JPS61229152A - Homophone learning device - Google Patents

Abstract

PURPOSE:To realize Kana (Japanese syllabary)-to-Kanji (Chinese character) conversion which can learn the using frequency of a word among its homophones in a short period, by providing a word dictionary, homophone order storing section, and maximum-frequency word selecting section and retrieving them in the order of maximum-frequency word selecting section, homophone order storing section, and word dictionary. CONSTITUTION:Kana codes inputted from a Kana input section 1 are transmitted to a character converting section 3 where Kana-to-Kanji conversion is made. Firstly, a maximum-frequency word selecting section 9 which stores a word having the maximum referred frequency among its homophones in a homophone order storing section 4 is retrieved. When an aimed word is not contained in the section 9 or when the conversion is inappropriate, the homophone order storing section 4 is then retrieved. In this case, the retrieval is made in the order opposite to the storing order. If the conversion is inappropriate or the aimed word cannot be found even when the section 4 is retrieved, a word dictionary 2 is retrieved. When the dictionary 2 is retrieved, a frequency updating section 7 and homophone selecting section 5 are actuated and the retrieval is stored in the homophone order storing section 4 and its using frequency is updated.

Description

[Detailed Description of the Invention] Industrial Field of Application The present invention is a method for selecting a specific word from among a plurality of homonyms when inputting an input kana character string into kana-kanji conversion as a corresponding Japanese kanji-kana mixed notation. This is a homophone learning device for efficiently selecting homophones.

Background of the Invention In recent years, kana-kanji conversion devices have been particularly popular as devices that can easily create Japanese documents using kana and human power, and are equipped with a homophone learning mechanism to quickly select the correct word from homophones. Some have been developed.

For example, what is called a homophone usage frequency learning device changes the search order specification information for homophones in a word dictionary based on the usage frequency of words used in the word dictionary to quickly obtain frequently used words. is possible.

Still, what is called a homophone usage order learning device stores the usage history information of that word in order and inputs it into an input kana column every time a specific word is selected from words in a word dictionary that have multiple homophones. On the other hand, by outputting the most recently selected homophone as the first search word, the previously selected homophone as the second search word, and so on, you can quickly obtain recently used words. .

Conventionally, when the functions of learning the frequency of use of homophones and learning the order of use of homophones were to be implemented together, corresponding mechanisms were provided for each.

For example, in homophone use order learning, word identifiers that can uniquely limit a word are stored in an order in a homophone order memory, and when searching for a word, this homophone order memory is sequentially searched to find homophones. It was possible to learn the order of use.

In addition, homophone usage frequency learning is done by updating the usage frequency information in the word dictionary every time a homophone is selected, and searching for words in homophones in the order sorted by this usage frequency. By having these two mechanisms, it is possible to perform both homophone usage order learning, which is meant for short-term learning, and homophone frequency learning, which is meant for long-term learning.

(For example, JP-A-52-132653, JP-A-5
5-153033, etc.) ゛ 1. However, in such conventional homophone learning devices,
Even if frequency learning and order learning are combined, priority is given to order learning in short-term learning, so it is not possible to perform short-term frequency learning according to individual documents.

A conventional homophone learning device will be explained below. FIG. 3 is a block diagram showing the configuration of a conventional homophone learning device.

1, a phonetic character string such as Japanese hiragana and katakana, a character conversion instruction code z1, and a hiragana display instruction code z2 are inputted in the ephemeral input section.

Reference numeral 2 is a word dictionary which stores Japanese word pronunciations and corresponding word notations in association with each other and arranged in alphabetical order.

Reference numeral 3 denotes a character conversion unit which sequentially searches the input kana character string from the kana input unit in the homophone order storage unit 4 (described later) and the word dictionary 2 to obtain the corresponding word notation. At this time, the search in the homophone order storage section 4 is performed in the opposite order to the order in which the words are stored in the homophone order storage section 4. Furthermore, in the search of the word dictionary 2, if there are a plurality of words with pronunciations that match the input kana string, the words are searched in descending order of frequency and output together with the input kana string to the homophone selection section 5, which will be described later.

Reference numeral 4 denotes a homophone order storage unit which stores homophones selected from a homophone selection unit 6, which will be described later, in the order in which they are selected, with their pronunciations and spellings associated with each other, in a first-in, first-out manner.

Reference numeral 5 denotes a homophone selection unit, which selects the next candidate instruction code z3 from the homophone specification device 9, which will be described later, when there are multiple word expressions output from the character conversion unit 3, that is, when there are homophones and allographs. Depending on the input, the display unit 6 (described later)
This allows the operator to select the desired notation.

A display section 6 displays the word notation selected by the word selection section 4 on the screen.

Reference numeral 7 denotes a frequency updating unit which adds 1 to the frequency of the specified word in the word dictionary 2 from the homophone selection unit 6.

Reference numeral 8 denotes a homophone selection device which outputs a next candidate specification code z3 and a confirmation code z4 indicating completion of selection among homophones to the homophone selection device 6.

The problems with this conventional homophone learning device are shown below using FIG. Figures 4d and 4b show the contents of a document appearing on the display while the document is being created using a conventional homophone learning device. Further, C in the figure shows the contents of the frequency storage section for frequency learning.

That is, for items that have not been learned in order, they are output in the order of frequency listed in this frequency storage unit. In this example, "publication", "tourism", "practice", "completion", "
Output in the order of "Dare".

FIG. 4a shows the display state of a document in the middle of creation. At this time, if we focus on the word with the pronunciation "Kanko" stored in the homophone order memory, we can see that the words "sightseeing", "sightseeing", and "custom" have been stored in the homophone order memory so far.
Homophones are stored in this order. In normal document creation, it often happens that multiple types of homophones are selected, but generally there is a slight deviation between these homophones in terms of frequency of use. In the document shown in Figure 4, ``tourism'' is used with priority over ``practice.''

Now, at this time, when the operator uses the word "sightseeing" again, the conventional homophone learning device simply prioritizes the most recently used homophone, based on the homophone order learning process that takes priority in short-term learning. For output, "Practical" is selected as the first candidate and displayed on the screen. Naturally, the operator must perform processing to select another candidate.

Figure 4 shows that although there is a considerable difference in the frequency of use of each homophone that appears in the table document of the document at the end of this process, conventional homophone learning devices Since homophones are displayed with priority, there is a problem in that words that are originally desired to be displayed are postponed. The present invention has been made in view of the above problems, and has a simple configuration that performs short-term usage frequency learning according to the usage status of homophones during document creation, and efficiently prioritizes output of homophones. The purpose of the present invention is to provide a homophone learning device that makes it possible.

Means for Solving the Problems In order to achieve the above-mentioned goals, the present invention provides a word dictionary that stores at least the pronunciations and spellings of Japanese words, and a kana input section for inputting Japanese pronunciations in kana. a homophone selection unit that selects any one from a plurality of homophone selection units; a homophone order storage unit that stores the selection order of words selected by the homophone selection unit; and a homophone order storage unit that stores the selection order of the words selected by the homophone selection unit; a maximum frequency word selection section that counts the number of duplicates of stored words and searches for the word with the largest number; and a maximum frequency word selection section that counts the number of duplicates of stored words and searches for the word with the largest number; The homophone learning device includes a homophone order storage unit, and a character conversion unit that searches the word dictionary in order and converts an input kana string into a kanji-kana-mixed sentence.

According to the above-mentioned configuration, the present invention can be easily implemented by using only the word usage information stored in the homophone order storage unit by the function of the maximum frequency word selection unit that can count the number of duplications of words in the homophone order storage unit. short-term frequency learning can be performed. This makes it possible to combine short-term homophone usage frequency learning with homophone usage frequency learning.

Embodiment FIG. 1 is a block diagram of a homophone learning device according to an embodiment of the present invention. 1 In the ephemeral input section, Japanese hiragana, katakana, etc. phonetic character strings and character conversion instruction code z1,
The hiragana display code z2 is input as a code.

Reference numeral 2 is a word dictionary which stores Japanese word pronunciations and corresponding word notations in association with each other and arranged in the order of the pentagrams.

3 is a character conversion unit which sequentially searches the input kana character string from the kana input unit in the maximum frequency word selection unit 9 (to be described later), the homophone order storage unit 4 (to be described later), and the word dictionary 2 to find the corresponding word notation. It's something you get. At this time, the search in the homophone order storage section 4 is performed in the opposite order to the order in which the words are stored in the homophone order storage section 4. In addition, when searching the word dictionary 2, if there are multiple words with pronunciations that match the input Kana string, these are searched in descending order of frequency and are sent to the homophone selection section 6, which will be described later, along with the input Kana string. .

Reference numeral 4 denotes a homophone order storage unit which stores homophones selected from a homophone selection unit 6, which will be described later, in the order in which they are selected, with their pronunciations and spellings associated with each other, in a first-in, first-out manner.

Reference numeral 6 denotes a homophone selection unit, which inputs the next candidate instruction code z3 from the homophone specification device 9, which will be described later, when there are multiple word expressions output from the character conversion unit 3, that is, when there are homophones and allographs. are sequentially displayed on a display section 6 to be described later, allowing the operator to select a desired notation.

A display section 6 displays the word notation selected by the word selection section 4 on the screen.

Reference numeral 7 denotes a frequency updating unit which adds 1 to the frequency of the specified word in the word dictionary 2 from the homophone selection unit 5.

Reference numeral 8 denotes a homophone selection device which outputs a next candidate specification code z3 and a confirmation code z4 indicating completion of selection among homophones to the homophone selection device 6.

Reference numeral 9 denotes a maximum frequency word selection section which counts the number of homophones in the homophone order storage section 4, obtains the one with the largest value, and outputs it to the character conversion section 3.

The operation of the homophone learning device of this embodiment configured as described above will be explained below using the flowchart shown in FIG.

〔Character conversion〕

(1) The kana string entered by the operator through the kana input section 1 is sent to the character conversion section 3 in the form of a conversion instruction code z1 or a hiragana display instruction code z2 appended to the end.

('4 Character converter 3 If the end of the input kana string from the ephemeral kana input unit 1 is a hiragana display instruction code z2, the hiragana notation of k is output to the homophone selection unit 6 as a word notation. In other words, the control is Move on to (6).

(3) The character conversion unit 3 sequentially searches the homophone order storage unit 4 and the word dictionary 2 for the input kana string k, and sends words with the corresponding pronunciation to the homophone selection unit 6 along with the input kana string in the order of search. Put it out. At this time, if there are multiple homophones, all of them are output to the homophone selection section 5.

(4) The homophone selection unit 5 first determines whether the word received from the character conversion unit 3 includes a homophone or allograph.

If it is included, control is transferred to process (@). If not, the following process (6) is performed.

(6) The display unit 6 outputs the notation of the word delivered from the homophone selection unit 6 to the display unit 6 and displays it on the screen, and then enters an input waiting state.

(6) The homophone selection unit 6 outputs the first received word notation among the handed over homophones to the display unit 6 and displays it on the screen, and also provides guidance for the operator to select the homophone on the display screen. After issuing the command, it enters the input waiting state.

[Homophone selection]

(7) When the operator inputs the next candidate instruction code z3 from the homophone selection device 9 while the guidance for homophone selection instruction is displayed, the homophone selection section 6 selects the sentence (8).The operator selects the desired word notation on the screen. Once obtained, the confirmation code z4 is input from the homophone designation device 9. Then, the homophone selection unit 6 outputs the selected word notation to the homophone order storage unit 4 along with the input kana string.

(9) The homophone order storage unit 4 stores the selected homophones in the order in which they were selected. At this time, the homophone order storage unit 4 stores latest word position information indicating which word is the latest stored word.

After storing the word, it will enter the input waiting state again.

Here, the operation of the present homophone learning device will be further explained using an input example. What is shown in FIG. 5a is only sentences with the pronunciation of ``kanko'' extracted from the contents of the homophone order storage unit 4 at the time when the input of example sentences shown in FIG. 4a is completed. Furthermore, what is shown in FIG. 6 is only those with the pronunciation of "Kanko" extracted from the contents of the homophone order storage unit 4 immediately after the example sentence shown in FIG. 4 is input.

At the point in Figure 6a, the operator, with the reading ``Kanko'', aims to scan the character string ``Tourism''.
When the corresponding pronunciation is entered using the kana input device 1, the character conversion unit 3 first obtains the notation “Kanko J” from the maximum frequency word selection unit 9. Among the words with the reading ``sightseeing'' is 2.
This is because there was only one ``practice''. If the operator wants to input the word "custom" that was input just before, by giving a command to the homophone selection unit 6, the character string "custom" can be obtained as the next candidate.

In this way, due to the effect of short-term frequency learning, not only can appropriate words be prioritized and displayed as the first candidate, but also the most recently used word can usually be selected as the second or third candidate, so the usage order It can be said that it also has a learning effect.

Furthermore, according to this embodiment, conventional frequency learning is also performed at the same time, so if there is no word to refer to in the homophone order storage unit 4, homophones can be output preferentially based on their long-term frequency of use. I can do it.

Effects of the invention As is clear from the above explanation, this homophone learning device has the following effects:
Homophone usage frequency learning and short-term homophone usage frequency learning are realized in a compact configuration without having frequency information in the word dictionary or preparing a special storage device for frequency learning. You can.

Furthermore, as shown in the embodiment, since the method is based on the method of "prioritizing and outputting words that have been frequently used recently", desired homophones can be accurately and preferentially displayed.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a homophone learning device according to an embodiment of the present invention, FIG. 2 is a flowchart showing the operation of the homophone learning device according to an embodiment of the present invention, and FIG. 3 is a conventional homophone learning device. Figures 4a and b are diagrams showing how the display unit displays during document creation, Figure 4c is a diagram showing the frequency of homophones in the word dictionary, and Figures 6a and b
1 is a diagram showing extracted contents of a homophone order storage unit during document creation. 1...Kana input section, 2...Word dictionary,
3... Character conversion unit, 4... Homophone order storage unit, 6... Homophone selection unit, 6...
Display unit, 7... Frequency update unit, 8... Homophone selection instruction device, 9... Maximum frequency word selection unit. Name of agent: Patent attorney Toshio Nakao and 1 other person No. 1
Figure 2 Figure 3 Figure 4 (a-) (b) Figure 5 (a) I (b) [-*

Claims (1)

[Claims] A word dictionary that stores at least the pronunciation and notation of Japanese words, a kana input section that inputs Japanese pronunciations in kana, and a homophone selection that selects any one from multiple homophones. a homophone order storage unit that stores the selection order of words selected by the homophone selection unit; and a homophone order storage unit that stores the selection order of words selected by the homophone selection unit; and a word with the largest number by counting the number of duplicates of words stored in the homophone order storage unit. A maximum frequency word selection unit that searches for a substring of the input kana string from the kana input unit searches and inputs a pronunciation that matches a subsequence of the input kana string in the order of the maximum frequency word selection unit, the homophone order storage unit, and the word dictionary. A homophone learning device characterized by having a character conversion unit that converts a kana sequence into a sentence containing kanji and kana.