１０は、取得部１１と、予測部１２と、出力部１３と、生成部１４と、記憶部１５を備える。 The configuration of the state prediction system 10 will be explained. FIG. 2 is a diagram showing an example of the configuration of the state prediction system 10. As shown in FIG. The state prediction system 10 includes an acquisition section 11 , a prediction section 12 , an output section 13 , a generation section 14 , and a storage section 15 .
　取得部１１は、競走馬に関する情報を取得する。競走馬に関する情報は、例えば、競走馬の将来の状態に影響し得る情報である。取得部１１は、競走馬に関する情報として、親馬の情報、馬体の生体情報と、育成履歴とを取得する。予測対象の競走馬がレースに出走している場合には、取得部１１は、予測対象の競走馬の戦績を取得してもよい。親馬の情報は、例えば、名称、性別、生産者、調教師、厩舎、怪我の有無、調教タイムの履歴、レース距離ごとのタイム、戦績、獲得賞金、脚質、スタミナ評価およびオークション価格のうち少なくも１つ以上である。戦績は、例えば、過去に出走したレースにおける、レース場、レース距離、天候、順位、レース展開、場特性、出走頭数、格付けおよび獲得賞金のうち、少なくも１つ以上である。戦績には、初出走時の年齢および引退年齢が含まれていてもよい。親馬の情報には、予測対象の競走馬の兄弟馬の情報が含まれていてもよい。兄弟馬の情報は、例えば、名称、性別、生産者、調教師、厩舎、怪我の有無、年齢ごとの調教タイム、レース距離ごとのタイム、戦績、獲得賞金、脚質およびオークション価格のうち少なくも１つ以上である。親馬の情報および兄弟馬の情報は、上記に限られない。生体情報は、性別、体重、体重変化、血液データおよび筋肉量のうち少なくも１つ以上である。生体情報は、上記に限られない。育成履歴は、例えば、生産者、調教師、調教タイムの履歴のうち少なくも１つ以上である。調教師および調教タイムの履歴は、調教が始まっている場合に取得される。育成履歴は、上記に限られない。 The acquisition unit 11 acquires information regarding racehorses. Information regarding racehorses is, for example, information that can affect the future state of racehorses. The acquisition unit 11 acquires information about the parent horse, biological information about the horse, and breeding history as information about the racehorse. When the racehorse to be predicted is running in a race, the acquisition unit 11 may acquire the race record of the racehorse to be predicted. Information on the parent horse includes, for example, name, gender, breeder, trainer, stable, presence of injuries, history of training times, times for each race distance, race results, winnings, leg quality, stamina evaluation, and auction price. At least one or more. The race results include, for example, at least one of the following: race course, race distance, weather, ranking, race development, field characteristics, number of runners, ranking, and earned prize money in races in which the racer ran in the past. The race record may include the age at the time of first race and the age at retirement. The parent horse information may include information on sibling horses of the racehorse to be predicted. Information on sibling horses includes, for example, name, gender, breeder, trainer, stable, presence or absence of injuries, training time by age, time by race distance, race record, winnings, leg quality, and auction price. One or more. Parent horse information and sibling horse information are not limited to the above. The biological information includes at least one of sex, weight, weight change, blood data, and muscle mass. Biometric information is not limited to the above. The training history is, for example, at least one of the history of the producer, trainer, and training time. The history of the trainer and training time is acquired when training has started. The training history is not limited to the above.
　競走馬の将来の状態の予測に複数の予測モデルが用いられる場合に、取得部１１は、予測モデルの選択結果を利用者端末装置２０から取得してもよい。予測モデルの選択結果は、例えば、予測結果を利用する人物の操作によって利用者端末装置２０に入力される。また、取得部１１は、予測結果を利用する人物の操作によって利用者端末装置２０に入力される表示項目の選択を、利用者端末装置２０から取得してもよい。 When a plurality of prediction models are used to predict the future state of the racehorse, the acquisition unit 11 may acquire the selection results of the prediction models from the user terminal device 20. The selection result of the prediction model is input to the user terminal device 20 by, for example, an operation by a person who uses the prediction result. Further, the acquisition unit 11 may acquire from the user terminal device 20 a selection of display items that is input to the user terminal device 20 through an operation by a person who uses the prediction results.
　状態予測システム１０が予測モデルを生成する場合に、取得部１１は、予測モデルを生成するための教師データとして、競走馬に関する情報および競走馬の将来の状態を取得してもよい。取得部１１は、例えば、記憶部１５に、取得した競走馬に関する情報および競走馬の将来の状態を保存する。 When the state prediction system 10 generates a prediction model, the acquisition unit 11 may acquire information regarding the racehorse and the future condition of the racehorse as training data for generating the prediction model. The acquisition unit 11 stores, for example, the acquired information regarding the racehorse and the future state of the racehorse in the storage unit 15.
　予測部１２は、競走馬に関する情報から競走馬の将来の状態を予測する予測モデルを用いて、取得した競走馬に関する情報から、競走馬の将来の状態を予測する。また、予測部１２は、予測モデルが、競走馬の将来の状態を予測した理由を、予測理由として抽出する。予測部１２は、例えば、予測モデルが競走馬の将来の状態を予測した際のパラメータを取得する。そして、予測部１２は、競走馬の将来の状態の予測に寄与が大きいパラメータから予測の理由を抽出する。 The prediction unit 12 predicts the future state of the racehorse from the acquired information about the racehorse using a prediction model that predicts the future state of the racehorse from the information about the racehorse. Furthermore, the prediction unit 12 extracts the reason why the prediction model predicted the future state of the racehorse as the prediction reason. For example, the prediction unit 12 acquires parameters used when the prediction model predicts the future state of the racehorse. The prediction unit 12 then extracts the reason for the prediction from the parameters that make a large contribution to predicting the future state of the racehorse.
　予測部１２は、例えば、競走馬の将来の状態として、競走馬のオークション価格と、購入した競走馬の維持費用と、購入した競走馬が獲得する賞金額の少なくとも１つを予測する。競走馬のオークション価格は、競走馬のオークションにおける落札価格である。競走馬の維持費用は、競走馬を継続的に保有するために支払う費用である。競走馬の維持費用は、例えば、競走馬の給餌、健康維持および調教のための費用である。競走馬の維持費用は、購入した競走馬を厩舎に委託する費用であってもよい。競走馬の維持費用は、上記に限られない。購入した競走馬が獲得する賞金の額は、購入した競走馬が将来、レースに出走した際に、獲得する賞金の額である。予測部１２は、競走馬の将来の状態として、購入した競走馬が獲得する賞金の額および維持費用を、馬の年齢ごとに予測してもよい。予測部１２は、購入した競走馬が獲得する賞金の額および維持費用を月ごとに予測してもよい。購入した競走馬が獲得する賞金の額および維持費用を予測する期間は、適宜、設定され得る。 For example, the prediction unit 12 predicts at least one of the auction price of the racehorse, the maintenance cost of the purchased racehorse, and the prize money that the purchased racehorse will win as the future state of the racehorse. The auction price of a racehorse is the winning bid price in an auction of a racehorse. Racehorse maintenance costs are expenses paid to continue owning a racehorse. The maintenance cost of a racehorse is, for example, the cost of feeding, maintaining health, and training the racehorse. The cost of maintaining a racehorse may be the cost of entrusting a purchased racehorse to a stable. The maintenance costs for racehorses are not limited to the above. The amount of prize money that the purchased racehorse will receive is the amount of prize money that the purchased racehorse will receive when it runs in a race in the future. The prediction unit 12 may predict, as the future state of the racehorse, the amount of prize money and maintenance costs that the purchased racehorse will receive for each age of the horse. The prediction unit 12 may predict the amount of prize money and maintenance costs that the purchased racehorse will receive on a monthly basis. The period for predicting the amount of prize money and maintenance costs that a purchased racehorse will receive may be set as appropriate.
　予測部１２は、競走馬の将来の状態として、競走馬のオークション価格と、購入した競走馬の維持費用と、購入した競走馬が獲得する賞金額とを基に、将来における収支を予測してもよい。予測部１２は、例えば、競走馬のオークション価格と、将来のある時点までの購入した競走馬の維持費用、および購入した競走馬が獲得する賞金額を予測する。そして、予測部１２は、購入した競走馬が獲得する賞金額の総額から、競走馬のオークション価格と、購入した競走馬の維持費用の積算額を引くことで、将来のある時点における収支を予測する。予測部１２は、例えば、月ごとに将来における収支を予測する。将来における収支を予測する間隔は、月ごとに限られず、状態予測システム１０の予測結果を利用する人物の好み等に応じて適宜定められればよい。 The prediction unit 12 predicts the future income and expenditure as the future state of the racehorse based on the auction price of the racehorse, the maintenance cost of the purchased racehorse, and the prize money that the purchased racehorse will win. Good too. The prediction unit 12 predicts, for example, the auction price of the racehorse, the maintenance cost of the purchased racehorse up to a certain point in the future, and the prize money that the purchased racehorse will win. The prediction unit 12 then predicts the income and expenditure at a certain point in the future by subtracting the auction price of the racehorse and the accumulated amount of maintenance costs of the purchased racehorse from the total prize money that the purchased racehorse will win. do. The prediction unit 12 predicts future income and expenditure, for example, on a monthly basis. The interval for predicting future income and expenditure is not limited to monthly intervals, and may be determined as appropriate according to the preferences of the person using the prediction results of the state prediction system 10.
　購入済みの競走馬の将来における収支を予測する場合には、予測部１２は、競走馬の購入価格と、購入した競走馬の維持費用と、購入した競走馬が獲得する賞金額を基に、将来における収支を予測してもよい。購入した競走馬の維持費用は、月ごとまたは年ごとにあらかじめ設定された額であってもよい。また、予測モデルが予測する競走馬の将来の状態は、上記の例に限られない。 When predicting the future income and expenditure of a purchased racehorse, the prediction unit 12 uses the purchase price of the racehorse, the maintenance cost of the purchased racehorse, and the prize money that the purchased racehorse will earn. You can also predict future income and expenditure. The maintenance cost for the purchased racehorse may be a predetermined amount per month or year. Further, the future state of the racehorse predicted by the prediction model is not limited to the above example.
　予測部１２は、予測の目的に応じた予測モデルを用いて、競走馬の将来の状態を予測してもよい。予測部１２は、例えば、予測の目的がオークション価格の予測の場合には、競走馬の将来の状態としてオークション価格を予測する予測モデルを用いて、競走馬に関する情報からオークション価格を予測する。また、予測部１２は、例えば、予測の目的が獲得賞金の予測の場合には、競走馬の将来の状態として獲得賞金を予測する予測モデルを用いて、競走馬に関する情報から獲得賞金を予測する。 The prediction unit 12 may predict the future state of the racehorse using a prediction model depending on the purpose of prediction. For example, when the purpose of prediction is to predict an auction price, the prediction unit 12 predicts the auction price from information regarding the racehorse using a prediction model that predicts the auction price as the future state of the racehorse. Further, for example, when the purpose of prediction is to predict the winning prize money, the prediction unit 12 predicts the winning prize money from information about the racehorse using a prediction model that predicts the winning prize money as the future state of the racehorse. .
　予測部１２は、例えば、競走馬の将来の状態として、競走馬の購入の推奨度を予測する予測モデルを用いて、競走馬に関する情報から推奨度を予測してもよい。購入の推奨度は、例えば、競走馬を評価する専門家によってあらかじめ設定された指標を用いて算出される。 For example, the prediction unit 12 may predict the recommendation level from information regarding the racehorse using a prediction model that predicts the recommendation level for purchasing the racehorse as the future state of the racehorse. The degree of recommendation for purchase is calculated using, for example, an index set in advance by an expert who evaluates racehorses.
　予測部１２は、利用者が重視する項目に重みを置いた予測モデルを用いて、競走馬の将来の状態を予測してもよい。複数の予測モデルを用いて競走馬の将来状態を予測する場合に、予測部１２は、各予測モデルの予測結果に重みを付けて予測結果を決定してもよい。予測部１２は、例えば、複数の予測モデルのうち、どの予測モデルの結果を重視するかに応じて、各予測モデルの予測結果の重み付けを行う。また、予測部１２は、予測結果を利用する人物が重視する項目に応じて競走馬の購入の推奨度を予測する予測モデルを用いて、競走馬に関する情報から推奨度を予測してもよい。予測部１２は、例えば、予測結果を利用する人物が血統を重視する場合に、予測の際の血統に関するデータに重みを置いた予測モデルを用いて、競走馬に関する情報から推奨度を予測する。 The prediction unit 12 may predict the future state of the racehorse using a prediction model that places weight on items that are important to the user. When predicting the future state of a racehorse using a plurality of prediction models, the prediction unit 12 may determine the prediction result by weighting the prediction result of each prediction model. The prediction unit 12 weights the prediction results of each prediction model, for example, depending on which of the plurality of prediction models the results of which prediction model is prioritized. Furthermore, the prediction unit 12 may predict the recommendation level from information regarding the racehorse using a prediction model that predicts the recommendation level for purchasing a racehorse according to the items that are important to the person using the prediction results. For example, when a person using the prediction results places emphasis on pedigree, the prediction unit 12 predicts the recommendation level from information about the racehorse using a prediction model that places weight on data related to the pedigree at the time of prediction.
　予測部１２は、予測対象の競走馬の年齢に応じた予測モデルを用いて、競走馬の将来の状態を予測してもよい。予測部１２は、例えば、０歳用の予測モデルと、１歳用の予測モデルと、２歳用の予測モデルとを用いて、競走馬の将来の状態を予測する。年齢に応じた予測モデルを用いて予測を行う場合の年齢の区分は、上記に限られない。 The prediction unit 12 may predict the future state of the racehorse using a prediction model according to the age of the racehorse to be predicted. The prediction unit 12 predicts the future state of the racehorse using, for example, a prediction model for 0-year-olds, a prediction model for 1-year-olds, and a prediction model for 2-year-olds. The age classification when making a prediction using a prediction model according to age is not limited to the above.
　予測部１２は、予測結果を利用する人物の属性に応じた予測モデルを用いて、競走馬の将来の状態を予測してもよい。予測結果を利用する人物の属性は、例えば、初心者、中級者または上級者の区別である。予測結果を利用する人物の属性は、オークション主催者、オークション参加者、馬主、調教師または記者の区分でもよい。予測結果を利用する人物の属性の区分は、上記の例に限られない。 The prediction unit 12 may predict the future state of the racehorse using a prediction model according to the attributes of the person using the prediction results. The attribute of the person who uses the prediction result is, for example, a beginner, an intermediate, or an expert. The attributes of the person who uses the prediction results may be classified as an auction organizer, an auction participant, a horse owner, a trainer, or a reporter. The classification of attributes of persons using prediction results is not limited to the above example.
　出力部１３は、競走馬の将来の状態の予測結果と、予測理由とを出力する。出力部１３は、例えば、利用者端末装置２０に、競走馬の将来の状態の予測結果と、予測理由とを出力する。出力部１３は、状態予測システム１０と接続されている、図示しない表示装置に、予測結果と、予測理由とを出力してもよい。また、出力部１３は、例えば、オークション情報を配信するサーバに、競走馬の将来の状態の予測結果と、予測理由とを出力してもよい。 The output unit 13 outputs the prediction result of the racehorse's future state and the reason for the prediction. The output unit 13 outputs the prediction result of the future state of the racehorse and the reason for the prediction to the user terminal device 20, for example. The output unit 13 may output the prediction result and the reason for prediction to a display device (not shown) connected to the state prediction system 10. Further, the output unit 13 may output the prediction result of the future state of the racehorse and the reason for the prediction to a server that distributes auction information, for example.
　競走馬の将来の状態について複数の項目の予測が行われる場合に、出力部１３は、複数の項目の予測結果を出力してもよい。出力部１３は、例えば、競走馬の将来の状態の予測結果として、オークション価格と、獲得賞金の額を出力する。競走馬の将来の状態の予測結果として、オークション価格と、獲得する賞金額を出力する場合に、出力部１３は、例えば、予測理由として、オークション価格の予測理由と、獲得賞金の額の予測理由とを出力する。また、出力部１３は、オークション価格の価格帯別に、オークション価格以外の項目についての将来の状態の予測結果と、予測理由とを出力してもよい。 When a plurality of items are predicted regarding the future state of the racehorse, the output unit 13 may output the prediction results of the plurality of items. The output unit 13 outputs, for example, the auction price and the amount of the won prize money as a result of predicting the future state of the racehorse. When outputting the auction price and the amount of prize money to be won as the prediction result of the future state of the racehorse, the output unit 13 outputs, for example, the reason for predicting the auction price and the reason for predicting the amount of prize money to be won as the prediction reason. Outputs . Further, the output unit 13 may output prediction results of future states of items other than the auction price and reasons for the prediction for each price range of the auction price.
　予測部１２が将来の収支の予測をする場合に、出力部１３は、将来の収支に関する予測結果を時間経過ごとの情報として出力してもよい。予測部１２が将来の収支の予測をする場合に、出力部１３は、例えば、馬の年齢ごとの収支に関する情報をグラフとして出力する。馬の年齢ごとの収支に関する情報をグラフとして出力する場合に、出力部１３は、例えば、馬の年齢ごとに支出の積算額と、収入の積算額とを示すグラフを出力する。出力部１３は、馬の年齢ごとに支出の積算額と、収入の積算額との差を収支として示したグラフを出力してもよい。また、予測結果の利用者が一口馬主である場合に、出力部１３は、オークション価格に代えて、一口馬主としての出資額を用いて、出資額に応じた支出と、収入の情報を出力してもよい。 When the prediction unit 12 predicts the future income and expenditure, the output unit 13 may output the prediction result regarding the future income and expenditure as information for each lapse of time. When the prediction unit 12 predicts the future income and expenditure, the output unit 13 outputs, for example, information regarding the income and expenditure for each age of the horse as a graph. When outputting information regarding income and expenditure for each age of the horse as a graph, the output unit 13 outputs, for example, a graph showing the cumulative amount of expenditure and the cumulative amount of income for each age of the horse. The output unit 13 may output a graph showing the difference between the cumulative amount of expenditure and the cumulative amount of income for each age of the horse as the income and expenditure. Further, when the user of the prediction result is a single horse owner, the output unit 13 uses the investment amount as a single horse owner instead of the auction price, and outputs information on expenditures and income according to the investment amount. It's okay.
　出力部１３は、予測結果とともに、競走馬に関する情報のうち、競走馬の将来の状態の予測において重視する項目の情報を参考情報として予測結果に付加して出力してもよい。競走馬の将来の状態の予測において重視する項目は、例えば、競走馬の将来の状態に影響を及ぼす頻度が高い項目である。また、競走馬の将来の状態の予測において重視する項目の情報を出力する場合に、出力部１３は、予測理由に該当する項目を強調表示するデータを出力してもよい。出力部１３は、競走馬に関する情報のうち、予測結果を利用する人物が選択した項目を参考情報として出力してもよい。 The output unit 13 may output information on items that are important in predicting the future state of the racehorse, which is added to the prediction result as reference information, among the information regarding the racehorse, along with the prediction result. Items to be emphasized in predicting the future state of a racehorse are, for example, items that have a high frequency of influencing the future state of the racehorse. Furthermore, when outputting information on items that are important in predicting the future state of a racehorse, the output unit 13 may output data that highlights items that correspond to the reason for the prediction. The output unit 13 may output, as reference information, items selected by a person who uses the prediction results from among the information regarding racehorses.
　出力部１３は、予測対象の競走馬の親馬の情報を出力してもよい。出力部１３は、予測対象の競走馬の年齢が基準未満の場合に、予測対象の競走馬の親馬の情報を出力してもよい。出力部１３は、例えば、予測対象の競走馬に関する戦績および育成履歴の情報が少ない年齢の場合に、参考情報として、親馬の戦績および育成履歴の情報を出力する。年齢の基準および出力する項目は、あらかじめ設定される。また、出力部１３は、参考情報として、予測対象の競走馬と属性が類似した競走馬の実績データを出力してもよい。属性が類似した競走馬は、例えば、予測対象の競走馬と、オークション価格と育成履歴が類似している競走馬である。属性が類似した競走馬は、予測対象の競走馬と、血統が類似している競走馬でもよい。属性が類似した競走馬は、上記の例に限られない。出力部１３は、参考情報として、予測対象の競走馬と属性が類似した競走馬が出走したレースの情報と獲得賞金とを出力する。 The output unit 13 may output information about the parent horse of the racehorse to be predicted. The output unit 13 may output information about the parent horse of the racehorse to be predicted, when the age of the racehorse to be predicted is less than a standard. The output unit 13 outputs information on the parent horse's race record and breeding history as reference information, for example, when the racehorse to be predicted is at an age where there is little information on the race record and breeding history. The age standard and the items to be output are set in advance. Furthermore, the output unit 13 may output performance data of racehorses having similar attributes to the racehorse to be predicted, as reference information. A racehorse with similar attributes is, for example, a racehorse whose auction price and breeding history are similar to the racehorse to be predicted. The racehorse with similar attributes may be a racehorse with a similar pedigree to the racehorse to be predicted. Racehorses with similar attributes are not limited to the above example. The output unit 13 outputs, as reference information, information on races in which racehorses having attributes similar to those of the racehorse to be predicted have run, and the winning prize money.
　出力部１３は、参考情報のうち、予測結果を利用する人物が選択した項目を強調表示するデータを出力してもよい。強調表示は、例えば、色、文字の大きさ、文字の太さ、文字の周囲の装飾の少なくとも１つを他の項目と変えることで行われる。出力部１３は、例えば、年齢、性別、体重、父馬、母馬、生産者および調教履歴のうち少なくとも１つ以上の項目を参考情報として出力する。また、出力部１３は、参考情報の項目のうち、予測理由に該当する項目がある場合に、予測理由に該当する強調表示するデータを出力してもよい。 The output unit 13 may output data that highlights items selected by the person using the prediction results among the reference information. Highlighting is performed, for example, by changing at least one of the color, font size, font thickness, and decoration around the font from other items. The output unit 13 outputs, for example, at least one item among age, sex, weight, father horse, mother horse, producer, and training history as reference information. Moreover, the output unit 13 may output highlighted data corresponding to the prediction reason when there is an item corresponding to the prediction reason among the reference information items.
　出力部１３は、予測理由として、予測結果に対して負の要因となる項目を出力してもよい。負の要因は、競走馬の将来の状態に影響が大きい項目のうち、競走馬としての評価が下がる予測への影響が大きい項目である。影響が大きいとは、例えば、ある項目を変化させた場合に、他の項目より、競走馬の将来の状態の予測結果の変動が大きいことをいう。公営競技が競馬の場合に、例えば、親馬に怪我または病気の履歴がある場合に、健康状態に関する項目が負の要因として抽出される。 The output unit 13 may output an item that is a negative factor for the prediction result as the reason for the prediction. Negative factors are items that have a large influence on the prediction that the racehorse's evaluation will be lowered, among the items that have a large influence on the future state of the racehorse. A large influence means, for example, that when a certain item is changed, the fluctuation in the prediction result of the future state of the racehorse is larger than that of other items. When the publicly managed competition is horse racing, for example, if the parent horse has a history of injury or illness, items related to health status are extracted as negative factors.
　出力部１３は、予測理由を文章として出力してもよい。出力部１３は、例えば、競走馬の将来の状態の予測への影響度が高い競走馬に関する情報に基づいて、予測理由を示す文章を出力する。例えば、競走馬の将来の状態の予測への影響度が高い競走馬に関する情報と、予測理由を示す文章との関係は、あらかじめ設定されている。公営競技が競馬の場合において、予測理由が獲得賞金の場合に、出力部１３は、例えば、予測理由として「多くの獲得賞金を見込める馬のためおすすめです。」のような文章を出力してもよい。 The output unit 13 may output the prediction reason as a text. The output unit 13 outputs a sentence indicating the reason for the prediction, for example, based on information regarding a racehorse that has a high degree of influence on the prediction of the future state of the racehorse. For example, the relationship between information about a racehorse that has a high degree of influence on the prediction of the future state of the racehorse and a sentence indicating the reason for the prediction is set in advance. In the case where the publicly managed competition is horse racing, and the reason for prediction is winning prize money, the output unit 13 may output a sentence such as "This horse is recommended because it is expected to win a lot of prize money" as the reason for prediction. good.
　出力部１３は、予測結果と、予測理由とともに、競走馬の画像を出力してもよい。また、出力部１３は、予測結果と、予測理由とともに、競走馬の調教時の画像を出力してもよい。また、出力部１３は、予測対象の競走馬の親馬または兄弟馬の画像を出力してもよい。画像は、静止画または動画のいずれであってもよい。 The output unit 13 may output an image of the racehorse along with the prediction result and the reason for the prediction. Further, the output unit 13 may output an image of the racehorse being trained together with the prediction result and the reason for the prediction. Further, the output unit 13 may output an image of the parent horse or sibling horse of the racehorse to be predicted. The image may be a still image or a moving image.
　出力部１３は、競走馬の画像に、予測結果と、予測理由とを重畳して出力してもよい。また、出力部１３は、予測結果と、予測理由の一方を、競走馬の画像上に重畳して出力してもよい。 The output unit 13 may output the prediction result and the reason for the prediction superimposed on the image of the racehorse. Further, the output unit 13 may output either the prediction result or the prediction reason superimposed on the image of the racehorse.
　図３は、競走馬の将来の状態の予測結果の表示画面の例を示す。図３の表示画面の例は、競走馬の将来の状態の予測結果として、オークション価格を予測する場合における予測結果の表示画面である。図３の表示画面の例では、予測対象の競走馬について、馬名、予想価格および理由が一覧として表示されている。図３の表示画面の例において、予想価格は、予測モデルが予測したオークション価格である。図３の表示画面の例において、予測モデルが予測結果とともに出力する予測理由である。予測理由は、複数の項目であってもよい。 FIG. 3 shows an example of a display screen of a prediction result of a racehorse's future state. The example of the display screen in FIG. 3 is a display screen of a prediction result when an auction price is predicted as a prediction result of a racehorse's future state. In the example of the display screen in FIG. 3, horse names, predicted prices, and reasons are displayed as a list for racehorses to be predicted. In the example display screen of FIG. 3, the predicted price is the auction price predicted by the prediction model. In the example of the display screen in FIG. 3, this is the prediction reason that the prediction model outputs together with the prediction result. The prediction reason may include multiple items.
　図４は、図３の表示画面の例において、競走馬が獲得する賞金の額をさらに表示する表示画面の例を示す。図４の表示画面の例では、オークション価格の予測結果が予想価格、獲得する賞金額の予測結果が獲得金額予想として表示されている。 FIG. 4 shows an example of a display screen that further displays the amount of prize money won by a racehorse in the example of the display screen of FIG. 3. In the example of the display screen in FIG. 4, the prediction result of the auction price is displayed as the expected price, and the prediction result of the prize amount to be won is displayed as the prediction of the amount to be won.
　図５は、図３の表示画面の例において、予測対象の競走馬に関する情報をさらに表示する表示画面の例を示す。図５の表示画面の例では、予測対象の競走馬の生産者、父馬および母馬の名称がさらに表示されている。予測結果ともに表示する競走馬に関する情報は、上記の例に限られない。 FIG. 5 shows an example of a display screen that further displays information regarding the racehorse to be predicted in the example of the display screen of FIG. 3. In the example of the display screen in FIG. 5, the names of the breeder, father horse, and mother horse of the racehorse to be predicted are further displayed. The information regarding racehorses that is displayed together with the prediction results is not limited to the above example.
　図６は、競走馬に関する収支の変化をグラフとして表示する表示画面の例を示す。図６の表示画面の例では、横軸を予測対象の競走馬の年齢を示す馬齢、縦軸を支出と収入の額を示す金額として設定したグラフが表示されている。図６の表示画面の例では、オークション価格と、維持に必要な経費の積算額の合計値を示す支出と、獲得賞金の積算額を示す収入がグラフとして表示されている。図６の表示画面の例のグラフのうち、破線で示されるグラフは、オークション価格と、維持に必要な経費との合計の積算額を示す。図６の表示画面の例のグラフのうち、実線で示されるグラフは、獲得賞金の積算額を示す。図６の表示画面の例のグラフ上には、予測理由が重畳して表示されてもよい。また、図６の表示画面の例のグラフは、他の表示画面の例に組み合わされて表示されてもよい。 FIG. 6 shows an example of a display screen that displays changes in income and expenditure regarding racehorses as a graph. In the example of the display screen in FIG. 6, a graph is displayed in which the horizontal axis represents the age of the racehorse to be predicted, and the vertical axis represents the amount of money representing expenditures and income. In the example of the display screen shown in FIG. 6, the auction price, expenditures representing the cumulative amount of expenses necessary for maintenance, and income representing the cumulative amount of won prizes are displayed as a graph. Among the graphs in the example display screen of FIG. 6, the graph indicated by a broken line indicates the cumulative amount of the auction price and the expenses necessary for maintenance. Among the graphs in the example of the display screen in FIG. 6, the graph indicated by a solid line indicates the cumulative amount of earned prize money. The reason for prediction may be displayed in a superimposed manner on the graph in the example display screen of FIG. Furthermore, the graph in the example display screen of FIG. 6 may be displayed in combination with another example display screen.
　図７は、図３の表示画面の例において、予測理由として負の要因をさらに表示する表示画面の例を示す。図７の表示画面の例において、プラス要因は、正の要因を示す。正の要因は、オークション価格の予測結果に影響が大きい項目のうち、価格が高くなる予測への影響が大きい項目である。図７の表示画面の例では、負の要因がマイナス要因として示されている。負の要因は、オークション価格の予測結果に影響が大きい項目のうち、価格が低くなる予測への影響が大きい項目である。 FIG. 7 shows an example of a display screen that further displays negative factors as prediction reasons in the example of the display screen of FIG. 3. In the example of the display screen in FIG. 7, the plus factor indicates a positive factor. A positive factor is an item that has a large influence on the prediction that the price will be high among the items that have a large influence on the prediction result of the auction price. In the example display screen of FIG. 7, negative factors are shown as negative factors. Negative factors are items that have a large influence on the prediction that the price will be low, among the items that have a large influence on the prediction result of the auction price.
　図８は、図３の表示画面の例において、オークション価格が低価格な競走馬について、オークション価格の予想結果と、オークション価格が低価格である予測理由とともに、予測理由とは反対の高い評価につながる要因を示す表示画面の例を示す。図８の表示画面の例は、予測価格が低価格な競走馬について、低価格である理由と、評価が高くなる要因とを表示する表示画面である。図８の表示画面の例では、オークション価格の予測結果を示す予想価格と、予測理由を示す価格理由が表示されている。また、図８の表示画面の例では、予測結果が高くなる要因が評価理由として表示されている。評価が高くなる要因は、例えば、オークション価格が低いのに対し、獲得賞金が高くなる要因になる項目である。例えば、健康状態がオークション価格を下げているが、調教タイムが伸びていることで獲得賞金が叩くなることが期待されている場合に、評価が高くなる要因として、調教タイムが表示される。 FIG. 8 shows, in the example of the display screen in FIG. 3, the predicted auction price of a racehorse with a low auction price, the predicted reason why the auction price is low, and the high evaluation that is opposite to the predicted reason. An example of a display screen showing related factors is shown. The example of the display screen in FIG. 8 is a display screen that displays, for a racehorse with a low predicted price, the reason why the price is low and the factors that increase the evaluation. In the example of the display screen in FIG. 8, an expected price indicating the prediction result of the auction price and a price reason indicating the reason for the prediction are displayed. Further, in the example of the display screen in FIG. 8, factors that increase the prediction result are displayed as evaluation reasons. A factor that increases the evaluation is, for example, an item that causes the winning prize to be high while the auction price is low. For example, if the health condition is lowering the auction price, but the increased training time is expected to increase the prize money, the training time will be displayed as a factor that will increase the evaluation.
　図９は、図３の表示画面の例において、参考情報をさらに表示する表示画面の例を示す。図９の表示画面の例では、左の枠内に、予測結果と、予測理由が表示されている。図９の表示画面の例では、右の枠内に、Ｂの競走馬に関する情報が参考情報として示されている。図９の表示画面の例では、参考情報として、馬齢、生産者、調教師、血統、体重および調教履歴が表示されている。参考情報として表示する項目は、上記に限られない。また、図９の表示画面の例において、着順予測の欄で、いずれかの出走馬が選択されると、参考情報の欄に、選択された出走馬に関する参考情報が表示されるようにしてもよい。 FIG. 9 shows an example of a display screen that further displays reference information in the example of the display screen of FIG. 3. In the example of the display screen in FIG. 9, the prediction result and the reason for the prediction are displayed in the left frame. In the example of the display screen in FIG. 9, information regarding racehorse B is shown as reference information in the right frame. In the example of the display screen in FIG. 9, the horse's age, breeder, trainer, pedigree, weight, and training history are displayed as reference information. Items displayed as reference information are not limited to the above. In addition, in the example of the display screen in FIG. 9, when any runner is selected in the finish prediction column, reference information about the selected runner is displayed in the reference information column. Good too.
　図１０は、図９の表示画面の例において、競走馬の画像を出力する表示画面の例を示す。図１０の表示画面の例では、下段の左側に競走馬の画像が表示されている。競走馬の画像は、予想結果の欄で、いずれかの競走馬が選択されると、画像の表示箇所に、選択された競走馬の画像が表示されるようにしてもよい。また、予測対象の競走馬の画像に代えて、予測対象の競走馬の親馬の画像が表示されるようにしてもよい。予測対象の競走馬の親馬の画像は、過去に親馬がレースに出走した時の画像であってもよい。また、競走馬の画像として、調教時の映像が表示されるようにしてもよい。競走馬の画像は、例えば、情報管理サーバ３０から取得される。 FIG. 10 shows an example of a display screen that outputs an image of a racehorse in the example of the display screen of FIG. 9. In the example of the display screen in FIG. 10, an image of a racehorse is displayed on the left side of the lower row. When any racehorse is selected in the prediction result column, the image of the selected racehorse may be displayed in the image display area. Further, instead of the image of the racehorse to be predicted, an image of the parent horse of the racehorse to be predicted may be displayed. The image of the parent horse of the racehorse to be predicted may be an image when the parent horse ran in a race in the past. Furthermore, a video during training may be displayed as the image of the racehorse. The image of the racehorse is acquired from the information management server 30, for example.
　図１１は、図１０の表示画面の例において、競走馬の画像上に、予測結果と予測理由が重畳して表示されている表示画面の例を示す。図１１の表示画面の例では、予測理由として、血統が良好であることが表示されている。競走馬の画像上に表示される予測理由は、複数の項目であってもよい。また、予測結果と予測理由のいずれか一方が競走馬の画像上に重畳して表示されてもよい。 FIG. 11 shows an example of a display screen in which a prediction result and a reason for prediction are displayed superimposed on an image of a racehorse in the example of the display screen of FIG. 10. In the example of the display screen in FIG. 11, a good pedigree is displayed as the reason for prediction. The prediction reason displayed on the image of the racehorse may be a plurality of items. Further, either the prediction result or the prediction reason may be displayed superimposed on the image of the racehorse.
　状態予測システム１０において予測モデルを生成する場合に、生成部１４は、競走馬に関する情報から競走馬の将来の状態を予測する予測モデルを生成する。生成部１４は、例えば、競走馬に関する情報と、競走馬の将来の状態との関係を学習し、競走馬に関する情報から競走馬の将来の状態を予測する予測モデルを生成する。生成部１４は、競走馬の将来の状態について、複数の項目を予測する予測モデルを生成してもよい。生成部１４は、例えば、オークション価格と、獲得賞金とを予測する予測モデルを生成する。 When generating a prediction model in the state prediction system 10, the generation unit 14 generates a prediction model that predicts the future state of the racehorse from information regarding the racehorse. For example, the generation unit 14 learns the relationship between information about racehorses and the future state of the racehorse, and generates a prediction model that predicts the future state of the racehorse from the information about the racehorse. The generation unit 14 may generate a prediction model that predicts a plurality of items regarding the future state of the racehorse. The generation unit 14 generates, for example, a prediction model that predicts an auction price and an acquired prize.
　予測部１２が複数の予測モデルのうちいずれかによって競走馬の将来の状態を予測する場合に、生成部１４は、予測部１２が予測に用いる複数の予測モデルをそれぞれ生成してもよい。 When the prediction unit 12 predicts the future state of the racehorse using one of the plurality of prediction models, the generation unit 14 may generate each of the plurality of prediction models used by the prediction unit 12 for prediction.
　０歳用の予測モデルを生成する場合に、生成部１４は、例えば、親馬の戦績を含む競走馬に関する情報と、競走馬の将来の状態との関係を学習して、予測モデルを生成する。１歳用の予測モデルを生成する場合に、生成部１４は、例えば、競走馬の生体情報を含む競走馬に関する情報と、競走馬の将来の状態との関係を学習して、予測モデルを生成する。競走馬の生体情報は、例えば、性別、毛並み、歩容、体重、体長、体重の変化、体調の変化、血液データおよび性格のうち１つ以上である。競走馬の生体情報は、上記に限られない。２歳用の予測モデルを生成する場合に、生成部１４は、例えば、調教履歴を含む競走馬に関する情報と、将来の状態との関係を学習して、予測モデルを生成する。 When generating a prediction model for 0-year-olds, the generation unit 14, for example, learns the relationship between information about the racehorse, including the race record of the parent horse, and the future state of the racehorse, and generates the prediction model. . When generating a prediction model for a one-year-old, the generation unit 14, for example, learns the relationship between information about the racehorse, including biological information of the racehorse, and the future state of the racehorse, and generates the prediction model. do. The biological information of the racehorse includes, for example, one or more of sex, coat, gait, weight, body length, change in weight, change in physical condition, blood data, and personality. The biological information of racehorses is not limited to the above. When generating a prediction model for a 2-year-old horse, the generation unit 14 generates the prediction model by learning, for example, the relationship between information about the racehorse including its training history and its future state.
　予測結果の利用者が重視する項目に応じた予測モデルを生成する場合に、生成部１４は、例えば、重視する項目に関する情報を含む競走馬に関する情報と、競走馬の将来の状態との関係を学習して、予測モデルを生成する。血統を重視する予測モデルを生成する場合に、生成部１４は、例えば、親馬の生体情報と、親馬の戦績とを含む競走馬に関する情報と、競走馬の将来の状態との関係を学習して、予測モデルを生成する。予測結果の利用者が重視する項目に応じた予測モデルを生成する場合に、生成部１４は、利用者が重視する項目の重みを大きくして予測モデルを生成してもよい。利用者が重視する項目の重みを大きくする場合に、生成部１４は、例えば、利用者が重視する項目に関する特徴量の係数が、他の項目に関する特徴量の係数よりも大きくなるように予測モデルを生成する。 When generating a prediction model according to the items that the user of the prediction results places importance on, the generation unit 14 generates, for example, the relationship between information about the racehorse, including information on the items that the user places emphasis on, and the future state of the racehorse. Learn and generate predictive models. When generating a predictive model that emphasizes pedigree, the generation unit 14 learns, for example, the relationship between the biological information of the parent horse, information about the racehorse including the parent horse's race record, and the future state of the racehorse. and generate a predictive model. When generating a prediction model according to the items of the prediction results that the user considers important, the generation unit 14 may generate the prediction model by increasing the weight of the items that the user considers important. When increasing the weight of the item that the user places importance on, the generation unit 14 generates a prediction model such that, for example, the coefficient of the feature amount regarding the item that the user places importance on is larger than the coefficient of the feature amount regarding other items. generate.
　生成部１４は、例えば、因子化漸近ベイズ推論を基にした学習アルゴリズムを用いて予測モデルを生成する。因子化漸近ベイズ推論を基にした学習アルゴリズムを用いて学習を行う際に、生成部１４は、競走馬に関する情報を入力データ、競走馬の将来の状態を正解データとして決定木形式のルールによって場合分けする。そして、生成部１４は、各場合で異なる説明変数を組み合わせた線形モデルを用いて競走馬の将来の状態を予測する学習モデルを生成する。生成部１４は、データの場合分け条件の最適化、説明変数の組み合わせの最適化のよる予測モデルの生成、および不要な予測モデルの削除の処理を順に行うことで学習モデルを生成する。このような学習モデルの生成方法は、異なる説明変数の組み合わせによる予測モデルを組み合わせて予測するため、異種混合学習とも呼ばれる。異種混合学習によって予測モデルを生成することで、予測結果への影響が強い場合分けの条件を用いて競走馬の将来の状態の予測結果を説明することが可能になるため、予測結果の説明性が向上する。異種混合学習の手法は、例えば、米国特許出願公開第２０１４／０２２２７４１号明細書に開示されている。 The generation unit 14 generates a predictive model using, for example, a learning algorithm based on factorized asymptotic Bayesian inference. When performing learning using a learning algorithm based on factorized asymptotic Bayesian inference, the generation unit 14 uses information about the racehorse as input data and the future state of the racehorse as correct answer data, and generates a case according to decision tree-style rules. Divide. The generation unit 14 then generates a learning model that predicts the future state of the racehorse using a linear model that combines different explanatory variables in each case. The generation unit 14 generates a learning model by sequentially optimizing data case classification conditions, generating a predictive model by optimizing combinations of explanatory variables, and deleting unnecessary predictive models. This method of generating a learning model is also called heterogeneous mixture learning because it makes predictions by combining prediction models based on combinations of different explanatory variables. By generating a prediction model using heterogeneous mixture learning, it becomes possible to explain the prediction results of the future state of the racehorse using conditions that have a strong influence on the prediction results, which improves the explainability of the prediction results. will improve. A method of heterogeneous mixture learning is disclosed in, for example, US Patent Application Publication No. 2014/0222741.
　予測モデルを生成する機械学習に用いる学習アルゴリズムは、上記の例に限られない。生成部１４は、例えば、競走馬に関する情報から競走馬の将来の状態を予測する学習モデルを、ニューラルネットワークを用いた深層学習によって生成してもよい。このような学習モデルを生成する際に、生成部１４は、例えば、各項目のデータを変動させ、競走馬の将来の状態の変化を基に、競走馬の将来の状態への影響の大きい項目を予測の理由として抽出する予測モデルを生成する。そして、生成部１４は、各項目のデータを変動させ、競走馬の将来の状態への影響が大きい項目を予測の理由として抽出する。 The learning algorithm used for machine learning to generate a predictive model is not limited to the above example. For example, the generation unit 14 may generate a learning model that predicts the future state of a racehorse from information about the racehorse by deep learning using a neural network. When generating such a learning model, the generation unit 14, for example, varies the data of each item and selects items that have a large impact on the future condition of the racehorse based on changes in the future condition of the racehorse. Generate a prediction model that extracts as the reason for prediction. Then, the generation unit 14 changes the data of each item and extracts an item that has a large influence on the future state of the racehorse as the reason for prediction.
　記憶部１５は、例えば、予測モデルを保存する。複数の予測モデルが用いられる場合には、記憶部１５は、複数の予測モデルを保存する。また、状態予測システム１０が予測モデルを生成する場合に、記憶部１５は、教師データとして用いる、競走馬に関する情報と、競走馬の将来の状態とを関連付けたデータを保存してもよい。また、予測結果に参考情報を付加する場合に、記憶部１５は、参考情報として用いるデータを保存してもよい。なお、予測部１２が用いる予測モデルは、記憶部１５以外の記憶手段に保存されていてもよい。 The storage unit 15 stores, for example, a prediction model. When a plurality of prediction models are used, the storage unit 15 stores the plurality of prediction models. Further, when the state prediction system 10 generates a prediction model, the storage unit 15 may store data that is used as teacher data and associates information regarding the racehorse with the future state of the racehorse. Further, when adding reference information to the prediction result, the storage unit 15 may store data used as the reference information. Note that the prediction model used by the prediction unit 12 may be stored in a storage means other than the storage unit 15.
　利用者端末装置２０は、状態予測システム１０から、予測結果と、予測理由とを取得する。そして、利用者端末装置２０は、例えば、図示しない表示装置に、予測結果と、予測理由と出力する。 The user terminal device 20 acquires the prediction result and the reason for the prediction from the state prediction system 10. Then, the user terminal device 20 outputs the prediction result and the reason for the prediction to, for example, a display device (not shown).
　利用者によって予測モデルの選択が行われる場合に、利用者端末装置２０は、例えば、予測モデルの選択結果として、利用者の操作によって入力される予測モデルの名称を取得する。そして、利用者端末装置２０は、状態予測システム１０に、入力された予測モデルの名称を出力する。 When a prediction model is selected by the user, the user terminal device 20 obtains, for example, the name of the prediction model input by the user's operation as the selection result of the prediction model. Then, the user terminal device 20 outputs the name of the input prediction model to the state prediction system 10.
　利用者端末装置２０には、例えば、スマートフォン、タブレット型コンピュータ、ノート型コンピュータまたはデスクトップ型コンピュータが用いられる。利用者端末装置２０に用いられる端末装置は、上記の例に限られない。 For example, a smartphone, a tablet computer, a notebook computer, or a desktop computer is used as the user terminal device 20. The terminal device used for the user terminal device 20 is not limited to the above example.
　情報管理サーバ３０は、例えば、競走馬に関する情報を保存または管理するサーバである。情報管理サーバ３０は、競走馬に関する情報の内容に応じて設置されている複数のサーバであってもよい。競走馬に関する情報は、情報管理サーバ３０が管理する記憶装置に保存されてもよい。また、情報管理サーバ３０は競走馬の画像を保存してもよい。 The information management server 30 is, for example, a server that stores or manages information regarding racehorses. The information management server 30 may be a plurality of servers installed according to the content of information regarding racehorses. Information regarding racehorses may be stored in a storage device managed by the information management server 30. Further, the information management server 30 may store images of racehorses.
　競走馬予測システムの状態予測システム１０において、競走馬の将来の状態を予測する際の動作について説明する。図１２は、状態予測システム１０が競走馬の将来の状態を予測する際の動作フローの例を示す図である。 In the state prediction system 10 of the racehorse prediction system, the operation when predicting the future state of a racehorse will be explained. FIG. 12 is a diagram showing an example of an operation flow when the state prediction system 10 predicts the future state of a racehorse.
　取得部１１は、競走馬の将来の状態を予測する対象となる競走馬に関する情報を取得する（ステップＳ１１）。取得部１１は、例えば、情報管理サーバ３０から、競走馬に関する情報を取得する。 The acquisition unit 11 acquires information regarding a racehorse whose future state is to be predicted (step S11). The acquisition unit 11 acquires information regarding racehorses from the information management server 30, for example.
　競走馬に関する情報が取得されると、予測部１２は、競走馬に関する情報から競走馬の将来の状態を予測する予測モデルを用いて、取得部１１が取得した競走馬に関する情報から、競走馬の将来の状態を予測する（ステップＳ１２）。 When the information about the racehorse is acquired, the prediction unit 12 uses the prediction model that predicts the future state of the racehorse from the information about the racehorse to predict the future state of the racehorse from the information about the racehorse acquired by the acquisition unit 11. The future state is predicted (step S12).
　競走馬の将来の状態が予測されると、出力部１３は、予測結果と、予測理由とを出力する（ステップＳ１３）。出力部１３は、例えば、利用者端末装置２０に、予測結果と、予測理由とを出力する。 When the future state of the racehorse is predicted, the output unit 13 outputs the prediction result and the reason for the prediction (step S13). The output unit 13 outputs the prediction result and the reason for the prediction to the user terminal device 20, for example.
　予測結果と、予測理由とを受け取った利用者端末装置２０は、例えば、表示装置に、予測結果と、予測理由とを表示する。 The user terminal device 20 that has received the prediction result and the prediction reason displays the prediction result and the prediction reason on the display device, for example.
　状態予測システム１０において、予測モデルを生成する際の動作について説明する。図１３は、状態予測システム１０が予測モデルを生成する際の動作フローの例を示す図である。 The operation when generating a prediction model in the state prediction system 10 will be explained. FIG. 13 is a diagram illustrating an example of an operation flow when the state prediction system 10 generates a prediction model.
　取得部１１は、競走馬に関する情報と、競走馬の将来の状態とを取得する（ステップＳ２１）。競走馬に関する情報と、競走馬の将来の状態とを取得すると、生成部１４は、競走馬に関する情報と、競走馬の将来の状態との関係を学習し、競走馬に関する情報から競走馬の将来の状態を予測する予測モデルを生成する（ステップＳ２２）。予測モデルを生成すると、生成部１４は、生成した予測モデルを記憶部１５に保存する（ステップＳ２３）。 The acquisition unit 11 acquires information regarding the racehorse and the future state of the racehorse (step S21). When the information about the racehorse and the future state of the racehorse are acquired, the generation unit 14 learns the relationship between the information about the racehorse and the future state of the racehorse, and calculates the future state of the racehorse from the information about the racehorse. A prediction model that predicts the state of is generated (step S22). After generating the prediction model, the generation unit 14 stores the generated prediction model in the storage unit 15 (step S23).
　本実施形態の競走馬予測システムの状態予測システム１０は、競走馬に関する情報を取得し、予測モデルを用いて、競走馬の将来の状態を予測している。そして、状態予測システム１０は、例えば、利用者端末装置２０に、予測結果と、予測理由を出力する。競走馬の将来の状態の予測結果とともに、予測理由を出力することで、予測結果を利用する人物は、競走馬の将来の状態の予測結果を容易に解釈することができる。よって、状態予測システム１０を用いることで、競走馬の将来の状態の予測結果の解釈を容易にすることができる。 The state prediction system 10 of the racehorse prediction system of this embodiment acquires information about racehorses and uses a prediction model to predict the future state of the racehorses. Then, the state prediction system 10 outputs the prediction result and the reason for the prediction to the user terminal device 20, for example. By outputting the reason for the prediction together with the prediction result of the future state of the racehorse, a person using the prediction result can easily interpret the prediction result of the future state of the racehorse. Therefore, by using the condition prediction system 10, it is possible to easily interpret the results of predicting the future condition of the racehorse.
　競走馬の収支に関する情報を出力する場合に、状態予測システム１０は、例えば、オークション価格および維持費用と、獲得賞金との差を収支として示すことで、予測結果を利用する人物が競走馬の儲けを検討する際に参考とする情報を出力することができる。また、競走馬の収支に関する情報をグラフとして出力する場合に、状態予測システム１０は、利用者が視覚的に理解しやすい形式で、予測結果である収支に関する情報を出力することができる。 When outputting information regarding the racehorse's income and expenditure, the condition prediction system 10 can, for example, show the difference between the auction price and maintenance costs and the won prize money as the income and expenditure, so that a person using the prediction results can see the racehorse's profit. It is possible to output information for reference when considering. Further, when outputting information regarding the income and expenditure of a racehorse as a graph, the state prediction system 10 can output information regarding the income and expenditure, which is a prediction result, in a format that is easy for the user to visually understand.
　複数の予測モデルを用いる場合に、状態予測システム１０は、例えば、予測結果を利用する人物が重視する項目に応じた予測モデルを用いることで、予測結果を利用する人物が重視する項目に応じた予測結果とともに、予測理由を出力することができる。また、競走馬の年齢に応じた予測モデルを用いる場合には、状態予測システム１０は、例えば、予競走馬の年齢に応じた適切な予測結果と、予測理由を出力することができる。 When using a plurality of prediction models, the state prediction system 10 can, for example, use a prediction model that corresponds to the items that the person using the prediction results places importance on, so that the state prediction system 10 can use the prediction model that corresponds to the items that the person using the prediction results places importance on. The reason for the prediction can be output together with the prediction result. Moreover, when using a prediction model according to the age of the racehorse, the state prediction system 10 can output an appropriate prediction result and the reason for the prediction according to the age of the preliminary racehorse, for example.
　予測結果とともに、参考情報を出力する場合に、状態予測システム１０が、予測結果と、予測理由に、さらに参考情報を付加して出力することで、予測結果を利用する人物は、予測理由と参考情報を参照して、より容易に予測理由を解釈することができる。 When outputting reference information together with the prediction result, the state prediction system 10 adds and outputs the reference information to the prediction result and the reason for the prediction, so that a person using the prediction result can understand the reason for the prediction and the reference information. You can refer to the information to more easily interpret the reason for the prediction.
　（第２の実施形態）
　本発明の第２の実施形態について図を参照して詳細に説明する。図１４は、本実施形態の状態予測システム１００の構成の例を示す図である。状態予測システム１００は、取得部１０１と、予測部１０２と、出力部１０３を備える。 (Second embodiment)
A second embodiment of the present invention will be described in detail with reference to the drawings. FIG. 14 is a diagram showing an example of the configuration of the state prediction system 100 of this embodiment. The state prediction system 100 includes an acquisition section 101, a prediction section 102, and an output section 103.
　取得部１０１は、競走馬に関する情報を取得する。予測部１０２は、競走馬に関する情報から競走馬の将来の状態を予測する予測モデルを用いて、取得した競走馬に関する情報から、競走馬の将来の状態を予測する。出力部１０３は、予測の結果と、予測の理由とを出力する。 The acquisition unit 101 acquires information regarding racehorses. The prediction unit 102 predicts the future state of the racehorse from the acquired information about the racehorse using a prediction model that predicts the future state of the racehorse from the information about the racehorse. The output unit 103 outputs the prediction result and the reason for the prediction.
　ここで、第１の実施形態の取得部１１は、取得部１０１の一例である。また、取得部１０１は、取得手段の一態様である。第１の実施形態の予測部１２は、予測部１０２の一例である。また、予測部１０２は、予測手段の一態様である。第１の実施形態の出力部１３は、出力部１０３の一例である。また、出力部１０３は、出力手段の一態様である。 Here, the acquisition unit 11 of the first embodiment is an example of the acquisition unit 101. Further, the acquisition unit 101 is one aspect of an acquisition means. The prediction unit 12 of the first embodiment is an example of the prediction unit 102. Furthermore, the prediction unit 102 is one aspect of prediction means. The output unit 13 of the first embodiment is an example of the output unit 103. Furthermore, the output unit 103 is one aspect of output means.
　状態予測システム１００の動作について説明する。図１５は、状態予測システム１００の動作フローの例を示す図である。 The operation of the state prediction system 100 will be explained. FIG. 15 is a diagram illustrating an example of the operation flow of the state prediction system 100.
　取得部１０１は、競走馬に関する情報を取得する（ステップＳ１０１）。競走馬に関する情報が取得されると、予測部１０２は、競走馬に関する情報から競走馬の将来の状態を予測する予測モデルを用いて、取得した競走馬に関する情報から、競走馬の将来の状態を予測する（ステップＳ１０２）。競走馬の将来の状態が予測されると、出力部１０３は、予測の結果と、予測の理由とを出力する（ステップＳ１０３）。 The acquisition unit 101 acquires information regarding racehorses (step S101). When the information regarding the racehorse is acquired, the prediction unit 102 predicts the future condition of the racehorse from the acquired information regarding the racehorse using a prediction model that predicts the future condition of the racehorse from the information regarding the racehorse. Predict (step S102). When the future state of the racehorse is predicted, the output unit 103 outputs the prediction result and the reason for the prediction (step S103).
　本実施形態の状態予測システム１００は、予測モデルを用いて競走馬の将来の状態を予測する。そして、状態予測システム１０は、競走馬の将来の状態の予測結果と、予測理由とを出力する。その結果、状態予測システム１０は、競走馬の将来の状態の予測結果の解釈を容易にすることができる。 The state prediction system 100 of this embodiment predicts the future state of a racehorse using a prediction model. Then, the condition prediction system 10 outputs the prediction result of the future condition of the racehorse and the reason for the prediction. As a result, the condition prediction system 10 can facilitate the interpretation of prediction results of the racehorse's future condition.
　第１の実施形態の状態予測システム１０および第２の実施形態の状態予測システム１００における各処理は、コンピュータプログラムをコンピュータで実行することによって実現することができる。図１６は、第１の実施形態の状態予測システム１０および第２の実施形態の状態予測システム１００における各処理を行うコンピュータプログラムを実行するコンピュータ２００の構成の例を示したものである。コンピュータ２００は、ＣＰＵ（Central Processing Unit）２０１と、メモリ２０２と、記憶装置２０３と、入出力Ｉ／Ｆ（Interface）２０４と、通信Ｉ／Ｆ２０５を備える。 Each process in the state prediction system 10 of the first embodiment and the state prediction system 100 of the second embodiment can be realized by executing a computer program on a computer. FIG. 16 shows an example of the configuration of a computer 200 that executes a computer program that performs each process in the state prediction system 10 of the first embodiment and the state prediction system 100 of the second embodiment. The computer 200 includes a CPU (Central Processing Unit) 201, a memory 202, a storage device 203, an input/output I/F (Interface) 204, and a communication I/F 205.
　ＣＰＵ２０１は、記憶装置２０３から各処理を行うコンピュータプログラムを読み出して実行する。ＣＰＵ２０１は、複数のＣＰＵの組み合わせによって構成されていてもよい。また、ＣＰＵ２０１は、ＣＰＵと他の種類のプロセッサの組み合わせによって構成されていてもよい。例えば、ＣＰＵ２０１は、ＣＰＵとＧＰＵ（Graphics Processing Unit）の組み合わせによって構成されていてもよい。メモリ２０２は、ＤＲＡＭ（Dynamic Random Access Memory）等によって構成され、ＣＰＵ２０１が実行するコンピュータプログラムや処理中のデータが一時記憶される。記憶装置２０３は、ＣＰＵ２０１が実行するコンピュータプログラムを記憶している。記憶装置２０３は、例えば、不揮発性の半導体記憶装置によって構成されている。記憶装置２０３には、ハードディスクドライブ等の他の記憶装置が用いられてもよい。入出力Ｉ／Ｆ２０４は、作業者からの入力の受付および表示データ等の出力を行うインタフェースである。通信Ｉ／Ｆ２０５は、利用者端末装置２０および情報管理サーバ３０との間でデータの送受信を行うインタフェースである。また、利用者端末装置２０および情報管理サーバ３０も同様の構成としてもよい。 The CPU 201 reads computer programs for performing each process from the storage device 203 and executes them. The CPU 201 may be configured by a combination of multiple CPUs. Further, the CPU 201 may be configured by a combination of a CPU and other types of processors. For example, the CPU 201 may be configured by a combination of a CPU and a GPU (Graphics Processing Unit). The memory 202 is configured with a DRAM (Dynamic Random Access Memory) or the like, and temporarily stores computer programs executed by the CPU 201 and data being processed. The storage device 203 stores computer programs executed by the CPU 201. The storage device 203 is configured by, for example, a nonvolatile semiconductor storage device. Other storage devices such as a hard disk drive may be used as the storage device 203. The input/output I/F 204 is an interface that receives input from a worker and outputs display data and the like. The communication I/F 205 is an interface that transmits and receives data between the user terminal device 20 and the information management server 30. Further, the user terminal device 20 and the information management server 30 may also have similar configurations.
　各処理の実行に用いられるコンピュータプログラムは、プログラムのデータを非一時的に記録するコンピュータ読み取り可能な記録媒体に格納して頒布することもできる。記録媒体としては、例えば、データ記録用磁気テープや、ハードディスクなどの磁気ディスクを用いることができる。また、記録媒体としては、ＣＤ-ＲＯＭ（Compact Disc Read Only Memory）等の光ディスクを用いることもできる。不揮発性の半導体記憶装置を記録媒体として用いてもよい。 The computer program used to execute each process can also be stored and distributed in a computer-readable recording medium that non-temporarily records program data. As the recording medium, for example, a magnetic tape for data recording or a magnetic disk such as a hard disk can be used. Further, as the recording medium, an optical disc such as a CD-ROM (Compact Disc Read Only Memory) can also be used. A nonvolatile semiconductor memory device may be used as the recording medium.
　以上、上述した実施形態を例として本発明を説明した。しかしながら、本発明は、上述した実施形態には限定されない。即ち、本発明は、本発明のスコープ内において、当業者が理解し得る様々な態様を適用することができる。 The present invention has been described above using the above-described embodiment as an example. However, the invention is not limited to the embodiments described above. That is, the present invention can apply various aspects that can be understood by those skilled in the art within the scope of the present invention.
　１０　　状態予測システム
　１１　　取得部
　１２　　予測部
　１３　　出力部
　１４　　生成部
　１５　　記憶部
　２０　　利用者端末装置
　３０　　情報管理サーバ
　１００　　状態予測システム
　１０１　　取得部
　１０２　　予測部
　１０３　　出力部
　２００　　コンピュータ
　２０１　　ＣＰＵ
　２０２　　メモリ
　２０３　　記憶装置
　２０４　　入出力Ｉ／Ｆ
　２０５　　通信Ｉ／Ｆ 10 State prediction system 11 Acquisition unit 12 Prediction unit 13 Output unit 14 Generation unit 15 Storage unit 20 User terminal device 30 Information management server 100 State prediction system 101 Acquisition unit 102 Prediction unit 103 Output unit 200 Computer 201 CPU
202 Memory 203 Storage device 204 Input/output I/F
205 Communication I/F

Claims (10)

  1. 　競走馬に関する情報を取得する取得手段と、
    　競走馬に関する情報から競走馬の将来の状態を予測する予測モデルを用いて、取得した競走馬に関する情報から、競走馬の将来の状態を予測する予測手段と、
    　前記予測の結果と、前記予測の理由とを出力する出力手段と
    　を備える状態予測システム。     an acquisition means for acquiring information about racehorses;
    A prediction means for predicting the future state of the racehorse from the obtained information about the racehorse using a prediction model that predicts the future state of the racehorse from the information about the racehorse;
    A state prediction system comprising: output means for outputting a result of the prediction and a reason for the prediction.
  2. 　前記予測手段は、競走馬のオークション価格と、購入した競走馬の維持費用と、購入した競走馬が獲得する賞金額の少なくとも１つを予測する、
    　請求項１に記載の状態予測システム。     The prediction means predicts at least one of the auction price of the racehorse, the maintenance cost of the purchased racehorse, and the prize money that the purchased racehorse will win.
    The state prediction system according to claim 1.
  3. 　前記予測手段は、競走馬のオークション価格と、購入した競走馬の維持費用と、購入した競走馬が獲得する賞金額とを基に、将来における収支を予測する、
    　請求項２に記載の状態予測システム。     The prediction means predicts future income and expenditure based on the auction price of the racehorse, the maintenance cost of the purchased racehorse, and the prize money that the purchased racehorse will win.
    The state prediction system according to claim 2.
  4. 　前記予測手段は、予測対象の競走馬の年齢に応じた予測モデルを用いて、競走馬の将来の状態を予測する、
    　請求項１から３いずれかに記載の状態予測システム。     The prediction means predicts the future state of the racehorse using a prediction model according to the age of the racehorse to be predicted.
    A state prediction system according to any one of claims 1 to 3.
  5. 　前記出力手段は、予測対象の競走馬の年齢が基準未満の場合に、予測対象の競走馬の親馬の情報を出力する、
    　請求項１から４いずれかに記載の状態予測システム。     The output means outputs information on the parent horse of the racehorse to be predicted when the age of the racehorse to be predicted is less than a standard;
    A state prediction system according to any one of claims 1 to 4.
  6. 　前記出力手段は、予測対象の競走馬と属性が類似した競走馬の実績データを出力する、
    　請求項１から５いずれかに記載の状態予測システム。     The output means outputs performance data of racehorses having similar attributes to the racehorse to be predicted.
    A state prediction system according to any one of claims 1 to 5.
  7. 　前記競走馬に関する情報は、予測対象の競走馬の生体情報、調教履歴および親馬の情報のうち、少なくとも１つを含む、
    　請求項１から６いずれかに記載の状態予測システム。     The information regarding the racehorse includes at least one of biological information, training history, and parent horse information of the racehorse to be predicted.
    A state prediction system according to any one of claims 1 to 6.
  8. 　競走馬に関する情報と、競走馬の将来の状態との関係を学習し、予測対象の競走馬に関する情報から競走馬の将来の状態を予測する予測モデルを生成する生成手段をさらに備える、
    　請求項１から７いずれかに記載の状態予測システム。     Further comprising a generation means for learning the relationship between information about the racehorse and the future state of the racehorse, and generating a prediction model that predicts the future state of the racehorse from the information about the racehorse to be predicted.
    A state prediction system according to any one of claims 1 to 7.
  9. 　競走馬に関する情報を取得し、
    　競走馬に関する情報から競走馬の将来の状態を予測する予測モデルを用いて、取得した競走馬に関する情報から、競走馬の将来の状態を予測し、
    　前記予測の結果と、前記予測の理由とを出力する、
    　状態予測方法。     Get information about racehorses,
    Using a prediction model that predicts the future state of a racehorse from information about the racehorse, predicting the future state of the racehorse from the obtained information about the racehorse,
    outputting the result of the prediction and the reason for the prediction;
    State prediction method.
  10. 　競走馬に関する情報を取得する処理と、
    　競走馬に関する情報から競走馬の将来の状態を予測する予測モデルを用いて、取得した競走馬に関する情報から、競走馬の将来の状態を予測する処理と、
    　前記予測の結果と、前記予測の理由とを出力する処理と
    　をコンピュータに実行させる状態予測プログラムを非一時的に記録する記録媒体。     A process of acquiring information about racehorses;
    A process of predicting the future state of the racehorse from the obtained information about the racehorse using a prediction model that predicts the future state of the racehorse from the information about the racehorse;
    A recording medium that non-temporarily records a state prediction program that causes a computer to execute a process of outputting the prediction result and the reason for the prediction.
PCT/JP2022/012105 2022-03-17 2022-03-17 Situation prediction system, situation prediction method, and recording medium WO2023175809A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/JP2022/012105 WO2023175809A1 (en) 2022-03-17 2022-03-17 Situation prediction system, situation prediction method, and recording medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/JP2022/012105 WO2023175809A1 (en) 2022-03-17 2022-03-17 Situation prediction system, situation prediction method, and recording medium

Publications (1)

Publication Number Publication Date
WO2023175809A1 true WO2023175809A1 (en) 2023-09-21

Family

ID=88022544

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2022/012105 WO2023175809A1 (en) 2022-03-17 2022-03-17 Situation prediction system, situation prediction method, and recording medium

Country Status (1)

Country Link
WO (1) WO2023175809A1 (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006004362A (en) * 2004-06-21 2006-01-05 Nec Mobiling Ltd Joint investment type auction system using communication network, joint bid method, server, and program
WO2019142597A1 (en) * 2018-01-19 2019-07-25 ソニー株式会社 Information processing device, information processing method and program
JP2020149583A (en) * 2019-03-15 2020-09-17 株式会社インター通信社 Program, method, and system for predicting potential ability of race horse, seed horse candidate presentation program, method for presenting seed horse candidate, seed horse candidate presentation system, and prediction career earnings database preparation program used for the same
WO2021171591A1 (en) * 2020-02-28 2021-09-02 裕造 園生 Information processing method, information processing device, and program

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006004362A (en) * 2004-06-21 2006-01-05 Nec Mobiling Ltd Joint investment type auction system using communication network, joint bid method, server, and program
WO2019142597A1 (en) * 2018-01-19 2019-07-25 ソニー株式会社 Information processing device, information processing method and program
JP2020149583A (en) * 2019-03-15 2020-09-17 株式会社インター通信社 Program, method, and system for predicting potential ability of race horse, seed horse candidate presentation program, method for presenting seed horse candidate, seed horse candidate presentation system, and prediction career earnings database preparation program used for the same
WO2021171591A1 (en) * 2020-02-28 2021-09-02 裕造 園生 Information processing method, information processing device, and program

Similar Documents

Publication Publication Date Title
Marwala et al. Artificial intelligence and economic theory: skynet in the market
Seip et al. Anger motivates costly punishment of unfair behavior
Bergevoet et al. Entrepreneurial behaviour of Dutch dairy farmers under a milk quota system: goals, objectives and attitudes
Gifford Jr Emotion and self-control
US20220114903A1 (en) Systems and methods for providing tailored educational materials
Xu et al. Entrepreneurial learning and disincentives in crowdfunding markets
CN113570395A (en) Information processing method and device, computer readable medium and electronic equipment
CN113570397A (en) Model training device, method, device and medium
Bredow et al. Market value, quality of the pool of potential mates, and singles' confidence about marrying
JP5702422B2 (en) Information processing apparatus, information processing method, and information processing program
Brooks et al. The importance of staying positive: The impact of emotions on attitude to risk
Foxall Perspectives on Consumer Choice From Behavior to Action, from Action to Agency
Heidenreich et al. If at first you don't adopt-Investigating determinants of new product leapfrogging behavior
Mengov Decision science: A human-oriented perspective
WO2023175809A1 (en) Situation prediction system, situation prediction method, and recording medium
US20210089945A1 (en) Method and Machine for Predictive Animal Behavior Analysis
JP7301341B2 (en) Racehorse potential ability prediction program, racehorse potential ability prediction method, racehorse potential ability prediction system, stallion candidate presentation program, stallion candidate presentation method, and stallion candidate presentation system , and a program for creating a predicted lifetime winning prize database used for these
Hu et al. Innovation contests: attracting new solvers and new high‐quality solutions
WO2023170832A1 (en) Prediction system, prediction method, and recording medium
Carlberg Excel Sales Forecasting for Dummies
WO2023175910A1 (en) Decision support system, decision support method, and recording medium
Linde et al. The effect of quantity and quality of information in strategy tournaments
WO2023162081A1 (en) Time discount rate estimation device, machine learning method, time discount rate analysis method, and program
Yuan Supply side optimisation in online display advertising
Mueller et al. Excited and aroused: The predictive importance of simple choice process metrics.

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22932085

Country of ref document: EP

Kind code of ref document: A1