CN109840062B - Input support device and recording medium - Google Patents

Abstract

An input support device and a recording medium are provided, which can properly determine the item value input to the variable slot of the format data even if the recognition result based on the voice recognition contains an error. The input support device according to an embodiment includes a template storage unit, an acquisition unit, and a determination unit. The template storage unit stores a format template having format data of 1 or more variable slots for inputting item values in association with item names, and the item names of the respective variable slots and options of option type variable slots for selecting item values from a plurality of options are described together with their respective reading methods. The acquisition unit acquires recognition result data obtained by voice-recognizing a voice uttered by a user, including a description and a reading. The determination unit determines an item value to be input to the variable slot of the format data based on the reading of the recognition result data and the reading of the item name and the item option described in the format template.

Description

Input support device and recording medium

Technical Field

Embodiments of the present invention relate to an input assisting apparatus and a recording medium.

Background

Conventionally, a technique for supporting an input operation by a user using a voice recognition technique is known. Speech recognition engines that perform speech recognition are roughly classified into grammar type and dictation (transcription) type. The grammar type is a method of performing voice recognition based on a rule (grammar) created in advance, and is easy to improve recognition accuracy, but has a feature that a speech which does not follow the rule cannot be recognized. Moreover, the creation of the rule requires professional knowledge related to voice recognition, and the task of appropriately creating and managing the rule is complicated. On the other hand, the dictation type is a method for sequentially recognizing an arbitrary utterance without depending on a specific rule, and has a feature of high versatility, but has a lower recognition accuracy than the grammar type.

Here, it is considered that the operation of inputting the item value to the variable slot (slot) of the format data (formdata) is assisted by the voice recognition technology. In this specification, format data is data whose format is specified, and which has 1 or more variable slots into which item values are input in correspondence with item names. For example, when assisting the input of format data to a work report by a worker who performs a field (field) work, it is not realistic to expect a statement to follow a rule for all the workers, and it is considered desirable to use a dictation-type voice recognition engine for voice recognition of the statement. However, since the recognition accuracy of the dictation-type speech recognition engine is low as described above, it is often impossible to determine which part of the recognition result is input as the item value of which variable slot. Therefore, it is desirable to provide a new technique for appropriately determining an item value to be input to a variable slot of format data even when an error is included in a recognition result.

Disclosure of Invention

An object of the present invention is to provide an input assisting apparatus and a recording medium capable of appropriately determining an item value to be input to a variable slot of format data even when an error is included in a recognition result by voice recognition.

An input support device according to an embodiment includes a template storage unit, an acquisition unit, and a determination unit. The template storage unit stores a format template having format data of 1 or more variable slots for inputting item values in association with item names, and the item names of the variable slots and options of option type variable slots for selecting item values from a plurality of options are described together with their respective reading methods. The acquisition unit acquires recognition result data obtained by voice-recognizing a voice uttered by a user, including a description and a reading. The determination unit determines an item value to be input to the variable slot of the format data based on the reading of the recognition result data and the reading of the item name and the option described in the format template.

According to the input assisting apparatus configured as described above, even when an error is included in the recognition result by the voice recognition, it is possible to appropriately determine the item value to be input to the variable slot of the format data.

Drawings

Fig. 1 is a diagram showing an example of format data.

Fig. 2 is a diagram showing an example of a format template.

Fig. 3 is a diagram showing an example of recognition result data.

Fig. 4 is a diagram showing an example of format data after entry of an item value.

Fig. 5 is a block diagram showing an example of a functional configuration of the input assisting apparatus.

Fig. 6 is a flowchart showing an example of the processing procedure performed by the extraction unit.

Fig. 7 is a diagram showing an example of the candidate list.

Fig. 8 is a flowchart illustrating an example of the procedure of processing performed by the option item value determining unit.

Fig. 9 is a flowchart showing an example of a processing procedure performed by the free-description type item value determining unit.

Fig. 10 is a flowchart showing an example of the processing procedure performed by the template management section.

Fig. 11 is a diagram showing an example of format data in which entry values of some of the variable slots have been input.

Fig. 12 is a diagram showing an example of a format template.

Fig. 13 is a diagram showing an example of the recognition result data.

Fig. 14 is a diagram showing an example of format data after entry of an item value into an uninputed variable bin.

Fig. 15 is a block diagram showing an example of the hardware configuration of the input assisting apparatus.

Detailed Description

Hereinafter, an input support device and a program according to an embodiment will be described in detail with reference to the drawings.

< summary of the embodiments >

The input support device of the present embodiment determines an item value to be input to a variable slot of format data using recognition result data obtained by voice-recognizing the utterance of a user. It is assumed that a dictation-type voice recognition engine capable of performing voice recognition without following a specific rule is used for voice recognition performed on speech of a user, and an error is included in recognition result data. In order to appropriately determine the item value to be input to the variable slot of the format data from the recognition result data including such an error, a format template previously created by a template creator is used in the present embodiment. In the following, the format data of the job report for reporting the job result of the field job is shown as an example of the format data to be processed, but the format data that can be applied is not limited to this.

The format data is data whose format is specified, and has 1 or more variable slots into which item values are input, in correspondence with item names. Fig. 1 shows an example of format data. The format data a shown in fig. 1 is an example of format data of a job report to which no item value has been input, and includes 4 variable slots B, that is, a variable slot (variable slot No. 1) with an item name of "work place", a variable slot (variable slot No. 2) with an item name of "witness", a variable slot (variable slot No. 3) with an item name of "replacement part", and a variable slot (variable slot No. 4) with an item name of "special matter".

The variable slot B includes 2 types, i.e., an option-type variable slot B1 and a free-description variable slot B2. The option type variable slot B1 is a variable slot B for selecting an item value from a plurality of options. The free-description variable slot B2 is a variable slot B into which an item value is freely written. In the format data a illustrated in fig. 1, 3 variable slots B of the variable slots No.1 to No.3 are option-type variable slots B1, and the variable slot B of the variable slot No.4 is a free-description variable slot B2. As the item value of each variable slot B, a character string based on the utterance of the operator is input.

The format template is a template of format data a, and items of item names of the variable slots B and items of item values of the option-type variable slots B1 included in the format data a are described together with their respective reading methods (shown in hiragana in the japanese specification corresponding to the present invention, but shown in lower-case roman letters in the present invention). Fig. 2 shows an example of a format template. The format template C shown in fig. 2 is an example of a format template corresponding to the format data a shown in fig. 1, and the 3 option type variable slots B1 from the variable slot No.1 to the variable slot No.3 have their respective readings described together with the option of the item name and the item value, and the freely-described variable slot B2 of the variable slot No.4 has its item name and its reading described. A plurality of readings may be described for 1 item name or option. In the format template C, a type of variable slot indicating whether each variable slot B is an option type or a freely written type is described.

The format template C is previously created by the template creator as described above. The format template C is used as clue information for determining item values to be input to the variable slots B of the format data a from recognition result data including errors, but it is not necessary to perform detailed parameter setting related to voice recognition such as rule creation in a grammar-type voice recognition engine. That is, even if the template creator does not have professional knowledge relating to voice recognition, the template creator can easily create the format template C and manage it simply by setting the item name of each variable slot B of the corresponding format data a and the option of the option-type variable slot B1 together with the reading method.

Fig. 3 shows an example of recognition result data obtained by voice-recognizing the utterance of the user. Fig. 3 (a) shows an example of utterance D for creating a work report by an operator who performs a live work, and fig. 3 (b) shows an example of recognition result data E obtained by voice-recognizing utterance D in fig. 3 (a) by a voice recognition engine. As shown in fig. 3 (b), the recognition result data E includes description and reading. In the example shown in fig. 3, "kawasaki" in utterance D is erroneously recognized as "shikaki", and "air filter" in utterance D is erroneously recognized as "filter-then", and an error is included in recognition result data E.

In addition, the reading of the recognition result data E may be performed by providing a reading information generating unit for generating a reading based on the description of the recognition result data E output from the voice recognition engine, instead of using the voice recognition engine, and using the output of the reading information generating unit. In the present embodiment, the utterance D including a plurality of words and phrases as shown in fig. 3 (a) is input to the speech recognition engine at one time, and the recognition result data E as shown in fig. 3 (b) is obtained, but the timing of the speech recognition is arbitrary. For example, the following structure may be adopted: each sentence constituting utterance D is sequentially input to the speech recognition engine, and speech recognition is performed in real time.

The input assisting apparatus according to the present embodiment acquires, for example, the recognition result data E as illustrated in fig. 3 (B), and determines the item value to be input to each variable slot B of the format data a as illustrated in fig. 1 based on the reading of the recognition result data E and the reading of the item name and the option described in the format template C as illustrated in fig. 2. Fig. 4 shows an example of format data a into which the item values determined by the input support device of the present embodiment are input. As shown in fig. 4, the process of the input assisting apparatus of the present embodiment is such that "kawasaki" is input as an item value to the option type variable tank B1 (variable tank No. 1) having the item name "work place", that "zhongcun" is input as an item value to the option type variable tank B1 (variable tank No. 2) having the item name "witness", and that "air filter" is input as an item value to the option type variable tank B1 (variable tank No. 3) having the item name "replacement part". In the free-description type variable slot B2 having the item name "special event", the item value "special event: stumbled and nearly fallen off during movement by the holding school that requires the membrane.

The item value "kawasaki" input to the option type variable slot B1 (variable slot No. 1) having the item name "work place" is selected as the item value of the option type variable slot B1 because the reading of the recognition result data E includes "wa-sa-ki" similar to the reading "ka-wa-sa-ki" of the option "kawasaki" described in the format template C. However, "kawasaki" is included in the option of option type variable slot B1 (variable slot No. 2) whose project name is "witness", and therefore it is necessary to determine "kawasaki" corresponding to "wa-sa-ki" as the project value of option type variable slot B1 of which variable slot No.1 and variable slot No.2 are. The item value "Zhongcun" input to the option-type variable slot B1 (variable slot No. 2) with the item name "witness" is selected as the item value of the option-type variable slot B1 because the reading of the recognition result data E includes "ka-mu-ra" similar to the reading "na-ka-mu-ra" of the option "Zhongcun" described in the formatting template C. However, since the option of the option type variable slot B1 (variable slot No. 2) having the item name "witness" also includes "kawasaki" corresponding to "wa-sa-ki" included in the reading of the identification result data E, it is necessary to determine which of "kamura" and "kawasaki" is to be the item value of the option type variable slot B1 of the variable slot No. 2. The input assisting apparatus according to the present embodiment can appropriately determine the item value to be input to each variable slot B even when there are a plurality of candidates for matching the item values to be input to the variable slots B in this manner.

< example of input assisting device >

Next, a specific example of the input assisting apparatus according to the present embodiment will be described in detail. Fig. 5 is a diagram showing an example of a functional configuration of the input assisting apparatus 100 according to the present embodiment. As shown in fig. 5, for example, the input assisting apparatus 100 of the present embodiment includes a format data storage unit 10, a template storage unit 20, an acquisition unit 30, a determination unit 40, and a template management unit 50.

The format data storage unit 10 stores format data a. For example, when a template creator creates a format template C (see fig. 2), format data a (see fig. 1) corresponding to the format template C, to which no item value has been input, is generated and stored in the format data storage unit 10. Further, the item value determined by the determination unit 40 is input to each variable bin B of the format data a stored in the format data storage unit 10, thereby generating the format data a to which the item value is input (see fig. 4). The format data a to which the item value is input can be read from the format data storage unit 10 as necessary, and output in any output form such as display by a display, print output to a recording paper sheet, and data transmission to an external device.

The template storage unit 20 stores the format template C. The template creator can create a template C using the template management unit 50, and the template C created by the template creator is stored in the template storage unit 20. The format template C stored in the template storage unit 20 is read by the determination unit 40. When the template storage unit 20 stores a plurality of format templates C, the determination unit 40 may select a format template C read from the template storage unit 20, for example, based on an operation by a user who selects a format template C to be used, or based on a degree of matching with the recognition result data E acquired by the acquisition unit 30. The template creator can also edit or delete the format template C stored in the template storage unit 20 using the template management unit 50.

The acquisition unit 30 acquires recognition result data E (see fig. 3) obtained by voice-recognizing utterances of users (in the present embodiment, operators who perform field work) by the voice recognition engine. The recognition result data E acquired by the acquisition unit 30 includes the description and the reading method as described above. The acquisition unit 30 may receive an input of a voice uttered by a user and acquire the recognition result data E by performing voice recognition on the input voice using an internal voice recognition engine, or may acquire the recognition result data E by performing voice recognition on a voice uttered by the user using an external voice recognition engine. That is, the voice recognition engine that performs voice recognition on the voice uttered by the user may be a configuration provided in the input assisting apparatus 100 according to the present embodiment, or may be a configuration provided in an apparatus external to the input assisting apparatus 100, such as a terminal apparatus used by the user or an external server that provides a voice recognition service. The recognition result data E acquired by the acquisition unit 30 is delivered to the determination unit 40.

The determination unit 40 determines the item value to be input to each variable slot B of the format data a based on the reading of the recognition result data E acquired by the acquisition unit 30 and the reading of the item name and option described in the format template C read from the template storage unit 20. As shown in fig. 5, the determination unit 40 includes an extraction unit 41, an option item value determination unit 42, and a free-description item value determination unit 43, and performs processing in this order.

The extracting unit 41 extracts a partial character string similar to the reading method of the item name or the option described in the format template C from the character string of the reading method of the recognition result data E acquired by the acquiring unit 30. Further, the extraction unit 41 registers a combination of a partial character string extracted from the character string of the reading of the recognition result data E and the item name or option of the format template C used for extracting the partial character string as a candidate of the corresponding portion between the recognition result data E and the format template C in the candidate list. The details of the candidate list will be described later.

Fig. 6 is a flowchart showing an example of the processing procedure performed by the extraction unit 41. When the processing shown in fig. 6 is started, the extraction unit 41 first initializes the candidate list (step S101). Next, the extracting unit 41 sets 1 of the item name and the option described in the format template C to the variable TEXT (step S102). At this time, the reading of the item name or option set at the variable TEXT is also maintained.

Next, the extracting unit 41 extracts a partial character string that completely or partially matches the reading of TEXT from the character string of the reading of the recognition result data E (step S103). In this case, the extraction unit 41 may perform the processing of step S103 after performing preprocessing such as removing a filler word or removing a portion other than a portion considered to be a noun by morphological analysis on the recognition result data E. Further, the extraction unit 41 calculates the similarity of the partial character string extracted in step S103 with respect to the reading method of TEXT (step S104). When a plurality of partial character strings are extracted in step S103, the similarity is calculated for each partial character string in step S104.

The similarity used here includes, for example, the ratio of the number of characters in a partial character string to the number of characters in the reading of TEXT. For example, in the case where the option "air filter" of the variable slot No.3 of the format template C shown in fig. 2 is set at TEXT, "fi-ru-ta" (line 1) and "fi-ru" (line 3) partially coinciding with the reading "E-a-fi-ru-ta" of the option "air filter" are extracted from the TEXT column of the reading of the recognition result data E shown in (b) of fig. 3. In this case, since the number of characters expressed by hiragana of "e-a-fi-ru-ta" is 6, the number of characters expressed by hiragana of "fi-ru-ta" is 4, and the number of characters expressed by hiragana of "fi-ru" is 3, it is possible to calculate that the degree of similarity of "fi-ru-ta" is 4/6=0.67, and the degree of similarity of "fi-ru" is 3/6=0.5. In this way, the similarity of the extracted partial character string is calculated to have a value that is larger as the similarity with the reading method of TEXT.

The method of calculating the similarity is not limited to the above method, and a known technique can be used. For example, the similarity may be calculated using the matching rate of the edit (levenshtein) distance (how many times the operations of inserting, deleting, and replacing characters are applied to 2 character strings to make the same character string) and the N-gram (a continuous element of a combination of N adjacent character strings or words in a certain character string).

Next, the extraction unit 41 determines whether or not the similarity calculated in step S104 is equal to or greater than a predetermined threshold (for example, 0.5) with respect to the partial character string extracted in step S103 (step S105). Further, when the similarity of the partial character string is equal to or greater than the threshold (yes in step S105), the extraction unit 41 registers a combination of the partial character string and the item name or the option set in TEXT as a candidate in the candidate list (step S106).

On the other hand, when the similarity of the partial character string is less than the threshold (NO in step S105), the process in step S106 is skipped and the registration to the candidate list is not performed. In this way, by adopting a configuration in which only partial character strings having a similarity equal to or greater than the threshold are registered in the candidate list, for example, short partial character strings having a number of characters represented by hiragana of about 1 character or 2 characters can be removed from the candidates, and a problem that the processing becomes complicated due to an enormous number of candidates can be effectively avoided.

Then, the extraction unit 41 determines whether or not all the item names and options described in the format template C have been processed (step S107), and if there are any unprocessed item names or options (no in step S107), it returns to step S102 and repeats the subsequent processing. Further, when the processing from step S102 to step S106 has been performed for all the item names and options described in the format template C (step S107: YES), the series of processing shown in FIG. 6 is ended.

Fig. 7 is a diagram showing an example of the candidate list F generated by the above-described processing performed by the extraction unit 41. The candidate list F shown in fig. 7 is an example of the candidate list F generated based on the format template C shown in fig. 2 and the recognition result data E shown in fig. 3 (b). In the candidate list F, a combination of a partial character string extracted from the character string of the reading of the recognition result data E and the item name or option of the format template C used for extracting the partial character string is registered as a candidate. The registered data of the candidate list F includes, for example, as shown in fig. 7, a candidate type, a variable slot No., an item name, an item value, a partial character string, an offset, a similarity, a type of a variable slot, and the like.

The candidate category indicates whether the candidate is to be the item name or the option (item value). The variable bin No. indicates which variable bin B the candidate is targeted for. The item name indicates the item name of the corresponding variable slot B. The item value indicates an option (item value) in the case where the candidate is targeted at the option (item value). The offset indicates the appearance position of the partial character string in the recognition result data E, and the first character indicating the partial character string is the first character from the beginning of the character string in the reading of the recognition result data E. The similarity indicates the similarity calculated in step S104 of fig. 6. The type of the variable slot indicates whether the corresponding variable slot B is the option-type variable slot B1 or the freely-described variable slot B2 when the candidate is targeted at the item name. The registered data of each candidate including the information is stored in association with a candidate ID uniquely assigned to each candidate. Hereinafter, among the candidates registered in the candidate list F, a candidate (candidate category of which is an item value) that is targeted at an option (item value) is referred to as an "item value candidate", and a candidate (candidate category of which is an item name) that is targeted at an item name is referred to as an "item name candidate".

The processing of the extraction unit 41 shown in fig. 6 and the configuration of the candidate list F shown in fig. 7 are merely examples, and are not limited thereto. The extraction unit 41 may be configured to extract a partial character string similar to the item name of the format template C or the reading method of the option from the character string of the reading method of the recognition result data E and register the partial character string in the candidate list F. The candidate list F may be configured to be able to know the correspondence between the partial character string extracted by the extraction unit 41 and the item name or the option. When the template storage unit 20 stores a plurality of format templates C, the candidate list F may be configured to store the respective pieces of registered data in association with a template ID for identifying the format template C to be used, in addition to the candidate IDs.

The option-type item value determining unit 42 determines an item value to be input to the option-type variable slot B1 of the format data a based on at least one of the degree of similarity of the partial character string similar to the reading of the option described in the format template C and the positional relationship between the partial character string and the other partial character string.

Fig. 8 is a flowchart showing an example of the processing procedure performed by the option item value determination unit 42. When the process shown in fig. 8 is started, the optional item VALUE determining unit 42 first sets one of the item VALUE candidates included in the candidate list F to the variable VALUE (step S201).

Next, the optional item VALUE determining unit 42 determines whether or not there is a matching candidate having the same item name or item VALUE for the item VALUE candidates set in the VALUE (step S202). For example, when the item VALUE candidate of the candidate ID "001" in the candidate list F shown in fig. 7 is set to VALUE, the item VALUE candidate of the candidate ID "003" having the same item VALUE "kawasaki" becomes a competition candidate. When the item VALUE candidate of the candidate ID "003" is set to VALUE, the item VALUE candidate of the candidate ID "001" having the same item VALUE "kawasaki" and the item VALUE candidate of the candidate ID "004" having the same item name "witness" are competition candidates. When the candidate ID "006" candidate for the item VALUE is set to VALUE, the candidate ID "007" having the same item VALUE "air filter" becomes a candidate for competition. When the candidate ID "007" is set to VALUE, the candidate ID "006" having the same item VALUE "air filter" becomes a candidate for competition.

If no match candidate exists as a result of the determination in step S202 (no in step S202), the optional item VALUE determination unit 42 determines the item VALUE of VALUE as the item VALUE of the corresponding variable slot B (step S203). On the other hand, if there is a match candidate (yes in step S202), the optional item VALUE determining unit 42 next determines whether or not the similarity of VALUE is greater than the similarity of the match candidate (step S204). Further, when the similarity of the VALUE is greater than the similarity of the competition candidates (yes in step S204), the process proceeds to step S203, and the item VALUE of the VALUE is determined as the item VALUE of the corresponding variable slot B. For example, when the item VALUE candidate of the candidate ID "006" in fig. 7 is set to VALUE, the similarity of VALUE 0.67 is greater than the similarity of item VALUE candidate of the candidate ID "007" as the competition candidate, which is 0.5, and therefore, the item VALUE "air filter" of VALUE is determined as the item VALUE of the option variable slot B1 of the variable slot No. 3.

On the other hand, if the similarity of the VALUE is not greater than the similarity of the competition candidates (no in step S204), the option item VALUE determining unit 42 next determines whether the similarity of the VALUE is smaller than the similarity of the competition candidates (step S205). When the similarity of the VALUE is smaller than the similarity of the competition candidates (yes in step S205), the process for the item VALUE candidates set in the VALUE is ended, and the process proceeds to step S211, which will be described later.

On the other hand, when the similarity of the VALUE is not smaller than the similarity of the competition candidates, that is, when the similarity of the VALUE is equal to the similarity of the competition candidates (no in step S205), the option item VALUE determining unit 42 next determines whether or not there is an item name candidate corresponding to the item name of the VALUE (step S206). Further, if there is an item name candidate corresponding to the item name of the VALUE (yes in step S206), the option-type item VALUE determining unit 42 next determines whether or not the difference in offset between the VALUE and the item name candidate is smaller than the difference in offset between the competition candidate and the item name candidate (step S207), and if the difference in offset is smaller than the competition candidate (yes in step S207), it proceeds to step S203 to determine the item VALUE of the VALUE as the item VALUE of the corresponding variable slot B.

For example, when the item VALUE candidate of the candidate ID "004" in fig. 7 is set to VALUE, there are item name candidates of the candidate ID "002" corresponding to the item VALUE "witness" of VALUE, and the difference in offset with respect to VALUE of the item name candidate is 25 to 17=8, whereas the difference in offset with respect to competition candidates of the item name candidates (the item VALUE candidate of the candidate ID "001" and the item name candidate of the candidate ID "003") is 17 to 2=15, and the difference in offset is smaller than the competition candidates. Therefore, the item VALUE "zhongcun" of VALUE is determined as the item VALUE of the option type variable slot B1 of the variable slot No. 2.

On the other hand, when the difference between the offsets of the VALUE and the item name candidates is equal to or greater than the difference between the offsets of the competition candidate and the item name candidate (no in step S207), the process for the item VALUE candidate set in the VALUE is ended, and the process proceeds to step S211, which will be described later.

If there is no item name candidate corresponding to the item name of the VALUE (no in step S206), the optional item VALUE determining unit 42 then determines whether there is an item name candidate corresponding to the item name of the competition candidate (step S208). Further, if there is an item name candidate corresponding to the item name of the competition candidate (yes in step S208), the option-type item value determination unit 42 next determines whether or not there is another item value candidate having the same item name as the competition candidate (step S209). If there is such an item VALUE candidate (yes in step S209), the option-type item VALUE determining unit 42 then determines whether or not the difference in offset between the competition candidate and the item name candidate is larger than the difference in offset between the other item VALUE candidate and the item name candidate (step S210), and if the difference in offset between the competition candidate and the item name candidate is larger than the difference in offset between the other item VALUE candidate and the item name candidate (yes in step S210), it proceeds to step S203 to determine the item VALUE of VALUE as the item VALUE of the corresponding variable slot B.

For example, when the item VALUE candidate of the candidate ID "001" in fig. 7 is set to VALUE, there is an item name candidate of the candidate ID "002" as an item name candidate corresponding to the item VALUE candidate of the candidate ID "003" as a competition candidate. Further, as another item value candidate having the same item name as the item value candidate of the candidate ID "003" as the competition candidate, there is an item value candidate of the candidate ID "004". Further, while the difference between the shift between the item value candidate of the candidate ID "003" as the competition candidate and the item name candidate of the candidate ID "002" is 17-2=15, the difference between the shift between the item value candidate of the candidate ID "004" as the other item value candidate and the item name candidate of the candidate ID "002" is 25-17=8, and the difference between the shift between the competition candidate and the item name candidate is larger than the difference between the shift between the other item value candidate and the item name candidate. Therefore, the VALUE "kawasaki" of VALUE is determined as the VALUE of the option type variable slot B1 of the variable slot No. 1.

When there is no item VALUE candidate corresponding to the item name of the competition candidate (no in step S208), or when there is no other item VALUE candidate having the same item name as the competition candidate (no in step S209), or when the difference in offset between the competition candidate and the item name candidate is equal to or less than the difference in offset between the other item VALUE candidate and the item name candidate (no in step S210), the process for the item VALUE candidate set in VALUE is ended, and the process proceeds to step S211.

Then, the optional item value determining unit 42 determines whether or not all the item value candidates included in the candidate list F have been processed (step S211), and if there are unprocessed item value candidates (no in step S211), returns to step S201, and repeats the subsequent processing. Further, after the processing from step S201 to step S210 has been performed on all the item value candidates included in the candidate list F (YES in step S211), the determined item value is input to the corresponding option type variable slot B1 of the format data A (step S212), and the series of processing shown in FIG. 8 is ended.

The above-described processing is performed by the option-type item value determining unit 42, and as shown in fig. 4, for example, "kawasaki" is input as an item value in the option-type variable slot B1 of the variable slot No.1 of the format data a. Then, "zhongcun" is input as an item value in the option-type variable slot B1 of the variable slot No. 2. In the option type variable tank B1 of the variable tank No.3, "air filter" is input as an item value.

The processing of the option item value determination unit 42 shown in fig. 8 is an example, and is not limited to this. The optional item value determination unit 42 may be configured to determine the item value of the optional variable slot B1 based on at least one of the similarity of the item value candidates (the similarity of the partial character string included in the item value candidate with respect to the item value) and the positional relationship between the item value candidate and another item value candidate or item name candidate (the difference in the offset of the partial character string included in each candidate). For example, in the above example, the degree of similarity is prioritized for comparison with the competition candidates, but a score for each candidate may be calculated by comprehensively determining the similarity and the positional relationship, and the item value of the option variable slot B1 may be determined based on the score of each candidate. In the above example, the positional relationship between the candidates is expressed by using the difference in the offsets of the partial character strings included in each candidate, but the positional relationship between the candidates may be expressed by using another index such as the number of other candidates existing among 2 candidates.

The free-written item value determination unit 43 determines an item value to be input to the free-written variable slot B2 of the format data a based on the description of the recognition result data E, based on the similarity of the partial character string similar to the reading of the item name of the free-written variable slot B2 described in the format template C and the position of the partial character string.

Fig. 9 is a flowchart showing an example of the processing procedure performed by the free-description type item value determining unit 43. When the process shown in fig. 9 is started, the free-description type item value determining unit 43 first refers to the type of the variable slot of the item name candidate included in the candidate list F, and determines whether or not the item name candidate of the free-description type variable slot B2 exists in the candidate list F (step S301). If there is no item name candidate for the free description type variable slot B2 in the candidate list F (no in step S301), the process ends as it is.

On the other hand, if there are item name candidates of the free description type variable slot B2 in the candidate list F (yes in step S301), the free description type item VALUE determining unit 43 then sets one of the item name candidates of the free description type variable slot B2 included in the candidate list F to the variable VALUE (step S302). Then, the free-description item VALUE determining unit 43 determines whether or not there is a match candidate having the same item name for the item name candidates in the free-description variable slot B2 set in the VALUE (step S303).

If there is no competition candidate having the same item name (no in step S303), the free-description item VALUE determining unit 43 determines that the description of the recognition result data E corresponding to the portion from the partial character string of VALUE to the immediately preceding partial character string corresponding to the item name or item VALUE of the other variable slot B is the item VALUE of the corresponding free-description variable slot B2 (step S304). In the reading of the identification result data E, when there is no partial character string corresponding to the item name or the item VALUE of the other variable slot B after the partial character string of the VALUE, the description of the identification result data E corresponding to the portion from the partial character string of the VALUE to the end of the reading is determined as the item VALUE of the free-description variable slot B2.

For example, the item name candidate of the free-description type variable slot B2 included in the candidate list F shown in fig. 7 is only the item name candidate of the candidate ID "008", and there is no competition candidate having the same item name as the item name candidate. In the reading of the recognition result data E, after the partial character string "to-kki-ji-ko-u" of the item name candidate of the candidate ID "008", there is no partial character string corresponding to the item name or the item value of the other variable slot B. Therefore, the description of the recognition result data E corresponding to the portion from "to-kki-ji-ko-u" to the end of the reading of the recognition result data E is "special matter: stumbled by a standing school that requires a film while moving and fallen almost. "is determined as the item value of the free-description type variable slot B2 of the variable slot No. 4.

On the other hand, when there is a match candidate having the same item name (yes in step S303), the free-description item VALUE determination unit 43 next determines whether or not the similarity of VALUE is greater than the similarity of the match candidates (step S305). Further, when the similarity of VALUE is greater than the similarity of the competition candidates (yes in step S305), the process proceeds to step S304, and the description of the recognition result data E corresponding to the portion from the partial character string of VALUE to the immediately preceding partial character string corresponding to the item name or item VALUE of the other variable slot B is determined as the item VALUE of the corresponding free-description variable slot B2.

On the other hand, when the similarity of the VALUE is equal to or less than the similarity of the competition candidates (no in step S305), the process for the item name candidate set in the freely-described variable slot B2 of the VALUE is ended, and the process proceeds to step S306.

Then, the free-description item value determining unit 43 determines whether or not all of the item name candidates of the free-description variable slot B2 included in the candidate list F have been processed (step S306), and if there is an unprocessed item name candidate of the free-description variable slot B2 (step S306: no), returns to step S302 and repeats the subsequent processing. Further, after the processing from step S302 to step S305 has been performed on all the item name candidates of the free-description variable slot B2 included in the candidate list F (yes in step S306), the determined item value is input to the corresponding free-description variable slot B2 of the format data a (step S307), and the series of processing shown in fig. 9 is ended.

As shown in fig. 4, for example, in the free-description type variable slot B2 of the variable slot No.4 of the format data a, "special items: stumbled by a standing school requiring a film while moving and almost fallen down. ".

The processing of the free-description type item value determining unit 43 shown in fig. 9 is an example, and is not limited thereto. The free-description item value determining unit 43 may be configured to determine the item value to be input to the free-description variable slot B2 based on the similarity of the item name candidates in the free-description variable slot B2 (the similarity of the partial character string included in the item name candidates in the free-description variable slot B2 to the item name) and the position of the item name candidates (the offset of the partial character string included in the item name candidates) according to the description of the recognition result data E. For example, in the above example, when there is a match candidate, if the similarity is greater than the match candidate, it is determined that the item VALUE candidate of the free-description variable slot B2 set in the VALUE is the item name of the free-description variable slot B2, and the description of the recognition result data E of the part following the partial character string corresponding to the VALUE is determined as the item VALUE of the free-description variable slot B2. However, the following configuration may be adopted: even when the similarity of VALUE is greater than the similarity of the competition candidates, in the reading of the recognition result data E, if a partial character string corresponding to the item name or the item VALUE of another variable slot B exists at a position close to the partial character string after the partial character string of VALUE, it is determined that the competition candidate is the item name of the free-description variable slot B2, and the description of the recognition result data E of the portion after the partial character string corresponding to the competition candidate is determined as the item VALUE of the free-description variable slot B2.

The template management unit 50 performs creation, editing, deletion, and the like of the variable slot B in the format template C stored in the template storage unit 20 in accordance with an operation of a template creator, and manages the format template C stored in the template storage unit 20. As shown in fig. 5, the template management unit 50 includes a variable slot type setting unit 51, an item name setting unit 52, and an option setting unit 53.

The variable slot type setting unit 51 sets any one of an option type and a free description type for each variable slot B included in the format data a as the type of the variable slot B. The item name setting unit 52 sets the item name and reading method of each variable slot B included in the format data a. The option setting unit 53 sets options of item values of the option-type variable slot B1 included in the format data a and the reading thereof.

Fig. 10 is a flowchart showing an example of the processing procedure performed by the template management section 50. When the process shown in fig. 10 is started, the template management unit 50 first determines the operation type of the operation performed by the template creator (the operation of the template creator is one of the creation, editing, and deletion of the variable slot B) (step S401). Further, when the operation type is the creation of the variable slot B, the item name setting unit 52 sets the item name and the reading of the variable slot B to be created, based on the operation of the template creator (step S402), and the variable slot type setting unit 51 sets the type of the variable slot B (step S403). Further, it is determined whether the type of the variable slot B to be created is a free description type or an option type (step S404), and if the variable slot B to be created is an option type variable slot B1, the option setting unit 53 sets an option and a reading of the item value of the option type variable slot B1 according to the operation of the template creator (step S405). In addition, for setting the reading, a reading information generating unit for generating the reading based on the item name and item value input by the template creator may be separately provided, and the output of the reading information generating unit may be set. Furthermore, a plurality of readings may be set for 1 item name and item value.

When the operation type of the template creator is the editing of the variable slot B, the template management unit 50 selects the variable slot B to be edited (step S407), and edits the item name, the type of the variable slot B, the item value, and the like of the selected variable slot B according to the operation of the template creator (step S408). When the operation type of the template creator is deletion of the variable slot B, the template management unit 50 selects the variable slot B to be deleted (step S409), and deletes the selected variable slot B in accordance with the operation of the template creator (step S410). Note that, although the example of deleting the variable slot B is described here, when deleting the format template C, the format template C to be deleted may be selected in step S409.

Then, the template management unit 50 determines whether or not there is any more variable slot B to be operated by the template creator (step S406). Further, if there is a variable slot B to be operated (YES in step S406), the process returns to step S401 and the subsequent processes are repeated, and if there is no variable slot B to be operated (NO in step S406), the series of processes shown in FIG. 10 is ended. The processing of the template management unit 50 shown in fig. 10 is an example, and is not limited to this. The template management unit 50 may be configured to manage the format template C stored in the template storage unit 20 by creating, editing, deleting, and the like the variable slot B in the format template C stored in the template storage unit 20 according to an operation of a template creator.

< effects of the embodiment >

As described above in detail by way of specific examples, the input support apparatus 100 according to the present embodiment uses the format template C as cue information when determining the item values to be input to the variable slots B of the format data a by using the recognition result data E obtained by voice recognition of the utterance of the user. Further, the item value to be input to each variable slot B of the format data a is determined based on the reading of the identification result data E and the reading of the item name and option described in the format template C. Therefore, according to the input assisting apparatus 100 of the present embodiment, even when an error is included in the recognition result data E, the item value to be input to each variable slot B of the format data a can be appropriately determined. Further, the user can input an appropriate item value to each variable slot of the format data a based on a relatively flexible utterance, instead of an utterance that does not follow a rule.

In the format template C used as the cue information in the input support apparatus 100 of the present embodiment, only the item name of each variable slot B of the format data a and the option of the item value of the option-type variable slot B1 may be set together with the reading of each item, and detailed parameter setting for voice recognition such as rule creation in a speech recognition engine of a grammar type is not necessary. Thus, the template creator can easily create the format template C without having professional knowledge about voice recognition, and the management thereof is easy.

Further, according to the input assisting apparatus 100 of the present embodiment, even when the format data a has 2 types of variable slots B, that is, the option-type variable slot B1 and the free-description variable slot B2, the item values of the option-type variable slot B1 and the free-description variable slot B2 can be appropriately determined in accordance with 1 utterance made by the user.

< modification 1 >

The above-described embodiments are merely examples, and various modifications can be made. For example, in the above-described embodiment, the option-type item value determining unit 42 determines the item value to be input to the option-type variable slot B1 based on at least one of the similarity of the item value candidates (the similarity of the partial character string included in the item value candidate with respect to the item value) and the positional relationship between the item value candidate and the other item value candidates or item name candidates (the difference between the offsets of the partial character strings included in the candidates). In the above-described embodiment, the free-description item value determining unit 43 determines the item value to be input to the free-description variable slot B2 based on the description of the recognition result data E, based on the similarity of the item name candidates in the free-description variable slot B2 (the similarity of the partial character string included in the item name candidates in the free-description variable slot B2 with respect to the item name) and the position of the item name candidates (the offset of the partial character string included in the item name candidates). However, when an item value is already input to a part of the variable slot B of the format data a, the option-type item value determining unit 42 or the free-description item value determining unit 43 may determine an item value to be input to the option-type variable slot B1 or an item value to be input to the free-description variable slot B2 using the input state of the format data a.

Fig. 11 is a diagram showing an example of format data a in which entry values of a part of the variable slots B have been input, and fig. 12 is a diagram showing an example of a format template C corresponding to the format data a shown in fig. 11. Fig. 13 is a diagram showing an example of the recognition result data E used when determining the item value to be input to the variable slot B of the format data a shown in fig. 11.

In the example shown in fig. 11 to 13, "i-chi-ji-ka-n" is extracted as a partial character string similar to the item value (option) "1 hour (i-chi-ji-ka-n)" of the option-type variable slot B1 of the variable slot No.1 by the processing of the extraction unit 41 from the character string of the reading of the recognition result data E, and a combination of the item value (option) and the partial character string is registered as an item value candidate in the candidate list F. Then, by the processing of the extraction unit 41, "i-chi-ji-ka-n" is extracted as a partial character string similar to the item value (option) "primary side pipe (i-chi-ji-ga-wa-ha-i-ka-n)" of the option type variable tank B1 of the variable tank No.2 from the character string of the reading of the recognition result data E, and a combination of the item value (option) and the partial character string is registered as an item value candidate in the candidate list F.

In the present modification, when a plurality of item value candidates including the same partial character string are registered in the candidate list F as described above, the option-type item value determination unit 42 determines the item value of the option-type variable slot B1 of the format data a, using not only the comparison of the degree of similarity of the item value candidates described above or the comparison of the positional relationship with other candidates, but also the input status of the format data a. For example, when the variable tank B which is not input is prioritized as the item value, although the item value candidates corresponding to the item value (item) "1 hour" of the option type variable tank B1 of the variable tank No.1 are more similar than the item value candidates corresponding to the item value (primary side pipe) of the option type variable tank B1 of the variable tank No.2, as shown in fig. 11, the input of the item value of the option type variable tank B1 of the variable tank No.1 is completed, and the item value of the option type variable tank B1 of the variable tank No.2 is not yet input. Therefore, the optional item value determining unit 42 determines the item value of the variable tank No.2 as "primary-side pipe" by using the item value candidates corresponding to the item value (primary-side pipe) of the optional variable tank B1 of the variable tank No. 2.

In the example shown in fig. 11 to 13, "chu-u-i" is extracted as a partial character string similar to the item name "notice" of the free-description type variable slot B2 of the variable slot No.3 from the character string of the reading of the recognition result data E by the processing of the extraction unit 41, and the combination of the item name and the partial character string is registered as an item name candidate in the candidate list F.

In the present modification, when the item name candidates of the free-description variable slot B2 are registered in the candidate list F as described above, the free-description item value determining unit 43 determines the item values of the free-description variable slot B2 of the format data a using not only the similarity or position of the item name candidates but also the input status of the format data a. For example, when the variable slot B that has not been input is prioritized for the item value, even if the item name candidate corresponding to the free-description variable slot B2 of the variable slot No.3 is included in the candidate list F, the item value of the free-description variable slot B2 of the variable slot No.3 is already input, and therefore, the item value using the item name candidate is not determined.

Fig. 14 is a diagram showing an example of format data a in which an item value is input to an uninputed variable slot B according to the present modification. In comparison with the format data a shown in fig. 11, "primary-side piping" determined by the optional item value determining unit 42 is input as an item value to the optional variable tank B1 of the variable tank No.2 to which the item value has not been input. In addition, although the example in which the item value is prioritized over the variable slot B that is not input has been described here, it is considered that it is important to update the variable slot B in which the item value is input based on the utterance that is input later, and the variable slot B in which the item value is input may be prioritized.

As described above, in the present modification, when an item value is already input to a part of the variable slot B of the format data a, the option-type item value determining unit 42 or the free-description item value determining unit 43 also determines an item value of the option-type variable slot B1 or an item value of the free-description variable slot B2 using the input status of the format data a. Therefore, according to the present modification, the item value to be input to each variable slot B of the format data a can be determined more appropriately.

In addition, although the example in which the input state of the format data a is used in addition to the similarity and position of the item value candidates is assumed above, the following configuration may be adopted, for example: when only one item value candidate of a certain variable slot B is included in the candidate list F, if the item value of the variable slot B in the format data a has not been input, the item value of the variable slot B that has not been input is determined using the item value candidate.

< modification 2 >

When the format template C contains supplemental information in addition to the information such as the item name and item value of each variable slot B, the reading thereof, and the type of the variable slot B, the option-type item value determining unit 42 or the free-description item value determining unit 43 may determine the item value to be input to the option-type variable slot B1 or the item value to be input to the free-description variable slot B2, using the supplemental information contained in the format template C.

The supplementary information described in the format template C may be, for example, a priority input flag added to the variable slot B having a high priority. The priority input flag indicates that the entry value should be input to the variable slot B with priority. When a priority input flag is added to the option type variable slot B1, the option type item value determining unit 42 preferentially determines the item value of the option type variable slot B1 to which the priority input flag is added. When a priority input flag is added to the free-description type variable slot B2, the free-description item value determining unit 43 preferentially determines the item value of the free-description type variable slot B2 to which the priority input flag is added.

Another example of the supplementary information described in the format template C is information on the order of the variable slots B in the format data a. In this case, the optional item value determining unit 42 or the free-description item value determining unit 43 determines the item value of the optional variable slot B1 or the item value of the free-description variable slot B2, while taking into consideration the order of arrangement in each format data a, for the item value candidates or the item name candidates registered in the candidate list F by the extracting unit 41.

As described above, in the present modification, when supplemental information is described in the format template C, the option item value determination unit 42 or the free-description item value determination unit 43 determines the item value of the option variable slot B1 or the item value of the free-description variable slot B2 using the supplemental information described in the format template C. Therefore, according to the present modification, the item value to be input to each variable slot B of the format data a can be determined more appropriately.

In addition, although the above assumes an example in which the supplemental information described in the format template C is used in addition to the similarity and position of the item value candidates, the following configuration may be employed: the item value to be input to the variable slot B of the format data a is determined based on the item value candidates included in the candidate list F and the supplementary information described in the template C, without using the similarity or position of the item value candidates.

< modification 3 >

In the above-described embodiment, the virtual format data a has a format including both the option-type variable slot B1 and the free-description variable slot B2, and the configuration of the input assisting apparatus 100 including both the option-type item value determining section 42 that determines the item value input to the option-type variable slot B1 and the free-description item value determining section 43 that determines the item value input to the free-description variable slot B2 has been described. However, when the format data a in the form not including the free-description type variable slot B2 is to be processed, the input assisting apparatus 100 may not include the free-description type item value determining unit 43.

< supplementary notes >

The input assisting apparatus 100 of the present embodiment can be realized by, for example, hardware constituting a general computer and a program (software) executed by the computer in cooperation. For example, the acquiring unit 30, the determining unit 40 (the extracting unit 41, the option-type item value determining unit 42, and the free-description-type item value determining unit 43), and the template managing unit 50 (the variable slot type setting unit 51, the item name setting unit 52, and the option setting unit 53) can be realized by executing a predetermined program by a computer. The format data storage unit 10 and the template storage unit 20 can be realized by using a mass storage device provided in a computer.

Fig. 15 is a block diagram showing an example of the hardware configuration of the input assisting apparatus 100 according to the present embodiment. As shown in fig. 15, the input assisting apparatus 100 has a hardware configuration of a general computer including a hardware processor such as a CPU101, a storage device such as a RAM102 or a ROM103, a mass storage device such as a HDD104, a communication I/F105 for communicating with the outside via a network, and a device I/F106 for connecting peripheral devices.

In this case, the program is provided by being recorded on a magnetic disk, an optical disk, a semiconductor memory, or a similar recording medium, for example. The recording medium on which the program is recorded may be any recording medium that can be read by a computer system, and the storage form thereof may be any form. The program may be installed in advance in a computer, or the program distributed via a network may be installed in the computer as appropriate.

The program executed by the computer has a module configuration including the functional units such as the acquisition unit 30, the determination unit 40 (the extraction unit 41, the option item value determination unit 42, and the free-description item value determination unit 43), and the template management unit 50 (the variable slot type setting unit 51, the item name setting unit 52, and the option setting unit 53), and the program is appropriately read and executed by the processor, thereby generating the above-described units on the main memory such as the RAM 102.

The input assisting apparatus 100 according to the present embodiment may be configured such that a part or all of the above-described functional units are realized by dedicated hardware such as an ASIC (Application Specific Integrated Circuit) or an FPTA (Field-Programmable Gate Array).

The input assisting apparatus 100 according to the present embodiment may be configured as a network system in which a plurality of computers are communicably connected, and may be realized by dispersing the above-described components in the plurality of computers. The input support device 100 according to the present embodiment may be a virtual machine operating on a cloud system.

The embodiments of the present invention have been described above, but the above embodiments are merely presented as examples and are not intended to limit the scope of the invention. The new embodiment can be implemented in other various ways, and various omissions, substitutions, and changes can be made without departing from the spirit of the invention. The above-described embodiments and modifications thereof are also included in the scope and gist of the invention, and are also included in the invention described in the claims and equivalents thereof.

Claims (10)

1. An input assisting device is provided with: a template storage unit for storing a format template having format data of 1 or more variable slots for inputting item values in association with item names, the item names of the variable slots and the options of option-type variable slots for selecting item values from a plurality of options being described together with their respective reading methods; an acquisition unit that acquires recognition result data, which is obtained by voice-recognizing a speech of a user and includes a description and a reading; and a determination unit configured to determine an item value to be input to a variable slot of the format data based on a reading of the recognition result data and a reading of an item name and an option described in the format template, the determination unit including: an extracting unit that extracts a partial character string similar to the reading of the item name or the option described in the format template from the character string of the reading of the recognition result data, and registers a combination of the extracted partial character string, the item name or the option used for extracting the partial character string, and the similarity in a candidate list as candidates; and an option-type item value determining unit that determines an item value to be input to the option-type variable slot based on a positional relationship expressed by a difference between a deviation of the item value candidate and a deviation of the item name candidate or a deviation of another item value candidate different from the item name candidate and a similarity of each candidate, the item value candidates are partial character strings similar to the reading of the options described in the above-described format template, the item name candidates are partial character strings similar to the reading of the item name described in the above-described format template, the offset indicates that the first character of the partial character string in the recognition result data is the second character from the beginning of the character string in the reading of the recognition result data, the option-type item value determining unit sets one item value candidate of the candidate list as a 1 st item value candidate, determines a competition candidate having the same item name or item value as the 1 st item value candidate based on the candidate list, sets the competition candidate as a 2 nd item value candidate, when the similarity between the 1 st item value candidate and the 2 nd item value candidate is different, the one with the greater similarity between the 1 st item value candidate and the 2 nd item value candidate is determined as the item value to be input to the option-type variable slot, when the similarity between the 1 st item value candidate and the 2 nd item value candidate is equal, the difference between the offsets between the 1 st item value candidate and the item name candidate thereof and the difference between the 2 nd item value candidate and the item name candidate thereof are compared, and the smaller difference between the 1 st item value candidate and the 2 nd item value candidate is determined as the item value to be input to the option-type variable slot.

2. The input assisting apparatus according to claim 1, wherein the option-type item value determining unit determines the item value to be input to the option-type variable slot based on the item value candidates and an input state of the format data when an item value is input to a part of the plurality of variable slots included in the format data.

3. The input assisting apparatus according to claim 1, wherein the option-type item value determining unit determines an item value to be input to the option-type variable slot based on the item value candidates and supplementary information described in the template.

4. The input assisting apparatus according to any one of claims 1 to 3, wherein the determining unit further includes a freely-written item value determining unit that determines the item value to be input to the freely-written variable slot based on a degree of similarity of a partial character string similar to a reading of an item name of a freely-written variable slot in which the item value can be freely written, and a position of the partial character string, based on the description of the recognition result data.

5. The input assisting apparatus according to any one of claims 1 to 3, wherein the determination unit further includes a free-description type item value determination unit that determines the item value to be input to the free-description type variable slot based on a partial character string similar to a reading of the item name of the free-description type variable slot in which the item value can be freely written, and an input status of the format data when the item value is input to a part of the plurality of variable slots included in the format data.

6. The input assisting apparatus according to any one of claims 1 to 3, wherein the determining unit further includes a free-written item value determining unit that determines the item value to be input to the free-written variable slot based on a partial character string similar to a reading of an item name of the free-written variable slot in which the item value can be freely written, and the supplementary information described in the template.

7. The input assisting apparatus according to claim 4, further comprising a template management unit configured to manage the format template, the template management unit comprising: a variable slot type setting unit that sets, as a type of variable slot, either an option type or a free description type for each variable slot included in the format data; an item name setting unit that sets an item name and a reading method of each variable slot included in the format data; and an option setting unit that sets options and their reading methods for the item values of the option-type variable slots included in the format data.

8. The input assisting apparatus according to claim 5, further comprising a template management unit configured to manage the format template, the template management unit comprising: a variable slot type setting unit that sets, as a type of variable slot, either an option type or a free description type for each variable slot included in the format data; an item name setting unit that sets an item name and a reading method of each variable slot included in the format data; and an option setting unit that sets options and their reading methods for the item values of the option-type variable slots included in the format data.

9. The input assisting apparatus according to claim 6, further comprising a template management unit configured to manage the format template, the template management unit comprising: a variable slot type setting unit that sets, as a type of variable slot, either an option type or a free description type for each variable slot included in the format data; an item name setting unit that sets an item name and a reading method of each variable slot included in the format data; and an option setting unit that sets options and their reading methods for the item values of the option-type variable slots included in the format data.

10. A recording medium having recorded thereon a program for causing a computer to realize a function including a template storage unit for storing a format template having format data of 1 or more variable slots into which item values are input in correspondence with item names, the item names of the respective variable slots and options of option-type variable slots for selecting an item value from a plurality of options being described together with respective reading methods, the function comprising: a function of acquiring recognition result data, which is obtained by voice-recognizing a speech of a user and includes a description and a reading; a function of extracting a partial character string similar to the reading of the item name or the option described in the format template from the character string of the reading of the recognition result data, and registering a combination of the extracted partial character string, the item name or the option used in the extraction of the partial character string, and the similarity as a candidate in a candidate list; and a function of determining an item value to be input to the option-type variable slot based on a positional relationship expressed by a difference between the offset of the item value candidate and the offset of the item name candidate or the offset of another item value candidate different from the item value candidate, and the similarity of each candidate, the item value candidates are partial character strings similar to the reading of the options described in the above-described format template, the item name candidates are partial character strings similar to the reading of the item name described in the above-described format template, the offset indicates that the first character of the partial character string in the recognition result data is the second character from the beginning of the character string in the reading of the recognition result data, the function of determining the item value to be input to the option-type variable slot is to set one item value candidate of the candidate list as the 1 st item value candidate, determine a competition candidate having the same item name or item value as the 1 st item value candidate based on the candidate list, set the competition candidate as the 2 nd item value candidate, when the similarity between the 1 st item value candidate and the 2 nd item value candidate is different, the item value having the greater similarity between the 1 st item value candidate and the 2 nd item value candidate is determined as the item value to be input to the option-type variable slot, when the 1 st item value candidate and the 2 nd item value candidate have the same similarity, the difference between the offsets between the 1 st item value candidate and the item name candidate thereof and the difference between the 2 nd item value candidate and the item name candidate thereof are compared, and the smaller difference between the 1 st item value candidate and the 2 nd item value candidate is determined as the item value to be input to the option-type variable slot.

Images