Calibration device, calibration method and program
The proofreading device improves sentence proofreading by using morphological analysis to identify and correct word usage through candidate morpheme matching, ensuring accurate and effective proofreading.
Patent Information
- Application Number
- JP2023090935
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-06-01
- Publication Date
- 2024-12-12
AI Technical Summary
Existing proofreading devices struggle to accurately identify and correct inappropriate word usage in sentences, leading to suboptimal proofreading results.
A proofreading device that utilizes morphological analysis to identify candidate morphemes matching prohibited words, sets proofreading locations, and suggests recommended terms based on these matches, correcting word usage as necessary.
Enables accurate identification and correction of inappropriate word usage, ensuring proper proofreading by matching candidate morphemes with prohibited terms and recommending appropriate alternatives.
Smart Images

Figure 2024172894000001_ABST
Abstract
Description
[Technical field]
[0001] The present disclosure relates to a calibration device, a calibration method, and a program. [Background technology]
[0002] There are known devices that automatically proofread sentences. For example, Patent Document 1 describes that when it is determined that the analysis result of an input sentence to be proofread does not match the analysis result of an example sentence to be compared, the input sentence is presented to the user as a part to be proofread. [Prior art documents] [Patent documents]
[0003] [Patent Document 1] JP 2010-182181 A Summary of the Invention [Problem to be solved by the invention]
[0004] However, even when the parts to be corrected are detected based on a match between an example sentence and an input sentence as in Patent Document 1, there is still room for improvement in terms of appropriate proofreading, such as inability to properly identify the parts to be corrected.
[0005] An object of the present disclosure is to provide a proofreading device, a proofreading method, and a program capable of appropriately proofreading a text. [Means for solving the problem]
[0006] The proofreading device for a mobile body according to the present disclosure includes a term acquisition unit that acquires prohibited words and recommended terms associated with the prohibited words, a text acquisition unit that acquires a target text, which is text data to be proofread, an analysis unit that acquires morphemes contained in the prohibited words and morphemes contained in the target text, derived by morphological analysis performed on the prohibited words and the target text, and a selection unit that selects, as a candidate morpheme, a morpheme group including at least one morpheme contained in the target text, the morpheme group whose analysis result of the morpheme analysis matches the prohibited word, and a proofreading unit that sets a proofreading point of the target text and a proofreading candidate that is a term that is a candidate for proofreading at the proofreading point based on a comparison prohibited term that is the prohibited word whose analysis result matches the candidate morpheme and the candidate morpheme, and when the expression of the comparison prohibited term matches the expression of the candidate morpheme, the proofreading unit sets the candidate morpheme to the proofreading point, and sets the recommended term associated with the comparison prohibited word to the proofreading candidate.
[0007] The proofreading method according to the present disclosure includes the steps of: acquiring a prohibited term and a recommended term associated with the prohibited term; acquiring a target text, which is text data to be proofread; acquiring morphemes contained in the prohibited term and morphemes contained in the target text, which are derived by morphological analysis performed on the prohibited term and the target text; selecting, from a group of morphemes including at least one morpheme contained in the target text, a group of morphemes whose analysis result matches the prohibited term as a candidate morpheme; and setting a proofreading point of the target text and a proofreading candidate, which is a term that is a candidate for proofreading at the proofreading point, based on a comparison forbidden term, which is the prohibited term whose analysis result matches the candidate morpheme, and the candidate morpheme. In the step of setting the proofreading point and the proofreading candidate, if the expression of the comparison forbidden term matches the expression of the candidate morpheme, the candidate morpheme is set as the proofreading point, and the recommended term associated with the comparison forbidden word is set as the proofreading candidate.
[0008] The program of the present disclosure causes a computer to execute the steps of: acquiring a prohibited term and a recommended term associated with the prohibited term; acquiring a target text, which is text data to be proofread; acquiring morphemes contained in the prohibited term and morphemes contained in the target text, derived by morphological analysis performed on the prohibited term and the target text; selecting, from a group of morphemes including at least one morpheme contained in the target text, a group of morphemes whose analysis result matches the prohibited term as a candidate morpheme; and setting a proofreading point of the target text and a proofreading candidate, which is a term that is a candidate for proofreading at the proofreading point, based on a comparison forbidden term, which is the prohibited term whose analysis result matches the candidate morpheme, and the candidate morpheme. In the step of setting the proofreading point and the proofreading candidate, if the expression of the comparison forbidden term matches the expression of the candidate morpheme, the candidate morpheme is set to the proofreading point, and the recommended term associated with the comparison forbidden word is set to the proofreading candidate. Effect of the Invention
[0009] According to the present disclosure, text can be properly proofread. [Brief description of the drawings]
[0010] [Figure 1] FIG. 1 is a schematic block diagram of a calibration device according to this embodiment. [Diagram 2] FIG. 2 is a table showing an example of dictionary data. [Diagram 3] FIG. 3 is a table showing an example of the result of the morphological analysis. [Figure 4] FIG. 4 is a table showing an example of the result of the morphological analysis. [Diagram 5] FIG. 5 is a table showing an example of the result of the morphological analysis. [Figure 6] FIG. 6 is a table showing an example of the result of the morphological analysis. [Figure 7] FIG. 7 is a table showing an example of the result of the morphological analysis. [Figure 8] FIG. 8 is a flowchart illustrating a processing flow according to the first embodiment. [Figure 9] FIG. 9 is a table showing an example of dictionary data in the second embodiment. [Figure 10] FIG. 10 is a flowchart illustrating a processing flow according to the second embodiment. [Figure 11] FIG. 11 is a table showing an example of dictionary data in the third embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0011] Preferred embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings. Note that the present disclosure is not limited to these embodiments, and when there are multiple embodiments, the present disclosure also includes combinations of the embodiments.
[0012] (First embodiment) (calibration device) The proofreading device 10 according to this embodiment sets a proofreading point and proofreading candidates for the proofreading point from a target text to be proofread. A proofreading point refers to a character string that needs proofreading among character strings included in the target text. A proofreading candidate refers to a character string (term) that is a candidate for proofreading for the character string at the proofreading point.
[0013] FIG. 1 is a schematic block diagram of the calibration device according to the present embodiment. The calibration device 10 according to the present embodiment is, for example, a computer, and has an input unit 12, an output unit 14, a communication unit 16, a storage unit 18, and a control unit 20, as shown in FIG. 1. The input unit 12 is a device that accepts user operations, and may be, for example, a mouse, a keyboard, a touch panel, or the like. The output unit 14 is a device that outputs information, and may be, for example, a display that displays an image, or the like. The communication unit 16 is a module that communicates with an external device, and may include, for example, an antenna, or the like. In this embodiment, the communication method by the communication unit 16 is wireless communication, but the communication method may be any communication method. The calibration device 10 may not have the input unit 12, the output unit 14, and the communication unit 16. The calibration device 10 may be configured as a single device, may be configured integrally with other devices, or may be configured as a system combining various devices such as a calculation device and a data server, and is not particularly limited.
[0014] The storage unit 18 is a memory that stores various information such as the contents of calculations and programs of the control unit 20, and includes at least one of a main storage device such as a RAM (Random Access Memory), a ROM (Read Only Memory), and an external storage device such as an HDD (Hard Disk Drive). The programs for the control unit 20 saved in the storage unit 18 may be stored in a recording medium readable by the calibration device 10.
[0015] The control unit 20 is a calculation device and includes a calculation circuit such as a CPU (Central Processing Unit). The control unit 20 includes a term acquisition unit 30, a text acquisition unit 32, an analysis unit 34, a selection unit 36, a proofreading unit 38, and an output control unit 40. The control unit 20 realizes the term acquisition unit 30, the text acquisition unit 32, the analysis unit 34, the selection unit 36, the proofreading unit 38, and the output control unit 40 by reading and executing a program (software) from the storage unit 18, and executes the processes. The control unit 20 may execute these processes using one CPU, or may include multiple CPUs and execute the processes using the multiple CPUs. At least a part of the term acquisition unit 30, the text acquisition unit 32, the analysis unit 34, the selection unit 36, the proofreading unit 38, and the output control unit 40 may be realized by hardware.
[0016] The processing contents of the calibration device 10 will be explained below.
[0017] (Getting dictionary data) The term acquisition unit 30 acquires dictionary data to be used for proofreading. The dictionary data is a glossary of terms to be used for proofreading. In this embodiment, the dictionary data is data including prohibited terms and recommended terms associated with the prohibited terms. The prohibited terms are terms (character strings) that should not be used, and the recommended terms are terms (character strings) that should be used instead of the prohibited terms. In other words, when a prohibited term is used, it is recommended to use the recommended term instead of the prohibited term. If the associated prohibited term and recommended term are considered to be one set of corresponding terms, the term acquisition unit 30 acquires dictionary data including multiple sets of corresponding terms.
[0018] The dictionary data (prohibited terms and recommended terms in this embodiment) is set in advance, and the term acquisition unit 30 acquires the preset dictionary data. The contents of the dictionary data may be set as appropriate by the entity using the proofreading device 10. The method of acquiring the dictionary data may be arbitrary; for example, the term acquisition unit 30 may read out dictionary data stored in advance in the memory unit 18, or may acquire dictionary data from another device (such as a server) via the communication unit 16.
[0019] FIG. 2 is a table showing an example of dictionary data. In this embodiment, the prohibited words and recommended words included in the dictionary data (first dictionary data) have the same pronunciation, but at least some of the character strings have different character types. The character type refers to the type of character form, such as katakana, hiragana, kanji, and roman letters. In the example of the dictionary data in FIG. 2, the prohibited word "erase" is associated with the recommended word "erase", the prohibited word "great" is associated with the recommended word "great", and the prohibited word "show superiority" is associated with the recommended word "show superiority". However, the prohibited words and recommended words are not limited to those in the example in FIG. 2, and the number and contents of the prohibited words and recommended words may be set appropriately.
[0020] (Get target text) The text acquisition unit 32 acquires target text. The target text is text data (data including a character string) to be proofread. Any method of acquiring the target text may be used. For example, the text acquisition unit 32 may acquire the target text from another device (such as a server) via the communication unit 16, or may acquire text data input by the user to the input unit 12 as the target text.
[0021] (Morphological analysis) The analysis unit 34 performs morphological analysis on the target text and obtains the analysis result. Morphological analysis refers to a process of dividing a character string to be analyzed into morphemes. In this embodiment, the analysis result of the morphological analysis includes morphemes included in the character string to be analyzed and information on the normal form of the morpheme. It is preferable that the analysis result of the morphological analysis further includes at least one (preferably all) of the part of speech, conjugation, and dictionary form of the morpheme. The morphological analysis may be performed using any morphological analyzer.
[0022] 3 to 7 are tables showing examples of the results of morpheme analysis. Fig. 3(A) shows an example of the results of morpheme analysis when the target text is the character string "drown out the speech." In this case, the character string of the target text is divided into four morphemes (surface forms) - "speech," "o," "drown out," and "." - and the analysis results for each morpheme (in this example, dictionary form, part of speech, normal form, type of conjugation, and conjugation form) are derived.
[0023] The analysis unit 34 also acquires the analysis results of the morphological analysis performed on the prohibited words. The analysis unit 34 acquires the analysis results of the morphological analysis for each prohibited word included in the dictionary data. The analysis unit 34 may acquire the analysis results of the prohibited words by any method, for example, may acquire the analysis results by performing a morphological analysis of the prohibited words when the target text is acquired, or may read out the analysis results performed in advance from the storage unit 18. Note that the prohibited words may include only one morpheme, or may include multiple morphemes. (B) of FIG. 3 shows an example of the result of the morphological analysis of the prohibited word "scuff out".
[0024] (Candidate morpheme selection) The selection unit 36 selects, as a candidate morpheme, a morpheme group whose analysis result of the morpheme analysis matches a prohibited word from among the morpheme groups contained in the target text. A morpheme group refers to one that includes at least one morpheme contained in the target text, and refers to one morpheme in the target text, or two or more morphemes that are consecutively arranged in the target text. For example, "kakekaesu" (to erase) shown in Fig. 3 is a morpheme group consisting of one morpheme, and "wei furira" (to be explained later in Fig. 7) is a morpheme group consisting of two consecutive morphemes ("wei" and "furira").
[0025] The selection of the candidate morpheme will be described in more detail. The selection unit 36 selects the candidate morpheme and the comparison forbidden term based on the analysis result of each morpheme included in the target text and the analysis result of each forbidden term. That is, the selection unit 36 selects, as a candidate morpheme, a morpheme group whose analysis result matches any of the forbidden terms from among the morpheme groups included in the target text. A comparison forbidden term can be said to be a forbidden term whose analysis result matches the candidate morpheme among the forbidden terms. The analysis result here refers to the normal form and part of speech of the morpheme (or morpheme group). That is, in this embodiment, the selection unit 36 determines that the analysis results match when both the normal form and the part of speech match, and determines that the analysis results do not match when at least one of the normal form and the part of speech does not match.
[0026] In the example of Figure 3, since the prohibited term "scrape out" is present, among the morphemes contained in the target text, the morpheme "scrape out" whose normal form and part of speech match those of the prohibited term is selected as a candidate morpheme.
[0027] (Setting the correction points and correction candidates) The proofreading unit 38 sets the proofreading portion of the target text and the proofreading candidates for the proofreading portion based on the candidate morpheme and the comparison forbidden term. Specifically, the proofreading unit 38 judges whether the expression of the candidate morpheme matches the expression of the comparison forbidden term, and sets the proofreading portion and the proofreading candidates based on the judgment result. The expression of the candidate morpheme refers to the surface form of the candidate morpheme, and the expression of the comparison forbidden term refers to the surface form of the comparison forbidden term.
[0028] (When the candidate morpheme and the expression of the prohibited comparison term match) When the expression of the candidate morpheme and the expression of the comparison prohibited term match, the proofreading unit 38 sets the character string corresponding to the candidate morpheme among the character strings of the target text as the proofreading point, and sets the recommended term associated with the comparison prohibited term as the proofreading candidate. In the example of FIG. 3, the expression (surface form) of the candidate morpheme, "scrape out", matches the expression (surface form) of the comparison prohibited term, "scrape out". Therefore, the proofreading unit 38 sets "scrape out" among the target text, "scrape out the voice", as the proofreading point. Then, the proofreading unit 38 sets "scrape out", the recommended term associated with the comparison prohibited term, "scrape out", as the proofreading candidate.
[0029] (When the candidate morpheme and the expression of the prohibited comparison term do not match) On the other hand, if the expression of the candidate morpheme and the expression of the comparison-prohibited term do not match, the proofreading unit 38 corrects the recommended term associated with the comparison-prohibited term based on the analysis result of the candidate morpheme. Then, the proofreading unit 38 sets the candidate morpheme as a proofreading point, and sets the corrected recommended term as a proofreading candidate. The method of correcting the recommended term based on the analysis result of the candidate morpheme may be arbitrary, but in this embodiment, the proofreading unit 38 corrects the expression of the recommended term based on the conjugation of the candidate morpheme. Specifically, the proofreading unit 38 corrects the recommended term so that it has the same conjugation form as the conjugation form of the candidate morpheme. Hereinafter, the corrected recommended term will be referred to as a corrected term as appropriate.
[0030] FIG. 4 shows an example in which the target text is "Drown out the voices.", where (A) in FIG. 4 shows the morpheme analysis result of the target text, and (B) shows the recommended term and the correction term. In the example of FIG. 4, "Drown out" is set as a candidate morpheme, and "Drown out" (not shown in FIG. 4), which has the same normal form as "Drown out", is set as a comparison prohibited term. In the example of FIG. 4, the surface form of the candidate morpheme ("Drown out") does not match that of the comparison prohibited term ("Drown out"). Therefore, the proofreading unit 38 corrects "Drown out", which is the recommended term associated with the comparison prohibited term ("Drown out"), to the conjugated form (general imperative form) of the candidate morpheme, and sets the corrected expression "Drown out" as the correction term. The proofreading unit 38 sets "Drown out" in the target text "Drown out the voices." as the part to be corrected, and sets "Drown out" as a candidate for correction.
[0031] In the example of FIG. 4, there is only one correction term, but there may be a plurality of correction terms. The process in this case will be described. As described above, the proofreading unit 38 sets the recommended term corrected to the same conjugation form as the candidate morpheme as the correction term. When there are a plurality of correction terms, that is, when there are correction terms with the same conjugation form but different expressions, the proofreading unit 38 judges whether there is a correction term whose final character matches the final character of the candidate morpheme among the plurality of correction terms. When there is a correction term whose final character matches the candidate morpheme, the proofreading unit 38 sets the correction term whose final character matches the final character among the plurality of correction terms as a proofreading candidate.
[0032] FIG. 5 shows an example in which the target text is "Don't erase the voice.", and FIG. 4 (A) shows the morpheme analysis result of the target text, and (B) shows the recommended term and the correction term. In the example of FIG. 5, "kakekasa" is set as the candidate morpheme, and "kakekasa" (not shown in FIG. 5), which has the same normal form as "kakekasa", is set as the comparison prohibited term. In the example of FIG. 5, the surface form of the candidate morpheme ("kakekasa") does not match that of the comparison prohibited term ("kakekasa"). Therefore, the proofreading unit 38 corrects "kakekasa", which is the recommended term associated with the comparison prohibited term ("kakekasa"), to the conjugated form (general potential form) of the candidate morpheme, and sets the correction term as the term. There are two general potential form conjugations of "kakekasa", "kakekasa" and "kakekasa", so there are two correction terms, "kakekasa" and "kakekasa". In this case, the proofreading unit 38 sets "kakeisa" (not kakeisarenai) in the target text "kakei wo kakeisarenai" as the part to be proofread, and sets "kakeisa" (not kakeisarenai) whose final character matches the candidate morpheme as a proofreading candidate.
[0033] If there is no correction term whose final character matches the final character of the candidate morpheme among the multiple correction terms, the proofreading unit 38 sets all correction terms as correction candidates while setting the candidate morpheme as the correction point. That is, for example, in the example of Fig. 5, if the final characters of both the correction terms "kakesa" and "kakeso" do not match the final characters of the candidate morpheme, both "kakesa" and "kakeso" are set as proofreading candidates.
[0034] (Result output) The output control unit 40 outputs the processing result by the proofreading unit 38. The output control unit 40 outputs information indicating the proofreading parts and the proofreading candidates. Any output method may be used by the output control unit 40. For example, the output control unit 40 may cause the output unit 14 to output the information indicating the proofreading parts and the proofreading candidates (for example, to display the information on a display). Also, for example, the output control unit 40 may transmit (output) the information indicating the proofreading parts and the proofreading candidates to another device via the communication unit 16.
[0035] (If calibration is not performed) Although the case where the calibration part and the calibration candidates are set has been described above, there are also cases where the calibration device 10 does not perform calibration (does not set the calibration part and the calibration candidates). The case where calibration is not performed will be described below.
[0036] If the target text does not contain any candidate morphemes (i.e., if the target text does not contain any morphemes whose normal forms match the prohibited words), the proofreading unit 38 decides not to proofread and does not need to set the proofreading points and proofreading candidates.
[0037] If the target text contains a candidate morpheme, but the candidate morpheme is a stem, the proofreading unit 38 may determine not to proofread and may not set the proofreading part and the proofreading candidate. A stem refers to a part of a word that does not change even when it is conjugated, and whether it is a stem is analyzed by morphological analysis. That is, if the analysis result of the selected candidate morpheme indicates that the candidate morpheme is a stem, the proofreading unit 38 determines not to proofread and does not set the proofreading part and the proofreading candidate. FIG. 6 shows an example in which the target text is "Not very powerful for his age." and the prohibited word is "I". (A) in FIG. 6 is an example of the morphological analysis result of the target text, and (B) is an example of the morphological analysis result of a comparison prohibited word. In the example of FIG. 6, "I" in the target text is set as a candidate morpheme, but since "I" is a stem, the proofreading unit 38 determines not to proofread.
[0038] If the target text contains a candidate morpheme but the expression of the candidate morpheme matches the expression of a recommended term, the proofreading unit 38 may determine not to proofread and may not set the proofreading portion and the proofreading candidate. Also, if the target text contains a candidate morpheme but the expression of the candidate morpheme matches the expression of a preset permitted term, the proofreading unit 38 may determine not to proofread and may not set the proofreading portion and the proofreading candidate. The permitted term is a term that is permitted to be used, i.e., a whitelist, and is preset.
[0039] (If the prohibited term contains multiple morphemes) In the examples shown in Figs. 3 to 6, the prohibited term and the candidate morpheme are each composed of one morpheme. However, the prohibited term may contain multiple morphemes, and in this case, the candidate morpheme also contains multiple morphemes. Even when multiple morphemes are included, the correction candidates and the correction points are set in the same manner as described above, so a detailed description is omitted, but an example is described below based on Fig. 7. Fig. 7 shows an example in which the target text is "Not very powerful for his age." and the prohibited term is "Powerful." Fig. 6 (A) shows the morpheme analysis result of the target text, (B) shows the morpheme analysis result of the comparative prohibited term, and (C) shows the recommended term and the correction term. In the example of Fig. 7, a group of morphemes ("Powerful and Ra") whose analysis results (normal form and part of speech) match each morpheme of the comparative prohibited term ("Powerful") in the target text are selected as candidate morphemes. In the example of FIG. 7, since the expressions of the comparison-prohibited term ("i furu") and the candidate morpheme ("i fura") do not match, "i bura", which is a correction of the recommended term "i buru", is selected as a candidate for correction.
[0040] (Processing flow) The process flow of setting the proofreading part and the proofreading candidate as described above will be described. FIG. 8 is a flowchart for explaining the process flow according to the first embodiment. As shown in FIG. 8, the proofreading device 10 acquires the target text by the text acquisition unit 32 (step S10), and executes morphological analysis on the target text by the analysis unit 34 (step S12). The proofreading device 10 uses the selection unit 36 to determine whether the target text includes a candidate morpheme whose analysis result matches the prohibited term (step S14). If there is no candidate morpheme (step S14: No), the proofreading device 10 proceeds to step S16 and determines not to proofread. On the other hand, if there is a candidate morpheme (step S14; Yes), the proofreading device 10 uses the proofreading unit 38 to determine whether the candidate morpheme is a stem (step S18). If it is a stem (step S18; Yes), the proofreading device 10 proceeds to step S16. If it is not a stem (step S18; No), the proofreading device 10 determines whether the candidate morpheme matches an expression of a recommended term or an allowed term (step S20). If the candidate morpheme matches the recommended term or permitted term (step S20; Yes), the process proceeds to step S16. If the candidate morpheme does not match the recommended term or permitted term (step S20; No), the process proceeds to step S22. In step S22, the proofreading unit 38 judges whether the expression of the candidate morpheme and the comparison-prohibited term match (step S22). If the expression of the candidate morpheme and the comparison-prohibited term match, the proofreading unit 38 sets the candidate morpheme as the proofreading part and the recommended term corresponding to the comparison-prohibited term as the proofreading candidate (step S24). On the other hand, if the expression of the candidate morpheme and the comparison-prohibited term do not match (step S22; No), the proofreading unit 38 corrects the recommended term to be the inflected form of the candidate morpheme (step S26), sets the candidate morpheme as the proofreading part, and sets the corrected recommended term (corrected term) as the constituent candidate (step S28). In step S28, if there are multiple corrected recommended terms (corrected terms), proofreading candidates are set according to the case, as described above.
[0041] After executing any of steps S16, S24, or S28, the process proceeds to step S30, and if the process is to be ended (step S30; Yes), the process ends, and if the process is not to be ended (step S30; No), the process returns to step S10 and continues. However, if the target text contains multiple candidate morphemes, the process from step S18 onwards is executed for all candidate morphemes before proceeding to step S30.
[0042] (effect) As described above, in this embodiment, the target text is compared with the analysis result of the morpheme analysis of the prohibited word, a candidate morpheme is extracted, and the candidate morpheme is set as the correction point. Then, if the extracted candidate morpheme matches the expression of the prohibited word (comparatively prohibited word), the recommended word corresponding to the prohibited word is set as a correction candidate. In this embodiment, since the candidate morpheme is extracted based on the analysis result of the morpheme analysis, it is possible to find the correction point even if the expression does not completely match the prohibited word, and the correction point can be suitably detected. Therefore, according to this embodiment, it is possible to properly extract the character string to be corrected from the target text, and to properly correct the extracted character string.
[0043] Furthermore, in this embodiment, when the expression of the extracted candidate morpheme does not match that of the comparison prohibited word, that is, when the analysis result is the same but the expression is different, the recommended term (corrected term) corrected based on the analysis result of the candidate morpheme is set as a proofreading candidate. Therefore, even if the expression does not completely match the prohibited word, it is possible to find the part to be proofread, and the proofreading part can be more suitably detected. In addition, since the recommended term corrected to match the candidate morpheme is set as a proofreading candidate, it is possible to proofread while matching the expression to the original target text.
[0044] Second embodiment Next, the second embodiment will be described. In the first embodiment, candidate morphemes are selected from the analysis results of morphological analysis, but in the second embodiment, a morpheme group including expressions of prohibited terms is selected as candidate morphemes, which is different from the first embodiment. In the second embodiment, the description of the configuration common to the first embodiment will be omitted. Note that the second embodiment and the first embodiment may be combined, in other words, both the process described in the first embodiment and the process described in the second embodiment may be performed.
[0045] (Dictionary data) FIG. 9 is a table showing an example of dictionary data in the second embodiment. In the second embodiment, the term acquisition unit 30 acquires a plurality of sets of corresponding terms (corresponding prohibited terms and recommended terms) as dictionary data. The dictionary data (second dictionary data) in the second embodiment includes target terms in which the recommended terms include the same character string as the corresponding prohibited term and characters not included in the prohibited term. FIG. 9 shows an example of such corresponding terms in which the prohibited term is "potential energy" and the recommended term is "potential energy".
[0046] Furthermore, the dictionary data of the second embodiment may also include corresponding terms (prohibited terms and recommended terms) in which the readings of the prohibited terms and the recommended terms are the same but the type of characters is different at least in part. For corresponding terms that are read the same but the type of characters is different at least in part, it is preferable to set the part of speech of the prohibited terms (or recommended terms). In FIG. 9, an example of such corresponding terms is a prohibited term "saru" (monkey), a recommended term "saru" (monkey), and a part of speech that is a noun. Another example is a prohibited term "akuru" (scarlet) and a recommended term "akikuru" (light), and a prohibited term "roshi" (roshi) and a recommended term "roshi" (paper).
[0047] (Morphological analysis) In the second embodiment, the analysis result of the morphological analysis of the target text is obtained in the same manner as in the first embodiment. In the second embodiment, it is not necessary to obtain the analysis result of the morphological analysis of the prohibited words.
[0048] (Candidate morpheme selection) In the second embodiment, the selection unit 36 selects, from among the morphemes included in the target text, a morpheme group including an expression of a prohibited term as a candidate morpheme. Specifically, the selection unit 36 selects, from among the morphemes included in the target text, a morpheme including an expression of any prohibited term as a candidate morpheme, and selects a prohibited term whose expression is included in the candidate morpheme as a comparatively prohibited term. Note that the expression here refers to a surface form, as in the first embodiment.
[0049] (Setting the correction points and correction candidates) In the second embodiment, the proofreading unit 38 sets the proofreading portion of the target text and the proofreading candidates for the proofreading portion based on the analysis result of the morpheme analysis of the candidate morpheme and the comparison prohibited word. A specific setting method will be described below.
[0050] (When the candidate morpheme contains multiple morphemes) When the number of morphemes included in the candidate morpheme indicated in the analysis result of the candidate morpheme is multiple (two or more), the proofreading unit 38 sets the candidate morpheme as the proofreading point, and sets the recommended term associated with the comparison prohibited term as the proofreading candidate. For example, when the target text is "The peak of Mt. Fuji has a large potential energy," and there is a prohibited term "potential energy," the proofreading unit 38 sets "potential energy" in the target text as the candidate morpheme, and sets the prohibited term "potential energy" as the comparison prohibited term. In this case, the candidate morpheme "potential energy" includes two morphemes, "position" and "energy." Therefore, the proofreading unit 38 sets "potential energy" in the target text as the proofreading point, and sets "potential energy" associated with the comparison prohibited term as the proofreading candidate.
[0051] (When the number of morphemes included in the candidate morpheme is one) When the number of morphemes included in the candidate morpheme is one, the proofreading unit 38 judges whether the part of speech of the candidate morpheme matches the part of speech of the comparison-prohibited term. When the part of speech of the candidate morpheme matches the part of speech of the comparison-prohibited term, the proofreading unit 38 sets the candidate morpheme as a proofreading point, and sets the recommended term associated with the comparison-prohibited term as a proofreading candidate. On the other hand, when the part of speech of the candidate morpheme does not match the part of speech of the comparison-prohibited term, the proofreading unit 38 does not proofread the candidate morpheme, and does not set a proofreading point or a proofreading candidate. For example, when the target text is "Saru ga saru." and there is a prohibited term "saru" (part of speech is noun), the proofreading unit 38 sets "saru" and "saru" in the target text as candidate morphemes, and sets the prohibited term "saru" as a comparison-prohibited term. In this case, the first candidate morpheme "saru" is a noun and matches the part of speech of the prohibited word, so the first candidate morpheme is selected as the part of correction, and the recommended term "saru" that corresponds to the comparative prohibited word "saru" is selected as the candidate for correction. On the other hand, the second candidate morpheme "saru" is a verb and does not match the part of speech of the prohibited word, so the second candidate morpheme is not corrected.
[0052] In addition, if a part of speech is not set for the comparison prohibited term in the dictionary data, it is not necessary to judge whether the parts of speech match. That is, in this case, if the number of morphemes included in the candidate morpheme is one and a part of speech is not set for the comparison prohibited term, the proofreading unit 38 sets the candidate morpheme as the proofreading point and sets the recommended term associated with the comparison prohibited term as the proofreading candidate. For example, if the target text is "The morning of the next day" and there is a prohibited term "Akuru" (no part of speech is set), the proofreading unit 38 sets "Akuru" in the target text as the candidate morpheme and the prohibited term "Akuru" as the comparison prohibited term. In this case, since a part of speech is not set for the comparison prohibited term, the proofreading unit 38 sets "Akuru" in the target text as the proofreading point and "Akuru" associated with the comparison prohibited term as the proofreading candidate.
[0053] (If calibration is not performed) In the second embodiment, the case where the candidate morpheme is not proofread will be described below.
[0054] If the target text does not contain any candidate morphemes (i.e., if the target text does not contain any morphemes that contain prohibited word expressions), the proofreading unit 38 decides not to proofread and does not need to set the proofreading points and proofreading candidates.
[0055] The proofreading unit 38 may determine not to proofread and not set proofreading points and proofreading candidates when the target text contains a candidate morpheme but the expression of the candidate morpheme matches the expression of a recommended term. Also, the proofreading unit 38 may determine not to proofread and not set proofreading points and proofreading candidates when the target text contains a candidate morpheme but the expression of the candidate morpheme matches the expression of a preset allowed term.
[0056] If the target text contains a candidate morpheme, but both the first and last characters of the candidate morpheme are different from the comparison prohibited word, the proofreading unit 38 may determine that proofreading is not performed, and may not set the proofreading portion and the proofreading candidate. That is, if the first character of the candidate morpheme is different from the first character of the comparison prohibited word and the last character of the candidate morpheme is different from the last character of the comparison prohibited word, the proofreading unit 38 does not perform proofreading. For example, if the target text is "yoroshi onegaishimasu" and the prohibited word includes "roshi", the proofreading unit 38 sets "yoroshi" in the target text as a candidate morpheme and the prohibited word "roshi" as a comparison prohibited word. In this case, since both the first and last characters of the candidate morpheme are different from the first and last characters of the prohibited word, the proofreading unit 38 does not perform proofreading. Note that if only one of the first and last characters is different, proofreading is performed by the method described above.
[0057] When a candidate morpheme contains a plurality of types of comparison forbidden terms, the proofreading process is performed in the above-described manner for each morpheme (group) that is included in the candidate morpheme and corresponds to the comparison forbidden terms.
[0058] (Processing flow) The process flow of the proofreading described above will be explained. FIG. 10 is a flow chart for explaining the process flow according to the second embodiment. As shown in FIG. 10, the proofreading device 10 acquires the target text by the text acquisition unit 32 (step S40), and executes morphological analysis on the target text by the analysis unit 34 (step S42). The proofreading device 10 judges by the selection unit 36 whether the target text includes a candidate morpheme including an expression of a prohibited term (step S44), and if there is no candidate morpheme (step S44: No), proceeds to step S46 and decides not to proofread. On the other hand, if there is a candidate morpheme (step S44; Yes), the proofreading device 10 judges by the proofreading unit 38 whether the candidate morpheme matches an expression of a recommended term or an allowed term (step S48). If the candidate morpheme matches the recommended term or the allowed term (step S48; Yes), proceeds to step S46, and if there is no match (step S48; No), proceeds to step S50. In step S50, the proofreading unit 38 judges whether both the first character and the last character of the candidate morpheme are different from the comparison prohibited term (step S50). If both are different (step S50; Yes), the process proceeds to step S46. If both are not different (step S50; No), the process proceeds to step S52. In step S52, the proofreading unit 38 judges whether the number of morphemes included in the candidate morpheme is two or more (step S52). If the number is two or more, the candidate morpheme is set as the proofreading point, and the recommended term corresponding to the comparison prohibited term is set as the proofreading candidate (step S54). On the other hand, if the number is not two or more (step S52; No), the proofreading unit 38 judges whether the part of speech of the candidate morpheme and the comparison prohibited term match (step S56). If the part of speech is the same (step S56; Yes), the process proceeds to step S54. If the part of speech is not the same (step S56; No), the process proceeds to step S46. If the part of speech is not set for the comparison prohibited term, the process may proceed to step S54 without executing step S56.
[0059] After executing either step S46 or S54, the process proceeds to step S58. If the process is to be ended (step S58; Yes), the process ends. If the process is not to be ended (step S58; No), the process returns to step S40 and continues.
[0060] (effect) As described above, in the second embodiment, candidate morphemes including expressions of prohibited words are extracted, and if the number of candidate morphemes is two or more, the recommended term associated with the prohibited word is set as a proofreading candidate. If the number of candidate morphemes is one, the recommended term associated with the prohibited word is set as a proofreading candidate only if the parts of speech match. Therefore, it is possible to detect terms to be proofread based on the parts of speech, such as the above-mentioned "saru ga saru.", and it is possible to prevent terms with different parts of speech from being proofread. Therefore, according to this embodiment, proofreading can be performed appropriately.
[0061] Third embodiment The third embodiment differs from the first embodiment in that the dictionary data includes a target term and the format of an additional term located before or after the target term. In the third embodiment, a description of parts of the configuration common to the first embodiment will be omitted. Note that the third embodiment may be combined with at least one of the first and second embodiments. In other words, both the process described in the third embodiment and the process in at least one of the first and second embodiments may be performed.
[0062] FIG. 11 is a table showing an example of dictionary data in the third embodiment. In the third embodiment, the term acquisition unit 30 acquires data including a target term and a format of an additional term as dictionary data (third dictionary data). The target term is a term that is a target of proofreading. The additional term is a character located before or after (immediately before or immediately after) the target term, and the format of the additional term refers to the type of characters of the additional term (katakana, hiragana, kanji, roman letters, numbers, etc.). The formats of the target term and the additional term may be appropriately set by the entity using the proofreading device 10. Furthermore, the dictionary data in the third embodiment may include multiple sets of associated target terms and formats of additional terms.
[0063] In the example of FIG. 11, the target term is "secondary stress", the additional term is the character immediately before the target term, and the form of the additional term is a number.
[0064] (Morphological analysis) Also in the third embodiment, in the same manner as in the first embodiment, the analysis result of morphological analysis for the target text is obtained.
[0065] (Selection of candidate morphemes) In the third embodiment, the selection unit 36 selects, as candidate morphemes, a group of morphemes whose expression matches the target term from among the group of morphemes included in the target text. Specifically, the selection unit 36 selects, as a group of morphemes of the target text and a target term whose expressions match each other, based on each morpheme included in the target text and each target term, as a candidate morpheme and a comparison target term. That is, the selection unit 36 selects, as candidate morphemes, a group of morphemes whose expression matches any of the respective prohibited terms from among the group of morphemes included in the target text. Here, the expression refers to the surface form in the same manner as in the first embodiment. That is, for example, when the target text is "Calculate the primary stress." and the target term is "secondary stress", the correction unit 38 selects "secondary stress" in the target text as a candidate morpheme.
[0066] (Setting of correction location and correction candidate) In the third embodiment, the correction unit 38 sets the correction location of the target text and the correction candidate at that correction location based on the candidate morpheme and the form of the additional term (comparison additional term) corresponding to the comparison target term. Hereinafter, a specific setting method will be described.
[0067] The correction unit 38 extracts, as an additional morpheme, a morpheme at the same position as the position of the comparison additional term with respect to the comparison target term for the candidate morpheme. That is, for example, in the example of FIG. 11, since the position of the comparison additional term with respect to the comparison target term is immediately before the comparison target term, the correction unit 38 extracts "one" immediately before the candidate morpheme in the target text ("Calculate the primary stress.") as an additional morpheme.
[0068] The correction unit 38 acquires the form (type of characters) of the additional morpheme from the analysis result of morphological analysis of the target text. The correction unit 38 determines whether the form of the additional morpheme is different from the form of the comparative addition term shown in the dictionary data. If they are different, the additional morpheme is set as the correction location, and the term (characters) obtained by correcting the form of the additional morpheme to the form of the comparative addition term is used as the correction candidate. That is, in the above example, the additional morpheme "一" is a Chinese character, and the form of the comparative addition term is a number, so the forms are different. Therefore, the correction unit 38 sets "一" in the target text as the correction location and "1" converted to a number as the correction candidate. Note that a term composed of an additional morpheme and a candidate morpheme may be set as the correction location, and a term obtained by correcting the additional morpheme part of the term composed of the additional morpheme and the candidate morpheme to the form of the comparative addition term may be used as the correction candidate. That is, for example, "一次応力" in the target text may be set as the correction location, and "1次応力" may be used as the correction candidate.
[0069] Note that if the form of the additional morpheme is the same as the form of the comparative addition term shown in the dictionary data, no correction is performed.
[0070] In the third embodiment, candidate morphemes that match the target term are extracted, and it is determined whether to perform correction by comparing the form of the additional morpheme of the candidate morpheme with the form of the comparative addition term. Therefore, without uniformly correcting the form, characters before and after the target term can be selected for correction, so that the term to be corrected can be appropriately detected and corrected.
[0071] (Effect) As described above, the proofreading device 10 according to the first embodiment of the present disclosure includes a term acquisition unit 30, a text acquisition unit 32, an analysis unit 34, a selection unit 36, and a proofreading unit 38. The term acquisition unit 30 acquires prohibited terms and recommended terms associated with the prohibited terms, and the text acquisition unit 32 acquires a target text, which is text data to be proofread. The analysis unit 34 acquires morphemes included in the prohibited terms and morphemes included in the target text, which are derived by morphological analysis performed on the prohibited terms and the target text. The selection unit 36 selects, as a candidate morpheme, a morpheme group whose analysis result of the morphological analysis matches the prohibited term from among morpheme groups including at least one morpheme included in the target text. The proofreading unit 38 sets a proofreading point and a proofreading candidate of the target text based on the comparison prohibited term and the candidate morpheme. When the expression of the comparison prohibited term matches the expression of the candidate morpheme, the proofreading unit 38 sets the candidate morpheme as a proofreading point, and sets the recommended term associated with the comparison prohibited term as a proofreading candidate. According to the present disclosure, it is possible to properly extract a character string to be proofread from a target text, and to properly proofread the extracted character string.
[0072] The proofreading device 10 according to the second aspect of the present disclosure is the proofreading device 10 according to the first aspect, and when the expression of the comparison forbidden word does not match the expression of the candidate morpheme, the proofreading unit 38 corrects the recommended term associated with the comparison forbidden word based on the analysis result of the candidate morpheme, and sets the candidate morpheme as the proofreading point while setting the corrected recommended term (correction term) as a proofreading candidate. According to the present disclosure, it is possible to properly extract a character string to be proofread from the target text, and to properly proofread the extracted character string.
[0073] The proofreading device 10 according to the third aspect of the present disclosure is the proofreading device 10 according to the first or second aspect, and the proofreading unit 38 does not set a proofreading point or a proofreading candidate when the candidate morpheme is a stem in the analysis result of the candidate morpheme. Since proofreading the stem may change the meaning, not proofreading when it is a stem can prevent proofreading to an incorrect meaning.
[0074] A proofreading device 10 according to a fourth aspect of the present disclosure is the proofreading device 10 according to any one of the first to third aspects, in which the proofreading unit 38 does not set a proofreading portion or a proofreading candidate when the expression of the candidate morpheme matches a recommended term or a preset permitted term. By not proofreading the candidate morpheme when it matches a recommended term or permitted term, unintentional proofreading can be suppressed.
[0075] A proofreading device 10 according to a fifth aspect of the present disclosure is the proofreading device 10 according to any of the first to fourth aspects, in which the selection unit 36 determines that the analysis results match when the part of speech and normal form of the morpheme group match the part of speech and normal form of the forbidden word, and the proofreading unit 38 determines that the expressions match when the surface form of the comparative forbidden word matches the candidate morpheme. By using such criteria, the proofreading points and proofreading candidates can be appropriately set.
[0076] A proofreading device 10 according to a sixth aspect of the present disclosure is the proofreading device 10 according to any one of the first to fifth aspects, in which the selection unit 36 selects, from among the morpheme groups, a morpheme group including an expression of a prohibited term as a candidate morpheme. When the number of morphemes included in the candidate morpheme is two or more, the proofreading unit 38 sets the recommended term associated with the comparison prohibited term as a proofreading candidate. When the number of morphemes included in the candidate morpheme is one and the part of speech of the candidate morpheme and the part of speech of the comparison prohibited term match, the proofreading unit 38 sets the recommended term associated with the comparison prohibited term as a proofreading candidate. According to the present disclosure, it is possible to detect a term to be proofread based on the part of speech, and it is possible to prevent a term with a different part of speech from being proofread. Therefore, according to the present disclosure, proofreading can be performed appropriately.
[0077] The proofreading device 10 according to the seventh embodiment of the present disclosure is the proofreading device 10 according to the sixth embodiment, in which the proofreading unit 38 does not proofread when both the first and last characters of the candidate morpheme are different from the comparison prohibited term. If the first and last characters are proofread to a term that is different, there is a risk that the meaning will change. In contrast, by not proofreading in such a case, proofreading to an incorrect meaning can be suppressed.
[0078] A proofreading device 10 according to an eighth aspect of the present disclosure is the proofreading device 10 according to any one of the first to seventh aspects, in which the term acquisition unit 30 acquires a target term and a format of an added term located before or after the target term. The selection unit 36 selects, from among the morpheme groups, a morpheme group whose expression matches that of the target term as a candidate morpheme. When the format of the added morpheme located before or after the candidate morpheme does not match the format of the added term, the proofreading unit 38 sets the added morpheme as a proofreading point, while correcting the format of the added morpheme to the format of the added term, as a proofreading candidate. According to the present disclosure, it is possible to select characters before and after the target term and proofread them without uniformly proofreading the format, so that it is possible to appropriately detect and proofread the term to be proofread.
[0079] A proofreading method according to a ninth aspect of the present disclosure includes the steps of: acquiring a prohibited term and a recommended term associated with the prohibited term; acquiring a target text, which is text data to be proofread; acquiring a morpheme included in the prohibited term and a morpheme included in the target text, which are derived by a morphological analysis performed on the prohibited term and the target text; selecting a morpheme group, the analysis result of which matches the prohibited term, from a morpheme group including at least one morpheme included in the target text as a candidate morpheme; and setting a proofreading portion of the target text and a proofreading candidate, which is a term that is a candidate for proofreading at the proofreading portion, based on a comparison prohibited term, which is a prohibited term whose analysis result matches the candidate morpheme, and the candidate morpheme. In the step of setting the proofreading portion and the proofreading candidate, if the expression of the comparison prohibited term matches the expression of the candidate morpheme, the candidate morpheme is set as the proofreading portion, and the recommended term associated with the comparison prohibited term is set as the proofreading candidate. According to the present disclosure, proofreading can be performed appropriately.
[0080] A program according to a tenth aspect of the present disclosure causes a computer to execute the following steps: acquiring a prohibited term and a recommended term associated with the prohibited term; acquiring a target text, which is text data to be proofread; acquiring a morpheme included in the prohibited term and a morpheme included in the target text, which are derived by a morphological analysis performed on the prohibited term and the target text; selecting, as a candidate morpheme, a morpheme group whose analysis result of the morphological analysis matches the prohibited term from among a morpheme group including at least one morpheme included in the target text; and setting a proofreading portion of the target text and a proofreading candidate, which is a term that is a candidate for proofreading at the proofreading portion, based on a comparison prohibited term, which is a prohibited term whose analysis result matches the candidate morpheme, and the candidate morpheme. In the step of setting the proofreading portion and the proofreading candidate, if the expression of the comparison prohibited term matches the expression of the candidate morpheme, the candidate morpheme is set as the proofreading portion, and the recommended term associated with the comparison prohibited term is set as the proofreading candidate. According to the present disclosure, proofreading can be performed appropriately.
[0081] Although the embodiment of the present disclosure has been described above, the embodiment is not limited by the contents of this embodiment. In addition, the above-mentioned components include those that a person skilled in the art can easily imagine, those that are substantially the same, and those that are within the so-called equivalent range. Furthermore, the above-mentioned components can be appropriately combined. Furthermore, various omissions, substitutions, or modifications of the components can be made without departing from the gist of the above-mentioned embodiment. [Explanation of symbols]
[0082] 10 Calibration device 30 Terminology acquisition part 32 Text Acquisition Section 34 Analysis Department 36 Selection section 38 Proofreading Department 40 Output control section
Claims
1. a term acquisition unit that acquires prohibited terms and recommended terms associated with the prohibited terms; a text acquisition unit for acquiring a target text, which is text data to be proofread; an analysis unit that acquires morphemes contained in the prohibited words and morphemes contained in the target text, the morphemes being derived by a morphological analysis performed on the prohibited words and the target text; a selection unit that selects, from among a group of morphemes including at least one morpheme included in the target text, a group of morphemes whose analysis result of the morphological analysis matches the prohibited term as a candidate morpheme; a proofreading unit that sets a proofreading portion of the target text and a proofreading candidate that is a term that is a candidate for proofreading at the proofreading portion, based on the comparative forbidden term that is the forbidden term whose analysis result matches the candidate morpheme and the candidate morpheme; Including, When an expression of the comparison prohibited word matches an expression of the candidate morpheme, the proofreading unit sets the candidate morpheme as the proofreading point, and sets the recommended word associated with the comparison prohibited word as the proofreading candidate. Calibration device.
2. The proofreading device of claim 1, wherein when the expression of the comparison forbidden word does not match the expression of the candidate morpheme, the proofreading unit corrects the recommended term corresponding to the comparison forbidden word based on the analysis result of the candidate morpheme, sets the candidate morpheme as the proofreading point, and sets the corrected recommended term as the proofreading candidate.
3. 3. The proofreading device according to claim 1, wherein the proofreading section does not set the proofreading portion and the proofreading candidate when the candidate morpheme is a stem in the analysis result of the candidate morpheme.
4. 3. The proofreading device according to claim 1, wherein the proofreading unit does not set the proofread portion and the proofreading candidate when an expression of the candidate morpheme matches the recommended term or a preset allowed term.
5. the selection unit determines that the analysis results match when a part of speech and a normal form of the group of morphemes match a part of speech and a normal form of the prohibited word; 3. The proofreading device according to claim 1, wherein the proofreading section determines that the expressions match when a surface form of the comparison forbidden term matches the candidate morpheme.
6. the selection unit selects, from the group of morphemes, the group of morphemes including an expression of the prohibited term as the candidate morpheme; The calibration unit is When the number of morphemes included in the candidate morpheme is two or more, the recommended term associated with the prohibited term having an expression matching the candidate morpheme is set as the proofreading candidate; A proofreading device as described in claim 1 or claim 2, wherein when the number of morphemes contained in the candidate morpheme is one and the part of speech of the candidate morpheme matches the part of speech of the forbidden word whose expression matches that of the candidate morpheme, the recommended term corresponding to the forbidden word is set as the proofreading candidate.
7. The proofreading device according to claim 6 , wherein the proofreading unit does not perform proofreading when both the first character and the last character of the candidate morpheme are different from the prohibited word.
8. The term acquisition unit acquires a target term and a format of an additional term located before or after the target term, the selection unit selects, from the morpheme group, the morpheme group whose expression matches that of the target term as the candidate morpheme; The proofreading device according to claim 1 or claim 2, wherein when the format of an added morpheme located before or after the candidate morpheme does not match the format of the added term, the proofreading unit sets the added morpheme as the proofreading point, while setting a term in which the format of the added morpheme has been corrected to the format of the added term as the proofreading candidate.
9. obtaining prohibited terms and recommended terms associated with the prohibited terms; A step of obtaining a target text which is text data to be proofread; obtaining morphemes contained in the prohibited words and morphemes contained in the target text derived by a morphological analysis performed on the prohibited words and the target text; selecting, as a candidate morpheme, a morpheme group whose analysis result of the morphological analysis matches the prohibited term from among a morpheme group including at least one morpheme included in the target text; setting a proofreading portion of the target text and a proofreading candidate, which is a term that is a candidate for proofreading in the proofreading portion, based on the candidate morpheme and a comparative forbidden term that is a forbidden term whose analysis result matches the candidate morpheme; Including, In the step of setting the correction portion and the correction candidate, when an expression of the comparison prohibited word matches an expression of the candidate morpheme, the candidate morpheme is set as the correction portion, and the recommended word associated with the comparison prohibited word is set as the correction candidate. Calibration method.
10. obtaining prohibited terms and recommended terms associated with the prohibited terms; A step of obtaining a target text which is text data to be proofread; obtaining morphemes contained in the prohibited words and morphemes contained in the target text derived by a morphological analysis performed on the prohibited words and the target text; selecting, as a candidate morpheme, a morpheme group whose analysis result of the morphological analysis matches the prohibited term from among a morpheme group including at least one morpheme included in the target text; setting a proofreading portion of the target text and a proofreading candidate, which is a term that is a candidate for proofreading in the proofreading portion, based on the candidate morpheme and a comparative forbidden term that is a forbidden term whose analysis result matches the candidate morpheme; on the computer, In the step of setting the correction portion and the correction candidate, when an expression of the comparison prohibited word matches an expression of the candidate morpheme, the candidate morpheme is set as the correction portion, and the recommended word associated with the comparison prohibited word is set as the correction candidate. program.
Citation Information
Patent Citations
Document processing device and program
JP2010182181A
Cited By
Strategy evaluation method
CN121118884A
