Programs, information processing devices, methods, and systems

Abstract

To improve the accuracy of dependency parsing. [Solution] This is a program to be executed by a computer equipped with a processor and memory. The program causes the processor to perform the following steps: to restore case particles using case frames; to determine a modifying clause from among multiple clauses based on the restored case particles; and to present information regarding the dependency of the predicate identified based on the nature of the modifying clause. The step of determining the modifying clause includes the steps of extracting a part of speech or group of parts of speech from the subordinate clauses contained in the restored sentence; to determine whether the part of speech or group of parts of speech can connect with at least one of the subject and topic contained in the sentence based on connection information; and to determine the modifying clause according to the determination result. The connection information has a data structure that associates, for each part of speech or group of parts of speech, whether it can connect with at least one of the subject and topic and the content of the constraints on the connection.

Description

Technical Field

[0001] This disclosure relates to a program, an information processing apparatus, a method, and a system.

Background Art

[0002] Conventionally, research on techniques related to syntactic analysis of sentences has been underway. For example, Patent Document 1 discloses a technique for setting the dependency relationship between elements in a predicate list based on the dependency relationship between the components of a document estimated from the document structure. The predicate list is a list created with predicates as elements by dividing a document converted into a word sequence format into predicates. The technique disclosed in Patent Document 1 performs syntactic dependency structure analysis during the process of creating the predicate list. This analysis is performed, for example, by syntactic analysis using a case frame.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, the technique disclosed in Patent Document 1 did not have sufficient accuracy in syntactic analysis.

[0005] An object of this disclosure is to improve the accuracy of syntactic analysis.

Means for Solving the Problems

[0006] To solve the aforementioned problems, a program according to one aspect of this disclosure is a program to be executed on a computer having a processor and memory. The program causes the processor to perform a reception step of receiving text as input, a preprocessing step of dividing the received text into multiple clauses, a restoration step of restoring case particles using case frames for the text divided into multiple clauses, a determination step of determining a clause that modifies a predicate in the text as a modifying clause from among the divided multiple clauses based on the restored case particles, an identification step of identifying the recipient of the predicate based on the properties of the determined modifying clause, and a presentation step of presenting information about the identified recipient, wherein the determination step includes in the text after the case particles have been restored. The process includes: a first step of extracting parts of speech or groups of parts of speech that play a role in connecting clauses from among the subordinate clauses; a second step of determining, based on the connection information of the extracted parts of speech or groups of parts of speech, whether the extracted parts of speech or groups of parts of speech can connect with at least one of the subject and topic contained in the sentence after the case particles have been restored; and a third step of determining at least one of the subject and topic as a modifying clause according to the determination result. The connection information has a data structure that associates, for each part of speech or group of parts of speech, whether it can connect with at least one of the subject and topic and the content of the constraints on the connection. [Effects of the Invention]

[0007] According to this disclosure, the accuracy of dependency parsing can be improved. [Brief explanation of the drawing]

[0008] [Figure 1] This is a block diagram showing an example of the overall configuration of System 1. [Figure 2] This block diagram shows an example configuration of the terminal device 10 shown in Figure 1. [Figure 3] This block diagram shows an example configuration of server 20 as shown in Figure 1. [Figure 4] This figure shows the data structure of the frametable 2022, as shown in Figure 3. [Figure 5]This flowchart shows an example of how server 20 operates when performing dependency parsing. [Figure 6] This is a schematic diagram showing an example of the display screen of display 141 when analysis information is presented. [Figure 7] A block diagram showing the basic hardware configuration of Computer 90. [Modes for carrying out the invention]

[0009] The embodiments of this disclosure will be described below with reference to the drawings. In all the drawings illustrating the embodiments, common components are denoted by the same reference numerals, and repeated explanations are omitted. The following embodiments are not intended to unduly limit the content of this disclosure as described in the claims. Not all components shown in the embodiments are necessarily essential components of this disclosure. Also, each drawing is a schematic diagram and is not necessarily a strict illustration.

[0010] [1. Overview] The system according to this embodiment addresses issues such as when a case particle is overwritten by a linking particle, making it unclear which predicate phrase modifies which predicate phrase. In the system according to this embodiment, the server uploads the text input to the terminal device. The system according to this embodiment divides the text uploaded to the server into multiple phrases. For each of the divided phrases, the system according to this embodiment restores the case particle using a case frame. Based on the restored case particle, the system according to this embodiment determines a modifying clause from among the multiple phrases. In this specification, "modifying clause" refers to a phrase consisting of one or more phrases that modify a predicate. Based on the nature of the determined modifying clause, the system according to this embodiment determines the predicate's dependency and presents information regarding that dependency to the user.

[0011] The system according to this embodiment performs the following steps in the process of determining the modifying clause: extracting a part of speech or group of parts of speech that plays a role in connecting clauses from among the subordinate clauses contained in the sentence after the case particles have been restored; determining whether the extracted part of speech or group of parts of speech can be connected to at least one of the subject and topic contained in the sentence after the case particles have been restored, based on the connection information of the extracted part of speech or group of parts of speech; and determining at least one of the subject and topic as a modifying clause according to the determination result.

[0012] There are no particular limitations on the content and type of documents that the system according to this embodiment can handle. For example, in addition to documents written in reports, documents, manuals, or letters, documents written in novels, papers, essays, poems, or tanka poems can also be included.

[0013] [2. Overall System Configuration] Figure 1 is a block diagram showing an example of the overall configuration of System 1. System 1 shown in Figure 1 includes, for example, a terminal device 10 and a server 20. The terminal device 10 and the server 20 communicate with each other via, for example, a network 80.

[0014] In Figure 1, an example is shown in which System 1 includes one terminal device 10, but System 1 may include two or more terminal devices 10. Also, in this embodiment, a collection of multiple information processing devices may be considered as one server 20. The method of distributing the multiple functions required to realize the server 20 according to this embodiment to one or more hardware can be appropriately determined according to the processing capacity of each hardware and / or the specifications required for the server 20.

[0015] The terminal device 10 shown in FIG. 1 is, for example, an information processing device operated by a user who uses a dependency analysis service. The dependency analysis service is a service that analyzes the structure of a sentence by identifying the dependency relationships of each word or each clause in the sentence, and is provided, for example, by the server 20. The terminal device 10 is realized by, for example, a mobile terminal such as a smartphone or a tablet. In this embodiment, it is assumed that the terminal device 10 is a smartphone. The terminal device 10 may be realized by, for example, a stationary PC (Personal Computer), a laptop PC, or the like.

[0016] The terminal device 10 includes a communication IF (Interface) 12, an input device 13, an output device 14, a memory 15, a storage 16, and a processor 19. The input device 13 is a device for receiving an input operation from the user (for example, a pointing device such as a touch panel, a touch pad, a mouse, etc., a keyboard, etc.). The output device 14 is a device for presenting information to the user (a display, a speaker, etc.). In this embodiment, it is assumed that the terminal device 10 includes a touch panel in which the input device 13 and the output device 14 are integrated.

[0017] The server 20 is, for example, an information processing device that deploys a dependency analysis service, and is realized by a computer connected to the network 80. As shown in FIG. 1, the server 20 includes a communication IF 22, an input / output IF 23, a memory 25, a storage 26, and a processor 29. The input / output IF 23 functions as an interface for an input device for receiving an input operation from the user and an output device for outputting information to the user.

[0018] Each information processing device constituting the system 1 is configured by a computer 90 (see FIG. 7) including an arithmetic unit and a storage unit. The basic hardware configuration of the computer and the basic functional configuration of the computer realized by the hardware configuration will be described later. For each of the terminal device 10 and the server 20, descriptions overlapping with the basic hardware configuration of the computer and the basic functional configuration of the computer described later are omitted.

[0019] <2.1 Configuration of Terminal Device> FIG. 2 is a block diagram showing a configuration example of the terminal device 10 shown in FIG. 1. As shown in FIG. 2, the terminal device 10 includes a communication unit 120, an input device 13, an output device 14, a storage unit 180, and a control unit 190. Each block included in the terminal device 10 is electrically connected by, for example, a bus or the like. Note that the terminal device 10 may include an audio processing unit, a microphone, a speaker, a camera, a position information sensor, or a combination of at least two of these.

[0020] The communication unit 120 performs processes such as modulation / demodulation processing for the terminal device 10 to communicate with an external device. The communication unit 120 performs transmission processing on the signal generated by the control unit 190 and transmits it to the outside (for example, the server 20). The communication unit 120 performs reception processing on the signal received from the outside and outputs it to the control unit 190.

[0021] The input device 13 is a device for a user operating the terminal device 10 to input an instruction or information. The input device 13 is realized, for example, by a touch-sensitive device 131 that inputs an instruction by touching an operation surface. When the terminal device 10 is a PC or the like, the input device 13 may be realized by a reader, a keyboard, a mouse, or the like. The input device 13 converts the instruction input by the user into an electrical signal and outputs it to the control unit 190. Note that the input device 13 may include, for example, a reception port that receives an electrical signal input from an external input device.

[0022] The output device 14 is a device for presenting information to the user operating the terminal device 10. The output device 14 is implemented, for example, by a display 141. The display 141 displays various information according to the control of the control unit 190. The display 141 is implemented, for example, by an LCD (Liquid Crystal Display) or an organic EL (Electro-Luminescence) display.

[0023] The storage unit 180 is implemented, for example, by the memory 15 and storage 16 shown in Figure 1, and stores data and programs used by the terminal device 10. The programs include applications such as web browser applications.

[0024] The control unit 190 is realized when the processor 19 reads a program stored in the memory unit 180 and executes the instructions contained in the program. The control unit 190 controls the operation of the terminal device 10. By operating according to the program, the control unit 190 performs the functions of an operation reception unit 191, a transmission / reception unit 192, and a presentation control unit 193.

[0025] The operation reception unit 191 processes instructions or information input from the input device 13. Specifically, the operation reception unit 191 receives instructions or information input from the touch-sensitive device 131. The transmission / reception unit 192 processes data transmission and reception between the terminal device 10 and an external device such as the server 20, according to a communication protocol. Specifically, the transmission / reception unit 192 transmits instructions or information input by the user to the server 20. The transmission / reception unit 192 receives information transmitted from the server 20. The presentation control unit 193 controls the output device 14 to present various information transmitted from the server 20 to the user.

[0026] <2.2 Server Configuration> Figure 3 is a block diagram showing an example of the functional configuration of the server 20 shown in Figure 1. As shown in Figure 3, the server 20 performs the functions of a communication unit 201, a storage unit 202, and a control unit 203.

[0027] The communication unit 201 performs processing for the server 20 to communicate with external devices. The storage unit 202 is implemented, for example, by the memory 25 and storage 26 shown in Figure 1, and stores data and programs used by the server 20. The programs include applications such as web browser applications. The storage unit 202 stores, for example, the application 2021 and the frame table 2022.

[0028] App 2021 is an application for managing users' use of the dependency parsing service. App 2021 runs in the background of other applications installed on server 20, for example, and monitors processes performed by users. Users can access App 2021 on server 20 using a web browser application installed on terminal device 10.

[0029] The server 20 may, for example, monitor and manage the usage status of application 2021 and perform predetermined analysis processing. Alternatively, application 2021 may be installed on the terminal device 10 and stored in the storage unit 180. In this case, application 2021 may be installed, for example, as an extension of the document creation software installed on the terminal device 10.

[0030] The frame table 2022 is a table that stores information about each frame. Details about the frame table 2022 will be described later.

[0031] The tables held by the storage unit 202 are not limited to these. Furthermore, the individual frame table 2022 does not necessarily have to be stored in the storage unit 202. For example, the individual frame table 2022 may be stored in a server related to a service for sharing information about individual frames among multiple users.

[0032] The control unit 203 is realized when the processor 29 reads a program stored in the memory unit 202 and executes instructions contained in the program. The program that the control unit 203 reads from the memory unit 202 includes a programming language such as JavaScript® that is executed on a web browser application stored in the terminal device 10. By operating according to the program, the control unit 203 performs the functions of a receive control module 2031, a transmit control module 2032, a preprocessing module 2033, an analysis processing module 2034, and a presentation control module 2035.

[0033] The receive control module 2031 controls the process by which the server 20 receives signals from an external device according to a communication protocol. The transmit control module 2032 controls the process by which the server 20 transmits signals to an external device according to a communication protocol. The preprocessing module 2033 performs preprocessing on the text received by the server 20, dividing it into multiple phrases. Details of the preprocessing will be described later.

[0034] The parsing module 2034 uses case frames to reconstruct case particles in sentences that have been divided into multiple clauses by the preprocessing module 2033. A case frame organizes predicates and related items according to the usage of the predicate. An item is a noun accompanied by a case particle that indicates its syntactic role with respect to the predicate. In other words, a case frame summarizes information such as, for example, that the predicate "teach (verb)" takes "ni" or "o" as a case particle, and the predicate "eat (verb)" takes "o" as a case particle. In this embodiment, the case frames are embodied as a case frame table 2022 (case frame dictionary).

[0035] Based on the recovered case particles, the parsing module 2034 determines, from among the multiple phrases separated by the preprocessing module 2033, the phrase that modifies the predicate in the sentence as a modifying clause. Based on the properties of the determined modifying clause, the parsing module 2034 identifies the recipient of the predicate's dependency in the sentence. Details of each process performed by the parsing module 2034 will be described later.

[0036] The presentation control module 2035 controls the process of presenting various types of information to the user. For example, the presentation control module 2035 presents the user with information related to the dependencies of predicates identified by the analysis processing module 2034 (analysis information described later).

[0037] [3 Data Structure] Figure 4 shows the data structure of a table stored by server 20. Note that the data structure in Figure 4 is merely an example, and does not exclude data not shown in the figure. Furthermore, even data listed in the same table may be stored in separate memory areas within the storage unit 202.

[0038] Figure 4 shows the data structure of the case frame table 2022. The case frame table 2022 shown in Figure 4 is a case frame dictionary that serves as a "collection of verb-centered examples" describing the relationship between predicates such as verbs and their arguments. The case frame table 2022 shown in Figure 4 is a table that has columns for term, case particle, and example, with the case frame ID as the key.

[0039] The item "Case Frame ID" stores an identifier for identifying a case frame. The item "Term" stores specific examples of predicates. In the example in Figure 4, the item "Term" stores a verb as a specific example of a predicate, but the item "Term" may also store at least one of the following as specific examples of predicates: an adjective, an adjectival noun, or a combination of a noun and an auxiliary verb. The item "Case Particle" stores some or all of the case particles that a specific example of a predicate can take. The item "Example" stores some or all of the specific examples of nouns that make up an item that is related to a specific example of a predicate.

[0040] Note that the example in Figure 4 shows only a portion of the information stored in the case frame table 2022 for the sake of simplification. In reality, the case frame table 2022 stores multiple terms not shown in the figure, as well as case particles and examples associated with each of those terms. Furthermore, the contents of the case frame table 2022 may be updated as appropriate by, for example, the administrator of server 20.

[0041] [4 actions] This section describes an example of how Server 20 operates when performing dependency parsing. Figure 5 is a flowchart showing an example of how Server 20 operates when performing dependency parsing.

[0042] In step S11 shown in Figure 5, the server 20 accepts text input (reception step).

[0043] Specifically, for example, the operation reception unit 191 receives text input operations from the user and transmits text information indicating the input text to the transmission / reception unit 192. The transmission / reception unit 192 transmits the text information received from the operation reception unit 191 to the server 20. The reception control module 2031 receives text information transmitted from the terminal device 10, for example. As a result, the server 20 accepts text input from the user. The reception control module 2031 transmits the text information received from the terminal device 10 to the preprocessing module 2033, for example.

[0044] The server 20 may, for example, accept text input from a user by having the receiving control module 2031 receive text information that has been pre-stored in the storage unit 180. In this case, the operation reception unit 191 may, for example, accept a request from the user to send text information and send it to the transmission / reception unit 192. The transmission / reception unit 192, upon receiving the transmission request, may, for example, read the text information from the storage unit 180 and send it to the server 20. In this case, the text information may, for example, be text data of the text, or scanned data of a document containing the text.

[0045] In step S12, the server 20 performs preprocessing to divide the received text into multiple segments (preprocessing step).

[0046] Specifically, for example, preprocessing module 2033 breaks down the text information into words (morphological analysis). For example, if the sentence is "I eat curry", preprocessing module 2033 breaks this sentence down into "I / are / curry / eat".

[0047] Here, preprocessing module 2033 separates supplementary text and parentheses from a sentence if the sentence contains supplementary text and parentheses surrounding it. Supplementary text is supplementary explanation or note about a certain matter. Parentheses surrounding supplementary text are usually parentheses, but are not limited to this case; any type of parentheses surrounding supplementary text will be separated. For example, if the sentence is "I saw a lion (a large carnivorous mammal that lives in parts of Africa and India) at the zoo," preprocessing module 2033 will separate "(a large carnivorous mammal that lives in parts of Africa and India)" from the sentence.

[0048] On the other hand, in the case of sentences other than supplementary sentences and the parentheses surrounding such sentences, the preprocessing module 2033 does not separate them from the sentence. For example, if the sentence is "He greeted me with "hello", the preprocessing module 2033 does not separate "hello" from the sentence.

[0049] The preprocessing module 2033 performs preprocessing on the sentence after separating the supplementary sentence and parentheses, for example. For example, if the sentence is "I saw a lion (a large carnivorous mammal that lives in parts of Africa and India) at the zoo," the preprocessing module 2033 will break down the sentence "I saw a lion at the zoo" after separating the supplementary sentence and parentheses into "I / wa / zoo / de / lion / o / saw."

[0050] The preprocessing module 2033 also performs the preprocessing in step S12 on supplementary sentences separated from the main text, for example. For example, if the main text is "I saw a lion (a large carnivorous mammal that lives in parts of Africa and India) at the zoo," the preprocessing module 2033 will also perform the preprocessing in S12 on the separated supplementary sentence "a large carnivorous mammal that lives in parts of Africa and India" after breaking it down into "a large / carnivorous / mammal / that lives / in / parts / of / Africa / and / India / and

[0051] Note that the processes of separating supplementary text and parentheses, and preprocessing the supplementary text separated from the main text, are not mandatory. However, the accuracy of dependency parsing is improved if the preprocessing module 2033 performs these processes. Furthermore, the processing order of the processes of dividing the main text into words, separating supplementary text and parentheses, and preprocessing the supplementary text separated from the main text can be changed arbitrarily. Alternatively, the preprocessing module 2033 may perform these three processes simultaneously.

[0052] Next, the preprocessing module 2033 identifies predicates from the decomposed sentence, for example. In this embodiment, if the decomposed sentence contains multiple predicates, the preprocessing module 2033 identifies all of them. For example, if the decomposed sentence is "I / ga / katta / ta / karu / ni / karu", the preprocessing module 2033 identifies "katta" (verb and auxiliary verb) and "karu" (verb) as predicates from this sentence. As mentioned above, predicates consist of verbs, adjectives, adjectival nouns, or combinations of nouns, particles, and auxiliary verbs.

[0053] Next, preprocessing module 2033 groups the decomposed words into phrases, using particles in the sentence as clues, for example. For example, if the decomposed sentence is "I / am / curry / eating", it is grouped into multiple phrases such as "I / am / curry / eating". At this point, if there are places in the sentence where nouns are consecutive, preprocessing module 2033 determines whether that place is an idiom or whether it can be divided into multiple phrases. For example, in the case of "grilled meat set meal", preprocessing module 2033 determines that it is an idiom, and in the case of "last year's set meal", preprocessing module 2033 determines that it can be divided into two phrases "last year / set meal".

[0054] The order of processing the steps for identifying the predicate from the decomposed text and grouping the decomposed words into phrases can be changed. Alternatively, the preprocessing module 2033 may execute these two processes simultaneously.

[0055] By performing the above series of processes, the preprocessing module 2033 preprocesses the text that has been received as input. The preprocessing module 2033 then sends preprocessed text information, which shows the text after preprocessing, to the parsing module 2034.

[0056] It is not mandatory for the preprocessing module 2033 to perform preprocessing. For example, the server 20 may input the text it receives as input into an AI model, and the AI ​​model may output the text divided into multiple phrases. In this case, the AI ​​model may be, for example, a trained model that performs natural language processing, and may be a machine learning model or a generative AI model.

[0057] In step S13, the server 20 uses case frames to reconstruct case particles for the sentence that has been divided into multiple clauses (reconstruction step).

[0058] Specifically, for example, the parsing module 2034 determines whether a case particle has been omitted by a linking particle for each of the multiple clauses contained in the preprocessed text information received from the preprocessing module 2033. If it determines that a case particle has been omitted by a linking particle, the parsing module 2034 refers to the case frame table 2022 and restores the case particle that was omitted by the linking particle.

[0059] For example, if the preprocessed text information contains multiple phrases such as "I / eat / curry", the parsing module 2034 determines that the case particle that the noun "I" takes in the phrase "I" has been removed by the particle "wa". In this case, the parsing module 2034 refers to the case frame table 2022, for example, and overwrites the particle "wa" with the case particle that the predicate "eat" can take and that is suitable for the noun "I". In other words, the parsing module 2034 determines that of the case particles that the predicate "eat" can take, "wo" is already in use, so it overwrites the remaining case particle "ga" with the particle "wa". In this way, the parsing module 2034 restores the case particle that was removed by the particle.

[0060] For example, the parsing module 2034 determines whether or not a case particle has been omitted for each of the multiple clauses contained in the preprocessed text information received from the preprocessing module 2033. If it determines that a case particle has been omitted, the parsing module 2034 refers to the case frame table 2022 and restores the omitted case particle.

[0061] For example, if the preprocessed text information contains multiple phrases such as "I / went to Kyoto," the parsing module 2034 determines that the case particle that the noun "I" takes is omitted in the phrase "I." In this case, the parsing module 2034 refers to, for example, the case frame table 2022 and adds the case particle that the predicate "went" can take and that fits the noun "I." In other words, since "ni" is already used among the case particles that the predicate "went" can take, the parsing module 2034 adds the remaining case particle "ga" after the noun "I." In this way, the parsing module 2034 restores the omitted case particle.

[0062] In step S14, the server 20 determines, based on the recovered case particles and connection information, that the clause modifying the predicate in the sentence is the modifying clause from among the multiple separated clauses (determination step).

[0063] Conjunction information is a systematic description of the syntactic constraints of parts of speech or groups of parts of speech (hereinafter sometimes collectively referred to as "connecting elements") that play a role in connecting clauses, and it indicates whether or not a connecting element can be connected to at least one of the subject and topic in a sentence. Specifically, "parts of speech or groups of parts of speech that play a role in connecting clauses" refers to the elements that constitute the predicate of a subordinate clause, excluding the semantic core part such as the verb, noun, adjective, or adjectival noun that is located at the beginning of the predicate. Furthermore, "parts of speech or groups of parts of speech that play a role in connecting clauses" in this specification is a concept that encompasses single parts of speech such as "nagara" (conjunction) and "ga" (conjunction) included in subordinate clauses, as well as groups of parts of speech consisting of multiple parts of speech such as "no toki" (a combination of the particle "no" and the noun "toki").

[0064] For example, the conjunctive particle "nagara" indicates that it cannot connect to either the subject or the topic. Similarly, a combination of the particle "no" and the noun "toki," such as "no toki," indicates that it cannot connect to the topic but can connect to the subject. Furthermore, a conjunctive particle like "ga" indicates that it can connect to both the subject and the topic.

[0065] Such connection information is predefined information about various connection elements based on general linguistic knowledge. Specifically, the connection information has a table-like data structure that includes items such as the following: • Item 1 "Connecting elements": Parts of speech or groups of parts of speech that have a connecting function, such as the conjunctive particles "nagara," "node," and "ga," and "no toki," which includes formal nouns. • Item 2, "Connection Constraint Type": Indicates the target to which the constraint applies. Examples include "Subject Connection" and "Topic Connection." • Item 3 "Connection Availability": Indicates whether connection is possible with the target specified in the connection constraint type. Examples include "Possible," "Not Possible," "Conditional," etc. • Item 4 "Constraint Details": Describe the specific conditions and properties of the connection. For example, describe linguistic constraints such as "requires the subject to be the same as the subject of the predicate of the main clause."

[0066] For example, the connection information for the connection element "nagara" has a data structure associated with the connection constraint type "subject connection," the connection feasibility "not possible," and the constraint content "does not directly depend on the subject," and a data structure associated with the connection constraint type "topic connection," the connection feasibility "not possible," and the constraint content "requires to be the same as the subject of the predicate of the main clause." Also, for example, the connection information for the connection information "no toki" has a data structure associated with the connection constraint type "subject connection," the connection feasibility "possible," and a data structure associated with the connection constraint type "topic connection," the connection feasibility "not possible." Furthermore, the connection information for the connection information "ga" has a data structure associated with the connection constraint type "subject connection," the connection feasibility "possible," and a data structure associated with the connection constraint type "topic connection," the connection feasibility "possible."

[0067] In this embodiment, connection information is stored in a database (not shown) as a data structure that defines constraints regarding the connection between each connection element and the subject and topic. The database is stored, for example, in the storage unit 202. The server 20 searches this database according to the connection elements that appear in the text, reads the corresponding connection constraints, and uses them for analysis.

[0068] Furthermore, connection information may also include information indicating the type of connection between parts of speech or groups of parts of speech, such as sequential, contrasting, parallel / additive, comparative / selective, explanatory / supplementary, or transformative. Using such connection information can further improve the accuracy of dependency parsing.

[0069] Specifically, for example, the parsing module 2034 classifies the function of each of the multiple separated phrases. Here, the function of a phrase is classified into five types: subject, predicate, modifier, object, and independent word. For example, if the sentence after restoring the case particles is "Last week, while I was having lunch with a friend, I accidentally rang my cell phone," the parsing module 2034 classifies "Last week," as an adverbial clause (modifying the verb), "lunch," as a modifier, "while I was eating," as a predicate, and "I," as the subject, etc.

[0070] Next, the parsing module 2034 determines, for example, the phrases that modify the predicate in the text, based on the classification results of the function of the phrases, case frames, and connection information, and sets them as modifying clauses.

[0071] Specifically, first, the parsing module 2034 analyzes the sentence after the case particles have been restored and extracts parts of speech or groups of parts of speech (e.g., "nagara," "toki," etc.) that play a role in connecting clauses from the end of the predicate of the subordinate clause (Step 1). Next, the parsing module 2034 determines, based on connection information read from the database, whether the extracted parts of speech or groups of parts of speech can be connected to at least one of the subject and topic contained in the sentence after the case particles have been restored (Step 2). Then, the parsing module 2034 determines that at least one of the subject or topic is a modifying clause according to the determination result (Step 3). The parsing module 2034 performs this process of determining the modifying clause for each clause, from the end to the beginning of the sentence, as follows.

[0072] Specifically, the parsing module 2034 first sets the last phrase (including part of speech or group of parts of speech) of the sentence as a candidate predicate. Then, the parsing module 2034 iterates through the phrases one by one towards the beginning, determining whether the phrase being processed can modify (adapt) the candidate predicate. If the phrase being processed is a predicate, the parsing module 2034 switches the candidate predicate to the phrase being processed. More specifically, if the phrase being processed is a predicate that constitutes a subordinate clause, the parsing module 2034 limits the range of subjects / topics that the subordinate clause can modify based on the connection information contained in the connecting elements (conjunctions, formal nouns, etc.) at the end of the subordinate clause, and determines the candidate predicate. If the clause to be processed is the subject or topic (the first subject or topic located at the beginning of the sentence, prior to the part of speech or group of parts of speech), and the connection information determines that the predicate of the dependency candidate cannot be connected to that subject or topic, the parsing processing module 2034 reverts the current predicate of the dependency candidate back to the predicate that was previously the dependency candidate. The parsing processing module 2034 then determines the clause to be processed (subject or topic) as the modifying clause of the returned predicate. If the clause to be processed is the subject or topic, and the connection information determines that the predicate of the dependency candidate can be connected to that subject or topic, and it is determined that there is no subject or topic (second subject or second topic) at the beginning of the sentence prior to the clause to be processed, the parsing processing module 2034 determines the clause to be processed (subject or topic) as the modifying clause of the current predicate of the dependency candidate.

[0073] The parsing module 2034 can more accurately determine which predicate each clause modifies (i.e., the modifying clause) by repeating this process until the beginning of the sentence. This series of processes is not dependent on specific connection elements, but is executed generically for various sentences by referring to the aforementioned database and applying the connection constraints defined therein. Several specific examples are given below.

[0074] For example, if the sentence after classifying the function of the clauses is "I saw from afar that my brother was laughing a little as he fell," then the parsing module 2034 first determines the last predicate "saw" as the candidate predicate to which it should be modified. Next, the parsing module 2034 goes back to the clause "from afar" which is one clause earlier than "saw." Since "from afar" is not a predicate, the parsing module 2034 determines that "from afar" is a modifying clause to "saw." Next, the parsing module 2034 similarly determines that the clause "things" which is one clause earlier than "things" is also a modifying clause to "saw." Next, since the clause one clause earlier than "things" is the predicate "was laughing," the parsing module 2034 switches the candidate predicate to this predicate "was laughing." Next, since the phrase "a little" which is one position ahead of "was laughing" is not a predicate, the parsing module 2034 determines that "a little" is a modifying clause of "was laughing". Next, since the phrase "while falling" which is one position ahead of "a little" is a predicate, the predicate candidate to which it is modified is switched to this predicate "while falling".

[0075] Here, the parsing module 2034 extracts the conjunction "nagara" and reads connection information from the database indicating that it "cannot connect to a subject". This connection information for "nagara" includes information requiring that the subject of the predicate it connects to and the subject of the candidate predicate it modifies are the same. Based on this information, the module determines the subject of "korobinagara" (older brother) and the subject of "wawaraita" (was laughing). Due to the constraint indicated by this connection information, the current candidate "korobinagara" is determined not to be the target of the phrase "ani ga" (older brother), but rather to modify the subsequent predicate "wawaraita". Based on this determination, the parsing module 2034 removes "korobinagara" from the list of candidate targets and returns to the previous candidate target, the predicate "wawaraita". The parsing module 2034 then determines that "ani ga" is the modifying clause of the predicate "wawaraita". Finally, the parsing module 2034 reads connection information from the database stating that "the subject 'I' cannot be connected to the subordinate clause predicate 'was laughing'," and based on this connection information, it excludes "was laughing" from the list of possible modifiers for "I." Then, the parsing module 2034 returns the modifier candidates to the main clause predicate "saw," and makes "I" the modifying clause for "saw."

[0076] For example, if the sentence after classifying the function of the clauses is "I was laughing when my brother fell," then the noun "toki" in the clause "koronda toki," cannot connect to the clause "I (the subject)" due to its nature. Therefore, based on the connection information indicating that it cannot connect to the subject, the parsing module 2034 assumes that the predicate "koronda toki" is modified by the clause "my brother," and treats the clause "my brother" as the modifying clause.

[0077] For example, if the sentence after classifying the function of the clauses is "Grandpa mowed the lawn, but Grandma went to do the laundry," then the combination of the auxiliary verb "ta" and the conjunctive particle "ga" in the clause "shita," is, by its nature, connectable to the clause "Grandpa (topic)." Therefore, based on the connection information of "taga," which indicates that it can connect to the topic, the analysis processing module 2034 considers the predicate "shita" to be modified by the clause "Grandpa," and treats the clause "Grandpa" as a modifying clause.

[0078] For example, if the input sentence is "I watch TV while eating curry," the parsing module 2034 will determine the predicate "watch" from the latter part of the two predicates, "while eating" and "watch." The parsing module 2034 will also determine the modifying clauses based on whether or not they can be connected to the predicates.

[0079] In other words, if the sentence received as input is "I watch TV while eating curry," the parsing module 2034 refers to the case frame table 2022 to confirm that the predicate "to watch" takes the case particle "o," and recognizes that the phrase "television" can be connected to the predicate "to watch." Since the phrase "while eating," which is one phrase earlier than the phrase "television," is a predicate, the parsing module 2034 temporarily suspends its decision to identify the predicate "to watch."

[0080] Next, the parsing module 2034 refers to the case frame table 2022 to confirm that the predicate "tabenagara" takes the case particle "o", and recognizes that the phrase "kare o" can be connected to the predicate "tabenagara". Then, due to the nature of the conjunctive particle "nagara" in the predicate "tabenagara", it obtains information that the predicate "tabenagara" cannot be connected to the phrase "watashi wa (topic)". Based on this information, the parsing module 2034 determines that the phrase "kare o" is a modifying clause of the predicate "tabenagara".

[0081] Next, since the phrase "I" is the topic and cannot be connected to "while eating," the parsing module 2034 returns the connection destination to "see." Also, the particle "wa" in the phrase is restored to the case particle "ga" by the parsing module 2034. The parsing module 2034 refers to the case frame table 2022 to confirm that the predicate "see" takes the case particle "ga," and recognizes that the phrase "I" can be connected to the predicate "see." As a result, the parsing module 2034 determines that the phrases "I" and "television" are modifying clauses of the predicate "see."

[0082] In step S15, the server 20 identifies the predicate modified by the modified clause based on the determined properties of the modified clause (identification step).

[0083] Specifically, for example, the analysis processing module 2034 identifies the dependency based on (1) to (4) below and transmits the analysis information to the presentation control module 2035. The analysis information is information about the dependency identified by the analysis processing module 2034. (1) If a sentence contains multiple predicates, the parsing module 2034 identifies which clauses (recipients) each predicate other than the main clause modifies. (2) If the last part of a clause of a predicate modified by a modifying clause is a noun, the parsing module 2034 identifies that noun as the recipient of the modification of the predicate. For example, if the sentence after determining the modifying clause is "I get in the car I bought", the last part of the clause of the predicate "bought" modified by the modifying clause "I" is the noun "car", so the parsing module 2034 identifies the recipient of the modification of the predicate "bought" as the noun "car". (3) When the last part of a phrase of a predicate modified by a modifying clause is a quotative particle such as "to" or "tte", the analysis processing module 2034 recognizes the combination of the modifying clause and the predicate modified by that modifying clause as a quotation. The analysis processing module 2034 then identifies the quotative particle as the recipient of the modification of the predicate modified by the modifying clause. For example, if the sentence after determining the modifying clause is "I said, 'I'm tired'", the last part of the phrase of the predicate "tired" modified by the modifying clause "I" is the quotative particle "to", so the analysis processing module 2034 recognizes "I'm tired" as a quotation. The analysis processing module 2034 then identifies the recipient of the modification of the predicate "tired" as the quotative particle "to". (4) If the subordinate clause does not modify a noun or a quotative particle, the parsing module 2034 determines that the subordinate clause modifies the entire main clause that follows, and identifies the entire main clause as the target of the subordinate clause. For example, in "Grandpa went to mow the lawn, but Grandma went to the river to wash clothes," the subordinate clause "Grandpa went to mow the lawn," which includes the predicate "went," does not modify a noun or a quotative particle, so the parsing module 2034 determines that this subordinate clause modifies the entire main clause "Grandma went to the river to wash clothes."

[0084] Furthermore, if preprocessing is also performed on the supplementary text separated from the main text by the preprocessing module 2033, the analysis module 2034 will perform the processes from step S13 to step S15 on the supplementary text.

[0085] In step S16, the server 20 presents the analysis information (presentation step).

[0086] Specifically, for example, the presentation control module 2035 transmits the analysis information received from the analysis processing module 2034 to the terminal device 10. The presentation control module 2035 controls, for example, the transmitting / receiving unit 192 and the presentation control unit 193 to display the analysis information on the display screen of the display 141. As a result, the server 20 presents the analysis information to the user.

[0087] The parsing information presented to the user is not particularly limited in content, as long as it concerns the dependencies of predicates modified by modifying clauses. For example, the parsing information may present only the identified dependencies in a list format, or it may present the identified dependencies in a format that distinguishes them from other parts of the entire text. Alternatively, the parsing information may present each processing step leading up to the identification of the dependencies by the parsing module 2034, while showing the entire text.

[0088] The following describes an example of the display screen of the display 141 when analysis information is presented, with reference to Figure 6. Figure 6 is a schematic diagram showing an example of the display screen of the display 141 when analysis information is presented.

[0089] In the example shown in Figure 6, the analysis information 1411 is displayed on the display screen of the display 141. Specifically, for example, the original text display field 1412 displays the entire text (hereinafter referred to as the original text) that the server 20 has received as input. In Figure 6, the original text displayed in display field 1412 is "Last week, while having lunch with a friend (restaurant name: ☆☆ Ramen), I accidentally let my cell phone ring." Also, for example, the revised version display field 1413 displays the revised version of the original text after the analysis processing module 2034 has restored the case particles. In Figure 6, the revised version displayed in display field 1413 is "Last week, while having lunch with a friend (restaurant name: ☆☆ Ramen), I accidentally let my cell phone ring." Furthermore, the restored case particles "ni," "ga," and "wo" are displayed in bold and underlined. For example, the display field 1414 for the predicate's dependency shows the dependency identified by the parsing processing module 2034 and the predicate that modifies that dependency. In Figure 6, "Last week, I accidentally rang my cell phone." is displayed as the dependency, and "While I was eating," is displayed as the predicate that modifies that dependency, both in the display field 1414.

[0090] In the example shown in Figure 6, since case particles have been restored in three places in the original text, three correction fields are displayed on the display screen of display 141. Specifically, correction field 1415 shows that a case particle was omitted in the phrase "last week" and that the case particle "ni" has been restored. Correction field 1417 shows that in the phrase "watashi wa" the case particle was overwritten by the particle "wa" and that the case particle "ga" has been restored. Correction field 1419 shows that in the phrase "keitai mo" the case particle was overwritten by the particle "mo" and that the case particle "wo" has been restored.

[0091] Furthermore, a confirmation button 1416 for confirming the changes to the original text is displayed inside display field 1415, a confirmation button 1418 is displayed inside display field 1417, and a confirmation button 1420 is displayed inside display field 1419. In Figure 6, for example, when the operation reception unit 191 receives a tap operation for the confirmation button 1416, the presentation control module 2035 controls the presentation control unit 193 to confirm the restoration of the case particle "ni". Also, for example, when the operation reception unit 191 receives a tap operation for the confirmation button 1418, the presentation control module 2035 controls the presentation control unit 193 to confirm the restoration of the case particle "ga". Also, for example, when the operation reception unit 191 receives a tap operation for the confirmation button 1420, the presentation control module 2035 controls the presentation control unit 193 to confirm the restoration of the case particle "o".

[0092] Note that the example shown in Figure 6 is merely one example. Various variations are conceivable regarding the content and display method of the analysis information 1411 on the display screen of display 141.

[0093] [5 Summary] As described above, in this embodiment, the operation reception unit 191 receives text input operations from the user and transmits the text information to the server 20 via the transmission / reception unit 192. The reception control module 2031 receives the text information transmitted from the terminal device 10. The preprocessing module 2033 performs preprocessing to divide the text indicated by the text information into multiple clauses. The parsing processing module 2034 restores the case particles by referring to the case frame table 2022 for the text divided into multiple clauses by the preprocessing module 2033. Based on the restored case particles and connection information, the parsing processing module 2034 determines a modifying clause from among the divided clauses. Based on the properties of the determined modifying clause, the parsing processing module 2034 identifies the predicate's dependency. The presentation control module 2035 presents the parsing information to the user.

[0094] This prevents situations where, for example, case particles are omitted or overwritten by linking particles, making it unclear which predicate a clause modifies. As a result, the modification relationship between clauses and predicates can be clearly identified, improving the accuracy of dependency parsing. Furthermore, by presenting the parsing information, users can grasp various information about the predicates they modify. Therefore, users have more options for using the results of dependency parsing.

[0095] [6 Variations] In this embodiment, an example was described in which the server 20 determines the modifying clause using only a rule-based method. However, if the server 20 cannot determine the modifying clause using a rule-based method, for example, it may use other methods in combination to determine the modifying clause. Specifically, for example, if the server 20 cannot determine the modifying clause based on the restored case particles, it may determine the modifying clause by probabilistic judgment using case frames.

[0096] More specifically, for example, if the parsing processing module 2034 cannot determine the modifying clause based on the recovered case particle, it may determine the modifying clause by a known probabilistic decision while referring to the case frame table 2022. Various known probabilistic decisions can be envisioned, but for example, a probabilistic decision using the known open-source Cabocha or the known parsing system KNP may be adopted.

[0097] For example, if the sentence to be analyzed by the analysis processing module 2034 is "I also ate curry," both case particles are overwritten by the linking particles "wa" and "mo." In this case, since it is impossible to restore the case particles using a rule-based method as in this embodiment, the analysis processing module 2034 may use probabilistic judgment to restore the two case particles.

[0098] By using this probabilistic judgment in conjunction with other methods, dependency parsing can be reliably performed even in cases where multiple case particles are omitted or the sentences are complex and difficult to read due to grammatical relationships.

[0099] Furthermore, the method used in conjunction with rule-based determination when it is not possible to determine modifying clauses is not limited to the probabilistic judgment described above. For example, server 20 may use an AI model instead of probabilistic judgment. Specifically, for example, the analysis processing module 2034 may input a sentence in which it is not possible to determine modifying clauses based on the restored case particles into an AI model, and have the AI ​​model output a sentence in which the modifying clauses have been determined. In this case, the AI ​​model may be, for example, a trained model that performs natural language processing, and may be a machine learning model or a generative AI model.

[0100] [7. Basic Computer Hardware Configuration] Figure 7 is a block diagram showing the basic hardware configuration of computer 90. Computer 90 includes at least a processor 901, main memory 902, auxiliary storage 903, and a communication interface IF991. These are electrically connected to each other by a communication bus.

[0101] The processor 901 is hardware for executing the instruction set written in a program. The processor 901 consists of an arithmetic unit, registers, peripheral circuits, etc.

[0102] Main memory 902 is used to temporarily store programs and data processed by programs, etc. For example, it is a volatile memory such as DRAM (Dynamic Random Access Memory).

[0103] Auxiliary storage device 903 refers to a storage device for saving data and programs. Examples include flash memory, HDD (Hard Disc Drive), magneto-optical disk, CD-ROM, DVD-ROM, and semiconductor memory.

[0104] The IF991 communication interface is an interface for inputting and outputting signals for communication with other computers via a network using wired or wireless communication standards.

[0105] A network consists of various mobile communication systems built on the internet, LANs, wireless base stations, etc. For example, a network includes 3G, 4G, and 5G mobile communication systems, LTE (Long Term Evolution), and wireless networks that can connect to the internet via designated access points (e.g., Wi-Fi®). When connecting wirelessly, communication protocols include, for example, Z-Wave®, ZigBee®, and Bluetooth®. When connecting via a wired connection, the network also includes connections made directly via USB (Universal Serial Bus) cables, etc.

[0106] Furthermore, by distributing all or part of each hardware configuration across multiple computers 90 and connecting them to each other via a network, a computer 90 can be virtually realized. Thus, the concept of computer 90 includes not only a computer 90 housed in a single enclosure or case, but also a virtualized computer system.

[0107] [8. Basic Functional Configuration of Computer 90] The functional configuration of the computer realized by the basic hardware configuration of computer 90 (Figure 7) will be explained. The computer comprises at least one functional unit: a control unit, a memory unit, and a communication unit.

[0108] Furthermore, the functional units of computer 90 can also be realized by distributing all or part of each functional unit across multiple computers 90 interconnected via a network. The concept of computer 90 includes not only a single computer 90 but also a virtualized computer system.

[0109] The control unit is realized when the processor 901 reads various programs stored in the auxiliary storage device 903, loads them into the main memory device 902, and executes processing according to those programs. The control unit can realize various functional units that perform information processing depending on the type of program. In this way, the computer is realized as an information processing device that performs information processing.

[0110] The memory unit is implemented by the main memory 902 and the auxiliary memory 903. The memory unit stores data, various programs, and various databases. The processor 901 can also reserve memory areas corresponding to the memory unit in the main memory 902 or the auxiliary memory 903 according to the program. The control unit can also cause the processor 901 to perform operations such as adding, updating, and deleting data stored in the memory unit according to the various programs.

[0111] A database, specifically a relational database, is used to manage and link together tabular data sets called masters, which are structurally defined by rows and columns. In a database, tables are called tables, masters are called masters, the columns of tables are called columns, and the rows of tables are called records. In a relational database, relationships can be established and linked between tables and masters.

[0112] Typically, each table and master has a primary key column to uniquely identify records, but setting a primary key column is not mandatory. The control unit can instruct the processor 901 to add, delete, or update records in specific tables and masters stored in the memory unit, according to various programs.

[0113] Furthermore, by storing data, various programs, and various databases in the memory unit, the information processing device and information processing system related to this disclosure can be considered to have been manufactured.

[0114] Furthermore, the databases and masters in this disclosure may include any data structures (lists, dictionaries, associative arrays, objects, etc.) in which information is structurally defined. Data structures also include data that can be considered as data structures by combining data with functions, classes, methods, etc., written in any programming language.

[0115] The communication unit is implemented by the communication IF991. The communication unit provides the functionality to communicate with other computers 90 via the network. The communication unit can receive information transmitted from other computers 90 and input it to the control unit. The control unit can cause the processor 901 to perform information processing on the received information according to various programs. The communication unit can also transmit information output from the control unit to other computers 90.

[0116] Furthermore, each of the above-mentioned configurations, functions, processing units, processing means, etc., may be implemented in hardware, either partially or entirely, by designing them as integrated circuits, for example. The present invention can also be implemented by software program code that realizes the functions of the embodiment. In this case, a storage medium on which the program code is recorded is provided to a computer, and the processor of that computer reads the program code stored in the storage medium. In this case, the program code read from the storage medium itself realizes the functions of the embodiment described above, and the program code itself and the storage medium on which it is stored constitute the present invention. Examples of storage media used to supply such program code include flexible disks, CD-ROMs, DVD-ROMs, hard disks, SSDs, optical disks, magneto-optical disks, CD-Rs, magnetic tapes, non-volatile memory cards, ROMs, and the like.

[0117] Furthermore, the program code that implements the functions described in this embodiment can be implemented in a wide range of programming or scripting languages, such as assembler, C / C++, Perl, Shell, PHP, and Java (registered trademark).

[0118] Furthermore, the program code for the software that implements the functions of the embodiment may be distributed via a network and stored in a storage means such as a computer's hard disk or memory, or in a storage medium such as a CD-RW or CD-R, and the computer's processor may read and execute the program code stored in the storage means or storage medium.

[0119] The functions realized by the components described herein may be implemented in a circuit or processing circuitry, including general-purpose processors, application-specific processors, integrated circuits, ASICs (Application Specific Integrated Circuits), CPUs (a Central Processing Unit), conventional circuits, and / or combinations thereof, programmed to realize the functions described herein. A processor is considered to be a circuit or processing circuitry, including transistors and other circuits. A processor may be a programmed processor that executes a program stored in memory.

[0120] In this specification, circuitry, unit, and means are hardware programmed to perform or execute the functions described herein. Such hardware may be any hardware disclosed herein, or any hardware known to be programmed to perform or execute the functions described herein.

[0121] If the hardware is a processor that is considered to be a type of circuitry, then the circuitry, means, or unit is a combination of hardware and software used to constitute the hardware and / or processor.

[0122] Although several embodiments of this disclosure have been described above, these embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims and their equivalents.

[0123] [9. Addendum] The details described in the above embodiments and each of the modifications are noted below.

[0124] <Note 1> A program to be executed on a computer having a processor and memory, the program causes the processor to perform a reception step of receiving text as input, a preprocessing step of dividing the received text into multiple phrases, a restoration step of restoring case particles using case frames for the text divided into multiple phrases, a determination step of determining, based on the restored case particles, a phrase that modifies a predicate in the text from among the divided multiple phrases as a modifying clause, an identification step of identifying the recipient of the predicate based on the properties of the determined modifying clause, and a presentation step of presenting information about the identified recipient. The program includes the following steps: a first step of extracting parts of speech or groups of parts of speech that serve to connect clauses from among the subordinate clauses contained in the sentence after the case particles have been restored; a second step of determining, based on the connection information of the extracted parts of speech or groups of parts of speech, whether or not they can connect with at least one of the subject and topic contained in the sentence after the case particles have been restored; and a third step of determining at least one of the subject and topic as a modifying clause according to the determination result, wherein the connection information has a data structure that associates, for each part of speech or group of parts of speech, whether or not it can connect with at least one of the subject and topic and the content of the constraints on the connection.

[0125] <Note 2> In the decision step, if it is not possible to determine the modifying clause based on the restored case particle, the program determines the modifying clause by probabilistic judgment using a case frame (see Appendix 1).

[0126] <Note 3> In the third step, the program performs the process of determining the modifying clause from the end of the sentence to the beginning, and if it is determined that a part of speech or group of parts of speech contained in a subordinate clause or parallel clause cannot be connected to the first subject or first topic located earlier than that part of speech or group of parts of speech, the program returns the candidate to which the part of speech or group of parts of speech should be attached to the predicate that was previously the candidate to which it should be attached, and then determines the first subject or first topic as the modifying clause of the returned predicate. If it is determined that a part of speech or group of parts of speech contained in a subordinate clause or parallel clause can be connected to the first subject or first topic, and that there is no second subject or second topic located earlier than the first subject or first topic, the program determines the first subject or first topic as the modifying clause (as described in Appendix 1 or Appendix 2).

[0127] <Note 4> In the preprocessing step, if the input text contains supplementary text and parentheses enclosing the supplementary text, the program described in any of (Appendix 1) to (Appendix 3) separates the supplementary text and parentheses from the text.

[0128] <Note 5> The program described in (Appendix 4) divides the supplementary sentences separated from the input text into multiple clauses during the preprocessing step.

[0129] <Note 6> A program described in any of (Appendix 1) to (Appendix 5) that restores case particles when they have been obscured by linking particles during the restoration step.

[0130] <Note 7> The program described in (Appendix 1) to (Appendix 6) restores case particles when they are omitted during the restoration step.

[0131] <Note 8> An information processing device comprising a control unit and a storage unit, wherein the control unit executes all steps in any of the programs described in (Appendix 1) to (Appendix 7).

[0132] <Note 9> A method to be performed on a computer having a processor and memory, wherein the processor performs all steps in any of the programs described in (Appendix 1) to (Appendix 7).

[0133] <Note 10>

[0134] A system comprising one or more processors that execute all steps in any of the programs described in (Appendix 1) to (Appendix 7). [Explanation of Symbols]

[0135] 1... System 10…Terminal device 120... Communications Department 13…Input device 131…Touch-sensitive devices 14…Output device 141…Display 15…Memory 16…Storage 19… Processor 20... Server 22...Communication IF 23…Input / Output Interface 25…Memory 2 hours… storage 29… Processor

Claims

1. A program to be executed on a computer having a processor and memory, The program is provided to the processor: A reception step for accepting text input, A preprocessing step of dividing the received input text into multiple phrases, A restoration step in which case particles are reconstructed using case frames for the sentence divided into the aforementioned multiple clauses, Based on the restored case particle, a decision step is made to determine, from among the multiple separated phrases, the phrase that modifies the predicate in the sentence is designated as the modifying phrase. Based on the determined nature of the modifying clause, the process involves identifying the recipient of the preceding descriptive word, A presentation step that presents information regarding the identified dependency, and Make it run, The aforementioned decision step is, The first step is to extract from the subordinate clauses contained in the sentence after the case particles have been restored parts of speech or groups of parts of speech that serve to connect clauses, A second step is to determine, based on the connection information of the extracted parts of speech or group of parts of speech, whether the extracted parts of speech or group of parts of speech can be connected to at least one of the subject and topic included in the sentence after the case particle has been restored. The process includes a third step of determining, in accordance with the determination result, that at least one of the subject and the topic be the modifying clause, The connection information is a program having a data structure that associates, for each part of speech or group of parts of speech, whether or not it can be connected to at least one of the subject and the topic, with the content of the constraints on the connection.

2. The program according to claim 1, which, in the determination step, determines the modifying clause by probabilistic judgment using the case frame if it is not possible to determine the modifying clause based on the restored case particle.

3. In step 3, The process of determining the modifying clause is performed from the end of the sentence to the beginning, If the determination result is that the part of speech or group of parts of speech included in the subordinate clause cannot be connected to the first subject or first topic located earlier than the part of speech or group of parts of speech, the candidate to which the part of speech or group of parts of speech is to be modified is returned to the predicate that was previously the candidate to be modified, and the first subject or first topic is determined to be the modifying clause of the returned predicate. The program according to claim 1, wherein the part of speech or group of parts of speech included in the subordinate clause can be connected to the first subject or first topic, and if the determination result is that there is no second subject or second topic ahead of the first subject or first topic, the first subject or first topic is determined to be the modifying clause.

4. The program according to claim 1, wherein, in the preprocessing step, if the text received as input includes a supplementary sentence and parentheses enclosing the supplementary sentence, the program separates the supplementary sentence and parentheses from the text.

5. The program according to claim 4, wherein in the preprocessing step, the supplementary sentence separated from the text that received input is divided into a plurality of clauses.

6. The program according to claim 1, which, in the restoration step, restores the case particle if it has been deleted by a linking particle.

7. The program according to claim 1, which restores the case particle if it is omitted in the restoration step.

8. An information processing device comprising a control unit and a storage unit, wherein the control unit executes all steps in the program described in any one of claims 1 to 7.

9. A method to be performed on a computer comprising a processor and memory, wherein the processor performs all steps of a program according to any one of claims 1 to 7.

10. A system comprising one or more processors that perform all steps in the program according to any one of claims 1 to 7.

Images