System and program

Abstract

To provide a system and program configured to detect an input error of a user when inputting their first name in a family-name input field and inputting their family name in a first-name input field.SOLUTION: A system includes one or more computers. The one or more computers are configured to: acquire family name information indicating a first phrase and first name information indicating a second phrase; calculate a first score indicating the possibility that the first phrase indicated by the family name information is a family name or a first name; calculate a second score indicating the possibility that the second phrase indicated by the first name information is a family name or a first name; and determine whether the first phrase and the second phrase are wrong or not based on the first score and the second score. When they are wrong, the first phrase is a first name and the second phrase is a family name.SELECTED DRAWING: Figure 7

Description

[Technical Field]

[0001] The present invention relates to a system and a program. [Background technology]

[0002] An example of an invention related to a conventional system is an input error warning device described in Patent Document 1. The input error warning device calculates an error rate at which a correct word is misread as an error word based on a correct word, a first number of occurrences, a second number of occurrences, an error word, the first number of errors, the second number of errors, and a weighting coefficient. The correct word is predetermined character information. The first number of occurrences is the number of correct words included in processing target data, which is a processing target for detecting character information by pattern recognition processing. The second number of occurrences is the number of correct words included in character information input by a user. The error word is a result of a correct word being misread. The first number of errors is the number of correct words misread as error words by pattern recognition processing. The second number of errors is the number of correct words misread as error words by the user. The weighting coefficient indicates the degree of reliability of the second number of errors relative to the first number of errors.

[0003] When the input error warning device determines that the character information input by the user is stored as an error word and that the error rate calculated for the error word exceeds a predetermined threshold, it outputs the correct word corresponding to the error word. In this way, the input error warning device can detect and warn the user of an input error due to mishearing or misreading. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 5145287 Summary of the Invention [Problem to be solved by the invention]

[0005] Incidentally, there is a demand for detecting input errors when a user inputs a first name in a last name input field and a last name in a first name input field.

[0006] Therefore, an object of the present invention is to provide a system and program that can detect input errors when a user inputs a first name in a last name input field and a last name in a first name input field. [Means for solving the problem]

[0007] The first form is A system comprising one or more computers, The one or more computers obtain last name information indicating the first phrase and given name information indicating the second phrase; calculating a first score indicating a likelihood that the first term indicated by the surname information is a surname or a given name; calculating a second score indicating a likelihood that the second term indicated by the given name information is a surname or a given name; determining whether the first phrase and the second phrase are in an erroneous state based on the first score and the second score; In the error state, the first term is a first name and the second term is a last name.

[0008] A second aspect is the system according to the first aspect, In the process of calculating the first score and the process of calculating the second score, the one or more computers calculate the first score and the second score using a machine learning model that calculates a score indicating the likelihood that a word or phrase is a surname or a given name.

[0009] A third aspect is the system according to the second aspect, The machine learning model is a trained model that has been trained using training data that indicates a plurality of words labeled with surnames and a plurality of words labeled with given names.

[0010] A fourth aspect is the system according to any one of the first to third aspects, In the process of determining whether or not the state is erroneous, the one or more computers determine whether or not the first phrase and the second phrase are erroneous based on the magnitude relationship between the first score and the second score.

[0011] A fifth aspect is the system according to any one of the first to fourth aspects, the first phrase is a phrase entered as a last name for booking an airline ticket; The second phrase is the phrase entered as the first name for the airline ticket reservation.

[0012] A sixth aspect is the system according to any one of the first to fifth aspects, the first phrase and the second phrase are phrases input by a user into the one or more computers; If the first phrase and the second phrase are in the erroneous state, the one or more computers display a video indicating that the first phrase and the second phrase are in the erroneous state.

[0013] The seventh form is A system comprising one or more computers, The one or more computers obtain last name information indicating the first phrase and given name information indicating the second phrase; determining whether the first phrase and the second phrase are false states using a machine learning model; In the error state, the first term is a first name and the second term is a last name.

[0014] An eighth aspect is the system according to the seventh aspect, The machine learning model is a trained model that has been trained using training data that indicates a plurality of words labeled with surnames and a plurality of words labeled with given names.

[0015] The ninth form is A program executed on a system including one or more computers, The program is configured to: obtaining surname information indicating the first phrase and given name information indicating the second phrase; calculating a first score indicating that the first word indicated by the surname information is a surname or a first name; calculating a second score indicating that the second term indicated by the given name information is a surname or a given name; determining whether the first phrase and the second phrase are in an erroneous state based on the first score and the second score; In the error state, the first term is a first name and the second term is a last name.

[0016] The tenth form is A program executed on a system including one or more computers, The program is configured to: obtaining surname information indicating the first phrase and given name information indicating the second phrase; determining whether the first phrase and the second phrase are false states using a machine learning model; In the error state, the first term is a first name and the second term is a last name. [Effects of the Invention]

[0017] According to the present disclosure, it is possible to detect an input error in which a user inputs a first name in a last name input field and a last name in a first name input field. [Brief explanation of the drawings]

[0018] [Figure 1] FIG. 1 is an explanatory diagram of the system 1. [Figure 2] FIG. 2 is a block diagram of the user terminal 10. [Figure 3] FIG. 3 is a block diagram of the server 110. [Figure 4] FIG. 4 shows an image displayed on the user terminal 10. [Figure 5] FIG. 5 shows an image displayed on the user terminal 10. [Figure 6]FIG. 6 shows an image displayed on the user terminal 10. [Figure 7] FIG. 7 is an explanatory diagram of a machine learning model 300 used in the system 1. [Figure 8] FIG. 8 is a flowchart executed by the control unit 12 of the user terminal 10 and the control unit 112 of the server 110. DETAILED DESCRIPTION OF THE INVENTION

[0019] (Embodiment) A system 1 according to an embodiment of the present disclosure will be described with reference to the drawings.

[0020] [System Structure] First, the overall configuration of the system 1 will be described with reference to the drawings. Fig. 1 is an explanatory diagram of the system 1. Fig. 2 is a block diagram of the user terminal 10. Fig. 3 is a block diagram of the server 110.

[0021] 1 includes a user terminal 10 and a server 110. The user terminal 10 and the server 110 can communicate with each other via a communication network. The network may be the Internet, an intranet, or the like.

[0022] The user terminal 10 is an information processing device used by a user. The user terminal 10 is, for example, a smartphone, a tablet terminal, or a personal computer. As shown in FIG. 2, the user terminal 10 includes a control unit 12, a storage unit 14, a network interface 16, a graphics processing unit 18, a display 20, an operation unit 26, and a touch panel 28.

[0023] The storage unit 14 stores programs and data and is, for example, a combination of a read-only memory (ROM), a random access memory (RAM), and a storage (for example, a flash memory or a hard disk).

[0024] The programs include, for example, the following programs: OS (Operating System) programs - Programs for applications that process information (e.g., web browsers or target apps described below)

[0025] The data includes, for example, the following data: Databases referenced in information processing Data obtained by performing information processing (i.e., the results of performing information processing)

[0026] The control unit 12 executes the programs stored in the storage unit 14 to realize the functions of the user terminal 10. The control unit 12 is, for example, at least one of the following: ·CPU(Central Processing Unit) ·GPU(Graphic Processing Unit) ·ASIC(Application Specific Integrated Circuit) ·FPGA(Field Programmable Array)

[0027] The control unit 12 includes a first information acquisition unit 30, a transmission control unit 32, a second information acquisition unit 34, and a display control unit 36 ​​as functional blocks.

[0028] The network interface 16 controls communication between the user terminal 10 and an external device, which is a server 110.

[0029] The graphics processing unit 18 displays an image on the display 20 based on the image data generated by the control unit 12. The display 20 is a liquid crystal display or an organic EL (Electro Luminescence) display.

[0030] The operation unit 26 generates an operation signal based on a user's operation on the touch panel 28 and outputs the operation signal to the control unit 12 .

[0031] Server 110 is a web server that stores web page data. As shown in FIG. 3, server 110 includes a control unit 112, a storage unit 114, and a network interface .

[0032] The storage unit 114 stores programs and data and is, for example, a combination of a read-only memory (ROM), a random access memory (RAM), and a storage (for example, a flash memory or a hard disk).

[0033] The control unit 112 executes the programs stored in the storage unit 114 to realize the functions of the server 110. The control unit 112 is, for example, at least one of the following: ·CPU(Central Processing Unit) ·GPU(Graphic Processing Unit) ·ASIC(Application Specific Integrated Circuit) ·FPGA(Field Programmable Array)

[0034] The control unit 112 includes a third information acquisition unit 120, a first score calculation unit 122, a second score calculation unit 124, a determination unit 126, a result generation unit 128, and a transmission control unit 130 as functional blocks.

[0035] The network interface 116 controls communication between the server 110 and an external device, which is the user terminal 10.

[0036] [System 1 operation] Next, the operation of the system 1 will be described with reference to the drawings. Figures 4 to 6 show images displayed on the user terminal 10. Figure 7 is an explanatory diagram of a machine learning model 300 used in the system 1.

[0037] First, an overview of the operation of system 1 will be described. The images shown in FIGS. 4 to 6 are images for a user to input his or her first and last names when reserving an airline ticket. The images shown in FIGS. 4 to 6 are displayed on display 20 of user terminal 10. The images shown in FIGS. 4 to 6 include boxes 201 and 202. A user can input a first word / phrase W1 as a last name into box 201 using touch panel 28 of user terminal 10. Furthermore, the user can input a second word / phrase W2 as a first name into box 202 using operation unit 26 of user terminal 10. In this way, first word / phrase W1 is a word / phrase input as a last name for reserving an airline ticket. Second word / phrase W2 is a word / phrase input as a first name for reserving an airline ticket.

[0038] As shown in FIG. 5, the user inputs "YAMADA," which is a first phrase W1, into box 201, and "TARO," which is a second phrase W2, into box 202. "YAMADA" is a phrase that is likely to be a surname. "TARO" is a phrase that is likely to be a first name. In this case, the control unit 112 of the server 110 determines that it is likely that the user has input the surname and first name correctly. Specifically, the control unit 112 of the server 110 determines that the first phrase W1 and the second phrase W2 are in the correct state. In the correct state, the first phrase W1 is the surname, and the second phrase W2 is the given name. Then, no warning is displayed in the image of FIG. 5.

[0039] On the other hand, as shown in FIG. 6, the user inputs "TARO" which is the first phrase W1 into box 201 and "YAMADA" which is the second phrase W2 into box 202. In this case, the control unit 112 of the server 110 determines that there is a high possibility that the user has input the last name and first name incorrectly. Specifically, the control unit 112 of the server 110 determines that the first phrase W1 and the second phrase W2 are in an incorrect state. In the incorrect state, the first phrase W1 is the first name and the second phrase W2 is the last name. Then, a warning is displayed in the image of FIG. 6. As described above, the control unit 112 of the server 110 determines whether the first phrase W1 and the second phrase W2 are in an incorrect state using the machine learning model 300.

[0040] Next, the machine learning model 300 will be described with reference to Fig. 7. The storage unit 114 stores the machine learning model 300 shown in Fig. 7. The control unit 112 of the server 110 makes the above determination using the machine learning model 300.

[0041] The machine learning model 300 is a trained model trained using training data 302 indicating a plurality of phrases labeled with surnames and a plurality of phrases labeled with given names. In the training data 302, the plurality of phrases labeled with surnames are not included in the plurality of phrases labeled with given names. The plurality of phrases labeled with given names are not included in the plurality of phrases labeled with surnames. In other words, the training data 302 does not include phrases labeled with both surnames and given names.

[0042] Such machine learning model 300 calculates a score indicating the possibility that a word or phrase is a surname or a given name. More specifically, control unit 112 of server 110 causes machine learning model 300 to calculate a first score SC1 indicating the possibility that first word or phrase W1 is a surname. Furthermore, machine learning model 300 causes machine learning model 300 to calculate a second score SC2 indicating the possibility that second word or phrase W2 is a surname. Then, control unit 112 of server 110 determines whether first word or phrase W1 and second word or phrase W2 are in an incorrect state based on first score SC1 and second score SC2.

[0043] The machine learning program is a program for executing a machine learning algorithm to find certain rules from training data 302 and generate a trained machine learning model 300 that expresses the found rules. When the control unit 112 of the server 110 executes the machine learning program, multiple pieces of training data 302 are subjected to machine learning and the parameters of the inference program are adjusted. As a result, a trained machine learning model 300 is generated.

[0044] The machine learning algorithm is not particularly limited as long as it is supervised learning, and may be, for example, a decision tree, a nearest neighbor algorithm, a naive Bayes classifier, a support vector machine, or a neural network. Therefore, the trained machine learning model 300 includes a decision tree, a nearest neighbor algorithm, a naive Bayes classifier, a support vector machine, or a neural network. Backpropagation may be used in the machine learning used to generate the trained machine learning model 300.

[0045] For example, a neural network includes an input layer, one or more hidden layers, and an output layer. Specifically, the neural network is a deep neural network, a recurrent neural network, or a convolutional neural network, and performs deep learning. The deep neural network includes, for example, an input layer, multiple hidden layers, and an output layer.

[0046] Next, the operations executed by the control unit 12 of the user terminal 10 and the control unit 112 of the server 110 will be described in detail with reference to Fig. 8. Fig. 8 is a flowchart executed by the control unit 12 of the user terminal 10 and the control unit 112 of the server 110.

[0047] When the control unit 12 of the user terminal 10 reads out the programs stored in the storage unit 14, these programs cause the control unit 12 of the user terminal 10 to execute the operations described below. The programs then cause the control unit 12 of the user terminal 10 to function as a first information acquisition unit 30, a transmission control unit 32, a second information acquisition unit 34, and a display control unit 36.

[0048] When the control unit 112 of the server 110 reads out the programs stored in the memory unit 114, these programs cause the control unit 112 of the server 110 to execute the operations described below. The programs then cause the control unit 112 of the server 110 to function as a third information acquisition unit 120, a first score calculation unit 122, a second score calculation unit 124, a determination unit 126, a result generation unit 128, and a transmission control unit 130.

[0049] Display 20 of user terminal 10 displays the image of Fig. 4. The user uses touch panel 28 of user terminal 10 to input first phrase W1 into box 201 and second phrase W2 into box 202. In this manner, first phrase W1 and second phrase W2 are phrases input by the user via user terminal 10. As a result, control unit 12 (first information acquisition unit 30) of user terminal 10 acquires surname information I1 indicating first phrase W1 and given name information I2 indicating second phrase W2 (step S1).

[0050] Next, the control unit 12 (transmission control unit 32) of the user terminal 10 transmits the surname information I1 and the given name information I2 to the server 110 via the network interface 16 (step S2). In response, the network interface 116 of the server 110 receives the surname information I1 and the given name information I2 and outputs the surname information I1 and the given name information I2 to the control unit 112. As a result, the control unit 112 (third information acquisition unit 120) of the server 110 acquires the surname information I1 indicating the first word / phrase W1 and the given name information I2 indicating the second word / phrase W2 (step S11).

[0051] Next, the control unit 112 (first score calculation unit 122) of the server 110 calculates a first score SC1 indicating the possibility that the first term W1 indicated by the surname information I1 is a surname (step S12). Furthermore, the control unit 112 (first score calculation unit 122) of the server 110 calculates a second score SC2 indicating the possibility that the second term W2 indicated by the given name information I2 is a surname (step S13). In the process of calculating the first score SC1 (step S12) and the process of calculating the second score SC2 (step S13), the control unit 112 (first score calculation unit 122) of the server 110 calculates the first score SC1 and the second score SC2 using the machine learning model 300 that calculates a score indicating the possibility that a term is a surname. The first score SC1 and the second score SC2 are numerical values ​​indicating a probability between 0 and 1. As the first score SC1 approaches 1, the probability that the first term W1 is a surname increases. As the second score SC2 approaches 1, the probability that the second word W2 is a surname increases. 0 indicates 0%, and 1 indicates 100%. Details of the machine learning model 300 have already been explained, so further explanation will be omitted.

[0052] Next, the control unit 112 (determination unit 126) of the server 110 determines whether the first phrase W1 and the second phrase W2 are in an incorrect state based on the first score SC1 and the second score SC2 (step S14). In the process of determining whether the first phrase W1 and the second phrase W2 are in an incorrect state (step S14), the control unit 112 (determination unit 126) of the server 110 determines whether the first phrase W1 and the second phrase W2 are in an incorrect state based on the magnitude relationship between the first score SC1 and the second score SC2. In this embodiment, the control unit 112 (determination unit 126) of the server 110 determines whether the first score SC1 is smaller than the second score SC2. If the first score SC1 is smaller than the second score SC2, the possibility that the first phrase W1 is a surname is lower than the possibility that the second phrase W2 is a surname. Therefore, the control unit 112 (determination unit 126) of the server 110 determines that the first phrase W1 and the second phrase W2 are in an incorrect state. After this, the process proceeds to step S15. If the first score SC1 is not smaller than the second score SC2, the first word / phrase W1 is more likely to be a surname than the second word / phrase W2. Therefore, the control unit 112 (determination unit 126) of the server 110 determines that the first word / phrase W1 and the second word / phrase W2 are in the correct state. After this, the process proceeds to step S16.

[0053] If the first phrase W1 and the second phrase W2 are in the incorrect state, the control unit 112 (result generation unit 128) of the server 110 generates determination result information indicating that the first phrase W1 and the second phrase W2 are in the incorrect state (step S15). After that, the process proceeds to step S17.

[0054] If the first phrase W1 and the second phrase W2 are in the correct state, the control unit 112 (result generating unit 128) of the server 110 generates determination result information indicating that the first phrase W1 and the second phrase W2 are in the correct state (step S16). After that, the process proceeds to step S17.

[0055] In step S17, the control unit 112 (transmission control unit 130) of the server 110 transmits the determination result information to the user terminal 10 via the network interface 116 (step S17). In response, the network interface 16 of the user terminal 10 receives the determination result information and outputs the determination result information to the control unit 12. As a result, the control unit 12 (second information acquisition unit 34) of the user terminal 10 acquires the determination result information (step S3).

[0056] Next, the control unit 12 (display control unit 36) of the user terminal 10 causes the display 20 to display an image indicating the determination result information (step S4). If the determination result information indicates a correct state, the image shown in FIG. 5 is displayed on the display 20. The image shown in FIG. 5 does not include a warning indicating that the first phrase W1 and the second phrase W2 are in an incorrect state. On the other hand, if the determination result information indicates incorrect information, the image shown in FIG. 6 is displayed on the display 20. The image shown in FIG. 6 includes a warning indicating that the first phrase W1 and the second phrase W2 are in an incorrect state. In this way, if the first phrase W1 and the second phrase W2 are in an incorrect state, the control unit 12 of the user terminal 10 displays an image indicating that the first phrase W1 and the second phrase W2 are in an incorrect state. In addition, this warning has an input unit that indicates that the user has confirmed that the last name and first name have been entered correctly. The input unit is, for example, a checkbox. If the user checks the checkbox, even if the warning is included in the image, the control unit 112 of the server 110 determines that the state is not incorrect. After this, this process ends.

[0057] [effect] The system 1 can detect an input error (hereinafter simply referred to as an input error) in which a user inputs a first name in a last name input field and a last name in a first name input field. More specifically, the control unit 112 of the server 110 calculates a first score SC1 indicating the possibility that a first word / phrase W1 indicated by last name information I1 is a last name, and calculates a second score SC2 indicating the possibility that a second word / phrase W2 indicated by first name information I2 is a last name. Then, the control unit 112 of the server 110 determines whether the first word / phrase W1 and the second word / phrase W2 are in an incorrect state based on the first score SC1 and the second score SC2. This allows the control unit 112 of the server 110 to appropriately detect that an input error has occurred.

[0058] Furthermore, since the control unit 112 of the server 110 can appropriately detect that an input error has occurred, the user can recognize the input error. As a result, the user can correct the input error. This reduces the number of steps required to check and correct first and last names when arranging airline tickets.

[0059] System 1 can detect an input error in which a user inputs a first name in a last name input field and a last name in a first name input field. More specifically, control unit 112 of server 110 determines whether first phrase W1 and second phrase W2 are in an incorrect state using a machine learning model. This allows control unit 112 of server 110 to appropriately detect that an input error has occurred.

[0060] System 1 is effective when applied to an airline ticket reservation system. More specifically, if an input error occurs in an airline ticket reservation system, the user will be unable to board the aircraft. Therefore, in system 1, first phrase W1 is a phrase entered as a surname for reserving the airline ticket. Second phrase W2 is a phrase entered as a given name for reserving the airline ticket. As a result, when the user enters first phrase W1 and second phrase W2, control unit 12 of user terminal 10 displays the image shown in FIG. 6 on display 20. The image shown in FIG. 6 includes a warning, allowing the user to recognize that he or she has made an input error. As a result, the possibility that the user will be unable to board the aircraft due to an input error is reduced. As described above, system 1 is effective when applied to an airline ticket reservation system.

[0061] In the system 1, when the first phrase W1 and the second phrase W2 are in an incorrect state, the control unit 12 of the user terminal 10 displays an image indicating that the first phrase W1 and the second phrase W2 are in an incorrect state. This allows the user to recognize that an input error has occurred. Therefore, the user can correct the first phrase W1 and the second phrase W2.

[0062] (Other embodiments) The system according to the present invention is not limited to System 1, and can be modified within the scope of the gist thereof.

[0063] The control unit 112 of the server 110 may calculate a first score SC1 indicating the possibility that the first term W1 indicated by the surname information I1 is a first name. The control unit 112 of the server 110 may also calculate a second score SC2 indicating the possibility that the second term W2 indicated by the given name information I2 is a surname. In this case, in the process of calculating the first score SC1 and the process of calculating the second score SC2, the control unit 112 of the server 110 calculates the first score SC1 and the second score SC2 using a machine learning model that calculates a score indicating the possibility that a term is a first name.

[0064] In addition, when calculating the first score SC1 indicating the possibility that the first term W1 is a first name, the control unit 112 of the server 110 may calculate the score indicating the possibility that the first term W1 is a last name using a machine learning model, and then calculate the first score SC1 by subtracting this score from 1.

[0065] In addition, when calculating the first score SC1 indicating the possibility that the first term W1 is a surname, the control unit 112 of the server 110 may calculate the score indicating the possibility that the first term W1 is a given name using a machine learning model, and then calculate the first score SC1 by subtracting this score from 1.

[0066] In addition, when calculating the second score SC2 indicating the possibility that the second term W2 is a surname, the control unit 112 of the server 110 may calculate the score indicating the possibility that the second term W2 is a given name using a machine learning model, and then calculate the second score SC2 by subtracting this score from 1.

[0067] In addition, when calculating the second score SC2 indicating the possibility that the second term W2 is a first name, the control unit 112 of the server 110 may calculate the score indicating the possibility that the second term W2 is a last name using a machine learning model, and then calculate the second score SC2 by subtracting this score from 1.

[0068] The first score SC1 may be a score indicating that the first word / phrase W1 is a first name, and the second score SC2 may be a score indicating that the second word / phrase W2 is a first name. In this case, the control unit 112 of the server 110 compares the first score SC1 with 1 minus the second score to determine whether the first word / phrase W1 and the second word / phrase W2 are in an incorrect state. If the first score SC1 is greater than the value obtained by subtracting the second score SC1 from the second score, the possibility that the first word / phrase W1 is a first name is higher than the possibility that the second word / phrase W2 is a first name. In other words, the first word / phrase W1 and the second word / phrase W2 are in an incorrect state.

[0069] The first score SC1 may be a score indicating that the first word / phrase W1 is a last name, and the second score SC2 may be a score indicating that the second word / phrase W2 is a first name. In this case, the control unit 112 of the server 110 determines whether the first word / phrase W1 and the second word / phrase W2 are in the incorrect state by comparing the first score SC1 with a value obtained by subtracting the second score from 1. If the first score SC1 is smaller than the value obtained by subtracting the second score from 1, the possibility that the first word / phrase W1 is a last name is lower than the possibility that the second word / phrase W2 is a last name. In other words, the first word / phrase W1 and the second word / phrase W2 are in the incorrect state. The control unit 112 of the server 110 also determines whether the first word / phrase W1 and the second word / phrase W2 are in the incorrect state by comparing the value obtained by subtracting the first score SC1 from 1 with the second score. If the value obtained by subtracting the first score SC1 from 1 is greater than the second score, the first word W1 is more likely to be a given name than the second word W2 is a given name. In other words, the first word W1 and the second word W2 are in the false state.

[0070] The first score SC1 may be a score indicating that the first word / phrase W1 is a first name, and the second score SC2 may be a score indicating that the second word / phrase W2 is a last name. In this case, the control unit 112 of the server 110 determines whether the first word / phrase W1 and the second word / phrase W2 are in the incorrect state by comparing the first score SC1 with a value obtained by subtracting the second score from 1. If the first score SC1 is greater than the value obtained by subtracting the second score from 1, the possibility that the first word / phrase W1 is a first name is higher than the possibility that the second word / phrase W2 is a first name. In other words, the first word / phrase W1 and the second word / phrase W2 are in the incorrect state. The control unit 112 of the server 110 also determines whether the first word / phrase W1 and the second word / phrase W2 are in the incorrect state by comparing the value obtained by subtracting the first score SC1 from 1 with the second score. If the value obtained by subtracting the first score SC1 from 1 is less than the second score, the first word W1 is less likely to be a surname than the second word W2, i.e., the first word W1 and the second word W2 are in the false state.

[0071] Note that the control unit 112 of the server 110 does not need to output the first score SC1 and the second score SC2 when determining whether or not the first term W1 and the second term W2 are in an incorrect state using the machine learning model 300. In this case, the control unit 112 of the server 110 may use the surname information I1 indicating the first term W1 and the given name information I2 indicating the second term W2 as input information for the machine learning model 300, and output the determination result of whether or not the state is in an incorrect state from the machine learning model 300.

[0072] Specifically, the machine learning model 300 may acquire the first term W1 and the second term W2 and determine whether the first term W1 and the second term W2 are in an incorrect state by determining whether the first term W1 or the second term W2 is likely to be a surname. The machine learning model 300 may also acquire the first term W1 and the second term W2 and determine whether the first term W1 or the second term W2 is likely to be a given name by determining whether the first term W1 and the second term W2 are in an incorrect state.

[0073] The first phrase may be a phrase entered as a last name for a purpose other than reserving an airline ticket. The second phrase may be a phrase entered as a first name for a purpose other than reserving an airline ticket. Examples of purposes other than reserving an airline ticket include reserving a trip, reserving a limited express train such as a Shinkansen, or applying for various services.

[0074] Note that the training data 302 may include phrases labeled with both the surname and the given name.

[0075] The first score SC1 and the second score SC2 do not have to be numerical values ​​between 0 and 1. The first score SC1 and the second score SC2 may be numerical values ​​between 0 and 100, for example.

[0076] The first phrase W1 and the second phrase W2 may be written in characters other than Roman letters. Examples of characters other than Roman letters include hiragana, kanji, or a combination of these. The first phrase W1 and the second phrase W2 may also be the surname and given name of a non-Japanese person.

[0077] In the training data 302, multiple phrases labeled with a surname may be labeled with 1, indicating the surname, and multiple phrases labeled with a given name may be labeled with 0, indicating the given name. Also, multiple phrases labeled with a surname may be labeled with 0, indicating the surname, and multiple phrases labeled with a given name may be labeled with 1, indicating the given name.

[0078] Although the system 1 includes two computers, the user terminal 10 and the server 110, it may include one or more computers. When the system 1 includes one computer, the one computer is the user terminal 10 or the server 110. The user terminal 10 may calculate a first score SC1 and a second score SC2 to determine whether the first phrase W1 and the second phrase W2 are in an erroneous state. In this case, the memory unit 14 of the user terminal 10 stores the machine learning model 300.

[0079] In addition, in step S14, the control unit 112 (determination unit 126) of the server 110 may determine whether the first score SC1 is equal to or less than the second score SC2, rather than determining whether the first score SC1 is smaller than the second score SC2. [Explanation of symbols]

[0080] 1: System 10: User terminal 12: Control unit 14: Storage part 16: Network interface 18: Graphics processing unit 20: Display 26:Operation section 28: Touch panel 30: 1st Information Acquisition Department 32: Transmission control section 34:Second Information Acquisition Department 36: Display control unit 110: Server 112: Control unit 114: Storage section 116: Network interface 120:Third Information Acquisition Department 122: First score calculation unit 124: Second score calculation unit 126: Judgment section 128:Result generation part 130: Transmission control unit 201, 202: Box 300: Machine learning models 302: Teacher data I1: Last name information I2: Name Information SC1: First score SC2: Second score W1: 1st word W2: Second phrase

Claims

1. A system comprising one or more computers, The one or more computers obtaining last name information indicating the first word and given name information indicating the second word; calculating a first score indicating a likelihood that the first term indicated by the surname information is a surname or a given name; calculating a second score indicating a possibility that the second term indicated by the given name information is a surname or a given name; determining whether the first phrase and the second phrase are in an erroneous state based on the first score and the second score; In the error state, the first term is a first name and the second term is a last name. system.

2. In the process of calculating the first score and the process of calculating the second score, the one or more computers calculate the first score and the second score using a machine learning model that calculates a score indicating the possibility that a word or phrase is a surname or a given name. The system of claim 1 .

3. The machine learning model is a trained model trained using training data indicating a plurality of phrases labeled with surnames and a plurality of phrases labeled with given names. The system of claim 2 .

4. In the process of determining whether or not the first phrase and the second phrase are in the erroneous state, the one or more computers determine whether or not the first phrase and the second phrase are in the erroneous state based on a magnitude relationship between the first score and the second score. A system according to any one of claims 1 to 3.

5. the first phrase is a phrase entered as a last name for booking an airline ticket; the second phrase is the phrase entered as a given name for the airline ticket reservation; A system according to any one of claims 1 to 3.

6. the first phrase and the second phrase are phrases input by a user into the one or more computers; If the first phrase and the second phrase are in the erroneous state, the one or more computers display a video indicating that the first phrase and the second phrase are in the erroneous state. A system according to any one of claims 1 to 3.

7. A system comprising one or more computers, The one or more computers obtaining last name information indicating the first word and given name information indicating the second word; determining whether the first phrase and the second phrase are false states using a machine learning model; In the error state, the first term is a first name and the second term is a last name. system.

8. The machine learning model is a trained model trained using training data indicating a plurality of phrases labeled with surnames and a plurality of phrases labeled with given names. The system of claim 7.

9. A program executed in a system including one or more computers, The program is configured to: obtaining surname information indicating the first word and given name information indicating the second word; calculating a first score indicating that the first word or phrase indicated by the surname information is a surname or a given name; calculating a second score indicating that the second term indicated by the given name information is a surname or a given name; determining whether the first phrase and the second phrase are in an erroneous state based on the first score and the second score; In the error state, the first term is a first name and the second term is a last name. program.

10. A program executed in a system including one or more computers, The program is configured to: obtaining surname information indicating the first word and given name information indicating the second word; determining whether the first phrase and the second phrase are false states using a machine learning model; In the error state, the first term is a first name and the second term is a last name. program.

Images