JPS6118067A - Kana-to-kanji converter - Google Patents

Abstract

PURPOSE:To output with priority the description selected by an operator in the preceding time in case the same character string is supplied after the word learning information is registered, by providing each part to store, register and retrieve the word learning information containing the information on the maximum/longest coincident words and the longest coincident words. CONSTITUTION:A KANA (Japanese syllabary) string memory part 2 adds the KANA string supplied from a KANA input part 1 to the end of a stored KANA string and stores it there. A KANA string front part extracting part 7 receives a start action from a KANA string front part extraction start part 6 and designates a partial string starting from the head of the KANA train stored in the part 2 to send it to a word retrieving part 4 as well as to a KANA string front part extraction memory part 8. The words retrieved by the part 4 is stored to a longest coincident word temporary memory part 5. The information of both a maximum/longest coincident word memory part 9 and the part 5 are stored to a word learning information memory part 11. A word learning information retrieving part 12 retrieves a designated reading string out of the part 11 and outputs it to a KANJI (Chinese characters) output part 13. Thus the description selected in the preceding time can be outputted with priority in case the same character string is supplied.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a kana-kanji conversion device that inputs a Japanese sentence as a reading and outputs it as a kanji notation.

Structure of conventional examples and their problems What is known as a conventional kana-kanji conversion device is a keyboard for inputting the pronunciation of Japanese sentences, and a word dictionary that stores pairs of pronunciations and kanji notation for many Japanese words. , a word search unit that searches the word dictionary for a word with the input reading, and an output unit that outputs the kanji notation of the searched word. It outputs the kanji notation of a word or word string with a reading. However, in Japanese words that have the same pronunciation but different kanji writing (homonyms, 3, -, -
・Because there are homophones (hereinafter referred to as homophones), conventional kana-kanji conversion devices usually have a homophone selection unit that outputs multiple words with the same pronunciation and allows the operator to select one of them. Furthermore, once a word has been selected, it is output preferentially over other homophones during the next conversion, thereby reducing the need for homophone selection operations. For this purpose, a word learning dictionary is used that temporarily stores a pair of information that uniquely specifies a certain word (ie, reading and kanji notation) and information indicating that the word has been selected. And above 1
When performing a word search for ', a method is used in which a word having the reading input previously in the learning dictionary is searched, and if a corresponding word is found, the kanji notation of the word is output.

In Japanese grammar, homonyms usually refer to cases where the reading of one word matches the reading of another word, but conventional kana-kanji converters extend this concept and When the pronunciation of word A matches the pronunciation of a series of different word sequences B1B2...Bn, the series of words B1B is used as a homophone of the word A.
2...Bn may be included. For example, the reading of the proper noun ``Hikariben'' is the same as the reading of ``Red'' which is a combination of the adjective stem ``Red'' and the adjective final form ``i'', but with conventional kana-kanji conversion devices, the reading of the proper noun ``Hikariben'' is the same. There is one that displays the kanji notation and the kanji for "red" at the same time, making it possible to select one of them.

Furthermore, there is a certain word series C1...cm reading, and a different word series Dl...
Some devices allow the above-mentioned homophone selection operation even when the pronunciations of Dn match. For example, "Shinkishu"
In response to the pronunciation, it outputs ``new model'', which is a sequence of the noun ``new'' and one suffix, and ``new model'', which is a sequence of the prefix ``new'' and the noun ``model'', and the operator selects one of them. Choose one. In this case as well, the following measures are taken so that the word string selected by the operator is preferentially output next time. That is, when the operator selects "new model" in the above example, the operator selects word 5, which has the longest matching pronunciation sequence from the beginning to the pronunciation sequence "shinkish". temporarily stores three sets in a word learning dictionary: information that uniquely indicates the pair with the word "new" at the beginning of the word string "new model", and information that the word "new" has been selected. . Then, when the pronunciation sequence ``Shinkishu'' is input again, the word search unit detects that the word with the longest matching pronunciation from the beginning of the pronunciation sequence is 1 new word, and furthermore, it detects that the word with the longest matching pronunciation from the beginning of the pronunciation sequence is 1 new word. ” and detects that the word “new” has been previously selected for the word “new”. Then, after converting the first two pronunciations of the pronunciation "Shinkishu" to "shin", the input pronunciation was changed to "kishu", and by searching the word dictionary, the kanji notation "model" was obtained, and finally "new". We obtain the kanji notation for the word string "model".

The conventional kana-kanji conversion device realizes a selective learning function between words having different reading lengths as described above. However, the above-mentioned kana-kanji conversion device is premised on displaying different kanji outputs for the same input kana string.

On the other hand, when the kanji output of a kana-kanji conversion for a certain input kana string is incorrect, a method has been devised in which the erroneous one-word conversion result can be easily corrected by separating the input kana string itself. However, there is no such thing that has realized the word selection learning function described above, and even if the operator changes a certain conversion result to what he or she desires, the next conversion will result in an incorrect conversion again. It outputs the results.

Purpose of the Invention In view of the above-mentioned drawbacks of conventional kana-kanji converting devices, the present invention provides a system that, when an operator modifies the output kanji string of the device to another kanji string, the same reading string is inputted again after that. The purpose is to give priority to the kanji output of the word selected by the operator.

Structure of the Invention In order to achieve the above object, the kana-kanji conversion device according to the present invention stores the input kana by adding it to the end of the currently stored kana string. forward kana column designation means Noni for specifying a subsequence starting from the beginning of the kana column stored in the column temporary storage unit;
The kana strings stored in the kana string temporary storage section are also t, 1. word search means for searching for a word having the longest matching pronunciation from the beginning of the partial kana string; a longest matching word temporary storage section for storing the longest matching word; a maximum longest match jF-word temporary storage unit that stores a maximum reading length of the longest matching word when sequentially activating the word search means for the kana string or the partial kana string in the column temporary storage unit; a word learning information storage unit that stores word learning information including information about the maximum longest matching word and the longest matching word; a word learning information registration unit that registers the word learning information in the word learning information storage unit; searching for the read sequence in the word learning information storage unit;
Figure 1 is a block diagram of a kana-kanji conversion device according to an embodiment of the present invention. In the figure, 1 is a temporary input unit, such as a kana keyboard defined by the Japanese Industrial Standards: Ti SC-6233, a Romanji-kana conversion device, a monosyllable recognition device, or another kana character string storage device. 2. In the ephemeral sub-storage section, add the kana string input from kana input section 1 to the end of the currently stored kana string, and output the resulting kana string, its kana number, and the word search section to be described later. The number of kahikos of the partial kana sequence M from the beginning of the above kana sequence to be
I remember. Here, in the following explanation, the kana sequence input from the simple kana input section 1 is Kiyone (Ah,
ka, me, etc.) are also voiced and semi-voiced sounds (ga, ge, etc.).

It is assumed that the kana sub-storage unit 2 stores all of the sounds such as kya, ja, etc., encoded as one unit. An example of such a coding rule is shown in FIG.

Reference numeral 3 denotes a word storage section in which a large number of sets of Japanese word pronunciation information, kanji notation information, and part of speech information are stored. Here, the word pronunciation information indicates how to pronounce the word. Example 9 He〉゛The word ``lesson'' has the pronunciation information ``ju'', ``gyo'', and ``u''. The rules for encoding this reading information are the same as in the explanation of the kana sub-storage unit 2.
It is assumed that clear sounds, voiced sounds, semi-voiced sounds, and persistent sounds are each coded as one unit. The kanji notation information for a word indicates how to write the word. This is encoded, for example, by a Kanji code defined in the Japanese Industrial Standard JiS C-6226. Furthermore, when the word is normally written in hiragana, this is also referred to as kanji notation information.

4 is a word search unit which searches in the word storage unit 3 for the word whose pronunciation information matches longest with the gana sequence from the beginning of the partial kana sequence M of the kana sequence stored in the kana sub-storage unit 2. Search and obtain reading information V and kanji notation information W of the searched word.
The set is transferred to the longest matching word temporary storage unit, which will be described later. Note that in this embodiment, the word search section 5 is connected to the kana sub-storage section 2.
It is assumed that the operation is activated every time a value is set to the kana number N of the kana string or the kana number N of the partial kana string stored in .

6 is a longest matching word temporary storage unit, which is realized as a single key that stores reading information ■ and kanji notation information W of the word searched by the word search unit 4, and pressing this key causes a kana column front part extraction unit to be described later. Sends activation signal divided by the same. 7. When the ephemeral column front partial extraction section receives the activation signal from the gana column front partial extraction activation section 6, it selects the part starting from the beginning of the kana column stored in the kana sub-storage section 2 or the outer column as the next one. This is specified as follows and sent to the word search unit 4. That is, the value of the number N of kana in the partial kana sequence stored in the kana sub-storage unit 2 is set to a positive number smaller than the current value. In this embodiment, it is assumed that N=N-1. When sending out the partial kana string to the word search unit 4, the pre-column partial extraction unit 7 sends a signal indicating that the front portion of the kana sequence has been extracted to the pre-column, partial extraction storage unit. Send. It has 8 Hagana column front R, and this part Kana column extraction flag H has,
If the kana column front part extracting section 7 sends a signal to the word search section 4 to the effect that the front part kana column has been sent, the logic ゛1 is satisfied.
”, otherwise the logic “0”’ is stored. e is the maximum longest matching word storage unit and has the maximum longest matching reading string register U, and the longest matching word sent out by the word search unit 4 is temporarily stored. The reading sequence ■ of the longest matching word P stored in the storage unit 6 and the information R stored in the kana column front part extraction storage unit 8
is read, and the storage state of the internal longest matching reading sequence register U is set as follows. In other words, R is ○"
When R is 1M, the value of V is stored in U, and when R is 1M, the value of U is retained as before.

In other words, in a state where the kana column front partial extraction unit 7 is not activated, the contents of the maximum longest matching pronunciation string register U are updated to the pronunciation sequence of the longest matching word P, and a state where the kana column front partial extraction unit 7 is not activated. In this case, the above update is not performed.

Reference numeral 10 denotes a word learning information registration unit which, when receiving an activation signal from a word determining unit (described later), records the pronunciation sequence U of the maximum longest matching word stored in the maximum longest matching word temporary storage unit 9 and the temporary longest matching word. A set (U, V, W) of the pronunciation sequence V of the longest matching word stored in the storage unit 6 and its kanji notation information W
is transferred to a word learning information storage unit to be described later.

11 is a word learning information storage unit; a word learning information registration unit 10;
Word learning information (U, V.

W). Further, this information is referred to by a word learning information search unit, which will be described later.

Reference numeral 12 denotes a word learning information retrieval unit which retrieves a pronunciation sequence that matches the pronunciation sequence ■ of the longest matching word stored in the longest matching word temporary storage unit 5 from among the maximum longest matching pronunciation sequence U in the word learning information storage unit 11. search for. This search is successful and the maximum matching reading sequence is U. If you are bitten, read this reading sequence U. If the longest match corresponding to is unsuccessful, the pronunciation sequence of the word stored in the longest match word temporary storage unit 6 and its kanji notation information W are sent to the kanji output unit to be described later.
/・-7 Issue.

Reference numeral 13 denotes a kanji output unit, which outputs the kanji notation information of the word for the pronunciation sequence of the word notified by the word learning information retrieval unit 12 out of the kana sequence stored in the kana sequence temporary storage unit 2; For kana columns other than those listed above, the kanji notation (hiragana part of the kanji code) is output. This output format is, for example, CR
It may be displayed on the T screen.

Reference numeral 14 denotes a word determination unit, which is realized by one function key, and when this key is pressed, the word learning information registration unit 10
is started. Also, the contents of the kana column temporary storage section 2 are cleared. This key is pressed when the operator confirms that the word being displayed on the kanji output section 13 is the desired word.

In the apparatus configured as described above, the operation when performing kana-kanji conversion will be described below using the flowchart shown in FIG. If the process number following each process is not specified, the process proceeds to the process with the next process number in alphabetical order.

[Normal input]

14. (a) (Kana dynamic input/memory) Input a kana string from the kana input section 1 in the input waiting state. The input kana string is additionally stored in the kana sub-storage section 2.

The stored kana fiL of the kana string is set in the internal register N representing the number of kana in the partial kana string. Furthermore, the partial kana column extraction flag R of the kana pre-column partial extraction storage section 14 is set to logic "o".

(B) (Word search) Word search section 4 searches the word P that has the reading information that matches the longest kana string from the beginning of the partial kana string M stored in the ephemeral sub-storage section 2. The storage unit 3 is searched, and the pair of reading information V and kanji notation information W of the searched word P is stored in the longest matching word temporary storage unit 6.
Transfer to.

H (maximum longest matching word update) The maximum longest matching word temporary storage section 9 stores the reading information V of the longest matching word P stored in the longest matching word temporary storage section 6 and the kana column front part retrieval storage section 8
The information R stored in is read, and the longest internal match is 16 bases.

The storage state of the reading sequence register U is set as follows. That is, when R is °'0'', the value of ■ is set for U, and when R is 1'', the value of U is maintained as before.

(2) (Learning information search) The word learning information searching unit 12 converts the longest matching word pronunciation ■ stored in the longest matching word temporary storage unit 5 into the longest matching pronunciation sequence U in the word learning information storage unit 11.
Search within. This search was successful, and the longest matching reading string was U. If found, the second word pronunciation sequence V corresponding to this pronunciation sequence U0 and its kanji notation information W. is sent to the kanji output unit 13. On the other hand, if the above-mentioned search is unsuccessful, the pronunciation string (3) of the word stored in the longest matching word temporary storage section 5 and its kanji notation information W are sent to the kanji output section.

(4) (Kanji notation output) Of the kana strings stored in the kana string storage section 2 of the kanji output section 13, the pronunciation sequence of the word notified by the word learning information search section 12 is the kanji notation of this word. The information is output, and the kanji notation (hiragana) for the character strings other than the above is output. After this, it enters an input waiting state.

[Kana row front part taken out]

(F) (Kana front part extraction) When the Kana front part extraction starting unit 6 is activated in the input waiting state, the Kana front part extraction unit 7 is activated. And kana row front part extraction part 7
The value of the kana number N of the partial kana string stored in the ephemeral string storage section 2 is set to N-1. Further, the value of the partial kana column extraction flag H of the kana pre-column portion extraction unit 7 and the kana pre-column portion extraction storage unit 8 is set to logic II, II.

After this, proceed to (b).

[Word determination]

(G) (Word determination) In the input waiting state, the function key of the word determination section 14 is pressed.

(-1 (word learning information registration) At this time, word learning registration section 1
o is the pronunciation sequence U of the maximum longest matching word stored in the maximum longest matching word temporary storage unit 9, the pronunciation sequence of the maximum 17 page long matching word stored in the longest matching word temporary storage unit 5, and its kanji. Notation information W and silk (U
, V, W) are stored in the word learning information storage section 11.

Further, the word determining section 14 sends a clear signal to the kana string temporary storage section 2. After this, it enters an input waiting state.

A specific example of the operation of the kana-kanji conversion apparatus of this embodiment configured as described above will now be described.

Now, assume that the operator creates the Japanese sentence ``Don't run.'' To do this, the operator inputs the reading of the above Japanese sentence in kana from the kana input section. “Ha” “
The Honkana-Kanji converter operates as follows when the kana characters ``shi'', ``u'', ``su'', and ``i'' are input in sequence. Figure 4 shows the state of each part of the Honkana-Kanji converter during operation.

In FIG. 4, one example is the input kana column, the storage state of the kana column in the kana column storage section 2, the L column, the kana number 2M column in the kana column storage section 2, the storage state of the kana column in the same section, Column N is the kana number of the partial kana column in the same section, and column R is the partial kana column extraction flag in the partial extraction storage section 8 before the kana column.

The 7th column is the reading information V of the longest matching word P in the longest matching word temporary storage unit 5, the W column is the kanji output information W of the longest matching word P in the same unit, and the o column is the word learning information from the word learning information search unit 12. Single reading sequence of search results when searching the storage unit 11, Wo
The column is the word kanji notation information of the search result of the same section, and the 0 column is the output kanji notation information of the kanji output section.

The portions in parentheses following each description of the specific example below indicate the location in FIG. 4 that corresponds to that description.

In the initial state, the storage contents of the kana column storage section 2 are cleared, and the register H of the kana column front part fetching storage section 8 is set to logic "0".

■ First, when the operator inputs the kana ``ノ・'', the kana string K ``ノ・'' is stored in the kana string storage section 2.

Also, the partial kana column M becomes "No1" (column 2M in the first row of FIG. 4).

■ The word search unit 4 searches the kana string “ha” in the word information storage unit 3.
Search for words that have as a pronunciation sequence.

If the storage state of the word information storage unit 3 is as shown in FIG. 5, the longest matching word P is searched for the word with the reading 19-1 column "ha" and the kanji notation information "teeth".

(Figure 4, row 1, column V, column W) ■ The maximum longest matching word temporary storage unit 9 is the maximum longest matching reading string register because the value of register H of the kana column front partial extraction storage unit 8 is “o”. Let the value of U be "ha" (FIG. 4, row 1, column U).

(2) The word learning information search section 12 searches for the pronunciation string "ha" stored in the longest matching word temporary storage section 51/C among the longest matching pronunciation strings in the word learning information storage section 11. FIG. 6a shows an example of the storage state of the word learning information storage section 11. In this example, since "Ha" is not searched in the U-th column, the word learning information search section 12 searches for the pronunciation string "NO" of the word stored in the longest matching word temporary storage section 6 and its kanji notation information. "Tooth" is output to the kanji output unit 13 (FIG. 4, row 1, column V, column Wo).

■ The kanji output unit 13 outputs the kanji notation information "teeth".

■ After this, the next kana character ``shi'' is waiting for input.
When input, the processing shown in FIG. 3 (a) to (e) is performed.

As a result, the internal state of the Honkana-Kanji converter becomes as shown in the second line of FIG. 4.

(2) If ``u'', ``su'', or ``i'' are input after this, the processes shown in FIG. 3(a) to 0u are performed, respectively. The internal state of the apparatus after this process is completed is as shown in lines 3 to 5 of FIG. 4.

■ In this state, the output state of the kanji output unit 13 is "No Pillar"
and is different from what the operator desires. Thereupon, the operator presses the function key which is the kana column front part extraction starting section 6.

(2) At this time, the pre-kana column partial extractor 7 sets the value of the kana number in the partial kana column to 4, which is one less than the current value (FIG. 4, row 6, columns M and N).

■ After this, set the value of the partial kana column extraction flag R in the kana column front part extraction unit 7 in the kana column front part extraction storage unit 8 to logic ``1#'' (row 6, column R in Figure 4). ・0 This After that, this device performs the processes shown in Figure 3 (B) to (E). Here, in the process shown in Figure 3 (C), since the 21 page value of H above is 1", the maximum longest matching word reading sequence is maintained at its current value "Hashira" (row 6, column U in FIG. 4).

The internal state of the device after this process is completed is as shown in line 6 of FIG. 4.

■ The kanji output state at this point is ``J without pillars'', which is not what the operator desires.Therefore, the operator further presses the kana column front partial extraction function key.

0 At this time, the kana column front partial extraction section 7 sets the value of the kana number N of the partial kana column to a value 1 less than the current value 4, that is, 3 (row 7, column M, column N in FIG. 4).

0 After this, the device performs the processes (b) to (e) in FIG. The internal state of the apparatus after this processing is completed is as shown in line 7 of FIG. 4.

[Phase] Since the kanji output state in this state is "no pillar", which is not what the operator desires, the operator further presses the kana column front partial extraction function key.

[Phase] At this time, the value of N is 2, and the partial kana sequence becomes "hashi". After this, the apparatus performs the processes shown in FIGS. 3(b) to 3(e). What should be especially noted here is that in the process of Figure 3 (c), the value of H mentioned above is °°1'', so even though the reading of the longest matching word in this case is "hashi", the maximum The longest matching word pronunciation string is kept at its current value "Hashira".

After this process is completed, the internal state of the device becomes as shown in line 8 of FIG. 4.

■ The kanji output state at this point is "not running", which is what the operator desires. Then, the operator presses the word determination function key which is the word determination section 14.

■ At this time, the word learning registration unit 1o has the maximum longest matching word pronunciation U "Hashira" and the current longest matching word pronunciation ■
"Hashi" and its kanji notation information W "Hashi" pair ("Hashira"
, "Hashi", "Hashi") are stored in the word learning information storage section 11. At the end of this process, the word learning information storage unit 11
The storage state of is as shown in Figure 6.

23 b.

Further, the word determining section 14 sends a clear signal to the kana string temporary storage section 2 (line 9 in FIG. 4).

Next, we will explain the operation of the Honkana-Kanji converter when the operator inputs the kana letters ``ha'', ``shi'', ``u'', ``su'', and ``i'' sequentially in order to obtain the kanji string ``runai'' again. do.

@ When the operator inputs the kana ``ノ'', this device operates in the same manner as explained in the above ① to ①. The internal state of the device at the end of this process is as shown in line 10 of FIG. 4.

@ Furthermore, when the operator inputs "shi", the above ■
It works the same as described in . The internal state of the apparatus at the end of this process is as shown in line 11 of FIG. 4.

(2) If "U" is further input, the word search section 4 searches the word information storage section 3 for a word having the kana string "Hashira" as a pronunciation string. At this time, a word whose pronunciation is "hafuu" and whose kanji notation information is "pillar" is searched as the longest matching word P (FIG. 4, row 12, column V, column W).

The maximum longest matching word temporary storage unit 9 sets the value of the maximum longest matching reading string register U to ``hashira'' since the value of the register H of the kana sequence front part extraction storage unit 8 is "o" (see Fig. 4). 12th row, column U).

[Phase] The word learning information search section 12 searches for the pronunciation string "No Shira" stored in the longest matching word temporary storage section among the longest matching pronunciation strings in the word learning information storage section 11. 6th
In the storage state of the word learning information storage unit 11 shown in the figure, "Hashira" is detected in the third row of the U-th column. At this time, the word learning information search unit 12 searches V in the third row. Column "Hashi" and its kanji notation W. “Kan” is output to the kanji output unit 13.

■ The kanji output section 13 reads from the kana string storage section 2 that the current partial kana string is "Hashira, J," and further reads the kana string V sent from the word learning information search section 12. "Hashi" and its kanji. The notation W0 is read and the kanji notation information 1 is outputted from this information.

(Figure 4, row 12) 26 A-',: ■ After this, when "su" and "i" are input, the same processing as in the above ■ ~ @ is performed, and the output of the kanji output unit 13 is performed. ``Don't run'' respectively.

"Do not run" (column 13th row and 14th row in Figure 4)
rows and columns).

As described above, the operator's desired kanji notation ``runai'' can be obtained.

As described above, according to this embodiment, when the operator desires the kanji notation for ``runnai'', the operator first inputs the kana ``ha'', ``shi'', ``u'', ``su'', and ``i'' in sequence. The kanji output notation for is ``Hashira nai''. Following this, when the function key for extracting the front part of the kana column is pressed three times, the kanji output notation becomes ``no run''. If you press the word confirmation key here, the longest matching word reading string ``Hashira'' and the longest matching word reading string ``Hashi'' and its kanji notation information at the time the word confirmation key was pressed are displayed. The set is stored in the learning word information storage section. And after this “
When inputting kana such as ``c'', ``shibo'', ``uj'', ``su'', and ``i'' in sequence, the states of the above-mentioned pairs stored in the word information storage section are called up, and in this case, Even if you do not press the ephemeral column front partial extraction key, the kanji output notation will be ``do not run''.

Although the present embodiment describes a case where a kana string is converted into a kanji string every time one character is input from the ephemeral input section, the present invention is not limited to this. For example, the kana input section and the kana string temporary storage section may be operated asynchronously with other components.

In this embodiment, a word search is performed only once from the beginning of the kana string stored in the kana string temporary storage section, and no subsequent Japanese grammatical processing is performed. but,
The present invention is not limited to such a configuration, but may include another component that performs grammatical processing on word candidates obtained as a result of a word search.

In addition, in this example, narrow-sense homophones that have the same pronunciation, such as ``shu'', ``hashi'', and ``edge'', are discussed. It goes without saying that conventional techniques for storing, registering, and searching can be applied.

Effects of the Invention The kana-kanji conversion device of the present invention includes a forward kana column specifying means for specifying a partial kana string from the beginning of the kana string stored in the kana string storage section; a word search means for searching for a word having the longest matching pronunciation from , a longest matching word temporary storage section for storing the longest matching word, and a kana string or kana string in the kana string storage section whose length is different by at least one character; a maximum longest-number storage unit that stores the maximum value of the reading length of the longest matching word when the word search means is sequentially activated for this partial kana string; This device is equipped with parts that memorize, register, and search word learning information consisting of information, so that when the % character output result of this device is not what the operator desires, the operator can specify a part of the input kana string and read the kanji again. Displays the output result, and furthermore, when the operator notifies that the kanji output notation of the word being displayed is the desired one, the learning information of the word is registered in the word learning information storage unit, and the word search means The longest matching word obtained by is searched in the word learning information storage section, and if a matching word is found, the second word stored attached to the word is processed preferentially.

Therefore, once word learning information is registered, if the same character string is input, the kanji notation of the word selected by the operator last time will be output preferentially, and the practical effect is big.

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing the configuration of an embodiment of the kana-kanji device of the present invention, Fig. 2 is a diagram showing an example of the reading encoding rule in the embodiment, and Fig. 3 explains the operation of the embodiment. FIG. 4 is a flowchart showing a specific example of the state of each component of the same embodiment, and a diagram showing an example of a storage state of a storage section. 291, -L' Retrieval unit, 8...Kana row front part retrieval storage unit, 9.
...Maximum longest matching word temporary storage section, 1o...
. . . word learning information registration unit, 11 . . . word learning information storage unit, 12 . . . word learning dictionary search unit.

Claims (1)

[Claims] A kana string storage section that stores the input kana string by adding it to the end of the currently stored kana string, and a forward kana string that specifies the subsequence starting from the beginning of the kana string that is stored in the kana string storage section. a specifying means, a word search means for searching for a word having the longest matching pronunciation from the beginning of the kana string or the partial kana string stored in the kana string storage section, and a longest matching word temporary storage for storing the longest matching word. When sequentially activating the word search means for the kana strings or the partial kana strings in the kana string storage section whose kana string lengths differ by at least one character, the maximum value of the reading length of the longest matching word is determined. a maximum longest matching word temporary storage section to store; a word learning information storage section storing word learning information consisting of information of the maximum longest matching word and the longest matching word; and a word learning information storage section storing the word learning information. A kana-kanji conversion device comprising: a word learning information registration section for registering; and a word learning information retrieval section for searching the word learning information storage section for a designated pronunciation sequence and retrieving the corresponding word learning information.