JP2012194775A - Information processor control method, control program and information processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an information processor capable of efficiently generating a reply sentence in a style unique to a user.SOLUTION: An information processor includes a processing unit and a storage unit for storing reception sentences and plural reply sentence candidates associated with the reception sentences. In a method for controlling the information processor, the processing unit obtains a reception sentence similar to a divided mail obtained by dividing a received mail from the storage unit, selects a reply sentence candidate associated with the obtained reception sentence from the storage unit, stores conversion rule candidate information for editing the selected reply sentence candidate on the basis of a difference in the reply sentence candidate between before and after editing when the selected reply sentence candidate is edited, stores the same conversion rule candidate information as conversion rule information in the storage unit when the same conversion rule candidate information is stored a predetermined number of times or more, and refers to the conversion rule information to convert the selected reply sentence candidate.

Description

  The present invention relates to an information processing apparatus control method, a control program, and an information processing apparatus.

  At present, communication by e-mail is indispensable, and in particular, the frequency of using mobile phone e-mail is very high among young people. In addition, there are many users of mobile phone mail who think that they must reply immediately when receiving mail, and users who expect to receive a reply immediately when sending mail. It also contributes to the stress of using email. In such a situation, if there is a function that automatically prepares reply mail texts according to the contents of various received mails, the user can save time and effort to enter emails. It is possible to reduce the burden on the user.

  For example, there is a technique in which a user prepares a reply mail associated with a keyword included in a received mail in advance, and transmits the prepared reply mail when the received mail includes the keyword (Patent Document 1).

  In addition, for example, there is a technique for creating a reply mail by searching for a reply mail or a past reply mail prepared in advance using a keyword or the entire sentence in the received mail as a search key (Patent Document 2).

Japanese Patent Laid-Open No. 02-241241 Japanese Patent Laid-Open No. 11-275137

  By the way, in private e-mail, users often write e-mails with their own wording and usage of pictograms, and the recipient of the e-mail also feels the personality of the other party in terms of writer's wording and usage of pictograms. A text template has been prepared in advance for the mail function of recent mobile phone terminals, and using it can reduce the time and effort of creating a mail. However, such text has resistance because it loses its individuality and is used less frequently. There is no actual situation.

  A problem in the disclosure of the present technology is to provide an information processing apparatus that can efficiently create a reply sentence with a user-specific sentence style.

  In order to solve the above-described problem, a control method for an information processing apparatus having a processing unit and a storage unit that stores a received sentence and a plurality of reply sentence candidates associated with the received sentence is received by the processing apparatus. The mail is divided into divided mails, a received sentence similar to the divided mail is acquired from the storage unit, a reply sentence candidate corresponding to the acquired received sentence is selected from the storage unit, and the selected reply sentence candidate Is edited, the difference between the reply sentence candidate before editing and the reply sentence candidate after editing is extracted, and the conversion rule candidate information for converting the selected reply sentence candidate is converted based on the difference. When the same conversion rule candidate information is stored a predetermined number of times or more in the storage unit, the same conversion rule candidate information is stored in the storage unit as conversion rule information, and the storage unit is referred to by referring to the conversion rule information. It performs a conversion against al selected reply sentence candidate.

  According to one aspect of the technology of the present disclosure, the difference extracted from before and after editing a fixed phrase can be reflected in another fixed sentence, and thus a user-specific sentence reply can be efficiently created.

FIG. 2 is a first diagram illustrating a hardware configuration of an information processing apparatus. FIG. 6 is a diagram (part 1) illustrating a process for deriving a style conversion rule in the first embodiment. FIG. 6 is a diagram (part 2) illustrating a process of deriving a style conversion rule in the first embodiment. FIG. 4 is a diagram (part 1) illustrating a style conversion rule according to the first embodiment. FIG. 11 is a diagram (No. 3) for explaining the process of deriving the style conversion rule in the first embodiment. FIG. 11 is a diagram (No. 4) illustrating the process of deriving the style conversion rule in the first embodiment. FIG. 9 is a diagram (part 2) illustrating a style conversion rule according to the first embodiment. FIG. 10 is a diagram (No. 5) illustrating the process of deriving the style conversion rule in the first embodiment. FIG. 10 is a diagram (No. 6) illustrating the process of deriving the style conversion rule in the first embodiment. FIG. 11 is a diagram (No. 3) illustrating a style conversion rule in the first embodiment; FIG. 6 is a diagram (part 1) illustrating a style conversion rule application process in the first embodiment. FIG. 10 is a diagram (part 4) illustrating a style conversion rule according to the first embodiment; FIG. 10 is a diagram (part 2) illustrating the application process of the stylistic conversion rule in the first embodiment. FIG. 11 is a diagram (No. 7) illustrating the process of deriving the style conversion rule in the first embodiment. FIG. 10 is a diagram (No. 8) illustrating the process of deriving the style conversion rule in the first embodiment. FIG. 10 is a diagram (No. 5) illustrating a style conversion rule in the first embodiment. FIG. 10 is a diagram (No. 9) illustrating the process of deriving the style conversion rule in the first embodiment. FIG. 10 is a diagram (No. 10) illustrating the process of deriving the style conversion rule in the first embodiment. FIG. 6 is a diagram (No. 6) illustrating a style conversion rule in the first embodiment; FIG. 10 is a diagram (part 3) for explaining the application process of the style conversion rule in the first embodiment. FIG. 10 is a diagram (No. 7) illustrating a style conversion rule in the first embodiment; FIG. 10 is a diagram (part 4) illustrating the application process of the style conversion rule in the first embodiment. 3 is a flowchart (part 1) illustrating a process according to the first embodiment. 3 is a flowchart (part 2) illustrating a process according to the first embodiment. FIG. 2 is a second diagram illustrating a hardware configuration of the information processing apparatus. FIG. 10 is a diagram showing a received / reply text table in the second embodiment. It is a figure which shows the style conversion rule in Embodiment 2. FIG. It is FIG. (1) explaining the style conversion of the reply sentence candidate in Embodiment 2. FIG. It is FIG. (2) explaining the style conversion of the reply sentence candidate in Embodiment 2. FIG. 6 is a flowchart (part 1) illustrating a process according to the second embodiment. 10 is a flowchart (part 2) illustrating a process according to the second embodiment. FIG. 3 illustrates a hardware configuration of the information processing apparatus (part 3); FIG. 10 is a diagram showing a received / reply text table in the third embodiment. FIG. 10 is a diagram illustrating a style conversion rule in a third embodiment. FIG. 10 is a diagram for explaining style conversion of a reply sentence candidate in the third embodiment. 10 is a flowchart (part 1) illustrating a process according to the third embodiment. 10 is a flowchart (part 2) illustrating a process according to the third embodiment. FIG. 4 is a diagram illustrating a hardware configuration of the information processing apparatus (part 4);

  Hereinafter, the information processing apparatus in the present embodiment will be described.

<Embodiment 1>
FIG. 1 is a hardware configuration diagram of the information processing apparatus 100 according to the present embodiment. The information processing apparatus 100 includes a received sentence / reply sentence database 101, a mail division unit 102, a search unit 103, a reply sentence candidate extraction unit 104, a reply sentence candidate sentence conversion unit 201, a sentence conversion rule table 206, and a reply sentence candidate display unit 105. A reply sentence selection / editing unit 202, a difference analysis unit 203, a style conversion rule generation unit 204, a style conversion rule candidate storage unit 205, and a reply text combination unit 107.

  The received sentence / reply sentence database 101 is a database in which a large number of sets of received sentences and one or more reply sentences for the received sentences are stored.

  When a received mail is input, the mail dividing unit 102 divides the partial character string having a length of about a sentence using punctuation marks, pictograms, emoticons and the like as clues.

  The search unit 103 searches the received sentence / reply sentence database 101 for a received sentence similar to the divided partial character string.

  The reply sentence candidate extraction unit 104 extracts reply sentence candidates for the received sentences retrieved from the received sentence / reply sentence database 101.

  The reply sentence candidate style conversion unit 201 converts each reply sentence candidate extracted from the database according to the style conversion rules registered in the style conversion rule table 206.

  The reply sentence candidate display unit 105 presents the converted reply sentence candidates to the user, and when there are a plurality of reply sentence candidates, prompts the user to select one sentence.

  The reply sentence selection editing unit 202 accepts selection of a reply sentence by the user and accepts editing of the user's text with respect to the selected sentence.

  When the sentence selected by the user is edited, the difference analysis unit 203 extracts the difference between the texts before and after editing.

  The style conversion rule generation unit 204 generates a style conversion rule candidate. The stylistic conversion rule generation unit 204 stores the generated stylistic conversion rule candidate in the stylistic conversion rule candidate storage unit 205 and collates with the past stylistic conversion rule candidate. When the same rule is generated more than a predetermined number of times, that is, when the same correction is performed more than the predetermined number of times, the style conversion rule generation unit 204 sets the style conversion rule candidate as the style conversion rule, Stored in the conversion rule table 206. The rules stored in the style conversion rule table 206 are applied to the style conversion of subsequent reply text candidates. When the number of times that the same rule is generated is less than the predetermined number, the style conversion rule generation unit 204 reflects it in the style conversion rule candidate storage unit 205 as a style conversion rule candidate.

  The reply sentence combining unit 107 connects reply sentences selected and edited by the user and creates a reply mail.

  2 to 4 illustrate processing for extracting a difference from a reply sentence candidate correction result to generate a style conversion rule candidate and deriving a style conversion rule from a plurality of style conversion rule candidates in this embodiment. FIG.

  In FIG. 2, one of the reply sentence candidates extracted from the received sentence / reply sentence database 101 by the reply sentence candidate extraction unit 104 is “I am fine” of the text 1, and after the user selects the text 1, “ "I'm fine!" Here, the difference analysis unit 203 acquires that the difference “.” Has been replaced with “-!” By comparing the text before and after editing. Further, when the difference analysis unit 203 performs morphological analysis on the text before editing, “Genki” is an adjective verb stem, “da” is an adjective verb final form, “yo” is a final particle, and “.”. To get a punctuation mark. The difference analysis unit 203 generates the style conversion rule candidate 10 based on the morphological analysis result.

  The style conversion rule candidate 10 has condition information 20 and edit content information 30. The condition information 20 includes preceding word information 21, the word information 22 and subsequent word information 23. The editing content information 30 includes replacement / insertion / deletion information 31, pre-substitution character string information 32, and post-substitution character string information 33. In the present embodiment, the difference analysis unit 203 replaces “.” Of the word information 22 when the preceding word information 21 is the final particle “yo”, the word information 22 is the punctuation “.”, And the subsequent word information 23 is not present. A stylistic conversion rule candidate 10 to be replaced with “-!” In the subsequent character string information 33 is generated, “replacement” in the replacement / insertion / deletion information 31, “.” In the pre-replacement character string information 32, and the character after replacement “−!” Is stored in the column information 33.

  In FIG. 3, when the reply sentence candidate extraction unit 104 edits “I will go tomorrow” in the text 2 of the reply sentence candidate extracted from the received sentence / reply sentence database 101 to “I will go tomorrow!”. In addition, the style conversion rule generation unit 204 acquires the same style conversion rule candidate 10. As described above, after the reply sentence is selected and edited several times, the same editing, that is, the editing of replacing “.” With “-!” In the replaced character string information 33 in advance. If it occurs more than the predetermined number of times, the style conversion rule generation unit 204 adopts it as the style conversion rule 40 in association with the common condition and stores it in the style conversion rule table 206.

  FIG. 4 shows the style conversion rule 40 obtained from the processing described in FIGS. The style conversion rule 40 includes condition information 50 and editing content information 60. The condition information 50 includes preceding word information 51, the word information 52, and subsequent word information 53. The edited content information 60 includes replacement / insertion / deletion information 61, pre-substitution character string information 62, and post-substitution character string information 63. In the present embodiment, the stylistic conversion rule generation unit 204 replaces the word “.” When the preceding word information 51 is the final particle “yo”, the word information 52 is the phrase “.”, And the subsequent word information 53 is not present. The stylistic conversion rule to replace the character string information 63 with “-!” Is finalized.

  A process for deriving another style conversion rule 40 will be described with reference to FIGS. As shown in FIG. 5, the editing content is replacement in the processing examples of FIGS. 2 and 3, but in this processing, the editing content is deletion. Therefore, the editing content information 30 includes replacement / insertion / deletion information 31. , And deleted character string information 34. According to the editing of the user who changed “Yappayo Iyo” in text 1 to “Yappa Iyoi”, there is no preceding word information 21, the word information 22 is the adverb “Yappari”, and the subsequent word information 23 is the adjective verb stem “I”. At this time, the stylistic conversion rule candidate 10 for deleting “ri” of “after all” in the word information 22 is obtained.

  Further, as shown in FIG. 6, from the editing of “Tomorrow after all” in text 2, “Tomorrow after all” is edited, the preceding word information 21 is the common noun “Tomorrow”, and the word information 22 is the adverb “After all. ”, When the subsequent word information 23 is the first-stage verb stem“ stop ”in the line, the stylistic conversion rule candidate 10 for deleting“ ri ”of“ after ”of the word 21 is obtained. After the selection and editing of the reply sentence is repeated several times, if the editing of deleting “ri” of “after all” is performed a predetermined number of times or more, the edited content is used as the style conversion rule 40 together with the common condition. Determine.

  FIG. 7 shows the style conversion rule 40 obtained from the processing described in FIGS. 5 and 6. From the stylistic conversion rule 40 shown in FIG. 7, “ri” described in the deletion character string information 64 is deleted from the adverb “Yappari” of the word information 52 regardless of the preceding word information 51 and the succeeding word information 53. The rule to do is obtained. As described above, the editing of deleting a part of a word in an independent word such as an adverb can generally be ruled regardless of the preceding word information 51 and the succeeding word information 53.

  A process for deriving another style conversion rule 40 will be described with reference to FIGS. Since the editing content is insertion in this process, the editing content information 30 includes replacement / insertion / deletion information 31 and insertion character string information 35. As shown in FIG. 8, for the text 1 “I ’m looking forward to a drinking party”, from the edit that inserts a specific pictogram after the “drinking party”, the preceding word information 21 is the common noun “Drinking party”, When the word information 22 is not specified and the subsequent word information 23 is a common noun “fun”, a stylistic conversion rule candidate 10 is obtained in which a pictogram is inserted as the inserted character string information 35 at the position of the word information 22. In this way, when a character string is inserted between words in the text before editing, the word information 22 is treated as unspecified for convenience.

  As shown in FIG. 9, for the text 2 “Let's drink”, from the edit that inserts a specific pictogram after “Drink”, the preceding word information 21 is the common noun “Drink”, the word When the information 22 is not specified, and the subsequent word information 23 is the sub-variable-type “shi”, the stylistic conversion rule candidate 10 is obtained in which a pictogram is inserted as the insertion character string information 35 at the position of the word information 22. After selecting and editing a reply sentence several times, if editing for inserting the same pictograph is performed a predetermined number of times or more, the edited content is confirmed as a style conversion rule 40 together with the common conditions.

  FIG. 10 shows the style conversion rule 40 obtained from the processing described in FIGS. From the stylistic conversion rule 40 shown in FIG. 10, the rule that the same pictogram is inserted as the inserted character string information 65 immediately after the preceding word information 51 “drinking party” is obtained. The rule for inserting a character string between words as described above is based on a common condition from a plurality of editing cases in order to determine whether the rule depends on the preceding word information 51 or the following word information 53. Need to be extracted.

  Processing for converting the style of a reply sentence candidate by applying the style conversion rule 40 described with reference to FIGS. 2 to 10 will be described with reference to FIGS. 11 to 13.

  As shown in FIG. 11, in the received sentence / reply sentence database 101, for example, “I am free,” “I ’m sorry, I ’m a little busy” for the received sentence 70 “Today's time?” It is assumed that a reply sentence 80 saying "I haven't done anything" is stored. Similarly, the reply sentences “Let's go to drink tomorrow ~”, “OK, I ’m looking forward to it!” And “I ’m sorry, I ’m sorry,” and the received sentence “Waiting for a reply.” It is assumed that reply sentences “Thank you.”, “Please wait” and “That's why” are stored.

  When receiving an incoming mail saying “Tomorrow's time? Do you want to drink? Waiting for contact?”, First, the mail dividing unit 102 divides the mail text using punctuation marks, pictograms, emoticons, and the like as clues. In the present embodiment, the mail dividing unit 102 divides the mail sentence into “Tomorrow's time?”, “Do not go for a drink?” And “Waiting for contact”.

  The search unit 103 searches the received sentence / reply sentence database 101 for a received sentence similar to each of the partial character strings divided by the mail dividing unit 102. As a similar sentence search algorithm, for example, there is an algorithm for searching for a similar sentence using the number of matching characters, but the algorithm is not limited to this. In the present embodiment, the search unit 103 searches for the received sentence “Today's Hima?” In the received sentence / reply sentence database 101 for the received sentence “Tomorrow Hima?” In the received mail. Similarly, the search unit 103 responds “Let's go drinking for tomorrow” for “Would you like to drink?” And “Waiting for a reply!” For “Waiting for contact!” As a similar sentence.

  The reply sentence candidate extraction unit 104 extracts a reply sentence for the received sentence searched by the search unit 103. That is, the reply sentence candidate extraction unit 104 sends reply sentences such as “I'm free,” “I ’m sorry, I ’m too busy” and “I ’m not doing anything” associated with “Today's time?”. , Extracted as a reply sentence candidate for “Tomorrow Hima?” In the received mail sentence. Similarly, the reply sentence candidate extraction unit 104 seems to be unable to drink as a reply sentence candidate for the received sentence “Would you like to drink?”, “OK! I'm looking forward to it!” And “I'm sorry, after all. ”Is extracted as a reply sentence candidate for the received sentence“ Waiting for contact ”,“ Please wait ”, and“ That's why ”.

  FIG. 12 shows the style conversion rule table 206 in this embodiment. The style conversion rule table 206 stores the three types of style conversion rules 40 described with reference to FIGS. In the stored rules, when the preceding word information 51 is the final particle “yo”, the word information 52 is the punctuation “.”, And when there is no subsequent word information 53, “.” Of the word information 52 is “-”. The rule that deletes “ri” of the adverb “Yappari” as deleted character string information 64 regardless of the rule that replaces “!”, The preceding word information 51, and the succeeding word information 53, immediately after the preceding word information 51 “drinking party” This is a rule for inserting pictograms as the inserted character string information 65.

  As shown in FIG. 13, the reply sentence candidate style conversion unit 201 refers to the style conversion rule table 206 shown in FIG. For example, the reply sentence candidate “I'm free” is because the preceding word information 51 has the final particle “yo”, the word information 52 has the punctuation “.”, And the subsequent word information 53 does not exist. Refers to the stylistic conversion rule table 206 and converts it to “It's free!”. Similarly, the reply sentence candidate style conversion unit 201 converts “I'm not doing anything” to “I'm not doing anything!”. The reply sentence candidate “Drinking party is OK!” Has a preceding word “Drinking party”, so the reply sentence candidate sentence style conversion unit 201 refers to the style conversion rule table 206 and inserts an emoticon after “drinking party”. And convert it to "Drinking party is good!" In addition, the reply sentence candidate sentence style conversion unit 201 refers to the style conversion rule table 206, and the reply sentence candidate "I'm sorry, I can't do it." The rule to delete "" is applied and converted to "I'm sorry, it's impossible." Since there is no applicable style conversion rule for other reply sentence candidates, the reply sentence candidate style conversion unit 201 directly adopts the text in the received sentence / reply sentence database 101 as a reply sentence candidate.

  The reply sentence candidate display unit 105 presents the reply sentence candidate extracted from the received sentence / reply sentence database 101 and further applied with the style conversion rule to the user, and prompts the user to select a reply sentence candidate.

  The reply sentence selection editing unit 202 accepts selection of a reply sentence by the user and accepts editing of the user's text with respect to the selected sentence. Here, it is assumed that the user selects “It's free time!”, “Drinking party is good!” And “Let me know”. Here, when the selected reply sentence is edited, the difference analysis unit 203 and the style conversion rule generation unit 204 extract the style conversion rule candidates, derive the style conversion rules, and register the style conversion rules. Done.

  The reply message combining unit 107 combines the reply texts selected and edited by the user, and creates a reply mail of “It's free! Drinking party ☆ Good! Nice to meet you.” Here, a function of accepting edits such as addition, deletion, and correction of sentences by the user may be provided.

  Processing for generating another style conversion rule will be described with reference to FIGS. 14 to 16. In the above-described processing, it has been described that the notation and part of speech of the preceding word information 51, the word information 52, and the succeeding word information 53 are used as conditions for applying the style conversion rule. As described above, substitution / insertion / deletion within the character of an independent word can be performed without depending on the preceding word information 51 and the succeeding word information 53, but the auxiliary words such as particles and auxiliary verbs are Depending on the word, it may or may not be converted.

  As shown in FIG. 14, since “It's amazing” in text 1 is converted to “It's amazing”, the reply sentence candidate sentence conversion unit 201 indicates that the preceding word information 21 has the adjective end form “ When the answer is “Yes”, it is possible to derive a style conversion rule candidate for converting the final particle “Ne” with “Jan” in the replaced character string information 33.

  As shown in FIG. 15, since “very fast” in text 2 is converted to “very fast”, the reply sentence candidate sentence conversion unit 201 sets the preceding word information 21 to the adjective end form “ When the answer is “Yes”, it is possible to derive a style conversion rule candidate for converting the final particle “Ne” with “Jan” in the replaced character string information 33.

  FIG. 16 shows a style conversion rule 40 generated from the processing described with reference to FIGS. 14 and 15. When the preceding word information 21 is the adjective ending form “I”, a rule for converting the final particle “Ne” with “Jan” in the replaced character string information 63 can be derived. At this time, the subsequent word information 53 is not affected.

  Processing for generating another style conversion rule will be described with reference to FIGS. 17 and 18. In the above-described processing, the example of using the preceding word, the notation of the subsequent word, and the part of speech as the condition for applying the style conversion rule has been described. However, the method for setting the condition to apply the style conversion rule is not limited to this. As shown in the present embodiment, in the case of editing in which some pictogram is inserted at the end of a sentence, any of the words constituting the sentence may be used as a condition without depending on the immediately preceding word.

  In the example of FIG. 17, in the case of editing in which a specific pictograph is inserted at the end of the text 1 “I will go tennis”, the adverb “related word 1” 24, which is an independent word, is obtained from the morphological analysis result. For example, the common noun “tennis” of “Now”, “Related Word 2” 25, and the Ka-line five-level verb stem “Line” of “Related Word 3” 26 can be extracted, for example. Conversion that inserts the pictogram as the insertion character string information 35 at the end of the sentence on the condition of the extracted three words is generated as a stylistic conversion rule candidate.

  In the example of FIG. 18, when editing is performed in which a specific pictograph is inserted at the end of the text 2 “schedule for next week's tennis”, “related word 1” 24, which is an independent word, is obtained from the morphological analysis result. The common noun “next week”, the common noun “tennis” of “related word 2” 25 and the common noun “schedule” of “related word 3” 26 can be extracted, and the pictograms are inserted string information on the condition of these three words. Conversion for inserting the sentence end as 35 can be generated as a style conversion rule candidate. After selecting and editing a reply sentence several times, if the conversion for inserting the pictogram at the end of the sentence occurs more than a predetermined number of times, it is registered as a style conversion rule 40 together with those common conditions.

  FIG. 19 shows the style conversion rule 40 generated from the processing described with reference to FIGS. 17 and 18. A sentence style conversion rule 40 is derived that inserts the pictogram as the inserted character string information 65 for the sentence in which the common noun “tennis” as the related word 57 appears in the sentence.

  Processing for converting the style of a reply sentence candidate by applying the style conversion rule 40 described with reference to FIGS. 14 to 19 will be described with reference to FIGS.

  The style conversion rule application processing in this embodiment will be described with reference to FIG. As shown in FIG. 20, in the received sentence / reply sentence database 101, for example, for a received sentence 70 "Good morning", reply sentences 80 "Good morning", "Good morning" and "Good morning". Is stored. Similarly, it is assumed that the reply sentences “Tennis OK! Fun!”, “That's good, I ’m going” and “I ’m sorry, it seems a bit impossible” are stored for the received sentence “Let's play tennis on Sunday”.

  When the received mail “Good morning! Don't you play tennis next Sunday?” Is received, the mail dividing unit 102 first divides the mail text using punctuation marks, pictograms, emoticons and the like. In the present embodiment, the mail dividing unit 102 divides the mail sentence into “Good morning!” And “Do not play tennis next Sunday?”.

  The search unit 103 searches the received sentence / reply sentence database 101 for a received sentence similar to each of the partial character strings divided by the mail dividing unit 102. In this embodiment, the search unit 103 searches for the received sentence “Good morning” in the received sentence / reply sentence database 101 for the received sentence “Good morning!” In the received mail. Similarly, it is assumed that the retrieval unit 103 retrieves “Let's play tennis on Sunday” as a similar sentence in response to “Do you want to play tennis next Sunday?”

  The reply sentence candidate extraction unit 104 extracts a reply sentence for the received sentence searched by the search unit 103. That is, the reply sentence candidate extraction unit 104 sends reply sentences such as “good morning”, “good morning” and “good morning” associated with “good morning” to the reply to the received mail sentence “good morning!”. Extract as sentence candidates. Similarly, the reply sentence candidate extraction unit 104 has “tennis OK! I am looking forward to it!” And “I'm happy. . ”Is extracted as a reply sentence candidate for the received sentence“ Let's play tennis on Sunday ”.

  FIG. 21 shows a style conversion rule table 206 in the present embodiment. The style conversion rule table 206 stores the two types of style conversion rules 40 described with reference to FIGS. In the stored rules, the preceding word information 51 is the adjective “I”, the word information 52 is the final particle “Ne”, and the “Ne” in the word information 52 is replaced regardless of the subsequent word information 53. This is a rule for replacing with “Jan” in the subsequent character string information 63, and a rule for inserting pictograms as the insertion character string information 65 when the related word information 57 is the common noun “tennis”.

  In FIG. 22, using this style conversion rule, the reply sentence candidate “That is good. I will go” extracted from the received sentence / reply sentence database 101 is converted to “That is good. I will go.” Can do. On the other hand, among the reply sentence candidates, there is a sentence “I'm sorry, it seems a little impossible”, and the final particle “ne” exists in this sentence. Replacing it with “Jan” will result in “I ’m sorry, it ’s a little impossible”, and it ’s unnatural as Japanese, but because the preceding word information 51 has the adjective “I”, this book No rules are applied, and as a result, such unnatural conversion can be avoided. As described above, it is necessary to derive a rule with a condition that considers the preceding and succeeding words in the stylistic conversion rule relating to substitution / insertion / deletion within the attached word. Further, in the reply sentence candidate “tennis OK! Fun!” Extracted from the received sentence / reply sentence database 101, the pictograph is inserted at the end of the sentence “tennis OK!”, And “tennis OK! It is converted as follows.

Here, if the user selects “Ohayo ~” and “Tennis OK! ♪ Fun!” From the presented reply sentence candidates, the reply sentence combining unit 107 combines the reply sentences selected by the user. Then, a reply mail “Ohayo ~ Tennis OK! Furthermore, a function for accepting edits such as addition, deletion, and correction of sentences by the user may be provided.
23 and 24 show a processing flow in this embodiment.

  In step S <b> 101, when the received mail is input, the mail dividing unit 102 divides the received mail into partial character strings about the length of sentences using punctuation marks, pictograms, emoticons, and the like as clues. Let N be the number of partial character strings divided at this time. In the processing described with reference to FIG. 11, the mail dividing unit 102 divides the received mail into three partial character strings. In this case, N is “3”. The process proceeds to step S102.

  In step S102, the mail dividing unit 102 sets a counter value I = 0 and initializes a buffer for temporarily storing reply mail. The process proceeds to step S103.

  In step S103, the mail dividing unit 102 determines whether or not the number N of partial character strings is greater than the counter value I. In the processing described with reference to FIG. 11, the number N of partial character strings is “3”, so “3” and “0” are compared at the first time. When the number N of partial character strings is larger than the counter value I, the process proceeds to step S104. On the other hand, when the number N of partial character strings is not larger than the counter value I, the process proceeds to step S113.

  In step S104, the search unit 103 searches the received sentence / reply sentence database 101 for a received sentence similar to the partial character string. In the processing described with reference to FIG. 11, the search unit 103 first searches for “Tomorrow's time?” For “Today's time?”. The process proceeds to step S105.

  In step S105, the reply sentence candidate extraction unit 104 extracts reply sentence candidates associated with the retrieved received sentence. In the process described with reference to FIG. 11, the reply sentence candidate extraction unit 104 first corresponds to “Today's time?”, “I'm busy,” “I ’m sorry, I ’m busy,” and “I'm not doing anything.” Is extracted. The process proceeds to step S106.

  In step S106, the reply sentence candidate sentence style conversion unit 201 applies any of the sentence style conversion rules registered in the style conversion rule table 206 that can be applied to the reply sentence candidate, and extracts the reply sentence candidates extracted from the database. Convert. In the processing described with reference to FIG. 13, the reply sentence candidate style conversion unit 201 sets “I'm free” to “I'm free!” And “I'm not doing anything” to “I'm not doing anything.” ! " The process proceeds to step S107.

  In step S107, the reply sentence candidate display unit 105 presents the converted reply sentence candidate to the user, and when there are a plurality of reply sentence candidates, prompts the user to select one sentence from them. In the processing described with reference to FIG. 11, the reply sentence candidate display unit 105 responds to “Tomorrow's time?”, “I'm free!”, “I ’m sorry, I ’m a little busy. ! "To the user. The process proceeds to step S108.

  In step S <b> 108, the reply sentence selection / editing unit 202 receives one sentence selected from reply sentence candidates by the user. In the processing described with reference to FIG. 13, the reply sentence selection / editing unit 202 accepts “I am free!”. The process proceeds to step S109.

  In step S109, the reply sentence selection editing unit 202 applies the edit performed by the user to the selected sentence. The process proceeds to step S110.

  In step S110, the reply message selection editing unit 202 stores the created reply message in the reply mail buffer. The process proceeds to step S111.

  Here, the style conversion rule registration processing in step S111 will be described in detail with reference to FIG.

  In step S <b> 1111, the difference analysis unit 203 determines whether or not the user has edited the registered reply sentence candidate. When the user edits the registered reply sentence candidate, the process proceeds to step S1112. On the other hand, when the user has not edited the reply candidate selected for registration, the process proceeds to step S112.

  In step S <b> 1112, the stylistic conversion rule generation unit 204 generates a stylistic conversion rule candidate based on the text difference before and after editing extracted by the difference analysis unit 203. For example, in the processing described with reference to FIG. 2, the difference analysis unit 203 determines that when the preceding word information 21 is the final particle “yo”, the word information 22 is the punctuation mark “.”, And the subsequent word information 23 is not present, A stylistic conversion rule candidate 10 that replaces “.” With “-!” In the post-replacement character string information 33 is generated, “replace” in the replacement / insertion / deletion information 31, and “. "-!" Is stored in the replaced character string information 33. The processing moves to step S1113.

  In step S <b> 1113, the style conversion rule generation unit 204 refers to the style conversion rule candidate storage unit 205 and acquires the number of times that a similar style conversion rule candidate has been derived in the past. For example, in the processing described with reference to FIG. 3, the same editing as the processing described with reference to FIG. 2 is performed in which “.” Is replaced with “-!” In the character string information 33 after replacement. In this way, the style conversion rule generation unit 204 acquires the number of times that the same editing as the processing described in FIGS. 2 and 3 has been performed. When the style conversion rule candidate is derived more than a predetermined number of times, the process proceeds to step S1114. On the other hand, when the style conversion rule candidate has not been derived more than a predetermined number of times, the process proceeds to step S1115.

  In step S1114, the style conversion rule generation unit 204 stores the style conversion rule candidate in the style conversion rule table 206 as a style conversion rule. The style conversion rule generation unit 204 stores, for example, the style conversion rules as described with reference to FIG. The process proceeds to step S112.

  In step S1115, if the style conversion rule candidate is already registered in the style conversion rule candidate storage section 205, the style conversion rule generation unit 204 increments the number of appearances of the style conversion rule candidate by one. When the style conversion rule candidate is not already registered in the style conversion rule candidate storage unit 205, the style conversion rule generation unit 204 newly registers the rule candidate and sets the appearance count to 1. The process proceeds to step S112.

  In step S112, the mail dividing unit 102 increments I to “1” from “0”. The process returns to step S103 to move to the next partial character string process. In the present embodiment, since N is “3”, the processing from step S101 to step S112 is repeated three times. When the process is completed for all partial character strings in this way, the process proceeds to step S113.

  Next, the second flow starting from step S103 will be described.

  In step S103, the mail dividing unit 102 determines whether or not the number N of partial character strings is greater than the counter value I. In the processing described with reference to FIG. 11, the number N of partial character strings is “3”, so “3” and “1” are compared in the second flow. When the number N of partial character strings is larger than the counter value I, the process proceeds to step S104. On the other hand, when the number N of partial character strings is not larger than the counter value I, the process proceeds to step S113.

  In step S104, the search unit 103 searches the received sentence / reply sentence database 101 for a received sentence similar to the partial character string. In the processing described with reference to FIG. 11, the search unit 103 searches “Let's go for a drink tomorrow?” For “Would you like to go for a drink?”. The process proceeds to step S105.

  In step S105, the reply sentence candidate extraction unit 104 extracts reply sentence candidates associated with the retrieved received sentence. In the processing described with reference to FIG. 11, the reply sentence candidate extraction unit 104 associates “Let's drink tomorrow ~” with “Let's have a drinking party!”, “OK! "It seems impossible." The process proceeds to step S106.

  In step S106, the reply sentence candidate sentence style conversion unit 201 applies any of the sentence style conversion rules registered in the style conversion rule table 206 that can be applied to the reply sentence candidate, and extracts the reply sentence candidates extracted from the database. Convert. In the processing described in FIG. 13, the reply sentence candidate style conversion unit 201 changes “I want to have a drinking party!” To “I want to have a drinking party!” And “I ’m sorry, it ’s impossible” after all. . " The process proceeds to step S107.

  In step S107, the reply sentence candidate display unit 105 presents the converted reply sentence candidate to the user, and when there are a plurality of reply sentence candidates, prompts the user to select one sentence from them. In the process described with reference to FIG. 12, the reply sentence candidate display unit 105 responds to “Don't go for a drink?” With “Drinking party ☆ OK!”, “OK! I ’m looking forward to it!” And “I ’m sorry. "It seems impossible." The process proceeds to step S108.

  In step S <b> 108, the reply sentence selection / editing unit 202 receives one sentence selected from reply sentence candidates by the user. In the processing described with reference to FIG. 12, the reply message selection editing unit 202 accepts “Drinking party is good!”. The process proceeds to step S109.

  In step S109, the reply sentence selection editing unit 202 applies the edit performed by the user to the selected sentence. The process proceeds to step S110.

  In step S110, the reply message selection editing unit 202 stores the created reply message in the reply mail buffer. The process proceeds to step S111.

  In step S111, a style conversion rule registration process is performed. The process proceeds to step S112.

  In step S112, the mail dividing unit 102 increments I to “2” from “1”. The process returns to step S103 to move to the next partial character string process.

  Next, the third flow starting from step S103 will be described.

  In step S103, the mail dividing unit 102 determines whether or not the number N of partial character strings is greater than the counter value I. In the process described with reference to FIG. 11, the number N of partial character strings is “3”, so “3” and “2” are compared in the third flow. When the number N of partial character strings is larger than the counter value I, the process proceeds to step S104. On the other hand, when the number N of partial character strings is not larger than the counter value I, the process proceeds to step S113.

  In step S104, the search unit 103 searches the received sentence / reply sentence database 101 for a received sentence similar to the partial character string. In the processing described with reference to FIG. 11, the search unit 103 searches for “Waiting for reply!” For “Waiting for contact!”. The process proceeds to step S105.

  In step S105, the reply sentence candidate extraction unit 104 extracts reply sentence candidates associated with the retrieved received sentence. In the processing described with reference to FIG. 11, the reply sentence candidate extraction unit 104 displays “Thank you for your reply!”, “Thank you for waiting”, “Wait for you”, and “That's why.” Extract. The process proceeds to step S106.

  In step S106, the reply sentence candidate sentence style conversion unit 201 applies any of the sentence style conversion rules registered in the style conversion rule table 206 that can be applied to the reply sentence candidate, and extracts the reply sentence candidates extracted from the database. Convert. In the processing described with reference to FIG. 11, the reply sentence candidate sentence style conversion unit 201 does not have a style conversion rule that can be applied to the reply sentence corresponding to “Wait for a reply!”. 201 does not perform the conversion process. The process proceeds to step S107.

  In step S107, the reply sentence candidate display unit 105 presents the converted reply sentence candidate to the user, and when there are a plurality of reply sentence candidates, prompts the user to select one sentence from them. In the processing described with reference to FIG. 13, the reply message candidate display unit 105 presents “Let me know”, “Wait me”, and “That's right” to the user in response to “Waiting for contact”. To do. The process proceeds to step S108.

  In step S <b> 108, the reply sentence selection / editing unit 202 receives one sentence selected from reply sentence candidates by the user. In the process described with reference to FIG. 13, the reply message selection / editing unit 202 accepts “Let me see you”. The process proceeds to step S109.

  In step S109, the reply sentence selection editing unit 202 applies the edit performed by the user to the selected sentence. The process proceeds to step S110.

  In step S110, the reply message selection editing unit 202 stores the created reply message in the reply mail buffer. The process proceeds to step S111.

  In step S111, a style conversion rule registration process is performed. The process proceeds to step S112.

  In step S112, the mail dividing unit 102 increments I from “2” to “3”. The process returns to step S103.

    In step S103, the mail dividing unit 102 determines whether or not the number N of partial character strings is greater than the counter value I. In the processing described with reference to FIG. 11, the number N of partial character strings is “3”, so “3” and “3” are compared in the third flow. When the number N of partial character strings is larger than the counter value I, the process proceeds to step S104. On the other hand, when the number N of partial character strings is not larger than the counter value I, the process proceeds to step S113.

In step S113, the reply message combining unit 107 outputs the character string stored in the reply mail buffer as a reply mail. In the processing described with reference to FIG. 13, the reply message combining unit 107 outputs “I am free! Drinking party ☆ Good! Nice to meet you” as a reply mail. The process ends.
As described above, according to the present embodiment, in the mail apparatus having the function of automatically presenting a reply sentence candidate corresponding to the content of the received mail, the same editing as the sentence style performed in the past has been performed. Reply sentence candidates are prepared in the state, and it is possible to save the user from editing the reply sentence again. Further, by continuing to use this function, it is possible to make the device appear to learn the user's style and style, and the effect of increasing the attachment of the user to the device can be achieved.

<Embodiment 2>
When a user creates an e-mail, the style is generally different depending on the transmission partner. Therefore, by managing the style conversion rule in association with the reply destination and converting the reply text candidate using the style conversion rule corresponding to the transmission partner, it is possible to provide a more convenient function. Also, in the derivation and registration of the style conversion rule, a process for registering the style conversion rule in association with the returned reply sentence is prepared.

  FIG. 25 is a hardware configuration diagram of the information processing apparatus 100 in the present embodiment. Constituent elements included in the information processing apparatus 100 described in Embodiment 1 are denoted by the same reference numerals in the drawing, and description thereof is omitted here. The information processing apparatus 100 according to the present embodiment newly includes a sender information acquisition unit 401 and a reply destination information management unit 402.

  A feature of the present embodiment is that the sender information acquisition unit 401 acquires the sender information from the received mail, and the reply destination information management unit 402 manages this as reply destination information. The caller information may be a mail address of the caller, a name for managing the mail partner on the mail software, or the like. The reply destination information management unit 402 converts the reply sentence candidate using the sentence conversion rule corresponding to the reply partner of the currently created mail when the reply sentence candidate sentence conversion unit 201 converts the reply sentence candidate. . The reply destination information management unit 402 stores the generated style conversion rule candidate in the style conversion rule candidate storage unit 205 in association with the reply destination in the style conversion rule generation unit 204. Further, the reply destination information management unit 402 also stores the generated style conversion rule candidate in association with the reply destination when registering it in the style conversion rule table.

  FIG. 26 shows a received / reply text table 101 in the present embodiment. For example, for the received sentence 70 “Today's time?”, Reply sentences 80 “I'm free,” “I ’m sorry, I ’m a little busy” and “I ’m not doing anything” are stored. Similarly, the reply sentences “Let's go to drink tomorrow ~”, “OK, I ’m looking forward to it!” And “I ’m sorry, I ’m sorry,” and the received sentence “Waiting for a reply.” The reply sentences “Thanks!”, “Take me waiting”, and “That ’s why” are stored.

  FIG. 27 shows the style conversion rule table 206 in this embodiment. In the style conversion rule table 206, item number information 58 and reply destination information 66 are stored together with the rule application conditions and edit contents described so far. In this embodiment, the style conversion rules of item numbers 1 to 3 are rules for the reply destination A, and the style conversion rules of item numbers 4 to 6 are rules for the reply destination B. That is, according to the rule for the reply destination A, when the preceding word information 51 is the final particle “yo”, the word information 52 is the punctuation “.”, And the subsequent word information 53 is not present, “.” Of the word information 52 is replaced. The character string information 61 is replaced with “-!” (Item number 1), and the adverb “Yappari” “ri” is deleted as the deleted character string information 64 regardless of the preceding and following words (item number 2). There is a rule that a pictogram is inserted as the inserted character string information 65 immediately after the word “drinking party” (item number 3). In addition, the rule for the reply destination B is that the auxiliary verb “DA” is replaced with “D” in the post-replacement character string information 61 regardless of the preceding and following words (item number 4). Insert “I” in the inserted character string information 65 into “I'm sorry” (item number 5). When the preceding word information 51 is an adjective verb end form “I” and the following word information 53 is a final particle “Y”, the inserted character A rule for inserting “I” in the column information 65 (item number 6) is prepared.

  The style conversion of the reply sentence candidate for the reply destination A will be described with reference to FIG. When the reply destination is A, the reply sentence candidate “I'm free”, when the preceding word information 51 is the final particle “yo”, the word information 52 is the punctuation “.”, And there is no following word information 53, Since the condition of replacing “.” In the word information 52 with “-!” Is applied, it is converted into the text “I'm free!”. Similarly, "I'm not doing anything" is converted to "I'm not doing anything!" “Drinking party is good!” Applies the rule that an emoticon is inserted immediately after the preceding word “Drinking party”, and after “Drinking party”, an emoticon is inserted as insertion string information 65 and “Drinking party is good” ! " “I ’m sorry, it ’s impossible” is converted to “I ’m sorry, it ’s impossible” by applying the rule to delete “ri” in the adverb “Yappari” regardless of the previous or next word.

  The style conversion of the reply sentence candidate for the reply destination B will be described with reference to FIG. When the reply destination is B, there is a rule to replace the auxiliary verb “DA” with “I”, regardless of the previous or next word, for the reply sentence candidates “I'm busy” or “I'm a little busy”. Applied and converted to "I'm free" and "I'm a little busy". For “sorry,” the rule that “please” is inserted into the impression verb “sorry” is applied and converted to “sorry,” regardless of the preceding and following words. For "Drinking party is good!" And "I'm not doing anything", the rule is to insert "I" when the preceding word is the adjective verb end form "I" and the following word is the final particle "Yo" Is applied and converted to "I'm gonna drink!" And "I'm not doing anything." Although all the rules for the reply destination B are not shown in the drawing, the following conversion is performed by applying the style conversion rules for the reply destination B in the same manner. In other words, “I'm looking forward to it!” Is “I ’m looking forward to it!”, “I ’m not going to be able to do it.” Is “I ’m not going to be able to do it.” , “Wait me” is converted to “Wait me for you”, and “That ’s why” is converted to “That ’s why”. In this way, a reply mail can be prepared with a style suitable for the reply partner.

  30 and 31 show a processing flow in the present embodiment.

  In step S <b> 201, the caller information acquisition unit 401 acquires caller information from the received mail and inputs it as reply destination information to the reply destination information management unit 402. The process proceeds to step S202.

  In step S202, when the received mail is input, the mail dividing unit 102 divides the received mail into partial character strings that are about the length of sentences using punctuation marks, pictograms, emoticons, and the like as clues. Let N be the number of partial character strings divided at this time. The process proceeds to step S203.

  In step S203, the mail dividing unit 102 sets a counter value I = 0 and initializes a buffer for temporarily storing reply mail. The process proceeds to step S204.

  In step S204, the mail dividing unit 102 determines whether or not the number N of partial character strings is greater than the counter value I. If the number N of partial character strings is larger than the counter value I, the process proceeds to step S205. On the other hand, if the number N of partial character strings is not greater than the counter value I, the process proceeds to step S214.

  In step S205, the search unit 103 searches the received text / reply text database 101 for a received text similar to the partial character string. The process proceeds to step S206.

  In step S206, the reply sentence candidate extraction unit 104 extracts a reply sentence candidate associated with the retrieved received sentence. The process proceeds to step S207.

  In step S207, the reply sentence candidate style conversion unit 201 confirms the reply destination information when converting the style of the reply sentence candidate, and uses the style conversion rule corresponding to the reply destination of the currently created mail. The reply sentence candidate extracted from the database is converted. The process proceeds to step S208.

  In step S208, the reply sentence candidate display unit 105 presents the converted reply sentence candidate to the user, and when there are a plurality of reply sentence candidates, prompts the user to select one sentence from them. The process proceeds to step S209.

  In step S209, the reply message selection editing unit 202 receives one sentence selected from reply message candidates by the user. The process proceeds to step S210.

  In step S210, the reply sentence selection editing unit 202 applies the edit performed by the user to the selected sentence. The process proceeds to step S211.

  In step S211, the reply message selection editing unit 202 stores the created reply message in the reply mail buffer. The process proceeds to step S212.

  Here, the style conversion rule registration processing in step S212 will be described in detail with reference to FIG.

  In step S <b> 2121, the difference analysis unit 203 determines whether or not the user has edited the registered reply sentence candidate. When the user edits the reply sentence candidate selected for registration, the process proceeds to step S2122. On the other hand, when the user has not edited the reply candidate selected for registration, the process proceeds to step S213.

  In step S2122, the stylistic conversion rule generation unit 204 generates a stylistic conversion rule candidate based on the text difference before and after editing extracted by the difference analysis unit 203. The processing moves to step S2123.

  In step S2123, the style conversion rule generation unit 204 refers to the style conversion rule candidate storage unit 205, and acquires the number of times that the same style conversion rule candidate has been derived in the past for the same reply destination. When the style conversion rule candidate is derived more than a predetermined number of times, the process proceeds to step S2124. On the other hand, when the style conversion rule candidate has not been derived more than a predetermined number of times, the process proceeds to step S2125.

  In step S2124, the style conversion rule generation unit 204 stores the style conversion rule candidate as a style conversion rule in the style conversion rule table 206 together with the reply destination. The process proceeds to step S213.

  In step S2125, if the style conversion rule candidate is already registered in the style conversion rule candidate storage unit 205, the style conversion rule generation unit 204 increases the number of appearances by associating it with the reply destination. If the style conversion rule candidate is not already registered in the style conversion rule candidate storage unit 205, the style conversion rule generation unit 204 newly registers the rule candidate in association with the reply destination and appears. The number of times is 1. The process proceeds to step S112.

  In step S213, the mail dividing unit 102 increments I. The process returns to step S204 to move to the next partial character string process. When the process is completed for all partial character strings in this way, the process proceeds to step S214.

  In step S214, the reply message combining unit 107 outputs the character string stored in the reply mail buffer as a reply mail. The process ends.

<Embodiment 3>
Even for the same user, there is not only one type of wording or pictograms used for e-mails, but there are various styles for similar sentences depending on the time and occasion. Can save you the trouble of editing. Therefore, the style conversion rules are stored together with the number of times the rules are applied, and the reply sentence candidates are converted by applying the one with the highest number of applications preferentially, or the converted reply sentence candidates are applied It is possible to improve the usability of this function by presenting it to the user in the descending order.

  FIG. 32 shows a hardware configuration diagram of the information processing apparatus 100 in the present embodiment. Constituent elements included in the information processing apparatus 100 described in Embodiment 1 are denoted by the same reference numerals in the drawing, and description thereof is omitted here. The information processing apparatus 100 according to the present embodiment newly includes a rule application count management unit 601.

  The rule application frequency management unit 601 manages the application frequency of the style conversion rules stored in the style conversion rule table 206. When the reply sentence candidate style conversion unit 201 converts the reply sentence candidate style, the rules are applied in the order of the most frequently applied rules. In addition, when the user adopts or edits a reply sentence selected from reply sentence candidates, the number of application of the style conversion rule stored in the style conversion rule table is updated. A method for updating the number of times of application will be described later.

  FIG. 33 shows the received text / reply text table 101 in this embodiment. For example, reply sentences such as “I'm free,” “I ’m sorry, I ’m busy for a while” and “I'm not doing anything” are stored for the received sentence “Today's time?”. Similarly, the reply sentences “Let's go to drink tomorrow ~”, “OK, I ’m looking forward to it!” And “I ’m sorry, I ’m sorry,” and the received sentence “Waiting for a reply.” The reply sentences “Thanks!”, “Take me waiting”, and “That ’s why” are stored.

  FIG. 34 shows the style conversion rule table 206 in the present embodiment. The stylistic conversion rule table 206 stores the rule application count information 67 so far, together with the above-described rule application conditions and editing contents. For example, when the preceding word is a final particle “yo”, the word is a punctuation mark “.”, And there is no following word, the number of times of application of the rule that “.” Of the word is replaced with “-!” (Item number 1) Is “10 times”. For example, the number of times of applying the rule of deleting “ri” of the adverb “Yappari” (item number 2) is “8 times” regardless of the preceding and following words. In addition, for example, the number of application of the rule of inserting a pictograph immediately after the preceding word “drinking party” (item number 3) is “6 times”. For example, when the preceding word is the final particle “yo” and the succeeding word is the punctuation mark “.”, The number of application of the rule of inserting the character string “ぉ ˜” (item number 4) is “12 times”. . Also, for example, the number of times of applying the rule that replaces the word “banquet” (item number 5) when the word is the common noun “drinking party” regardless of the preceding and following words is “9 times”.

  With reference to FIG. 35, the style conversion of a reply sentence candidate in the present embodiment will be described. When performing stylistic conversion, the application is applied in descending order of the number of applications. For example, with respect to the reply sentence candidate “free time”, first, when the preceding word is the final particle “yo” and the succeeding word is the punctuation mark “.”, The rule that the character string “ぉ ˜” is inserted (item number 4). Is applied and converted to the sentence "I'm free". When the preceding word is a final particle "yo", the word is a punctuation mark ".", And there is no following word, the rule application condition of replacing the word "." With "-!" Although it applies to a reply sentence candidate, when the preceding word is the final particle “yo” and the succeeding word is the punctuation mark “.”, The rule that the character string “（˜” is inserted (item number 4) is not applied because it has priority. Similarly, the reply sentence candidate "I'm not doing anything" is converted to "I'm not doing anything". Similarly, for a reply sentence candidate “Drinking party is OK!”, When the word is a common noun “Drinking party”, it is replaced with “Banquet” regardless of the preceding and following words (No. 5). ) Is applied, and the banquet is good! The application condition of the rule that a pictogram is inserted immediately after the preceding word “drinking party” (No. 3) also applies to this reply sentence candidate. At the time of “meeting”, the rule of replacing this with “banquet” (item number 5) is given priority. In this example, the reply sentence candidates converted by the priority rules are presented, but all reply sentence candidates converted by the respective rules are prepared and presented to the user in the order of the number of times the rules are applied. It is also possible. In that case, the user can also select a reply sentence candidate converted according to a rule that is presented at a lower level and is not frequently applied.

  36 and 37 show a flow in the present embodiment.

  In step S301, when the received mail is input, the mail dividing unit 102 divides the received mail into partial character strings that are about the length of sentences using punctuation marks, pictograms, emoticons, and the like as clues. Let N be the number of partial character strings divided at this time. The process proceeds to step S302.

  In step S302, the mail dividing unit 102 sets the counter value I = 0 and initializes a buffer for temporarily storing the reply mail. The process proceeds to step S303.

  In step S303, the mail dividing unit 102 determines whether or not the number N of partial character strings is greater than the counter value I. If the number N of partial character strings is larger than the counter value I, the process proceeds to step S304. On the other hand, if the number N of partial character strings is not larger than the counter value I, the process proceeds to step S313.

  In step S304, the search unit 103 searches the received text / reply text database 101 for a received text similar to the partial character string. The process proceeds to step S305.

  In step S305, the reply sentence candidate extraction unit 104 extracts a reply sentence candidate associated with the retrieved received sentence. The process proceeds to step S306.

  In step S306, the reply sentence candidate style conversion unit 201 refers to the application counts so far in the style conversion rules registered in the style conversion rule table 206, and applies them in order from the rule with the highest application count, Convert reply sentence candidates extracted from the database. The process proceeds to step S307.

  In step S307, the reply sentence candidate display unit 105 presents the converted reply sentence candidate to the user, and when there are a plurality of reply sentence candidates, prompts the user to select one sentence. The process proceeds to step S308.

  In step S308, the reply message selection editing unit 202 receives one sentence selected from reply message candidates by the user. The process proceeds to step S309.

  In step S309, the reply sentence selection editing unit 202 applies the edit performed by the user to the selected sentence. The process proceeds to step S310.

  In step S310, the reply message selection / editing unit 202 stores the created reply message in the reply mail buffer. The process proceeds to step S311.

  Here, the style conversion rule registration processing in step S311 will be described in detail with reference to FIG.

  In step S <b> 3111, the difference analysis unit 203 determines whether or not the user has edited the registered reply sentence candidate. When the user edits the registered reply sentence candidate, the process proceeds to step S3112. On the other hand, when the user has not edited the reply candidate selected for registration, the process proceeds to step S3117.

  In step S3117, the style conversion rule generation unit 204 determines whether the reply text is a reply text to which an existing style conversion rule is applied. If the reply sentence is a reply sentence to which an existing style conversion rule is applied, the process proceeds to step S3118. On the other hand, if the reply sentence is not a reply sentence to which an existing style conversion rule is applied, the process proceeds to step S312.

  In step S3118, the style conversion rule generation unit 204 adds one to the number of times the style conversion rule is applied. The process proceeds to step S312.

  In step S3112, the style conversion rule generation unit 204 generates a style conversion rule candidate based on the difference between the texts before and after editing extracted by the difference analysis unit 203. The processing moves to step S3113.

  In step S3113, the style conversion rule generation unit 204 refers to the style conversion rule candidate storage unit 205, and determines whether or not the same style conversion rule as the style conversion rule candidate is already registered in the style conversion rule table. . If the same style conversion rule as the style conversion rule candidate is already registered in the style conversion rule table, the process proceeds to step S3118. On the other hand, when the same style conversion rule as the style conversion rule candidate is not already registered in the style conversion rule table, the process proceeds to step S3114.

  In step S3114, the style conversion rule generation unit 204 refers to the style conversion rule candidate storage unit 205, checks how many times the same style conversion rule candidate has occurred in the past, and the number of times exceeds the predetermined number. It is determined whether or not. If the number of past occurrences of the same stylistic conversion rule candidate is equal to or greater than the predetermined number, the process proceeds to step S3115. On the other hand, if the number of past occurrences of the same style conversion rule candidate is less than the predetermined number, the process proceeds to step S3116.

  In step S3115, the style conversion rule generation unit 204 stores the style conversion rule candidate in the style conversion rule table 206 as a style conversion rule together with the number of times it has been applied so far. The process proceeds to step S312.

  In step S3116, the style conversion rule generation unit 204 stores the style conversion rule candidate as a style conversion rule candidate together with the number of times of application so far. The process proceeds to step S312.

  In step S312, the mail dividing unit 102 increments I. The process returns to step S303 to move to the next partial character string process. When the process is completed for all partial character strings in this way, the process proceeds to step S313.

  In step S313, the reply message combining unit 107 outputs the character string stored in the reply mail buffer as a reply mail. The process ends.

  FIG. 38 is a block diagram schematically illustrating an example of a hardware configuration of the information processing apparatus 100 according to the present embodiment. The information processing apparatus 100 includes a central processing unit (CPU, central processing unit) 702, an operation unit 704, a display unit 706, a ROM 708, a RAM 710, an external storage device 714, and a network interface device 718. The bus 716 connects the CPU 702, the operation unit 704, the display unit 706, the ROM 708, the RAM 710, the external storage device 714, and the network interface device 718 to each other.

  The CPU 702 executes various arithmetic processes. The operation unit 704 receives data input from the user. The display unit 706 displays various information. The ROM 708 stores various programs such as an operating system (OS, operating system). The RAM 710 executes programs and stores data. The external storage device 714 stores a dedicated application and the like. The network interface device 718 transmits / receives data to / from other devices via the network. The dedicated application may be stored in a computer-readable recording medium such as a portable recording medium such as a CD-ROM, a DVD disk, or a USB memory, or a semiconductor memory such as a flash memory.

  When the user issues an instruction to create a document using the operation unit 704, the CPU 702 executes the dedicated application by reading the dedicated application from the external storage device 714 and developing it in the RAM 710. By executing the dedicated application, the CPU 702 can create and edit a document.

  As described above, the inventions of the second embodiment and the third embodiment have been described by adding functions to the invention of the first embodiment. However, the invention of the second embodiment and the invention of the third embodiment are implemented in combination. It is also possible to do.

  The information processing apparatus according to the exemplary embodiment of the present invention has been described above, but the present invention is not limited to the specifically disclosed embodiment, and does not depart from the scope of the claims. Various modifications and changes are possible.

10 Stylistic conversion rule candidates 20, 50 Condition information 21, 51 Preceding word information 22, 52 Word information 23, 53 Subsequent word information 24, 25, 26, 57 Related word information 30, 60 Editing content information 31, 61 Substitution / insertion / Delete information 32, 62 Pre-substitution character string information 33, 63 Post-substitution character string information 34, 64 Deletion character string information 35, 65 Inserted character string information 40 Stylistic conversion rule 70 Received sentence 80 Reply sentence 100 Information processing apparatus 101 Received sentence Reply text database 102 Mail division unit 103 Search unit 104 Reply text candidate extraction unit 105 Reply text candidate display unit 107 Reply text combination part 201 Reply text candidate text conversion part 202 Reply text selection edit part 203 Difference analysis part 204 Stylistic conversion rule generation Section 205 Stylistic conversion rule candidate storage section 206 Stylistic conversion rule table 401 Sender information acquisition section 40 Reply destination information management section 601 rules applicable number of times management unit 702 CPU
704 Operation unit 706 Display unit 708 ROM
710 RAM
714 External storage device 716 Bus 718 Network interface device

Claims (5)

In a control method of an information processing apparatus having a processing unit and a storage unit that stores a received sentence and a plurality of reply sentence candidates associated with the received sentence,
The processing device is
Divide received email into split emails,
A received sentence similar to the divided mail is acquired from the storage unit,
Select a reply sentence candidate corresponding to the acquired received sentence from the storage unit,
When the selected reply sentence candidate is edited, the difference between the reply sentence candidate before editing and the reply sentence candidate after editing is extracted,
Based on the difference, the conversion rule candidate information for converting the selected reply sentence candidate is stored in the storage unit,
When the same conversion rule candidate information is stored a predetermined number of times or more, the same conversion rule candidate information is stored in the storage unit as conversion rule information,
With reference to the conversion rule information, the reply sentence candidate selected from the storage unit is converted.
A method for controlling an information processing apparatus. Obtain the sender information of the received email,
The conversion rule information and the transmission source information of the received mail are associated with each other and stored in the storage unit,
With reference to the sender information of the received mail, obtain the conversion rule information corresponding to the sender information from the storage unit,
With reference to the conversion rule information corresponding to the transmission source information, the reply sentence candidate obtained from the storage unit is converted.
The method for controlling an information processing apparatus according to claim 1. Referring to the conversion rule information, obtain the number of times the reply sentence candidate obtained from the storage unit has been edited,
The acquired number of times and the conversion rule information are associated with each other and stored in the storage unit,
When converting the reply sentence candidate selected from the storage unit, the conversion is performed with reference to the conversion rule information having the largest number of times.
The method for controlling an information processing apparatus according to claim 1. A control program for controlling an information processing apparatus having a storage unit that stores a received sentence and a plurality of reply sentence candidates associated with the received sentence,
In the information processing apparatus,
Divide received email into split emails,
A received sentence similar to the divided mail is acquired from the storage unit,
Select a reply sentence candidate corresponding to the acquired received sentence from the storage unit,
When the selected reply sentence candidate is edited, the difference between the reply sentence candidate before editing and the reply sentence candidate after editing is extracted,
Based on the difference, the conversion rule candidate information for converting the selected reply sentence candidate is stored in the storage unit,
When the same conversion rule candidate information is stored a predetermined number of times or more, the same conversion rule candidate information is stored in the storage unit as conversion rule information,
With reference to the conversion rule information, the reply sentence candidate selected from the storage unit is converted.
A control program for an information processing apparatus. A storage unit that stores a received sentence and a plurality of reply sentence candidates associated with the received sentence;
A splitting unit that splits received mail into split mails;
A search unit that searches the storage unit for a received sentence similar to the divided mail;
An extraction unit that extracts a reply sentence candidate corresponding to the acquired received sentence from the storage unit;
When the selected reply sentence candidate is edited, an analysis unit that analyzes a difference between the reply sentence candidate before editing and the reply sentence candidate after editing;
Based on the difference, conversion rule candidate information for editing the selected reply sentence candidate is stored in the storage unit, and when the same conversion rule candidate information is stored a predetermined number of times, the same conversion rule candidate information is converted. A generating unit that stores the rule information in the storage unit;
A conversion unit that converts the reply sentence candidate selected from the storage unit with reference to the conversion rule information;
An information processing apparatus comprising: