2001 Volume 15 Issue 5
Published: 15 October 2001
-
Select all|
-
2001, 15(5): 2-8.Automatic word-segmentation of modern Chinese text is the base of Chinese information processing. So a general-purpose application interface for word segmentation is important . In this thesis , the goal of a general-purpose word segmentation system is presented , it s principle and schemes are discussed. The development of this system has been completed basically.
-
2001, 15(5): 9-15,52.Based on the conception of formal word , the author inquires into the non-grammatical factors for determination of segmentation element . They include the semantic factors ,phonetic factors. He put forward the corresponding approaches : the semantic simplicity test approach ,the semantic tightness test approach ,the semantic derivation test approach ,the pause test approach ,the disyllabism test approach. Some non-linguistic principles are studied : easy-read principle ,pluralism principle ,domain-oriented principle.
-
2001, 15(5): 16-19,58.According to the development of the information processing technology , this paper points out the rationality and the urgency of link writing for Chinese word in written text ,and also puts forward the necessary of bringing this thought into the standard of CIP and embodys it in the future information platforms (such as ,eBook ,WWW) . In addition ,it also provides a feasible method to the problem ,how to accustom to the people’s longtime habit of reading and writing , when implementing the link writing.
-
2001, 15(5): 20-26.It is often difficult to construct a rule-based parser and adapt it to largescale real text . So we tried a statistical approach to prosodic phrasing. At first the text was segmented into Chinese words ,then word sequences are tagged automatically by POS tagger. The boundary pattern and boundary distribution probabilities are used in the algorithm to predict phrase breaks. The boundary distribution probabilities are derived from hand-annotated corpus. The errors caused by statistical method are corrected by rules. Through close testing and open testing on about 1000 sentences ,the correct POS tagging rate is about 95% ,the recalling rate of prosodic phrasing is around 60% ,and the correct rate of prosodic phrasing is about 80%.
-
2001, 15(5): 27-33.A method associating main features of events with ontology categories in HowNet is presented.By defining role and semanteme lists for events ,we can link event classes in HowNet into semantic interpretation. The formal description of role and semanteme lists and an algorithm acquiring them are given ,then their applications are exemplified by instances.
-
2001, 15(5): 34-40.According to common error types in pre-proof reading text , this paper introduce the method to structure correcting knowledge sets and a automatic correcting algorithm based on this correcting knowledge sets. The algorithm makes a full use of the characteristics of wrong strings and context heuristic information. It can provide correcting suggestions for such errors as ghost word ,missed Chinese characters ,superfluous Chinese characters ,reversed Chinese characters and substituted Chinese characters etc. The method of sorting the correcting suggestions is also discussed.
-
2001, 15(5): 41-46.This paper introduces the method of tranforming the semantic dependent tree to SQL in Nchiql. First we describe the concept of setblock and its transforming method ,then we give the whole algorithmic of SQL transforming method based on concept of setblock.
-
2001, 15(5): 47-52.The out-of-vocabulary problem is one of the bottlenecks in Chinese Language Modeling. The problem is especially serious for domain-specific training data set . This paper presents a new statistical method to extract new words from the training data. This new method is based on association norm estimation , and searches for the word boundaries by right boundary expanding. Combining the new method with another word-merging method ,we can iteratively optimize the lexicon ,segmentation and language model. And very encouraging results are reported in our experiments.
-
2001, 15(5): 53-58.Presently the researches and literatures on the technology of watermarking are mainly focused on the protection of still images and video documents.But few researchers’interests are in the watermarking of text document . In China ,there is even no literature on this area.But in fact , some text documents are more necessary to be protected than images and video documents. And protecting digital text documents is very important for the work of government and electronic business in the era of Internet . In this paper we mainly introduce the base idea and present research actuality about watermarking for text documents. This paper shows some methods for embedding and detecting watermark. Then the research direction and possible prospect of the digital watermarking for Chinese text documents are analyzed.
-
2001, 15(5): 59-65.More recently ,tiny electronic products and mobile communication products are being developed towards digital and network. Through these information devices ,interactive information exchanges may be enabled between enterprises and users ,and among various users. Chinese input technology is very important for them to be applied in China. The paper introduces a kind of Chinese input technology including three input methods based on digital keyboards ,such as English , PinYin and Qian Dao Pin Yin ,for information devices. They can be used to information devices to input a mount of information composed by English and Chinese words and symbols easily and rapidly. The main idea , the performances ,and characteristics of the input methods are described and discussed in the paper.
