WO2011148659A1 - 特別日の登録のための情報処理方法 - Google Patents
特別日の登録のための情報処理方法 Download PDFInfo
- Publication number
- WO2011148659A1 WO2011148659A1 PCT/JP2011/050846 JP2011050846W WO2011148659A1 WO 2011148659 A1 WO2011148659 A1 WO 2011148659A1 JP 2011050846 W JP2011050846 W JP 2011050846W WO 2011148659 A1 WO2011148659 A1 WO 2011148659A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- date
- character string
- data
- name
- special day
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/10—Text processing
Definitions
- the present invention relates to an information processing method implemented for registering information related to special days that come periodically such as birthdays and wedding anniversaries in a memory of a computer, a program to which the method is applied, and an information processing apparatus.
- Some recent mobile phones have a function of notifying a few days in advance that a special date registered in advance is approaching.
- a special day such as each person's birthday can be registered in the telephone directory in the mobile phone, and a notice is displayed on the display unit several days before the registered special day. Are displayed, and detailed information on special days is displayed in response to the icon being selected.
- Patent Document 2 describes that a bulleted portion is extracted from an e-mail, and further, schedule information is extracted from the bulleted portion for each item such as a title, date / time, and place.
- Patent Document 1 in order to perform advance notification of a special day, it is necessary for the user to register the content and date of the special day in advance. Since this registration work is considerably complicated, there is a demand for automatic registration based on the result of analyzing document data such as an electronic mail. However, in the conventional information processing, as described in Patent Document 2, only the information itself extracted from the document data is registered, and information that takes into consideration the periodicity of special days is not registered.
- the present invention pays attention to various problems including the above-mentioned problems, recognizes the name and date of the special day with high accuracy using the analysis result of the document data in which the topic about the special day that is periodically visited is described, It is an object to register highly reliable information.
- a wording pattern table storing a plurality of wording patterns in which a special day that is periodically visited is expressed using a first character variable representing the name and a second character variable representing the time.
- a special date data table storing a plurality of special names for special dates are registered in advance in a memory of a computer having a function of transmitting and receiving document data.
- wording pattern table it is desirable to store a plurality of types of wording patterns that are commonly used on various special days.
- Many of the general names stored in the special day data table come in one-year cycles (for example, “birthday”, “marriage anniversary”, etc.) or ones that come in one-month cycles (for example, “ Salary date ”,“ Board of Directors ”, etc.), but a special date with a cycle of several months or a special date that cannot uniquely identify the cycle may be stored.
- morphological analysis and syntax analysis are performed on the document data transmitted or received by the computer.
- a first character string is extracted from the document data, which is composed of a general name stored in the special day data table and a phrase that modifies the general name.
- a search and a second search for extracting a character string representing the time are executed.
- the word pattern extracted by the first search is applied to the first character variable
- the character pattern extracted by the second search is applied to the second character variable.
- a range including character strings extracted by the first and second searches in the document data is collated.
- the process proceeds to the fourth step and is based on the concept of the character string applied to the first and second character variables.
- Estimate dates for special days For example, a specific date is derived from a character string assigned to the second character variable, and a period according to the concept of the character string assigned to the first character variable is applied to the date, Date data can be derived, for example, if the cycle is a unit, ⁇ month ⁇ day, if the cycle is monthly, ⁇ day.
- the character string assigned to the first character variable is used as a special name, and the combination of this name and the date estimated in the fourth step is registered in the memory of the computer.
- the length of the character string that qualifies the general name for the special day varies depending on the situation.
- the general name and the general name are extracted from the document data including the general name for the special day. Since a character string composed of words to be modified is extracted and the relationship between the entire character string and the character string indicating the time is compared with each wording pattern, the wording indicating the special day can be extracted with high accuracy. Furthermore, based on the extracted wording, it is possible to register information that appropriately indicates the name and date of the special day.
- a character string indicating a date relative to the current time is extracted by a second search as a character string indicating a time
- transmission of document data to be processed is performed.
- a date corresponding to the second variable is estimated by assigning a relative date represented by the character string to the date or the received date.
- document data representing the time is exchanged by relative expressions such as “Yesterday” and “Tomorrow”, it is reasonable to think that the time is expressed based on the transmission date or the reception date. Therefore, according to the above-described embodiment, it is possible to guarantee the estimated accuracy of the date of the special day even when the time of the special day is written in an ambiguous manner.
- the special day data table includes data obtained by combining the general name of the special day and the general name for replacement. Further, in the fifth step when the character string including the general name combined with the general name for replacement is extracted by the first search, the general name in the character string applied to the first character variable is replaced. The character string after the change is combined with the date estimated in the fourth step as the special day name.
- 2 Appropriate names can be registered for special days after the first. For example, if “wedding” is combined with “marriage anniversary” as a replacement name and this combination is stored in the special day data table, the word “Mr. A's wedding” is read from the document to be processed. In response to the extraction of the column, this character string can be replaced with the character string “Mr. A's wedding anniversary” and registered.
- the wording pattern table stores a wording pattern indicating an expression negating a special day and a wording pattern indicating an expression affirming the special day.
- the third step when a wording that matches the wording pattern indicating the expression denying the special day is found, the fourth step and the fifth step are not performed.
- the character string indicating the name of the special day and the character string indicating the time are included, but when these character strings are used in expressions to deny the special day, these It is possible to prevent incorrect information from being registered by the character string. Therefore, the reliability of registration information can be increased.
- the program to which the above method is applied is incorporated in a computer having communication means for transmitting and receiving document data, and is a storage means in which the wording pattern table and the special day data table are registered; communication Analysis means for executing morphological analysis and syntax analysis on the document data transmitted or received by the means; Search means for executing the first and second searches for the document data based on the analysis result by the analysis means The character string extracted by the first search is applied to the first character variable, and the character pattern extracted by the second search is applied to the second character variable by the wording pattern table.
- processing of word collation means for collating a range including character strings extracted by the first and second searches, processing of word collation means
- the computer functions as each means of registration processing means for registering a combination of this name and the date estimated by the date estimation means in the memory of the computer, using the character string assigned to the character variable of 1 as a special day name.
- the above program is installed in, for example, a computer or personal computer that operates as a control unit of a mobile terminal device.
- the computer in which the program is installed operates as an information processing apparatus including a storage unit, a communication unit, an analysis unit, a search unit, a word matching unit, a date estimation unit, and a registration processing unit.
- a morphological analysis and a syntax analysis are performed on the document data, and a character string that represents the name of the special day based on the analysis result Can be recognized and registered with high accuracy.
- FIG. 1 is a functional block diagram illustrating an application for information processing related to a special day.
- This application 100 is incorporated in a control unit of a mobile phone, and extracts an expression related to a special day from document data of the e-mail in response to an e-mail application (not shown) sending or receiving an e-mail. Then, information on the special day is registered based on the extraction result. Also, it has a function of notifying that a registered special date is approaching, and a function of outputting special day registration information to an external application (scheduler, telephone book, etc.).
- the application 100 includes a document analysis processing unit 11, a special day expression extraction unit 12, a date expression extraction unit 13, a matching processing unit 14, a date estimation processing unit 15, a registration processing unit 16, and an output processing unit 17.
- the processing unit includes data tables such as a special date data table T1, a date data table T2, a wording pattern table T3, a recognition result storage table T4, an output data storage table T5, and an output rule table T6.
- the tables T1, T2, and T3 store data necessary for recognizing expressions related to special dates, and the tables T4 and T5 are derived as recognition results.
- the table T6 stores definitions and setting data required for special day notification and output.
- each table of T1, T2, and T3 will be described with reference to FIGS.
- an identification number (1, 2, 3,...) Is attached to each piece of information for one record.
- the configurations of the tables T1 to T5 shown in FIGS. 2 to 4 and FIGS. 6 and 7 to be described later are specific examples when the language to be processed is Japanese.
- the data structure of each table and the contents of the registered information are changed as appropriate to the grammar of that language.
- FIG. 2 shows a configuration example of the special day data table T1.
- the information for one record in the table T1 includes an extracted name, a registered name, a period, and a priority.
- the concrete numerical value is actually set to the priority of each record, in FIG. 2, description of a numerical value is abbreviate
- “Extract name” is the name of a special day extracted from the document data.
- a plurality of words that are frequently used as special day names such as “birthday”, “salary day”, and “wedding” are registered.
- Each word except “salary date” is given a symbol (* or ⁇ ) indicating a character string that modifies the word.
- ⁇ means any number
- * indicates any character string excluding the number.
- “Sachiko-chan no” in the character string “Birthday of Sachiko-chan”, which will be described later corresponds to the * part of the first extracted name.
- “Foundation” in the character string “10th anniversary” corresponds to the * part of the extraction name of No. 6, and “10” corresponds to the ⁇ part of the extraction name.
- Some special days represented by the above extracted names may be better to change the name of the next corresponding day.
- the registered name indicates the name after the change, and is set only when the name needs to be changed. Also for this registered name, the character string of the qualifying part is set by the same symbols * and ⁇ as the extracted name. The character strings of * and ⁇ are maintained as they are in the extraction name.
- “* wedding anniversary” is set as the registered name for the extracted name “* wedding”.
- “* ( ⁇ + n) anniversary” (n is the number of years elapsed since the year when the special day was registered) is set as the registered name for the extracted name “** anniversary”.
- “* date of death” is set as the registered name for the extracted name “* funeral”. In general, the date of death does not coincide with the date of funeral, but if you do this, you can automatically register a date close to the actual date of death, so you only have to correct the registered date to the correct one, The labor is greatly reduced compared to manually entering all information.
- a cycle of one year unit or one month unit is set according to each concept. In some cases, both the year and month cycles are set, as in the fifth record in the figure.
- FIG. 3 shows a configuration example of the date data table T2.
- the information for one record in the table T2 includes a date expression and a conversion rule. Dates are expressed specifically for the month or day (numbered 1 to 3 in the figure), expressed in words that represent the date relative to the current time (numbered 4 to 9 in the figure), or a combination of multiple words And date (Nos. 10 and 11 in the figure). Among these, in the date expression that specifically represents the month and day, the numerical value portion is replaced with ⁇ as in the special day data table T1 in FIG. 2 ( ⁇ is an arbitrary numerical value).
- the conversion rule is for deriving a specific date from the date expression, and is used by the date estimation processing unit 15 in FIG.
- Many conversion rules use the date of mail transmission or reception (unified into a variable called [transmission / reception date]).
- a conversion rule For example, for the date expressions Nos. 4-7 in the figure (“Tomorrow”, “Corner”, “Yesterday”, “Ototoi”), an expression that adds an adjustment value based on the concept of the expression to [Send / Receive Date] is a conversion rule.
- the 8th and 9th date expressions (“next week” and "last week") in the figure are based on the respective concepts to derive the date for one week before or after the week to which [Send / Receive Date] belongs.
- the conversion rule is set.
- FIG. 4A shows a configuration example of the wording pattern table T3.
- the information for one record in the table T3 includes a wording pattern, an OK or NG flag, and a priority.
- the wording pattern represents a standard wording related to a special day with a character string representing the name of the special day as a variable [anniversary] and a character string representing the time of the special day as a variable [date].
- the OK flag indicates that the corresponding wording pattern is an expression that affirms the special day
- the NG flag indicates that the corresponding wording pattern is an expression that denies the special day.
- a higher priority than the word pattern for which the OK flag is set is set for the word pattern for which the NG flag is set.
- FIG. 4A only two types of negative wording patterns are shown, but more negative wording patterns can be set.
- FIG. 4 (2) shows each wording pattern in association with specific wording examples to which these patterns are applied.
- the character string “Birthday of Sachiko” is assigned to the variable [anniversary]
- the character string “Tomorrow” is assigned to the variable [date].
- the document analysis processing unit 11 separates and recognizes each word in the text by performing morphological analysis on the document data of the outgoing mail or the incoming mail. Further, the document analysis processing unit 11 creates the tree structure data representing the relationship between the words by performing the syntax analysis based on the result of the morphological analysis.
- the tree structure data is used by the special day expression extraction unit 12, the date expression extraction unit 13, and the matching processing unit 14.
- the special day expression extraction unit 12 pays attention to each extraction name in the special day data table T1 shown in FIG. 2 in descending order of priority, and based on the analysis result by the document analysis processing unit 11, Perform a search by the extracted name under consideration. In this search, a character string composed of a general name in the extracted name under consideration and a phrase (part corresponding to * or ⁇ ) that modifies the general name is extracted. In addition, when a modification part is not found, only a general name is extracted. When a character string that matches one of the extraction names is found, the special day expression extraction unit 12 sets the character string to [anniversary].
- the extracted name and registered name of the special day data table T1 are modified by the * symbol and the ⁇ symbol.
- the present invention is not limited to this, and a simple general name may be set as the extracted name or registered name.
- the special day expression extraction unit 12 needs to extract a character string including a general name and a word that modifies the general name, and set the entire extracted character string as a character variable [anniversary].
- the date expression extraction unit 13 pays attention to each date expression in the date data table T2 shown in FIG. 3 in order, and executes a search based on the date expression focused on the document data.
- the date expression extraction unit 13 sets the character string in the variable [date].
- priority is set for each record in the date data table T2 based on the same rules as those for the special date data table T1, and when a character string suitable for a plurality of date expressions is found by the above search. Selects the one with the highest priority among them. For example, for the character string “March 10”, the date representations of No. 1, No. 2, and No. 3 in the date data table T2 are suitable, but the date representation of No. 3 is hit. It is determined that
- the matching processing unit 14 starts the process shown in FIG. 5 in response to the setting of a specific character string in [anniversary] and [date].
- the matching processing unit 14 first reads the word pattern having the highest priority from the word pattern table T3 (step S1). Then, the tree structure data created by the document analysis processing unit 11 is collated with this wording pattern, and it is determined whether there is any wording that matches the pattern read in the range including the character string corresponding to each variable. (Steps S2 and S3). Thereafter, the wording patterns are read in descending order of priority until a matching wording is found (steps S3 to S5), and matching processing using the read patterns (step S2) is executed.
- step S3 when a wording that matches any wording pattern is found (step S3 is “YES”), the flag set in the wording pattern is checked. If the OK flag is set (step S6 is “YES”), the character strings set in [anniversary] and [date] are determined as extraction results (step S7).
- step S6 when the NG flag is set in the wording pattern in which the match is recognized (step S6 is “NO”), or when no match is found in any wording pattern (step S4 is “YES”). ”), Each character variable is cleared (step S8), and the process is terminated.
- this clearing process is performed, the date estimation processing unit 15 and the registration processing unit 16 do not operate, and the processing for the document data is finished.
- a wording pattern using either one of two types of character variables [anniversary] and [date] is used as an expression for negating the special day.
- a wording pattern using both character variables is set, and the process shown in FIG. 5 is executed.
- a higher priority is set for the negative wording pattern than for the positive wording pattern. This makes it possible to prevent erroneous information from being created and registered from expressions that deny special days.
- “Tomorrow” and “Sachiko's birthday” extracted from this expression are cleared in step S8. Therefore, erroneous information due to these will not be registered.
- the latter expression does not match the word pattern with the NG flag, “Sachiko's birthday” and “Tomorrow” can be determined as the extraction result when it matches the sixth word pattern.
- the date estimation processing unit 15 collates the date data table T2 with the character string assigned to [date], reads the conversion rule corresponding to the character string, and indicates a specific date based on the conversion rule Deriving data.
- the derived date data is combined with [anniversary] instead of the [date] character string.
- a calendar information database (not shown) is also used.
- the registration processing unit 16 searches each extraction name in the special day data table T1 using [anniversary]. When a record corresponding to [anniversary] is found by this search, the registration processing unit 16 changes the date data into a format suitable for the period included in the extracted record. For example, if the cycle is “every year”, the date data is set to “month” and “date” is set to “date” if the cycle is “monthly”.
- the registration processing unit 16 creates recognition result data based on a combination of [annniversary] and the period extracted by the search and date data. If it is found by search that the registered name corresponds to [anniversary], [anniversary] in the recognition result data is replaced with the registered name. Then, the contents of the recognition result storage table T4 and the output data storage table T5 are updated with the confirmed recognition result.
- the output processing unit 17 executes the special day notification process described above and output to other applications.
- the output rule table T6 stores setting information necessary for this notification and output. For example, regarding notification, for each type of special day name, notification time, message template used for notification, image information to be displayed together with the message, and the like are registered. For these pieces of information, in addition to default information, information input by the user can also be registered.
- FIG. 6 shows a configuration example of the recognition result storage table T4
- FIG. 7 shows a configuration example of the output data storage table T5.
- Each table T4, T5 stores the recognized special day name (by variable [anniversary]) and the period.
- the date data obtained by the processing of the date estimation processing unit 15 is stored as it is in the recognition result storage table T4 of FIG. 6, whereas the output data storage table T5 of FIG. A specific day is stored as the estimated date.
- the recognition result storage table T4 is set so that a plurality of date data can be stored in each record in preparation for the possibility of processing a plurality of document data for the same special date.
- each date data is set with a numerical value indicating the priority, and the estimated date of the output table T5 is determined based on the priority of these date data. It has become.
- two date data “July 14” and “July 16” are registered for “Ichiro Sato's wedding anniversary”. "July 14" is adopted in the output table T5. This is because “July 14” has a higher priority than “July 16” in the recognition result storage table T4.
- the date data registration process will be described in detail later.
- FIG. 8 shows a specific example of information processing for a mail document transmitted by a user and an output example of information after registration in a mobile phone in which the application 100 is installed.
- a character string “Sachiko-chan's birthday” indicating the contents of the special day and a character “Yesterday” indicating the time of the special day. Columns are included. “Sachiko's birthday” is extracted as a match with the first “* birthday” in the special day data table T1 in FIG. 2, and “Yesterday” is the sixth data in the date table T2 in FIG. Is extracted as a match.
- the search process for the special day data table T1 the period corresponding to “Sachiko's birthday” is read and the date data is changed.
- the data having the contents shown in (d) is determined as the recognition result and registered in the recognition result storage table T4.
- FIG. 8 shows an example of a notification screen displayed on the day before the registration date of the next year after the above registration.
- a message indicating the content of the registered special day (Sachiko's birthday) and its time (tomorrow) is displayed along with the character image. Yes.
- FIG. 9 shows an example in which a replacement process using a registered name is performed in the special day recognition process.
- the transmitted mail in the example of (a) includes the character string “This Sunday is my sister's wedding”. Therefore, “sister's wedding” and “this Sunday” are extracted as corresponding to the variables [anniversary] and [date], respectively, and, as shown in FIG. It turns out that the sixth wording pattern is included in the document data. Further, based on the conversion rule corresponding to “Kondo Sunday” and the transmission date of the mail document, the extracted data having the contents as shown in (c) is determined.
- the special day data table T1 is searched by “sister's wedding” and the corresponding record is read out.
- the read record includes the registered name (* wedding anniversary). It is. Therefore, based on this registered name, “My sister ’s wedding” is replaced with “My sister ’s wedding anniversary”, and periodic data is added or date data is changed to recognize the contents as shown in (d). The result is confirmed and registered in the recognition result storage table T4.
- FIG. 10 shows an example in which the date range is narrowed down based on the recognition results for a plurality of mail documents related to the same special date, using a case where the special date is recognized with a certain range.
- the mail A includes the expression “next week's payday” and the mail B includes the expression “last week's payday”.
- the expression of the mail A matches the fourth wording pattern of the wording pattern table T3
- the expression of the mail B matches the third wording pattern.
- each date from Monday to Sunday of the next week of the mail document transmission date (February 18, 2010) is derived based on the conversion rule corresponding to “next week” in the date data table T2. (Fig. 10 (1-2)). Furthermore, by applying the frequency data corresponding to “Salary Day” to this date range, the data “22nd to 28th of every month” is confirmed as a confirmation result and registered in the recognition result storage table T4 ( FIG. 10 (1-3)).
- each date from Monday to Sunday of the previous week of the mail document transmission date (April 28, 2010) is derived based on the conversion rule corresponding to “last week” in the date data table T2. (Fig. 10 (2-2)). Furthermore, by applying the frequency data corresponding to the “payday” to this date range, the data “19th to 25th of every month” is confirmed as the confirmation result (FIG. 10 (2-3)).
- the application 100 of this embodiment extracts a character string representing the contents of the special day and a character string representing the time from the body data of the mail document.
- the expression of the range including the character string is collated with each wording pattern, and each character string is determined as representing a special day only when it matches the wording pattern for which the OK flag is set.
- the entire character string consisting of the name indicating the special date and the word that modifies it is applied to the variable [anniversary], so the correct matching process is performed without being affected by the length of the modifier be able to.
- the application 100 has a function for performing registration by performing the change when a special day name extracted from document data is necessary, and a function for narrowing and correcting date data. Therefore, it becomes possible to register more appropriate information about the special day.
- Step S11 to S23 in FIG. 11 show the procedure of registration processing for the recognition result storage table T4.
- [anniversary] determined as the extraction result and date data derived by the date estimation processing unit 15 are acquired (step S11).
- the record corresponding to [anniversary] is extracted by searching the extraction name of the special day data table T1 by [anniversary] (step S12).
- [anniversary] is changed to the registration name (step S14).
- the date data is changed to data in a format corresponding to the period in the extracted record (step S15).
- the recognition data is determined by the processing up to step S15.
- step S16 the special date name in the recognition result storage table T4 is searched by [anniversary].
- [anniversary] is set as a new registered date name, and a new date that includes this name, cycle, and date data is included.
- the record is registered in the recognition result storage table T4 (step S18). In this new record, a fixed numerical value K is set as the priority of date data.
- step S17 if a special date name matching [anniversary] is found (if step S17 is “YES”), the date data in the registered data of that name is replaced with new date data (set in step S15). ) (Step S19).
- the registration data includes a plurality of date data, the above collation is performed for each date data.
- step S20 If, by this collation, a date that overlaps between date data is not recognized (step S20 is “NO”), new date data is added to the registered data. At this time, the above-mentioned priority K is set for the new date data (step S21).
- step S20 when duplication is recognized between date data (step S20 is "YES"), the process which integrates these date data (step S22), and the priority of the integrated date data are updated.
- step S23 is executed. For example, as shown in the example of FIG. 10, when date data in which a certain range of length is set is integrated, in step S22, an overlapping portion between the two is extracted and the registration data is rewritten thereby. In step S23, a value obtained by adding a predetermined numerical value to the priority set in the date data before rewriting is set as the priority of the date data after rewriting. If the date data to be integrated is the same, the registration data is maintained as it is in step S22, and priority addition processing is executed in step S23.
- an estimated date to be registered in the output data storage table T5 is determined based on the registered data (step S24).
- the date data represents a specific one day as in the second “Hanako Tanaka's birthday”, that date is used as the estimated date as it is. .
- date data with a range setting is set as in the first “Taro Yamada's birthday”
- the date corresponding to the middle day of the range is set as the estimated date.
- the recognition result storage table T4 as in the third “pay day” and the fourth “Ichiro Sato ’s wedding anniversary”
- the higher priority is given.
- the estimated date is determined by applying the above rule to the date data.
- step S25 registration processing to the output data storage table T5 is performed thereafter (step S25).
- the detailed procedure is not shown for this process, if new data is registered in the recognition result storage table T4 in step S18, the new registration process is executed in step S25 as well.
- the output data storage table T5 is searched by [anniversary], the corresponding registration data is extracted, and the estimated date is obtained. Rewrite to the one determined in step S24.
- the date expression is ambiguous, and there is a possibility that the correct date cannot be specified.
- the date data is set to be narrowed down or corrected while processing a plurality of document data related to the same special date. Can be increased.
- step S24 the estimated date of the corresponding information in the output data storage table T5 is changed to information ( ⁇ day) that does not specify the month.
- the date estimation processing unit 15 it is desirable to consider the character string set to [anniversary]. For example, when “Salary Day” is set to [anniversary], Saturdays, Sundays, and holidays are not included in the date data, or a string containing “Funeral” is set to [anniversary] In this case, it is preferable that the date corresponding to Tomoiki is not included in the date data.
- the date estimation processing unit 15 takes into account the possibility of estimating the date from an expression including two or more data stored in the date data table T2 (for example, “5th of next month”). It is desirable to set an algorithm that can be derived.
- the character string set in [anniversary] is used as it is, but if necessary,
- the character string of the decoration part may be changed. For example, when a character string having a first-person expression such as “my birthday” as a qualifier is extracted from mail document data, “I” is replaced with the name of the mail sender and registered.
- the nickname is included in the qualified part, the name of the person corresponding to the nickname may be extracted by searching the phone book file, and the qualified part of [anniversary] may be replaced by the name.
- the application 100 has been described as extracting and registering expressions related to special dates using e-mail document data as a processing target.
- the processing target of the application 100 is limited to e-mail. It is not something.
- the same processing can be performed on document data transmitted for posting to a blog or Twitter.
- FIG. 12 shows an example in which information related to special dates is extracted from document data of posted articles and registered, taking a blog about childcare as an example.
- FIG. 12A shows a part of the document data of the posted article.
- the date estimation processing unit 15 acquires the date of the blog posting date (April 22, 2010), and applies the conversion rule corresponding to the character string of [date] to this date. Thus, a specific date is derived. As a result, data having contents as shown in FIG. Furthermore, the recognition result as shown in (C) is obtained by reading and applying the period corresponding to [anniversary] from the special day data table T1.
- (D) of FIG. 12 shows an example of data registered in the output data storage table T5.
- information related to the special date extracted by the process for the article posted before is registered.
- the application 100 described above it is possible to accurately extract and register information representing a special date from transmitted or received document data. Therefore, in addition to notifying the day when the special day is approaching, the registration information can be reflected in the schedule information and the data of the telephone directory. It is also possible to obtain the specific date of the special date from the document data including only the special date name using the registration information of the recognition result storage table T4 and the output data storage table T5.
- anniversary names of various names freely expressed by the user can be accumulated without requiring any special registration work. It can be saved as a memory.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Document Processing Apparatus (AREA)
Abstract
Description
『昨日』『明日』などの相対的な表現により時期を表した文書データがやりとりされる場合には、その送信日または受信日を基準に時期を表現していると考えるのが妥当である。したがって、上記の実施形態によれば、特別日の時期が曖昧に表記されている場合でも、特別日の日付の推定確度を保障することができる。
たとえば、「結婚式」に対し、置き換え用の名称として「結婚記念日」を組み合わせてこの組み合わせを特別日データテーブルに格納しておけば、処理対象の文書から『Aさんの結婚式』という文字列が抽出されたことに応じて、この文字列を『Aさんの結婚記念日』という文字列に置き換えて登録することができる。
このアプリケーション100は、携帯電話の制御部に組み込まれるもので、図示しない電子メール用のアプリケーションが電子メールを送信または受信したことに応じて、当該電子メールの文書データから特別日に関する表現を抽出し、その抽出結果に基づき特別日に関する情報を登録する。また、登録されている特別日が近づいてきたことを報知する機能や、特別日の登録情報を外部のアプリケーション(スケジューラ、電話帳など)に出力する機能も具備する。
なお、各レコードの優先度には、実際には具体的な数値が設定されるが、図2では、各特別日を優先度が高い順に並べることにより、数値の記載を省略している。図4(1)に示す言い回しパターンテーブルT3についても同様とする。
7番のレコードでは、抽出名の「*葬儀」に対して、登録名として「*命日」が設定される。一般に、命日は葬儀の日とは一致しないが、このようにすれば、実際の命日に近い日付を自動的に登録することができるので、登録された日付を正しいものに修正するだけでよく、全ての情報を手入力する場合より労力が大幅に軽減される。
このテーブルT2の1レコード分の情報には、日付表現と変換ルールとが含まれる。
日付表現には、月や日を具体的に表すもの(図中の1~3番)、現時点に対する相対的な日にちを表す単語によるもの(図中の4~9番)、複数の単語を組み合わせて日付を表現したもの(図中の10,11番)などがある。これらのうち月や日を具体的に表す日付表現では、図2の特別日データテーブルT1と同様に、数値の部分が○に置き換えられている(○には任意の数値が入る。)。
このテーブルT3の1レコード分の情報には、言い回しパターン、OKまたはNGのフラグ、優先度が含まれる。
なお、図4(1)の例では、否定的な言い回しパターンを2種類しか示していないが、より多くの否定的な言い回しパターンを設定することもできる。
文書解析処理部11は、送信メールまたは受信メールの文書データに対して形態素解析を実施することにより、本文中の各単語を切り分けて認識する。さらに、文書解析処理部11は、形態素解析の結果に基づく構文解析を実施することにより、各単語の関係を表す木構造データを作成する。この木構造データは、特別日表現抽出部12,日付表現抽出部13、およびマッチング処理部14により使用される。
いずれかの抽出名に適合する文字列が見つかると、特別日表現抽出部12は、その文字列を[anniversary]にセットする。
なお、日付データテーブルT2の各レコードにも、特別日データテーブルT1と同様のルールに基づき優先度が設定されており、上記の検索により、複数の日付表現に適合する文字列が見つかった場合には、その中で最も優先度の高いものを選択する。たとえば、「3月10日」という文字列に対しては、日付データテーブルT2内の1番、2番、3番の各日付表現が適合するが、これらのうちの3番の日付表現にヒットしたものと判定する。
なお、この日付データの導出処理では、図示しないカレンダー情報のデータベースも利用される。
出力用ルールテーブルT6には、この報知や出力に必要な設定情報が格納される。たとえば、報知に関しては、特別日名称の種類毎に、報知の時期、報知に用いるメッセージのテンプレート、メッセージと共に表示する画像情報などが登録される。これらの情報については、デフォルトの情報のほか、ユーザにより入力された情報を登録することもできる。
各テーブルT4,T5には、それぞれ認識された特別日の名称(変数[anniversary]によるもの)および周期が格納される。一方、日付に関しては、図6の認識結果保存テーブルT4には、日付推定処理部15の処理により求められた日付データがそのまま格納されるのに対し、図7の出力用データ保存テーブルT5には、特定の1日が推定日付として格納される。
この実施例のメール文書には、図中の(a)に示すように、特別日の内容を示す『幸子ちゃんの誕生日』という文字列と、その特別日の時期を示す『昨日』という文字列とが含まれている。『幸子ちゃんの誕生日』は、図2の特別日データテーブルT1の1番目の『*誕生日』に合致するものとして抽出され、『昨日』は、図3の日付テーブルT2の6番目のデータに合致するものとして抽出される。
また、変数[anniversary]や[date]にあてはまる文字列が抽出された場合でも、これらの文字列による表現がNGフラグが設定された言い回しパターンに適合する場合や、いずれの言い回しパターンにも適合しない場合には、特別日に関する認識データは導出されないので、適切でない情報が登録されるのを防止することができる。
図11中のステップS11~S23は、認識結果保存テーブルT4に対する登録処理の手順を示すものである。この処理では、まず、抽出結果として確定した[anniversary]および日付推定処理部15により導出された日付データを取得する(ステップS11)。
たとえば、図10の例のように、ある長さの範囲が設定されている日付データを統合する場合には、ステップS22では、両者間の重複部分を抽出して、これにより登録データを書き換える。またステップS23では、書き換え前の日付データに設定されていた優先度に所定の数値を加算した値を、書き換え後の日付データの優先度として設定する。また、統合対象の日付データが同一である場合には、ステップS22では登録データをそのまま維持し、ステップS23において、優先度の加算処理を実行する。
また、日付推定処理部15には、日付データテーブルT2に格納される2以上のデータを含む表現(たとえば『来月の5日』)から日付を推定する可能性を考慮して、正しい日付を導出できるようなアルゴリズムを設定するのが望ましい。
図12の(A)は、投稿記事の文書データの一部を示す。この文書データに対し、文書解析処理部11、特別日表現抽出部12、日付表現抽出部13の処理を実行することにより、[anniversary]に「ハイハイができた記念日」が、[date]に「昨日」が、それぞれ設定される。さらに、これらの文字列を用いてマッチング処理部14による照合処理を行うことにより、文書データ中に図4の6番目の言い回しパターンに一致する表現が含まれることが判明し、[anniversary]および[date]に設定された文字列が確定される。
T1 特別日データテーブル
T2 日付データテーブル
T3 言い回しパターンテーブル
T4 認識結果保存テーブル
T5 出力用データ保存テーブル
T6 出力用ルールテーブル
11 文書解析処理部
12 特別日表現抽出部
13 日付表現抽出部
14 マッチング処理部
15 日付推定処理部
16 登録処理部
17 出力処理部
Claims (6)
- 周期的に訪れる特別日を、その名称を表す第1の文字変数と時期を表す第2の文字変数とを用いて表現した言い回しパターンが複数格納された言い回しパターンテーブルと、前記特別日の一般名称が複数格納された特別日データテーブルとを、文書データを送信および受信する機能を具備するコンピュータのメモリに登録し、
各テーブルが登録されたコンピュータにおいて、
当該コンピュータが送信または受信した文書データに対し、形態素解析および構文解析を実行する第1ステップ、
前記第1ステップによる解析結果に基づき、前記文書データに対し、前記特別日データテーブルに格納されている一般名称および当該一般名称を修飾する語句から成る文字列を抽出する第1の検索と、時期を表す文字列を抽出する第2の検索とを実行する第2ステップ、
前記第1の検索により抽出された文字列を前記第1の文字変数にあてはめると共に、第2の検索により抽出された文字列を前記第2の文字変数にあてはめた前記言い回しパターンテーブルにより、文書データ中の前記第1および第2の検索により抽出された文字列を含む範囲を照合する第3ステップ、
前記第3ステップの照合処理により、文書データからいずれかの言い回しパターンに適合する言い回しが抽出されたとき、前記第1および第2の文字変数にあてはめられた文字列の概念に基づいて特別日の日付を推定する第4ステップ、
前記第1の文字変数にあてはめられた文字列を特別日の名称として、この名称と第4ステップで推定された日付との組み合わせを前記コンピュータのメモリに登録する第5ステップ、
の各ステップを実行することを特徴とする、特別日の登録のための情報処理方法。 - 現在に対する相対的な日にちを示す文字列が時期を表す文字列として前記第2の検索により抽出された場合の第3ステップでは、処理対象の文書データの送信日または受信日に対して当該文字列が表す相対的な日にちをあてはめることにより、前記第2の変数に対応する日付を推定する、請求項1に記載された特別日の登録のための情報処理方法。
- 前記特別日データテーブルには、前記特別日の一般名称と置き換え用の一般名称とを組み合わせたデータが含まれており、
前記置き換え用の一般名称に組み合わせられた一般名称を含む文字列が前記第1の検索により抽出された場合の前記第5ステップでは、第1の文字変数にあてはめられた文字列中の一般名称を前記置き換え用の一般名称に変更し、この変更後の文字列を前記特別日の名称として前記第4ステップで推定された日付に組み合わせる、請求項1に記載された特別日の登録のための情報処理方法。 - 前記言い回しパターンテーブルには、特別日を否定する表現を示す言い回しパターンと、特別日を肯定する表現を示す言い回しパターンとが格納されており、
前記第3ステップにおいて、前記特別日を否定する表現を示す言い回しパターンに適合する言い回しが見つかったとき、第4ステップおよび第5ステップを実施しないようにした、請求項1に記載された特別日の登録のための情報処理方法。 - 文書データを送信および受信するための通信手段を具備するコンピュータに導入されるプログラムであって、
周期的に訪れる特別日を、その名称を表す第1の文字変数と時期を表す第2の文字変数とを用いて表現した言い回しパターンが複数格納された言い回しパターンテーブルと、前記特別日の一般名称が複数格納された特別日データテーブルとが登録された記憶手段、
前記通信手段が送信または受信した文書データに対し、形態素解析および構文解析を実行する解析手段、
前記解析手段による解析結果に基づき、前記文書データに対し、前記特別日データテーブルに格納されている一般名称および当該一般名称を修飾する語句から成る文字列を抽出する第1の検索と、時期を表す文字列を抽出する第2の検索とを実行する検索手段、
前記第1の検索により抽出された文字列を前記第1の文字変数にあてはめると共に、第2の検索により抽出された文字列を前記第2の文字変数にあてはめた前記言い回しパターンテーブルにより、文書データ中の前記第1および第2の検索により抽出された文字列を含む範囲を照合する言い回し照合手段、
前記言い回し照合手段の処理により文書データからいずれかの言い回しパターンに適合する言い回しが抽出されたとき、前記第1および第2の文字変数にあてはめられた文字列の概念に基づいて特別日の日付を推定する日付推定手段、
前記第1の文字変数にあてはめられた文字列を特別日の名称として、この名称と日付推定手段により推定された日付との組み合わせを前記コンピュータのメモリに登録する登録処理手段、
の各手段として前記コンピュータを機能させる、特別日の登録のための情報処理用のプログラム。 - 周期的に訪れる特別日を、その名称を表す第1の文字変数と時期を表す第2の文字変数とを用いて表現した言い回しパターンが複数格納された言い回しパターンテーブルと、前記特別日の一般名称が複数格納された特別日データテーブルとが登録された記憶手段、
文書データの送信および受信を行うための通信手段、
前記通信手段が送信または受信した文書データに対し、形態素解析および構文解析を実行する解析手段、
前記解析手段による解析結果に基づき、前記文書データに対し、前記特別日データテーブルに格納されている一般名称および当該一般名称を修飾する語句から成る文字列を抽出する第1の検索と、時期を表す文字列を抽出する第2の検索とを実行する検索手段、
前記第1の検索により抽出された文字列を前記第1の文字変数にあてはめると共に、第2の検索により抽出された文字列を前記第2の文字変数にあてはめた前記言い回しパターンテーブルにより、文書データ中の前記第1および第2の検索により抽出された文字列を含む範囲を照合する言い回し照合手段、
前記言い回し照合手段の処理により文書データからいずれかの言い回しパターンに適合する言い回しが抽出されたとき、前記第1および第2の文字変数にあてはめられた文字列の概念に基づいて特別日の日付を推定する日付推定手段、
前記第1の文字変数にあてはめられた文字列を特別日の名称として、この名称と日付推定手段により推定された日付との組み合わせを自装置のメモリに登録する登録処理手段、
の各手段を具備する情報処理装置。
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US13/578,288 US20130054644A1 (en) | 2010-05-28 | 2011-01-19 | Information processing method and program for registering special day and information processing apparatus |
JP2012517156A JP5482894B2 (ja) | 2010-05-28 | 2011-01-19 | 特別日の登録のための情報処理方法 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2010122748 | 2010-05-28 | ||
JP2010-122748 | 2010-05-28 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2011148659A1 true WO2011148659A1 (ja) | 2011-12-01 |
Family
ID=45003654
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2011/050846 WO2011148659A1 (ja) | 2010-05-28 | 2011-01-19 | 特別日の登録のための情報処理方法 |
Country Status (3)
Country | Link |
---|---|
US (1) | US20130054644A1 (ja) |
JP (1) | JP5482894B2 (ja) |
WO (1) | WO2011148659A1 (ja) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2016194828A (ja) * | 2015-03-31 | 2016-11-17 | 大日本印刷株式会社 | サーバ装置、プログラム及び商品情報提供方法、並びに通信システム |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140365879A1 (en) * | 2013-06-07 | 2014-12-11 | Microsoft Corporation | Using aliases for date entry |
US9535904B2 (en) * | 2014-03-26 | 2017-01-03 | Microsoft Technology Licensing, Llc | Temporal translation grammar for language translation |
JP6555178B2 (ja) * | 2016-04-14 | 2019-08-07 | 株式会社島津製作所 | 情報処理装置および電子カルテ表示装置 |
JP7293693B2 (ja) * | 2019-02-05 | 2023-06-20 | 富士フイルムビジネスイノベーション株式会社 | 情報処理装置及びプログラム |
JP2020144646A (ja) * | 2019-03-07 | 2020-09-10 | 富士ゼロックス株式会社 | 情報処理装置及びプログラム |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH05120345A (ja) * | 1991-05-31 | 1993-05-18 | Teremateiiku Kokusai Kenkyusho:Kk | キーワード抽出装置 |
JPH07249007A (ja) * | 1994-03-11 | 1995-09-26 | Matsushita Electric Ind Co Ltd | 携帯情報端末 |
JP2001101100A (ja) * | 1999-09-30 | 2001-04-13 | Oki Electric Ind Co Ltd | 個人情報管理装置 |
JP2005346416A (ja) * | 2004-06-03 | 2005-12-15 | Matsushita Electric Ind Co Ltd | 日時情報変換装置、日時情報変換方法、日時情報変換プログラムおよび日時情報変換装置の集積回路 |
JP2009259144A (ja) * | 2008-04-21 | 2009-11-05 | Kyocera Corp | 情報処理装置およびスケジュール管理方法 |
Family Cites Families (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO1998030963A1 (en) * | 1997-01-14 | 1998-07-16 | Benjamin Slotznick | System for calculating occasion dates and converting between different calendar systems, and intelligent agent for using same |
US7426714B1 (en) * | 2002-07-02 | 2008-09-16 | Principal Decision Systems International | Methods and apparatuses to determine dynamic dates |
US20050149858A1 (en) * | 2003-12-29 | 2005-07-07 | Stern Mia K. | System and method for managing documents with expression of dates and/or times |
US20070226204A1 (en) * | 2004-12-23 | 2007-09-27 | David Feldman | Content-based user interface for document management |
US7730013B2 (en) * | 2005-10-25 | 2010-06-01 | International Business Machines Corporation | System and method for searching dates efficiently in a collection of web documents |
US8060567B2 (en) * | 2006-04-12 | 2011-11-15 | Google Inc. | Method, system, graphical user interface, and data structure for creating electronic calendar entries from email messages |
-
2011
- 2011-01-19 WO PCT/JP2011/050846 patent/WO2011148659A1/ja active Application Filing
- 2011-01-19 JP JP2012517156A patent/JP5482894B2/ja active Active
- 2011-01-19 US US13/578,288 patent/US20130054644A1/en not_active Abandoned
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH05120345A (ja) * | 1991-05-31 | 1993-05-18 | Teremateiiku Kokusai Kenkyusho:Kk | キーワード抽出装置 |
JPH07249007A (ja) * | 1994-03-11 | 1995-09-26 | Matsushita Electric Ind Co Ltd | 携帯情報端末 |
JP2001101100A (ja) * | 1999-09-30 | 2001-04-13 | Oki Electric Ind Co Ltd | 個人情報管理装置 |
JP2005346416A (ja) * | 2004-06-03 | 2005-12-15 | Matsushita Electric Ind Co Ltd | 日時情報変換装置、日時情報変換方法、日時情報変換プログラムおよび日時情報変換装置の集積回路 |
JP2009259144A (ja) * | 2008-04-21 | 2009-11-05 | Kyocera Corp | 情報処理装置およびスケジュール管理方法 |
Non-Patent Citations (1)
Title |
---|
SHIN'ICHIRO TAKAGI ET AL.: "Mail secretary services realized for electronic mails", NTT GIJUTSU JOURNAL, vol. 10, no. 8, 1 August 1998 (1998-08-01), pages 75 - 79 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2016194828A (ja) * | 2015-03-31 | 2016-11-17 | 大日本印刷株式会社 | サーバ装置、プログラム及び商品情報提供方法、並びに通信システム |
Also Published As
Publication number | Publication date |
---|---|
JP5482894B2 (ja) | 2014-05-07 |
JPWO2011148659A1 (ja) | 2013-07-25 |
US20130054644A1 (en) | 2013-02-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP6812473B2 (ja) | メッセージ中のタスクの識別 | |
US9792356B2 (en) | System and method for supporting natural language queries and requests against a user's personal data cloud | |
JP5482894B2 (ja) | 特別日の登録のための情報処理方法 | |
US12002010B2 (en) | Event extraction systems and methods | |
JP5796496B2 (ja) | 入力支援システム、方法、およびプログラム | |
US10924604B2 (en) | System, a computer readable medium, and a method for providing an integrated management of message information | |
US20110314375A1 (en) | Personal Assistant for Task Utilization | |
KR101358084B1 (ko) | 정보처리장치 및 워크플로우 처리방법 | |
US20110313803A1 (en) | Social Task Lists | |
US8793574B2 (en) | Methods and systems for identification and transcription of individual ancestral records and family | |
US20130218836A1 (en) | Deep Linking From Task List Based on Intent | |
US20110145761A1 (en) | Interactive task management system and method | |
JP2008533576A (ja) | 電子デバイスのカレンダーアプリケーションのための情報の形成 | |
JP5429377B2 (ja) | 文字入力における候補の表示方法 | |
KR102340792B1 (ko) | 소셜 네트워크의 소셜 업데이트를 기초로 한 어플리케이션 제어 방법 및 장치 | |
US9165056B2 (en) | Generation and use of an email frequent word list | |
CN102890801A (zh) | 一种文字提醒的方法及设备 | |
CN104156363A (zh) | 通信录信息的搜索方法 | |
JP2008089825A (ja) | 音声認識装置、および音声認識プログラム | |
CN111241833A (zh) | 一种文本数据的分词方法、装置及电子设备 | |
CN103902572A (zh) | 移动终端及其数据管理方法 | |
JPWO2017056825A1 (ja) | 予定管理装置、電子機器、予定管理装置の制御方法、および制御プログラム | |
JP2012089974A (ja) | 電話帳検索装置、電話帳検索方法及び電話帳検索プログラム | |
WO2023100384A1 (ja) | 処理動作支援装置及びプログラム | |
US20210141996A1 (en) | System and method for note taking and management |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 11786359 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2012517156 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13578288 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 11786359 Country of ref document: EP Kind code of ref document: A1 |