专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

1. 发明专利

JPH01162978A DICTIONARY RETRIEVING METHOD FOR NATURAL LANGUAGE PROCESSING 失效
公开(公告)号：JPH01162978A
公开(公告)日：1989-06-27
申请号：JP17690588
申请日：1988-07-18
申请人： HITACHI LTD
发明人： NITTA YOSHIHIKO , OKAJIMA ATSUSHI , YAMANO FUMIYUKI
IPC分类号： G06F17/27 , G06F17/28 , G06F17/30
摘要： PURPOSE:To retrieve a word dictionary at a high speed by retrieving copula dictionary only in case a copula including a relevant word is detected through the retrieval of the word dictionary with the word used as a headword. CONSTITUTION:An input device 100 is used together with a translation processor 200, a dictionary memory 300, a work memory 400, and an output device 500. The memory 300 includes a word dictionary 301, a copula dictionary 302, etc., and also stores previously a table where the translation processing rules are defined. Each word information of the dictionary 301 includes the presence or absence of a copula consisting of the relevant word and the information on a pointer to the dictionary 302. Thus the copulas can be retrieved concurrently with the segmentation of words. The dictionary 302 is retrieved only when the copula consisting the relevant word is detected. Thus the dictionaries are retrieved at a high speed.

2. 发明专利

JPS6470871A MAINTENANCE SYSTEM FOR WORD ORDER RELATION DICTIONARY 失效
公开(公告)号：JPS6470871A
公开(公告)日：1989-03-16
申请号：JP22633587
申请日：1987-09-11
申请人： HITACHI LTD
发明人： TOMINAGA MASASUKE , YAMANO FUMIYUKI
IPC分类号： G06F17/28
摘要： PURPOSE:To reduce load at the time of formation and to reduce fluctuation in the quality of dictionary information, by converting a word order relation primary diction ary into an execution type word order relation secondary dictionary by means of a conversion rule having word order relation with a meaning dictionary. CONSTITUTION:An implementor selects the sort of processing from a keyboard 9, specifies a word order for executing the processing and sets up the specified pattern in a word order pattern table 62 in a main memory 6. A CPU 1 reads word order relation to be updated and referred from the word order relation primary dictionary 2 and a word order relation secondary dictionary 3 to a primary information area 64 and then a secondary information area 65 for word order relation in accordance with table information. Then, the implementor updates or newly form primary informa tion by means of a keyboard 9 based on information stored in an example text file 5. After ending said operation, a conversion program is started and the primary infor mation is converted into execution type secondary information in accordance with conversion rules in a conversion rule file 7 and a dictionary conversion meaning diction ary 4. After ending said conversion, whether a trouble exists in the result or not is checked, and when the result is OK, respective information 64, 65 are stored in the dictionaries 2, 3 respectively.

3. 发明专利

JPS6438875A METHOD FOR RECOGNIZING PHRASE CONSTRUCTION OF NATURAL LANGUAGE 失效
公开(公告)号：JPS6438875A
公开(公告)日：1989-02-09
申请号：JP17690788
申请日：1988-07-18
申请人： HITACHI LTD
发明人： NITTA YOSHIHIKO , OKAJIMA ATSUSHI , YAMANO FUMIYUKI
IPC分类号： G06F17/27
摘要： PURPOSE:To cause the construction analysis and sentence pattern recognition of a text sentence to be easy and correct by segmenting a phrase from the part column of the word and collocation of the next sentence described in a natural language and obtaining the part column in which a part is newly added to the phrase. CONSTITUTION:A pattern 304 for cutting a phrase element is stored in a dictionary memory 300. The pattern 304 is used at the time of segmenting a phrase element from the part column composed of the row of the part added to the word and collocation and preparing it in a phrase element part column by newly adding the part to the phrase element. Since, by using such a pattern 304, the part column shorter than the previous part column corresponding to the word and collocation can be obtained, and simultaneously, the part has a corresponding relation close to the construction of a sentence, the construction analysis and sentence pattern recognition of the text sentence can be easily and correctly executed.

4. 发明专利

JPS63261467A METHOD AND DEVICE FOR TRANSLATION 失效
公开(公告)号：JPS63261467A
公开(公告)日：1988-10-28
申请号：JP9512287
申请日：1987-04-20
申请人： HITACHI LTD
发明人： OKAJIMA ATSUSHI , YAMANO FUMIYUKI , YAMAZAKI NORIYUKI
IPC分类号： G06F17/27 , G06F17/28
摘要： PURPOSE:To improve the quality of translation by displaying possible copulas before executing the translation of copulas including known words and allowing a user to update a dictionary. CONSTITUTION:At the time of executing translation, the user specifies the range of an objective text. Then, objective text in the range is read out from a main memory 7 and a copula list is formed by a copula list forming part 4. The formed copula list is displayed to the user through a display device 2. When the copula list is displayed to the user, the user updates a dictionary file 6 for copulas to be preferably registered in a dictionary as translation or grammatical structure information. After updating the dictionary, a translation part 5 translates the objective text by using the updated dictionary file 6. thus, the quality of translation can be improved.

5. 发明专利

JPS63228374A TRANSLATION PROCESSOR 失效
公开(公告)号：JPS63228374A
公开(公告)日：1988-09-22
申请号：JP6101787
申请日：1987-03-18
申请人： HITACHI LTD
发明人： YAMANO FUMIYUKI , OKAJIMA ATSUSHI , YAMAZAKI NORIYUKI
IPC分类号： G06F17/28
摘要： PURPOSE:To improve the quality of a translation processing by restricting to the correspondence to objective document equivocation in dictionary data which is used in the translation processing. CONSTITUTION:A means extracting frequency information of appeared words, a KWIC list and furthermore, information on equivocation in dictionary data with a whole document as an object when the document of a restricted field is translated, a means 6 restricting dictionary data to correspondence to the document for translation with respect to the words which are expected to be effective on the precision improvement of the translation processing with more than a prescribed level and a means 5 merging restricted dictionary data and having restricted dictionary data. Since selected dictionary data can be generated for the correspondence to the document of inputted sentences, the number of equivocation in dictionary data can be reduced. As a result, one word-one part of speech-one translation is the best case in which there is no equivocation. Generally, the quality of the translation processing can be improved with the reduction of equivocation.

6. 发明专利

JPS62256080A TRANSLATION PROCESSING SYSTEM 失效
公开(公告)号：JPS62256080A
公开(公告)日：1987-11-07
申请号：JP9792686
申请日：1986-04-30
申请人： HITACHI LTD
发明人： YAMANO FUMIYUKI
IPC分类号： G06F17/28
摘要： PURPOSE:To translate an optional input sentence under a limited computer resource by providing a means for clearly indicating a part capable of being partly translated to a sentence to be translated together with a syntax element of the part, repeating a partial translation for a limited number of times and finally obtaining the translation of all the sentences. CONSTITUTION:Initially, the range of a part capable of being a basic element syntactically as the partial aggregation of the input sentence and the syntax element are designated, translated, thereby, the sentence of the limited length may be treated and an applied grammar rule can be limited since the syntax element of the sentence to be translated is clearly indicated and it is not necessary to set a complicate grammar rule. Accordingly, a memory capacity required for a translation processing can be reduced and the quality of the translation is easily kept equal since the translation of a single word sequence of the length below the limited length is carried out. Further, the translated sentence can be obtained by applying the translation processing of the limited times to the input sentence of the optional length by the means for applying the translation processing of the limited times.

7. 发明专利

JPS6162162A Method for text display 失效
标题翻译：文本显示方法
公开(公告)号：JPS6162162A
公开(公告)日：1986-03-31
申请号：JP18263584
申请日：1984-09-03
申请人： Hitachi Ltd
发明人： OKAJIMA ATSUSHI , YAMANO FUMIYUKI , OKAMOTO ERI , KUBO YOSHIORI
IPC分类号： G09G5/32 , G06F3/14 , G06F17/21 , G09G1/00
摘要： PURPOSE: To help recognize a constitution of the whole text by providing a characteristic character-string pattern for a check pattern table and displaying the part coincided with the pattern in the text.
CONSTITUTION: A processing program 41, pattern table 42, text area 43, work register 44 and included in the main memory 4, and a processing program 41 is previously stored before the translation processing is carried out. (1) is stocked in a parameter (i), which shows a row of the pattern table 42. Thereafter, in case of the pattern display is designated, (1) is stocked in a parameter (j) which shows matching of the pattern from the character of the text. A character-string of (i) row of the pattern table 42 and of (j) row of the text are compared. If they are the same, a color cord and a display cord is given to the part of the text memory and repeatedly processed by advancing one (j) to display in a CRT2 in accordance with the color cord stored by these processing.
COPYRIGHT: (C)1986,JPO&Japio
摘要翻译：目的：通过为检查图案表提供特征字符串图案并显示与文本中的图案一致的部分来帮助识别整个文本的构成。构成：在执行翻译处理之前，预先存储处理程序41，模式表42，文本区域43，工作寄存器44并且包括在主存储器4中以及处理程序41。（1）被存储在表示图形表格42的行的参数（i）中。此后，在指定图案显示的情况下，（1）存储在显示图案匹配的参数（j）中从文字的字符。比较文本的（i）行的图案表42和（j）行的字符串。如果它们相同，则将文字存储器的一部分给出彩色线和显示屏，并根据这些处理存储的颜色线，使一个（j）前进（j）显示在CRT2中，并重复处理。

8. 发明专利

JPS6089275A TRANSLATION SYSTEM 失效
公开(公告)号：JPS6089275A
公开(公告)日：1985-05-20
申请号：JP19598983
申请日：1983-10-21
申请人： HITACHI LTD
发明人： YAMANO FUMIYUKI , OKAJIMA ATSUSHI
IPC分类号： G06F17/28 , G06F15/38
摘要： PURPOSE:To enable to effectively substitute the character-strings of the translated sentences when the character-strings of the second language are collectively substituted by providing the relation corresponding to the text of the first language as the data in the text of the second language. CONSTITUTION:The substitute object character-strings in the translated sentence data on the disc play 1 is designated from the keyboard 2, the processor 3 extracts the front position and rear position of the substituted object character- strings in the picture buffer 5. Next, the front position corresponding to the substitute object character-strings in the translated sentence data (translated sentence physics data) in the memory 4 corresponding to the translated sentence data of the buffer 5 is obtained. In this case, the translated sentence physics data are 4 byte more compared with the translated sentence data. Consequently, when the translated sentence data are scanned, the position equal to the front position of the substitute object character-strings in the translated sentence data can be extracted by skipping the 4 byte. In this way, the character-strings of the translated sentences can be effectively substituted by providing the relation corresponding to the first language in the text of the second language as data.

9. 发明专利

JPS59214980A Language processing system 失效
标题翻译：语言处理系统
公开(公告)号：JPS59214980A
公开(公告)日：1984-12-04
申请号：JP8912783
申请日：1983-05-23
申请人： Hitachi Ltd
发明人： HIRAI AKIHIRO , OKAJIMA ATSUSHI , YAMANO FUMIYUKI
IPC分类号： G06F17/22 , G06F17/27 , G06F17/28 , G06F15/38
CPC分类号： G06F17/2735 , G06F17/2836
摘要： PURPOSE:To shorten a correcting time and to improve the efficiency by providing an idiom dictionary in an automatic translating system to aid a user in recollection. CONSTITUTION:An automatic translation processing is executed in a central computer 8. An English-language text and a Japanese-language text are transmitted to a terminal equipment 1. If a corrector does not keep clearly an idiom, which should be inserted to the translated sentence, in memory, he designates a part, which he has a clear memory of, and a part, which he has not a clear memory of, of this idiom in accordance with a certain format on the screen of a CRT and inputs them as an idiom pattern for retrieval. The system retrieves the idiom dictionary to find all idioms corresponding to the idiom pattern for retrieval from the idiom dictionary and displays the found idioms on the screen of the CRT. A selected word is inserted to the translated sentence on the screen of the CRT.
摘要翻译：目的：通过在自动翻译系统中提供成语字典来缩短校正时间并提高效率，以帮助用户进行回忆。构成：在中央计算机8中执行自动翻译处理。将英文文本和日文文本发送到终端设备1.如果校正器没有保持明确的成语，应将其插入到翻译中句子，在记忆中，他指定了一部分，他对CRT的屏幕上按照某种格式清楚记忆，并且没有清楚记忆的部分，并将其输入为用于检索的成语模式。系统检索成语词典，找出与成语词典相对应的成语模式，用于从成语词典中检索，并在CRT的屏幕上显示发现的成语。所选择的单词被插入CRT的屏幕上的翻译句子。

10. 发明专利

JPS5858579A FOREIGN LANGUAGE TRAINING APPARATUS AND METHOD 失效
公开(公告)号：JPS5858579A
公开(公告)日：1983-04-07
申请号：JP15627981
申请日：1981-10-02
申请人： HITACHI LTD
发明人： OKAJIMA ATSUSHI , NITSUTA YOSHIHIKO , YAMANO FUMIYUKI
IPC分类号： G09B7/00 , G06F17/28 , G06Q50/00 , G06Q50/10 , G06Q50/20 , G09B19/06

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式