JP2003108547A - Kana/kanji conversion device and method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a KANA/KANJI conversion device and a method for converting various expression used by a user into proper expression without clearly selecting the expression. SOLUTION: This KANA/KANJI conversion device is provided with a KANA /KANJI conversion part 103 for converting an inputted character string into a KANA/KANJI mixed character string, an expression frequency learning part 109 for learning a frequency of the converting expression, and a control part 101 for converting the expression of the KANA/KANJI mixed character string by an expression converting processing part 106 on the basis of the frequency of the expression learnt by the expression frequency learning part 109.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to kana-kanji conversion for inputting Japanese, particularly in mobile terminals such as mobile phones, and particularly for inputting various expressions in colloquial style used in e-mail, bulletin boards and the like. The present invention relates to a kana-kanji conversion device and method suitable for.

[0002]

2. Description of the Related Art World Wide Web Mobile terminals equipped with advanced functions such as browsing Web information, sending and receiving e-mails, managing schedules, etc. are rapidly spreading, and when using these mobile terminals, Kana-Kanji conversion is used. Opportunities to input Japanese have become very large.

Unlike a personal computer or the like, a portable terminal is made compact in order to emphasize portability, and the number of input keys is small. When writing on e-mails or bulletin boards, expressions such as “Yaaaaaaa!” And “Hayaro!” Are used to emphasize emotions, but there are few input keys on mobile terminals. For,
It takes a lot of time and effort to input the long reading expression as described above. Conventionally, as a technique for reducing the time and effort of such input, an invention relating to a document processing method and its apparatus disclosed in Japanese Patent Laid-Open No. 191987/191987 has been made.

The conventional document processing apparatus is provided with an expression dictionary and an expression conversion unit, and converts the Kana-Kanji conversion result of the input phonetic character string into the expression type designated by the user. FIG. 22 is a block diagram showing the configuration of a conventional document processing apparatus. In the figure, 2201 is a character input unit, 220
2 is a kana-kanji conversion unit, 2203 is a kana-kanji conversion candidate holding unit, 2204 is an expression dictionary, 2205 is an expression type holding unit,
Reference numeral 2206 is an expression conversion candidate selection unit, 2207 is an expression type selection unit, 2208 is an expression conversion candidate holding unit, 2209 is an expression conversion unit, 2210 is a display unit, and 2211 is a previous expression type holding unit.

Next, the operation will be described. The kana-kanji conversion unit 2202 converts the reading character string input by the user from the character input unit 2201 into a kana-kanji mixed character string. Next, the expression conversion unit 2209 uses the Kana-Kanji conversion unit 220.
The expression dictionary 2204 is searched for the kana-kanji mixed character string converted in 2, and candidates for other expression types are acquired. The expression type selected by the user using the expression type selecting unit 2207 from these candidates is stored in the expression conversion candidate holding unit 2208 as an expression conversion candidate. Further, the expression conversion candidate selection unit 2206 is used to allow the user to select an expression conversion candidate, and a character string mixed with kana / kanji desired by the user is acquired.

In the conventional document processing apparatus, "taaaa!" Is registered in the expression dictionary 2204 as a colloquial expression for the auxiliary verb "ta" in the past, and the expression type is entered when the reading "yata" is input. In the selection, a colloquial expression conversion can be performed to obtain a conversion result of "I did it!". In addition, the expression type selected by the user using the expression type selection unit 2207 is stored in the previous expression type holding unit 2
The expression 211 is held in the table 211 and converted into the held expression type with respect to the subsequent input, so that the expressions are unified.

[0007]

Since the conventional document processing apparatus is configured as described above, the user needs to select the expression type when performing the expression conversion processing, and the operation is complicated. . Further, although the expression types stored in the previous expression type holding unit 2211 are used to unify the expressions, the expression types of colloquial styles used in e-mails and bulletin boards are various, and for example, "zo". There are many types of emphasized expressions for the final particle, such as "Zoo", "Zoo", "Zoo", and "Zoo", and they can be unified with only one or several types of expressions. Since it is not available, it is difficult to satisfy the user. Further, by using the expression type stored in the previous expression type holding unit 2211, the expressions can be unified at a rough level such as “spoken language” or “honorific word” stored immediately before, but a detailed expression type is selected. In this case, the user has to select a desired expression type from a very large number of expression types each time, which makes the input operation troublesome. Further, since only the expression type selected by the user is stored, it is easy to use when the expression differs depending on the destination of the e-mail, or when the expression used in the frequently browsed Internet document is used by the user. There was a problem that I could not do it.

The present invention has been made to solve the above problems, and expresses expressions explicitly by learning the frequency of use of various expression types used by a user when inputting a colloquial style. An object of the present invention is to obtain a kana-kanji conversion device and method capable of performing appropriate expression conversion without selecting a type.

Further, by detecting the expression type used in the received electronic mail and learning the frequency of use, it is possible for the user to input the text of the electronic mail without explicitly selecting the expression type. An object is to obtain a kana-kanji conversion device and method capable of performing appropriate expression conversion.

Further, by detecting the expression type used in the browsed Internet document and learning the frequency of use, it is possible to explicitly select the expression type when inputting on the bulletin board on the Internet. An object is to obtain a kana-kanji conversion device and method capable of performing appropriate expression conversion.

[0011]

A kana-kanji conversion device according to the present invention comprises input means for inputting a character string, and kana-kanji conversion means for converting a character string input from the input means into a kana-kanji mixed character string. , An expression conversion means for converting the expression of a kana-kanji mixed character string converted by the kana-kanji conversion means, an expression frequency learning means for learning the frequency of expressions to be converted by the expression conversion means, and an expression learned by the expression frequency learning means The expression conversion means is provided with a control means for converting the expression of a character string mixed with kana and kanji based on the frequency.

In the kana-kanji conversion device according to the present invention, the expression conversion means converts the expression of the kana-kanji character string based on a predetermined expression conversion rule.

The kana-kanji conversion device according to the present invention comprises expression selection means for selecting a desired expression from the plurality of expressions converted by the expression conversion means, and the expression frequency learning means selects the expression frequency selected by the expression selection means. And the expression to be prioritized in the conversion of the expression conversion means performed later is determined from the frequency.

In the kana-kanji conversion device according to the present invention, the expression frequency learning means obtains the expression frequency from the usage frequency of the expression conversion rule used in the conversion of the expression conversion means.

The kana-kanji conversion device according to the present invention includes an input means for inputting a character string, a kana-kanji conversion means for converting a character string input from the input means into a kana-kanji mixed character string, and a received e-mail. An e-mail expression frequency learning means for searching the used expression based on a predetermined expression search key and learning the frequency of the searched expression, and a pseudonym based on the frequency of the expression learned by the e-mail expression frequency learning means. An electronic mail expression conversion means for converting the expression of a character string mixed with kanji and a control means for controlling the electronic mail expression conversion means to create an electronic mail to be transmitted.

The kana-kanji conversion device according to the present invention comprises a sending address detecting means for detecting a sending address of an electronic mail to be created, and the e-mail address received by the electronic mail expression frequency learning means and the e-mail used in this electronic mail. Stored in association with the frequency of the expressed expression, the control means collates the address detected by the sending address detection means with the address of the electronic mail stored in the electronic mail expression frequency learning means, and the matched electronic mail Based on the frequency of expressions associated with the address, the expression conversion means converts the expression into an expression suitable for electronic mail.

The kana-kanji conversion device according to the present invention includes an input means for inputting a character string, a kana-kanji conversion means for converting the character string input from the input means into a kana-kanji mixed character string, and an internet document browsed. A browsing document expression frequency learning means for searching the used expression based on a predetermined expression search key and learning the frequency of the searched expression, and a pseudonym based on the frequency of the expression learned by the browsing document expression frequency learning means. An input form expression conversion means for converting the expression of a character string mixed with kanji and a control means for controlling the input form expression conversion means and for inputting the converted characters to the input form.

The kana-kanji conversion device according to the present invention comprises input form URL detection means for detecting the URL of an input form for input operation, and the URL of the document browsed by the browsed document expression frequency learning means and the browsed document The expression frequency used is stored in association with it, and the control means collates the URL detected by the input form URL detection means with the URL of the browsed document stored in the browsed document expression frequency learning means, and they match. The expression conversion means is caused to convert into an expression suitable for the input form based on the frequency of expression related to the URL of the browsed document.

The kana-kanji conversion method according to the present invention includes the steps of inputting a character string, converting the input character string into a kana-kanji mixed character string, and converting the representation of the converted kana-kanji mixed character string. The process includes a process of learning, a process of learning the frequency of expressions to be converted, and a process of converting the expression of a kana-kanji mixed character string based on the learned expression frequency.

[0020]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below. Embodiment 1. 1 is a block diagram showing a kana-kanji conversion device 1 according to a first embodiment of the present invention. In the figure,
Reference numeral 1 is a kana-kanji conversion device, 101 is a control unit (control means) that controls each part of the kana-kanji conversion device 1, 102
Is an input unit (input means) for inputting a reading character string, 103
Is a kana-kanji conversion unit (kana-kanji conversion means) for converting a reading character string input from the input unit 102 into a kana-kanji mixed character string. Reference numeral 104 is a kana-kanji conversion dictionary composed of data such as a kana-kanji mixed character string used in the conversion performed by the kana-kanji conversion unit 103, and 105 is a conversion performed by the kana-kanji conversion unit 103. It is a display unit that displays a result.

Reference numeral 106 denotes an expression conversion processing section (expression conversion means) for converting the expression of the conversion result performed by the kana-kanji conversion section 103 according to the expression conversion rule 107 described later. 107
Is an expression conversion rule stored in the storage unit 111, 10
Reference numeral 8 denotes a candidate selection unit (candidate selection means) that selects a desired candidate from the conversion result composed of a plurality of candidates. 109 is the expression conversion rule 107 based on the candidates selected by the candidate selection unit 108.
The expression frequency learning unit (expression frequency learning unit) that learns the expression frequency from the usage frequency of 110, and the expression frequency learning unit 109.
It is an expression frequency holding unit (expression frequency learning means) that holds the expression frequency learning result obtained from. The storage means for storing the kana-kanji conversion dictionary 104 may be the same as the storage means 111 or may be separately provided.

Next, the operation will be described. The detailed processing operation of the kana-kanji conversion device 1 will be described with reference to the flowchart of FIG. In step ST201, the control unit 101
It is determined whether to end the kana-kanji conversion process. When performing conversion processing of Kana-Kanji, step ST described later.
The processing from 202 to step ST209 is repeated.
When ending the conversion process, the ending process is performed.

In step ST202, when the reading character string is input to the input unit 102 and, for example, the reading character string "saikouda" is input to the kana-kanji conversion device 1 by a predetermined operation, The kana-kanji conversion device 1 repeats the processing from step ST203 to step ST209. In the kana-kanji conversion processing, the kana-kanji conversion unit 103 refers to the kana-kanji conversion dictionary 104,
It converts the input phonetic character string into a character string containing Kana and Kanji. Common algorithms used for kana-kanji conversion include the longest match method, the two-segment longest match method, and the minimum cost method. The kana-kanji conversion device 1 does not depend on these kana-kanji conversion algorithms and can be implemented by adopting any kana-kanji conversion algorithm. In the following description, for the sake of simplicity, the Kana-Kanji conversion process using the longest matching method will be described as an example. Here, the longest matching method is an algorithm in which, when converting a reading character string to a kana-kanji mixed character string, a candidate having the longest matching with the reading character string is given priority, and this is adopted as the conversion result.

In step ST203, the control unit 101 determines whether or not all the input reading character strings have been processed, and determines whether or not to repeat the conversion process. If all the input reading character strings have been processed, the process proceeds to step ST207 and the next process is continued. Further, for example, when the reading character string “saikouda” is input and some or all of the reading character strings are unprocessed, step ST204
Go to.

In step ST204, the kana-kanji conversion unit 103 searches the kana-kanji conversion dictionary 104 for dictionary items that match the input phonetic string. FIG. 3 is an explanatory diagram showing a configuration example of dictionary items of the kana-kanji conversion dictionary 104. In the figure, 301 is a reading field for searching, 302 is a field of a written character string containing kana-kanji characters corresponding to the reading field 301, 303 is a field of part-of-speech information indicating grammatical properties, and 304 is an applicable expression conversion. This is a field describing the rule 107. The dictionary entry has at least these fields. Further, the kana-kanji conversion dictionary 104 is also searched for the character string that is the portion of the input phonetic character string.
For example, for the reading character string "saikouda", the corresponding dictionary item is searched for the reading of parts such as "saikou", "sai", and "sa".

Further, in the longest matching method, the candidate having the largest number of matching read characters is preferentially selected, so that "highest" and "reconsider" are selected for the reading "saikou". The candidate is, for example, the storage unit 1 shown in FIG.
11 is stored in the conversion work area. FIG. 4 is an explanatory diagram showing an example of the configuration of the conversion work area. In the figure, 401 is an area for storing a clause number indicating which clause, 402 is an area for storing a candidate number indicating a candidate in a certain clause, and 403 is an output rank of another expression of the candidate. An area for storing the expression frequency used for control, an area for storing the number of the applicable rule 404, an area for storing the number of words of the candidate, 405.
Numerals 410 to 410 are areas for storing the words constituting the candidate. The conversion work area may be dynamically secured during the conversion process, or may be statically secured by previously determining the maximum clause number and the maximum candidate number. In addition, although what is shown in FIG. 4 shows only the display information as the content of the word, this is for simplifying the illustration and facilitating the understanding. To be done.

In step ST205, the kana-kanji conversion processing unit 103 refers to a connection table (not shown) that describes whether or not the words can be connected, and performs a connection process for each word. Since the candidate having the candidate number 1 in the region 402 and the word “highest” in the region 406 and the candidate having the candidate number 2 and the word “reconsider” in the region 402 shown in FIG. To do.

In step ST206, the expression conversion processing unit 106 refers to the expression conversion rule 107 and performs the expression conversion processing on the candidates stored in the conversion work area. This expression conversion processing is performed when the expression conversion rule 107 is described in the applicable field 304 for each notation character string of the field 302 in the dictionary item shown in FIG. For example, since the notation character strings “highest” and “reconsider” in the field 302 have no description of the expression conversion rule 107 in the field 304, step ST2
At 06, no processing is performed and the process returns to step ST203. The expression conversion processing will be described later.

In step ST203, which is performed again, "saikou" in the input reading character string has already been processed, so it is determined to process the remaining "da", and the process proceeds to step ST204 to proceed to "da". The kana-kanji conversion dictionary 104 is searched for. After this step ST
At 205, word connection processing is performed. FIG. 5 is an explanatory diagram showing an example of the conversion work area in which the connection process of step ST205 is performed. In this connection processing, of the word strings that can be connected to "da", the candidate with the longest reading character number is adopted. The adjective verb "highest" has a region 501 in the figure.
The adjective verb inflection “da” stored in the area is connected, and the affirmative verb “da” stored in the area 502 is connected to the sahen noun “reconsider”.

Thereafter, in step ST206, the expression conversion processing unit 106 refers to the expression conversion rule 107 to perform the expression conversion processing. As shown in the dictionary item of FIG. 3, the applicable expression conversion rule 107 is applied to the adjective verb inflection “da”.
Since the rule R1 and the rule R2 are described as, the expression conversion processing is performed.

Details of the processing operation of the expression conversion processing unit 106 will be described below with reference to the flowchart of FIG. In step ST601, when it is determined that the expression conversion processing has ended, if there are no candidates for the expression conversion processing, the processing ends. If there is a candidate for the expression conversion process, the process proceeds to step ST602. In the example illustrated in FIG. 5, the expression conversion process is performed for two words, the adjective verb inflection “da” stored in the area 501 and the assertive auxiliary verb “da” stored in the area 502.

At step ST602, it is determined whether or not the expression conversion rule 107 is described in the dictionary item of FIG.
If the expression conversion rule 107 is not described, step S
Returning to the processing of T601, the next candidate word is processed.
In the example of FIG. 3, the adjective inflection ending “da” is described as the expression conversion rule 107 so that the rules R1 and R2 correspond to each other, and when processing the “da” in the region 501 of FIG. Advances to step ST603. On the other hand, when processing "da" in the area 502 in FIG. 5, there is no description of the expression conversion rule 107 corresponding to the assertive auxiliary verb "da" in FIG. 3, so steps ST603 to ST6 in FIG.
The process of 05 is skipped and the process returns to step ST601.

At step ST603, expression conversion rule 1
When a plurality of rules from 07 are sequentially applied, it is determined whether or not there is an unprocessed rule, and it is determined whether or not the process is repeated. For example, the area 50 shown in FIG.
Rule R1 and R
The process to which 2 is applied is performed, and if both rules are applied and the process is completed, the expression conversion process is ended. Also,
If either or both of the rule R1 and the rule R2 are unprocessed, the process proceeds to the next step ST604.

FIG. 7 is an explanatory diagram showing an example of the structure of the expression conversion rule 107. In the figure, 701 is a column in which a rule number for identifying a rule is stored, 702 is a column in which a pre-conversion expression indicating an expression pattern before expression conversion is stored, and 703 is a post-conversion expression indicating an expression pattern after expression conversion. This is the field to be stored. The expression conversion rule 107 includes at least a rule number field 701, a pre-conversion expression field 702, and a post-conversion expression field 703.
It is composed of each information stored in.

In the alternative expression generation process of step ST604 of FIG. 6, the rule R1 of the rule number column 701 shown in FIG. 7 is applied to the adjective verb inflection “da” in the region 501 of FIG. “$ A” in the pre-conversion expression column 702 corresponding to the rule R1 represents the last character of the column of the processing target word. When the rule R1 is applied to the adjective ending "da", "da" is generated according to "$ A ~" in the post-conversion expression column 703.

In step ST605, candidate expansion processing is performed. As can be seen from the content illustrated in FIG. 7, a plurality of words can share the same conversion pattern, so that a predetermined applicable word is converted into a similar expression. In this way, it is possible to handle various expressions with a few rules. For example, the rule R1 can be shared with the adjective verb inflection ending "da", the past auxiliary verb "ta", and the final particle "na". "Wow!" And "I told you."
Can be converted into expressions such as. In addition to the rules applicable to a plurality of words, the rule number field 7 shown in FIG.
Like the rule R4 of 01, the expression "-" shown in the pre-conversion expression column 702 is replaced with "~" shown in the post-conversion expression column 703.
A rule for performing a simple word replacement to be converted into "~" may be provided.

When the candidate expansion is performed, the expression frequency holding unit 11
The use frequency of each rule stored in 0 is referred to, and the use frequency of the rule is added to the expression frequency of the candidate newly generated in the conversion work area. FIG. 8 is an explanatory diagram showing a configuration of information stored in the expression frequency holding unit 110. In the figure, 801 is a rule number column, and 802 is a column showing the rule usage frequency. The expression frequency holding unit 110 uses the rule number column 80.
The rule shown in 1 and the usage frequency shown in the column 802 are stored. The contents stored in the rule number column 801 correspond one-to-one with the contents of the rule number column 701 of the expression conversion rule 107 shown in FIG. The frequency of use of a rule is expressed by using an integer value from −255 to +255, for example, when a candidate to which no rule is applied is used as a frequency of use “0” and the candidate is expressed and converted based on the rule. "+1" is added to the value of the frequency of use and is subtracted from the value of the frequency of use so as to be "-1" when it is not used in the expression conversion. The frequency of expression is known from the frequency of use of this rule, and the order of expressions to be preferentially output as a conversion result from the expression conversion processing unit 106 is controlled based on the frequency of expression. It should be noted that the calculation of the rule usage frequency is performed by the expression frequency learning unit 109 described later, and the numerical value indicating the rule usage frequency, the calculation, and the like may be any as long as similar operational effects can be obtained. It is not limited to what has been described.

Next, each processing step from step ST603 to step ST605 in FIG. 6 will be described. FIG. 9 is an explanatory diagram showing each state of the conversion work area in the determination process of step ST603, the separate display generation process of step ST604, and the candidate expansion process of step ST605. After it is determined in step ST603 that there is an unprocessed rule, in the alternative expression generation process in step ST604, the candidate 901 shown in FIG. The candidate 902 shown in (B) is generated. At this time, as shown in FIG.
The usage frequency “3” of the rule R1 stored in 10 is added to the expression frequency “0” shown in the column 901a of FIG. 9 (A), and as shown in the column 902a of FIG. 9 (B), Candidate 902
It is stored in the conversion work area. Together with this, column 90
The rule R1 is stored in 2b as the applicable rule. In the example shown in FIG. 9, one rule is applied to each of the candidate 902 and the candidate 903 to be described later, but if the number of words stored in the conversion work area is large, words 1 to 5 will be used. In this case, a plurality of rules may be applied, and in this case, a field for storing the applied rules, for example, a field 902 in FIG. 9B.
A plurality of rule numbers are stored in the column b or the column 903b in FIG. 9C. In addition, the sum of the usage frequencies of the plurality of applied rules is the expression frequency, and the expression frequency holding unit 1
Stored in 10.

Next, as the candidate expansion in step ST605, the candidate 901 is duplicated as shown in FIG. 9C, and the rule R2 is applied to this to generate the candidate 903. This means that the "da" of the candidate 901 is converted into "daaaa" by the rule R2 shown in FIG. 7, and in addition to the candidates 901 and 902,
Further, the candidate 903 is developed.

The processing after step ST207 in FIG. 2 will be described below. In the candidate display of step ST207, the candidates stored in the work area shown in FIG. 9C are displayed on the display unit 105 shown in FIG. 1 under the control of the control unit 101. The candidates are displayed for each clause and in the order of candidate numbers. The candidates having the same candidate number are displayed in descending order of expression frequency. 10, FIG. 11 and FIG. 12 are explanatory diagrams showing an example of display contents of the display unit 105. Although the display example shown in FIG. 10 shows a list of candidates, the next candidate may be sequentially displayed in the same line by a predetermined operation by the user. In the display example shown in FIG. 10, all the expressions are sequentially expressed for each candidate.
You may make it display each candidate for every expression type like the example shown in FIG.

In the candidate selection of step ST208, the user is made to select a desired candidate from the plurality of candidates displayed on the display section 105. For this selection, the candidate selection unit 108 shown in FIG. 1 is provided with two kinds of selection keys, a candidate selection key and an expression selection key, and the candidate selection and the expression selection are operated independently as shown in FIG. You may When the expression selection key is operated, the display on the display unit 105 is as shown in FIG.
Transitions from the display 1201 shown in to the display 1202, and further to the display 1203 when the expression selection key is operated.
If the candidate selection key is operated while the display 1201 is displayed, the display 1204 is displayed. When displayed as shown in FIG. 12 and operated by the user, compared to a case where all the various expressions are displayed in a list and the user is allowed to select them,
The operation load on the user is reduced.

In the expression frequency learning in step ST209,
Regarding the candidates selected by the user in step ST208,
The expression frequency learning unit 109 shown in FIG. 1 refers to each information stored in the conversion work area shown in FIG. 9 to perform the expression frequency learning process. FIG. 13 is a flowchart showing the expression frequency learning process. Step ST1301 is the end judgment of the iterative process for each clause when the conversion result is composed of a plurality of clauses. In the example shown in FIG. 9, since the conversion result is one phrase, the learning process is performed on the phrase “saikouda”.

In step ST1302, it is determined whether or not there is the expression conversion rule 107 applied to the selected candidate, that is, whether or not there is another expression. For example, in step ST208 of FIG. 2, the candidate 903 shown in FIG.
It is assumed that "The best" is selected. Referring to the contents of the conversion work area in FIG. 9C, it can be determined that there are other expression candidates 901 and 902 having the same candidate number, and the process proceeds to step ST1303. If another expression candidate having the same candidate number does not exist, the expression conversion process is not performed for the candidate, so that the step ST
The processing from step 1303 to step ST1307 is skipped, the processing returns to step ST1301, and the same processing is repeated for the next clause.

In step ST1303, it is determined whether the selected candidate is the original expression to be subjected to the expression conversion process. Whether or not it is the original expression is determined by whether or not the application rule field of the conversion work area is blank display "-". If the applicable rule is blank "-", the process proceeds to step ST1304. In the example of FIG. 9C, the selected candidate 903
Since the rule R2 is applied as shown in the column 903b, the process proceeds to step ST1306.

In step ST1306, overflow processing is performed. In this overflow processing, if the frequency is continuously added in the next step ST1307, it will eventually reach a predetermined maximum value. In order to prevent this,
Candidate 90 selected in step ST208 of FIG.
When the usage frequency of the rule R2 applied to No. 3 reaches a predetermined maximum value of 255 when one more is added, the value of the usage frequency in the column 802 shown in FIG. Replace. In the example shown in FIG. 8, since the usage frequency of rule R2 is 3, the division described above is not performed and step ST13 is performed.
Proceed to 07. Then, in step ST1307, the rule R
1 is added to the usage frequency 3 of 2 to obtain 4. After this, the process returns to step ST1301 and the same process is repeated for the next clause.

Below, in step ST208 of FIG.
The process of steps ST1304 and ST1305 shown in FIG. 13 will be described by exemplifying the case where “highest” of the candidate 901 of (C) is selected. The applicable rule column of the candidate 901 shown in FIG. 9C is blank "-", so that the next step ST1 is determined by the determination in step ST1303.
Proceed to 304.

In step ST1304, underflow processing is performed. In the underflow process, when the usage frequency of each applicable rule is continuously subtracted in the next step ST1305, a predetermined minimum value is eventually reached. Therefore, in order to prevent this, in step ST1304, for example, selection of the candidate 901 is performed. When the usage frequencies of the rules R1 and R2 applied to the candidates 902 and 903 that are expressions reach a predetermined minimum value of −255 when 1 is subtracted next, the expression frequency holding unit 110 illustrated in FIG. All the values of the usage frequency of the stored column 802 are replaced with a value obtained by dividing by 2. In the example shown in FIG. 8, since the usage frequencies of the rules R2 and R3 are 3, the division described above is not performed and the process proceeds to step ST1305. Subsequently, in step ST1305, 1 is subtracted from the usage frequencies 3 of the rules R2 and R3 to obtain 2 respectively. After this, the process returns to step ST1301 and the same process is repeated for the next clause.

When the process of step ST209 shown in FIG. 2 is completed, the kana-kanji conversion device 1 returns to the process of step ST201 again, and repeats each process described so far as necessary to sequentially read the read character strings. Convert to a desired kana-kanji mixed character string.

As described above, according to the first embodiment, the usage frequency of the expression conversion rule 107 is determined by the expression frequency learning unit 1.
09, and stores this in the expression frequency holding unit 110 to convert the input reading character string into a candidate with a high priority according to the usage frequency. Therefore, the expression used by the user immediately before is converted. I can remember and unify the style,
With respect to various expressions, it is possible to perform detailed conversion control, and it is possible to convert appropriate expression candidates.

Further, since the expression frequency learning unit 109 is made to learn the frequency of the expression, for example, the expression generated by the rule R2 of "the best".
When selected by the operation of the candidate selection unit 108 by the user, 1 is added to the usage frequency of the rule R2, and the input of the reading character string from the input unit 102 from the next time onward is not “highest” but “highest”. "Aaaa" takes precedence and display 1
No. 05 is displayed and the user does not need to perform an expression selection operation, so that the input operation can be facilitated. In addition, since candidates are displayed with priority for similar expressions such as “winned aaaa” and “said aaaa” to which the rule R2 is applied, the candidate selection unit 108 by the user.
There is an effect that the expression selecting operation is unnecessary and the input operation can be facilitated.

Further, since the expression frequency learning unit 109 is made to learn the frequency of the expressions, for example, if the candidate "highest" is repeatedly selected and frequently used, the usage frequency of the rules R1 and R2 is used. Is lower,
The value of the usage frequency in the column 802 shown in is gradually subtracted and becomes a negative value. Then, “highest” is preferentially displayed, which has the effect of reducing the load of the expression selection operation of the selection operation unit 108 by the user.

Further, since the expression frequency learning unit 109 is made to learn the frequency of the expressions, if two kinds of expressions, "highest" and "quick white" are frequently used, these can be used. 7, the value indicating the frequency of use of the rules R1 and R3 shown in FIG. 7 applied to each expression becomes large, and the similar expressions such as “winned” and “win” are obtained by applying these two types of rules R1 and R3. Since "I'm sleepy" is preferentially displayed, it is possible to reduce the load of the expression selection operation of the candidate selection unit 108 by the user.

Embodiment 2. FIG. 14 is a block diagram showing a kana-kanji conversion device according to the second embodiment. The same reference numerals are used for the same or corresponding parts of the kana-kanji conversion device 1 shown in FIG. 1, and the description thereof will be omitted. In the figure, 14
Reference numeral 01 is an e-mail storage unit that classifies received e-mails and stores them for each folder; 1402 is an e-mail that detects expressions used in the e-mails stored in the e-mail storage unit 1401 and learns the frequency of use. Expression frequency learning unit (e-mail expression frequency learning unit), 1403 is a transmission address detection unit (transmission address detection unit) that detects a transmission address of an email being created, and 1404 is an expression learned by the E-mail expression frequency learning unit 1402. E-mail expression frequency holding unit (e-mail expression frequency learning means) that stores and holds the frequency of 1400, and 1405 a transmission address detecting unit 1403.
Sending address and email expression frequency holding unit 1 detected by
An e-mail expression conversion processing unit (e-mail expression conversion means) that performs expression conversion using information indicating the expression frequency held in 404.

Next, the operation will be described. The processing operation of the kana-kanji conversion device 2 shown in FIG. 14 will be described with reference to the flowchart of FIG. Processing operations similar to or corresponding to those in the kana-kanji conversion device 1 are denoted by the same step symbols as in FIG. 2, and description thereof will be omitted. In the e-mail expression frequency learning process of step ST1501, the e-mail expression frequency learning unit 1402 learns the contents of the e-mail stored in the e-mail storage unit 1401 in advance, and the learned contents are stored in the e-mail expression frequency holding unit 1404. To be stored in. The processing in step ST1501 may be performed for all unlearned emails each time the mobile terminal is powered on, or may be performed sequentially when a new email is received.

Here, the e-mail expression frequency learning process will be described. FIG. 16 shows the expression conversion rule 10 shown in FIG.
7, an expression search key 1601 is added, and a description showing the relation between the information stored in each of the rule number field 701, the expression field before conversion 702, and the expression field after conversion 703 and each information of the expression search key 1601 is shown. It is a figure. Note that the one illustrated in FIG. 14 is provided with the expression search key 1601 in the storage means 111 similarly to the expression conversion rule 107, and is limited to the configuration illustrated here as long as the same action and effect can be obtained. Not a thing.

The expression search key 1601 shown in FIG. 16 is a character string that associates various expressions included in the body of an electronic mail with the rules used for expression conversion. The e-mail expression frequency learning unit 1402 includes an e-mail storage unit 140.
A predetermined expression search key 1601 is used to search the e-mail stored in No. 1 for a character string that matches the content of the expression search key 1601. A more matched string,
A character string including the character strings before and after it is extracted from the body of the electronic mail, and the expression of these series of character strings is associated with the expression search key 1601 used for the search and the converted expression field 70
If the content stored in the e-mail expression frequency holding unit 1404 is used, it is determined that the rule stored in the rule number column 701 related to the expression search key 1601 is used. Add 1 to the frequency.

FIG. 17 shows the electronic mail expression frequency holding unit 14.
It is explanatory drawing which showed the structure of the information stored in 04. The explanatory diagram of FIG. 17 shows the rule number column 80 shown in the explanatory diagram of FIG.
1 shows an example of the relationship between the information stored in the usage frequency column 802 and the information stored in the folder name column 1501. The folder name column 1501 stores information indicating the folder name in which the electronic mail is stored. In FIG. 17, the electronic mail expression frequency learning unit 1 is shown.
As a result of 402 searching the electronic mail, it is shown that the rule determined to be used and the folder name in which the electronic mail using the rule is stored are associated with each other and stored. There is.

In the electronic mail expression frequency holding unit 1404, an arbitrary rule performed by the electronic mail expression frequency learning unit 1402 and the folder name of the storage destination of the electronic mail using the rule are stored in association with each other. For example, expressions frequently used in e-mails with friends and expressions often used in e-mails related to work can be acquired for each e-mail address for each expression conversion rule 107. it can.

After the processing of step ST1501 is performed in this way, processing similar to the contents described with reference to FIG. 2 is performed from step ST201 to step ST201 shown in FIG.
The processing is performed up to 205, and then the electronic mail expression conversion processing in step ST1502 is performed.

Next, the electronic mail expression conversion processing in step ST1502 will be described. FIG. 18 is an electronic mail expression conversion processing flowchart executed by the electronic mail expression conversion processing unit 1405. This email representation conversion process
Since it is the same as the expression conversion processing performed by the expression conversion processing unit 106 described in the first embodiment, the same reference numerals are used in the flowchart of FIG. 18 as in the flowchart of the expression conversion processing shown in FIG. Detailed description thereof will be omitted.

In step ST601, it is determined whether the e-mail expression conversion process is finished. If there is no candidate for expression conversion, the conversion process is ended, and if there is a candidate, step ST60.
Then, the process proceeds to step 2 and it is determined whether or not the dictionary conversion item 107 is described in the dictionary item. If the expression conversion rule 107 is described, it is determined in step ST603 whether there is any unprocessed rule, and if it is determined that there is an unprocessed rule, step S603.
Another expression generation of T604 is performed. The above operation processing is
Similar to the operation process of the kana-kanji conversion device 1 described in the first embodiment, it is performed in the kana-kanji conversion process when creating an electronic mail.

In the candidate expansion of step ST1801 shown in FIG. 18, the control unit 101 shown in FIG. 14 causes the transmission address detection unit 1403 to detect the transmission address described in the electronic mail being created, and further the detected address. Folder that contains a lot of
The folder name is acquired from 401. Control unit 101
Uses the folder name as the e-mail expression conversion processing unit 140.
5 is notified, and the expression conversion processing using the relation between the folder name and the usage frequency of each rule shown in FIG. 17 is performed. Further, when the control unit 101 cannot specify the folder in which many addresses detected by the transmission address detection unit 1403 are stored, the control unit 101 converts the expression to the e-mail expression conversion processing unit 1405 using the default frequency of use. Let the process take place.

When the process of step ST1801 is completed, the process returns to the process of step ST603, and the predetermined process is repeated. As in the case described in the first embodiment with reference to FIG. The process ends.

Further, step ST1502 shown in FIG.
When the process of step ST2 is completed, the process returns to step ST203, and step ST2 is performed in the same manner as described with reference to FIG.
The processing from 03 to step ST205 is repeated, the processing from step ST207 to step ST209 is performed under a predetermined condition, and the conversion processing of the kana-kanji conversion device 2 is performed. In addition, regarding the creation of the body of the e-mail, the input unit 10
The kana-kanji conversion unit 103 converts the reading character string input from 2 while referring to the kana-kanji conversion dictionary 104. As described above, the kana-kanji conversion regarding the character string input from the input unit 102 is performed in the same manner as the kana-kanji conversion device 1 described in the first embodiment, and the user selects it using the candidate selection unit 108. Regarding the expressions, the expression frequency learning unit 109 learns, that is, the usage frequency of the rules is added and subtracted, and this value is stored in the e-mail display frequency holding unit 1404, and the usage frequency of the rules is e-mail expression conversion processing unit 1405. It is used for the expression conversion that is performed in 1., and Kana-Kanji conversion is performed in the creation of an e-mail document.

As described above, according to the second embodiment, the expression used in the received electronic mail is detected, and the frequency of use of the expression is collectively held for each folder in which the electronic mail is stored. In addition, when the user composes the electronic mail, there is an effect that the expression conversion suitable for the destination of the electronic mail can be performed based on the frequency of the expression.

Third Embodiment FIG. 19 is a configuration diagram of a kana-kanji conversion device 3 according to the third embodiment of the present invention. The same reference numerals are used for the same or corresponding parts as the kana-kanji conversion device 1 shown in FIG. 1 or the kana-kanji conversion device 2 shown in FIG. 2, and the description thereof is omitted. In the figure, 190
1 is the Unif of the site that contains the input form being entered
orm Resource Locator (hereinafter U
An input form URL detection unit (input form URL detection means) for detecting RL), and a browsed document expression frequency learning unit 1902 for learning an expression used in a browsed Internet document (browsed document expression frequency learning). Means), 1903 is a browsed document expression frequency holding unit (browsing document expression frequency learning unit) that holds the learned expression frequency of the browsed document, and 1904 uses the expression frequency held in the browsed document expression frequency holding unit 1903. It is an input form expression conversion processing unit (input form expression conversion means) that performs expression conversion.

Next, the operation will be described. The operation process of the kana-kanji conversion device 3 will be described with reference to the flowchart shown in FIG. The same step codes as those in FIG. 2 are attached to the operation processes similar to or corresponding to the kana-kanji conversion device 1,
The description is omitted. First, in the browsed document expression frequency learning process of step ST2001, the browsed document expression frequency learning unit 1902 learns in advance the frequency of expressions included in the document browsed by the user on the Internet, and the browsed document expression frequency holding unit 1903. To be stored in. This step ST20
The processing of 01 is performed each time the user browses a document on the Internet.

Next, the browsing document learning process will be described. An expression search key 1601 shown in FIG. 19 is used in the browsed document learning process. This expression search key 1601 is the same as that described with reference to FIG. 16 in the second embodiment, and a detailed description will be given here. Omit it. The expression search key 1601 used in the browsed document learning process is a character string that associates various expressions included in the browsed document with each rule shown in the rule number column 701. The browsed document expression frequency learning unit 1902 searches a document browsed by the user on the Internet for a predetermined character string that matches the content of the expression search key 1601 and includes a series of characters including the character strings before and after the character string. Extract columns. Next, the converted expression field 70 in FIG.
It is judged whether or not the expression of the extracted character string corresponds to the expression stored in 3. If they match, it is determined that the corresponding rule in the rule number column 701 shown in FIG. 16 is used, and 1 is added to the value of the usage frequency of the rule stored in the browsed document expression frequency holding unit 1903.

FIG. 21 shows a browsed document expression frequency holding unit 1903.
FIG. 6 is an explanatory diagram showing the structure of information indicating the frequency of expressions stored in. Rule number column 801 and usage frequency column 8 shown in FIG.
02 is the same as that shown in FIG. URL column 210
The URL stored in 1 is the URL of the site to which the browsed document in which the rule in the rule number column 801 is detected belongs. The one illustrated in FIG. 21 is U in the URL column 2101.
For each RL, the rules extracted from the search results using the expression search key 1601 and the usage frequencies thereof are summarized.
In the usage frequency column 802, the usage frequency of the corresponding rule is added by the browsing document expression frequency learning unit 1902 and stored. In this way, expressions used in communication on the Internet, such as expressions often used in input forms for browsed documents and bulletin boards, are associated with each rule of the expression conversion rule 107, and the frequency of use of each expression is used. To learn.

When the processing of step ST2001 is completed, the kana-kanji conversion device 3 proceeds to step ST201 shown in FIG.
The respective processes from to ST205 are sequentially performed. Since the contents of each process are the same as the corresponding processes shown in FIG. 2, description thereof will be omitted here.

Next, step ST2002 shown in FIG.
The input form expression conversion processing of will be described. The control unit 101 causes the input form URL detection unit 1901 to detect the URL of the site including the input form being input by the user operation, and the detected URL is stored in the URL column 2101 shown in FIG. Check the URL. When the two URLs match, the input form expression conversion processing unit 1904 is caused to perform the expression conversion processing using the usage frequency of the rule detected by the URL. In the expression conversion processing performed by the input form expression conversion processing unit 1904, the operation of the browsed document expression frequency holding unit 1903 will be described as an operation of the corresponding expression frequency holding unit 110, and is shown in FIG. 1 described in the first embodiment. The expression conversion processing is the same as the expression conversion processing performed by the expression conversion processing unit 106. If there is no URL obtained from the browsed document, the default value of the frequency of use is used for the expression conversion process.

When the processing of step ST2002 is completed, the process returns to step ST203, and steps ST203 to ST2 are performed in the same manner as described with reference to FIG. 2 in the first embodiment.
The process of 05 is repeated, and step ST2 is performed under predetermined conditions.
The processing from 07 to step ST209 is performed, and the kana-kanji conversion processing of the kana-kanji conversion device 3 is performed. Regarding the character input of the input form, the kana-kanji conversion unit 103 converts the reading character string input from the input unit 102 with reference to the kana-kanji conversion dictionary 104. As described above, the kana-kanji conversion regarding the character string input from the input unit 102 is performed in the same manner as the kana-kanji conversion device 1 described in the first embodiment, and the user selects the candidate selection unit 10.
For the expression selected using 8, the expression frequency learning unit 1
09, the frequency of use of the rule is obtained, this value is stored in the browsed document display frequency holding unit 1903, the frequency of use of the stored rule is set as the expression frequency, and the input form is calculated based on this expression frequency. Expression conversion processing unit 1904
Converts the expressions and performs kana-kanji conversion when entering characters in the input form on the site.

As described above, according to the third embodiment, when the expression used in the browsed Internet document is detected, the frequency of the expression is learned, and the expression is input to the bulletin board or the like on the Internet, The effect is that it can be converted into an appropriate expression based on the frequency of expression.

[0074]

As described above, according to the present invention, input means for inputting a character string, kana-kanji conversion means for converting a character string input from the input means into a kana-kanji mixed character string, and kana-kanji character The expression conversion means for converting the expression of the kana-kanji mixed character string converted by the conversion means, the expression frequency learning means for learning the frequency of expressions to be converted by the expression conversion means, and the expression frequency learned by the expression frequency learning means. Since the expression conversion means is provided with the control means for converting the expression of the kana-kanji mixed character string based on the above, it is possible to unify the expression of the document and convert it into an appropriate expression corresponding to various expressions. effective.

According to the present invention, the expression conversion means converts the expression of the kana-kanji character string based on a predetermined expression conversion rule, so that the expression of the document can be unified and various expressions can be dealt with. The effect is that it can be converted into an appropriate expression.

According to the present invention, the expression conversion means is provided with the expression selection means for selecting a desired expression from the plurality of expressions converted, and the expression frequency learning means obtains the frequency of the expression selected by the expression selection means. Since the expression to be prioritized in the conversion performed by the expression conversion means is determined from the frequency, it is possible to unify the expressions of the document and convert the expressions into appropriate expressions corresponding to various expressions. is there.

According to the present invention, the expression frequency learning means obtains the expression frequency from the usage frequency of the expression conversion rule used in the conversion of the expression conversion means, so that the expression of the document can be unified and various expressions can be obtained. There is an effect that it can be converted into an appropriate expression according to the expression.

According to the present invention, the input means for inputting the character string, the kana-kanji conversion means for converting the character string input from the input means into the kana-kanji mixed character string, and the kana-kanji conversion means used for the received electronic mail are used. An e-mail expression frequency learning means for searching an expression based on a predetermined expression search key and learning the frequency of the searched expression, and a kana-kanji mixed character string based on the frequency of the expression learned by the e-mail expression frequency learning means Since the electronic mail expression conversion means for converting the expression of and the control means for controlling the electronic mail expression conversion means to create the electronic mail to be transmitted are provided, the expression conversion operation is performed when the body of the electronic mail is input. There is an effect that an appropriate expression conversion can be performed without any need.

According to the present invention, the transmission address detecting means for detecting the transmission address of the electronic mail to be created is provided,
The e-mail expression frequency learning means stores the address of the e-mail received and the expression frequency used in this e-mail in association with each other, and the control means stores the address detected by the sending address detecting means and the e-mail expression frequency. The e-mail address stored in the learning means is collated, and the expression conversion means is caused to convert the expression suitable for the e-mail based on the frequency of expressions related to the matched e-mail address. When inputting the text of the mail, there is an effect that an appropriate expression conversion can be performed without performing an expression conversion operation.

According to the present invention, the input means for inputting the character string, the kana-kanji conversion means for converting the character string input from the input means into the kana-kanji mixed character string, and the used internet document are used. A browsing document expression frequency learning means for searching an expression based on a predetermined expression search key and learning the frequency of the searched expression, and a kana-kanji mixed character string based on the frequency of the expression learned by the browsing document expression frequency learning means. Input form expression conversion means for converting the expression of
Since it has a control means for controlling the input form expression conversion means and for inputting the characters converted into the input form, it is possible to perform conversion into an appropriate expression without performing the operation of expression conversion when inputting into the input form. There is an effect that can be.

According to the present invention, the input form URL detecting means for detecting the URL of the input form for input operation is provided, and the U of the document browsed by the browsed document expression frequency learning means is provided.
The RL and the expression frequency used in this browsed document are stored in association with each other, and the control unit detects the URL detected by the input form URL detection unit and the URL of the browsed document stored in the browsed document expression frequency learning unit. The expression conversion means is caused to convert the expression into a suitable expression for the input form based on the frequency of the expression related to the URL of the matching browsed document. Therefore, the expression conversion operation is performed when the input form is input. There is an effect that conversion can be performed to an appropriate expression without performing.

According to the present invention, a step of inputting a character string, a step of converting the input character string into a kana-kanji mixed character string, and a step of converting an expression of the converted kana-kanji mixed character string, It has a process of learning about the frequency of expressions to be converted and a process of converting the expression of kana-kanji mixed character strings based on the frequency of learned expressions, so that the expression of documents can be unified and various expressions can be created. There is an effect that the corresponding expression can be converted into an appropriate expression.

[Brief description of drawings]

FIG. 1 is a configuration diagram of a kana-kanji conversion device according to a first embodiment of the present invention.

FIG. 2 is a flowchart showing a processing operation of a kana-kanji conversion device.

FIG. 3 is an explanatory diagram showing dictionary items forming a kana-kanji conversion dictionary.

FIG. 4 is an explanatory diagram showing a configuration of a conversion work area.

FIG. 5 is an explanatory diagram showing a conversion work area in which connection processing is performed.

FIG. 6 is a flowchart showing a processing operation of an expression conversion processing unit.

FIG. 7 is an explanatory diagram showing an expression conversion processing rule.

FIG. 8 is an explanatory diagram showing a configuration of information stored in an expression frequency holding unit.

FIG. 9 is an explanatory diagram showing a state of a conversion work area.

FIG. 10 is an explanatory diagram showing display contents.

FIG. 11 is an explanatory diagram showing display contents.

FIG. 12 is an explanatory diagram showing display contents.

FIG. 13 is a flowchart showing an expression frequency learning process.

FIG. 14 is a configuration diagram of a kana-kanji conversion device according to a second embodiment of the present invention.

FIG. 15 is a flowchart showing a processing operation of a kana-kanji conversion device.

FIG. 16 is an explanatory diagram showing a relationship between an expression conversion rule and an expression search key.

FIG. 17 is an explanatory diagram showing a configuration of information stored in an electronic mail expression frequency holding unit.

FIG. 18 is a flowchart showing a processing operation of an electronic mail expression conversion processing unit.

FIG. 19 is a configuration diagram of a kana-kanji conversion device according to a third embodiment of the present invention.

FIG. 20 is a flowchart showing a processing operation of the kana-kanji conversion device.

FIG. 21 is an explanatory diagram showing a configuration of information stored in a browsed document expression frequency holding unit.

FIG. 22 is a block diagram showing the configuration of a conventional document processing apparatus.

[Explanation of symbols]

1, 2, 3 Kana-Kanji conversion device, 101 control unit (control unit), 102 input unit (input unit), 103 Kana-Kanji conversion unit (Kana-Kanji conversion unit), 104 Kana-Kanji conversion dictionary, 105 display unit, 106 representation Conversion processing unit (expression conversion unit), 107 expression conversion rule, 108 candidate selection unit (candidate selection unit), 109 expression frequency learning unit (expression frequency learning unit), 110 expression frequency holding unit (expression frequency learning unit), 111 storage 1401 E-mail storage unit, 1402 E-mail expression frequency learning unit (E-mail expression frequency learning unit), 1403 Sender address detection unit (Send address detection unit), 1404 E-mail expression frequency holding unit (E-mail expression frequency learning unit) ), 1405
E-mail expression conversion processing unit (e-mail expression conversion unit), 1601 expression search key, 1901 input form URL detection unit (input form URL detection unit), 19
02 browsing document expression frequency learning unit (browsing document expression frequency learning unit), 1903 browsing document expression frequency holding unit (browsing document expression frequency learning unit), 1904 input form expression conversion processing unit (input form expression conversion unit).

Continued front page    (72) Inventor Katsushi Suzuki             2-3 2-3 Marunouchi, Chiyoda-ku, Tokyo             Inside Ryo Electric Co., Ltd. F-term (reference) 5B009 MA08 MA14 ME16 MG01 MG04

Claims (9)

[Claims] 1. A kana-kanji conversion device for converting an input character string into a kana-kanji mixed character string, and input means for inputting the character string; and a character string input from the input means for kana-kanji mixed character string. Kana-kanji conversion means for converting into kana, expression conversion means for converting the expression of a kana-kanji mixed character string converted by the kana-kanji conversion means, and expression frequency learning means for learning the frequency of expressions to be converted by the expression conversion means And a control unit for causing the expression conversion unit to convert the expression of a character string mixed with Kana and Kanji based on the frequency of the expression learned by the expression frequency learning unit. 2. The kana-kanji conversion device according to claim 1, wherein the expression conversion means converts the expression of the kana-kanji character string based on a predetermined expression conversion rule. 3. An expression selecting means for selecting a desired expression from a plurality of expressions converted by the expression converting means, wherein the expression frequency learning means obtains a frequency of the expression selected by the expression selecting means, and is performed later. The kana-kanji conversion device according to claim 1, wherein an expression to be prioritized in the conversion of the expression conversion means is determined from the frequency. 4. The kana-kanji conversion device according to claim 3, wherein the expression frequency learning means obtains the expression frequency from the usage frequency of the expression conversion rule used in the conversion of the expression conversion means. 5. A kana-kanji conversion device for converting an input character string into a kana-kanji mixed character string, and input means for inputting the character string; and a character string mixed with the kana-kanji mixed character string input from the input means. A kana-kanji conversion means for converting into a kana-kanji conversion means, an e-mail expression frequency learning means for searching the expression used in the received e-mail based on a predetermined expression search key, and learning the frequency of the searched expression, E-mail expression conversion means for converting the expression of the kana-kanji mixed character string based on the frequency of expressions learned by the e-mail expression frequency learning means, and control means for controlling the e-mail expression conversion means and creating an e-mail to be transmitted Kana-to-Kanji converter with and. 6. An e-mail expression frequency learning means is provided with a sending address detecting means for detecting a sending address of an e-mail to be created, and the e-mail expression frequency learning means detects the address of the received e-mail and the expression frequency used in this e-mail. The control means collates the address detected by the transmission address detection means with the address of the electronic mail stored in the electronic mail expression frequency learning means, and stores it in the matched electronic mail address. The kana-kanji conversion device according to claim 5, wherein the expression conversion means converts the expression into a suitable expression for the electronic mail based on the frequency of the related expressions. 7. A kana-kanji conversion device for converting an input character string into a kana-kanji mixed character string, and input means for inputting the character string, and a character string mixed with the kana-kanji character string for inputting the character string input from the input means. A kana-kanji conversion means for converting into a kana-kanji conversion means, a browsing document expression frequency learning means for searching an expression used in the browsed internet document based on a predetermined expression search key, and learning the frequency of the searched expression, An input form expression conversion means for converting the expression of the kana-kanji mixed character string based on the frequency of the expression learned by the document expression frequency learning means, and the input form expression conversion means for controlling the input form expression conversion means A kana-kanji conversion device having a control means for inputting the characters converted by. 8. An input form URL detecting means for detecting a URL of an input form for performing an input operation, wherein the browsed document expression frequency learning means is a URL of a browsed document.
And the expression frequency used in this browsed document are stored in association with each other, and the control means controls the URL detected by the input form URL detection means and the browsed document stored in the browsed document expression frequency learning means. 8. The kana-kanji conversion according to claim 7, wherein the expression conversion means converts the expression into a suitable expression for the input form based on the frequency of the expression related to the URL of the browsed document that matches. apparatus. 9. A step of inputting a character string, a step of converting the input character string into a kana-kanji mixed character string, a step of converting an expression of the converted kana-kanji mixed character string, and an expression to be converted. A method of converting kana-kanji characters, the method comprising: learning about the frequency of kanji; and converting the expression of a kana-kanji mixed character string based on the learned expression frequency.