Data processing: speech signal processing – linguistics – language – Linguistics – Natural language
Reexamination Certificate
1998-01-23
2002-06-25
Edouard, Patrick N. (Department: 2747)
Data processing: speech signal processing, linguistics, language
Linguistics
Natural language
Reexamination Certificate
active
06411924
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates generally to computer-implemented information retrieval systems and methods and, more particularly, to such systems and methods for identifying and displaying information based upon the linguistic content of the information.
2. Related Art
Increasing collection and exchange of computer-readable or computer-accessible documents, such as electronic mail, technical documentation, publications, notes, correspondence, and so on, require improved methods and systems for retrieving particular documents and efficiently displaying them to a user. Various conventional techniques have been developed to search through collections of documents and display the results. However, these conventional techniques have many drawbacks that limit their effectiveness in allowing a user quickly and intuitively to find and examine documents of interest.
One group of such conventional techniques involves searching documents by keywords, topics, titles, or other subject-matter indicators inserted in the documents by their authors. Such techniques generally limit the documents found by a user because the user typically must choose the same, or similar, keywords as those chosen by the authors. There is thus no assurance that a document pertaining to a particular subject matter will, first, be described by the author by the use of a particular keyword, or by any keyword at all, or, second, that the user will choose to search by a keyword chosen by the author. Also, such techniques often require repeated inquiries by the user until a desired document is found. Each such inquiry possibly may require significant effort to devise appropriate keywords or combinations of keywords, and the user generally cannot be assured that the most appropriate keywords or combinations have been tried. Methods for combining keywords may be unintuitive, and unique to each search mechanism. Repeated inquiries may consume significant amounts of time, and it may be difficult or impossible to alter previous steps to change the current search results without repeating the entire process. In addition, such keyword-searching techniques typically do not display information about the subject matter of the entire group of documents that is being searched, as contrasted with information about the particular documents in the group that satisfy the search criteria. Similarly, such techniques often do not present the results of searches, or repeated searches, in a manner that enables a user quickly, efficiently, and intuitively to compare the documents retrieved by one search with documents retrieved by another search in order to choose the most promising direction for further searching.
Other conventional systems or methods may attempt to display limited information about the subject matter of the groups of documents that are being searched. Such systems or methods may allow a user to select among a list of keywords, topics, titles, or other subject-matter indicators. Such list may be presented as an index, for example. However, in such systems, it is not provided that such list in fact describes the subject matter of the particular group of documents being searched. Rather, the list may consist of a predetermined group of subjects that are presumed to describe the content of representative collections of documents in general, or in particular subject areas. Other lists may include author-supplied descriptors, but, as noted, various authors may not use the same keywords to describe the same subject matter, or may not use keywords that a user would look for, or recognize, as being descriptive of a desired subject matter.
Still other conventional systems or methods may apply limited linguistic analysis to a group of documents in order to attempt automatically to provide information about their subject matter; that is, without relying on author-supplied keywords. For example, such systems or methods may attempt to identify proper nouns that are categorized by comparing them to a dictionary of proper nouns. Such systems or methods typically have significant limitations, including the inability to identify recently coined proper nouns used, for example, in quickly evolving technological fields. Also, certain parts of speech, such as proper nouns, may be systematically underrepresented in certain types of documents, such as is often the case with respect to proper nouns in technical documentation. Further, such systems or methods may not be capable of distinguishing among various uses of the same proper noun. For example, the proper noun “Madonna” may be categorized as pertaining to music or religion, rather than to visual art, because the system or method does not analyze the full morphological and syntactic context in which the proper noun appears.
With respect to all such conventional systems or methods, a user generally may not efficiently and intuitively identify from an initial collection of documents a sub-collection of documents that are likely to pertain to a subject matter of interest. Similarly, a user generally may not efficiently and intuitively further identify a sub-sub-collection of the original document collection, and so on, until a manageably small number of documents remains to be examined. Moreover, information displayed to a user about the subject matter of a collection of documents generally is not presented in an efficient and intuitive manner such that the user may readily determine whether such collection of documents contains a subject matter of interest, or how such desired subject matter relates to other subject matter contained in the collection of documents.
Accordingly, what is needed is a system and method that comprehensively and automatically (i.e., without relying on keywords or other subject-matter indicators inserted by authors) displays to a user the subject matter of a collection of documents, and enables a user intuitively and efficiently to find sub-groups of such collection containing subject matter of interest. In particular, what is needed is a system and method that efficiently displays information about the subject matter of the groups of documents that are being searched. Also, such system and method should enable a user quickly, efficiently, and intuitively to examine and alter the display in order to compare the documents retrieved by one search with documents retrieved by another search, or to successively narrow a search, in order to choose the most promising direction for further searching or to display desired documents.
SUMMARY OF THE INVENTION
The present invention is a computer-implemented information analysis and display system and method that dynamically generates and displays topics representing a linguistic content of documents in a file system. In accordance with one aspect of the invention, referred to as a linguistic filter, the documents are user-selected. In accordance with one aspect, the user operates a user computer to select one or more of such dynamically generated and displayed topics, preferably using a graphical user interface. In some embodiments, the linguistic filter displays document identifiers corresponding to those documents that are described by one or more of the topics selected by the user. In such, and other, embodiments, the linguistic filter displays the place or places within a document, or group of documents, at which are located linguistic content giving rise to one or more selected topics.
In one embodiment, the file system is local to the user computer; that is, it is located within the user computer or directly connected to it. In an alternative embodiment, the file system may include one or more file systems that are remote to the user computer; that is, the remote file systems are connected to the user computer through a network, or networks of networks.
In one embodiment, the linguistic filter of the present invention includes an interface manager, a linguistic topic analyzer, and a display manager. The interface manager retrieves selected files from the file system and
Akker David van den
de Hita Carolina Rubio
Forest Philippe
Vemula Venu
Dinsmore & Shohl LLP
Edouard Patrick N.
Novell Inc.
LandOfFree
System and method for linguistic filter and interactive display does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for linguistic filter and interactive display, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for linguistic filter and interactive display will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2952752