JPS5835306B2 - Word information search output device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides a memorization method that is effective when inputting a conjugated form different from the original form of a word in a desired language and searching for information such as a translated word corresponding to the input information in, for example, an electronic dictionary. related to.

Traditionally, in regular dictionaries, regular inflections of words (such as verb endings) are not shown as headwords, and if the word you want to look up has a conjugated form, you can guess its original form. , I had to investigate it in its original form.

For example, if you do not know the meaning of the German word BMchern, you cannot look it up even if you look up the word itself in a dictionary.

In English, you cannot look up the word car ries in a dictionary.

The former is Buch's third case plural, and the latter is the third person singular present tense of carry (however, the part of speech of carry is only the verb).

In such cases, it may take some time to find out the meaning of the word you are looking for by searching the nearby headwords in order, but if you want to find out the meaning of the word in the above example, It is difficult to investigate the original form without knowing the changes in usage in advance.

In addition, devices such as electronic dictionaries that store words with such conjugated forms as digital information, detect matches with the spelling of the introduced word, and derive necessary information such as its translation. , since the detection is so-called digital, as mentioned above, it is rather complicated to randomly look up words with different spellings in the vicinity, like a person looking up a dictionary, so it is necessary to introduce conjugated forms into the device. Taking into consideration the situation in which these words occur, we have no choice but to introduce words that include their conjugated forms in headwords.

However, if you include the conjugated form in the headword, for example, English nouns require nearly twice as many headwords, and for verbs, in addition to the present tense, past/past participle (mostly present tense + ed form), present participle, third person singular present There are as many conjugations as there are shapes, and about four times as many headwords are required.

Considering the conjugations of other adjectives and adverbs, the total number of headwords is quite large.

In languages with many external conjugations, for example Nepanese verbs, the indefinite form of l has about 50 conjugations, so the number of headwords is enormous. To provide a word storage method suitable for an electronic dictionary, etc., which outputs information such as a translation corresponding to a conjugated form when an operator inputs a conjugated form without introducing a form.

Hereinafter, one embodiment of the present invention will be described in detail using the drawings.

FIG. 1 shows a block diagram of an electronic dictionary using a word memory formula according to the present invention.

Figure 2; shows the classification of regular changes in conjugated forms of verbs in English words.

As shown in the figure, the regular inflection endings of verbs can be classified into several rules.

In the same figure, there are five verbs A to E, but the number can be reduced depending on the verbs included (for example, if verbs with an e at the end of the present tense are not included, rule D in Figure 2 can be omitted). ).

Also, some English verbs have irregular conjugations of present tense, past tense, and past participle; in other words, the past tense and past participle forms can be formed by adding ed or d to the present tense. There are some verbs that do not exist, such as go -went -gone, but these verbs with irregular conjugations need to be recorded as headwords.

Next, rules A to B will be explained.

The present tense of C is used as a headword.

Rule D. E has a different present tense and headword.

For the third person singular present tense, rule A adds s1, rule C1D adds es, and rule E adds ies.

The present participle is ing for rules A, C, and D, and + for rule B.
: lng (: * is the last letter of the headword spelling), Rule E adds ying.

The past/past participle is ed in rules A, C, and D, and rule B
In this case, add I 1ed, and in rule E, add ied.

The figure also shows examples of words that follow the above rules of conjugation.

Here, example (■) is a verb whose forms other than past tense and past participle form change regularly, and only the irregular part needs to be provided as a headword.

Next, the block diagram of FIG. 1 will be explained.

In the figure, reference numeral 1 denotes a register that temporarily stores words to be searched that are introduced from the outside by means such as manual input.

In addition to the original form, the inflected form may be input as is into the register 1.

2 is a storage device for storing headwords, and 3 is the register 1 mentioned above.
This is a detection device that detects a match between the contents of the storage device 2 and the output contents of the storage device 2.

4 is an address counter that updates the addresses of the storage device 2 and storage devices 5 and 6; 7 is a signal that instructs to search for a word introduced into the register 1; when the signal 7 is introduced, the address counter 4 is updated; will be reset.

The above-mentioned storage device 5 stores the usage rules for headwords (rules in FIG. 2), and the storage device 6 stores necessary information such as the part of speech and translation of the headword.

Further, the detection circuit 3 outputs when the output contents of the storage device 2 are included in the contents of the register 1.

For example, when the storage device 2 is outputting tr information and the contents of the register 1 are trace, a match signal 8 is output.

If they do not match, a mismatch signal 9 is output.

Reference numeral 10 denotes a synthesizing device that creates inflected forms from the storage devices 2 and 5 in accordance with the coincidence signal 8.

Reference numeral 11 denotes a coincidence detection device for detecting coincidence between the output of the synthesis device 10 and the output of the register 1.

If they match, a match signal 12 is output, and if they do not match, a mismatch signal 13 is output.

The discrepancy signal 13 is sent to a synthesizer 10.

In the synthesizer 10, when the mismatch signal 13 is sent, the inflection is changed and the information is transferred to the match detection circuit 11 again.The synthesizer 10 further counts the number of signals 13, and uses the counted number as the signal 1.
4 and is sent to the display unit 15.

Further, when the synthesizer 10 receives the mismatch signal 13, a signal 16 is outputted to update the address counter 4.

Next, specific operations will be explained based on the above block diagram.

Suppose that the English word asked is introduced into register 1.

Next, the address counter 4 resets the address by the search instruction signal 7 and updates the address until the output of the storage device 2 is the same as or includes the contents of the register 1.

Assuming that the headword of the storage device 2 is as shown in FIG.

) When the output of the storage device 2 becomes ask, the content asked of the register 1 includes the output of the storage device 2, so the coincidence detection device 3 outputs a coincidence signal 8, and the address counter 4 stops counting.

The synthesis device 10 is driven by the coincidence signal 8 mentioned above.

That is, a conjugated form is created using the output ask of the storage device 2 and the rule master outputted from the storage device 5.

In the synthesizer 10 (without ending), (sL (ing).

(ecj) is added to the end of ask to create and output the conjugated form.

First, ask is outputted as is to the coincidence detection device 11, and a coincidence with asked, which is temporarily stored in the register 1, is detected.

In this case, since there is a mismatch, the coincidence detection device 11 sends a mismatch signal 1.
3 and instructs the synthesis device 10 to output the next conjugated form,
The synthesizer 10 outputs the next conjugated ask.

In this way, when asked is output to the coincidence detection device 11, a coincidence signal 12 is obtained from the coincidence detection device 11,
Search ends.

At this time, the output of the storage device 6 instructed by the address counter 4 is information regarding parts of speech, translations, etc., and the first inflected form output from the synthesizer 10 (if the present tense is the last output If a register 17 is provided to store the last one), it is possible to output the original form, and the function to count the number of times the signal 13 occurs (however, this count is reset by the signal 8) allows the output of the utilized form. By keeping the output order constant, the usage content can also be derived from the synthesis device 10.

On the other hand, for example, the word carriage is in register 1.
, the match detection device 3 detects a match with the unchanged part of carry because it contains carr, but since the output of the synthesizer 10 wants to match the contents of register 1, the synthesizer 10 outputs the signal 16. is output ca
The match detection device 3 detects a match between the word next to rr and "carriage" again.

Furthermore, in cases where the past participle cut is introduced into the register 1, a match is detected as the present tense, but in such a case, information regarding the content of usage may be stored in the storage device 6.

The above is a list of English verbs, but nouns (singular/plural),
The same goes for adjectives (original, comparative, and superlative); the number of headwords can be greatly reduced, and the storage space can be made smaller, which also shortens the average time required for retrieval.

For example, the conjugated form of fly is ny from the synthesizer.

It prints f 1ies, flying, fl ied, where flied is actually not the correct word but fly
The past tense of is flew1, and the past participle form of is flown.

However, in reality, the word "fried" is never input, so there is no matching word, and there is no problem in use.

Furthermore, the display contents of the display unit 15 include the original form of the word and its translation by receiving the contents of the storage devices 2 and 6, and what the original form of the searched word corresponds to by receiving the output signal 14 of the synthesis device 10. Display whether to do so.

For example, if you search for 5topped, you will find “stopped 5TOP”.
"Past/past participle of" is displayed.

In addition, although the case of an electronic dictionary was shown as the above example,
The present invention is not limited to this, but is also useful as a word search method for a translation machine.

Translation machines are used to translate sentences into other languages, and known translation methods include the neighbor-joining method, the pattern method, the right-to-left scan method, and the simultaneous analysis and synthesis method.

What is important in such machine translation is to determine the part of speech of each word, and the present invention can efficiently search for words other than the original form, so it is especially useful for translation machines that often search for modified forms of words. This is an effective method.

As explained above, according to the present invention, it is possible to obtain an extremely useful word information search and output device that uses fewer headwords, can be searched in conjugated form, etc., and requires less storage capacity. It is.

[Brief explanation of drawings]

Fig. 1 is a block diagram of an electronic dictionary which is an example of using the word storage method according to the present invention, Fig. 2 is a diagram showing the classification of regular changes in verb conjugations of English words, and Fig. 3 is 2 shows a diagram showing the contents of a storage device.

Claims (1)

[Scope of Claims] 1. A first storage means for sequentially storing information indicating an unchanged portion of a word whose inflection occurs and an inflection pattern of the word; a second storage means for storing a set of word endings to be added; and a second storage means for sequentially reading unchanged parts of the word from the first storage means, and detecting whether or not the read part is included in the input word. and when the first detection means detects that the input word includes the word read out from the first storage means, the inflection pattern associated with the word is detected. A set of corresponding word endings is selected based on the information, and word endings included in the selected set are sequentially added to the unchanged part read out from the first storage means to match/mismatch with the input word. and a means for outputting predetermined information such as a translated word corresponding to the input word based on the second detection means detecting a match. Word information search output island