Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
1999-08-13
2003-01-28
Channavajjala, Srirama (Department: 2177)
Data processing: database and file management or data structures
Database design
Data structure types
C707S793000, C707S793000, C707S793000, C707S793000, C707S793000, C345S960000
Reexamination Certificate
active
06513036
ABSTRACT:
A method and an apparatus for searching and presenting electronic information from one or more information sources.
TECHNICAL FIELD
The present invention relates to a method of searching and presenting electronic information from one or more information sources, said method comprising the steps of:
presenting one or more choices to a user,
registering one or more user specified choices,
using said one or more user specified choices to retrieve one or more references to information from said information sources, and
providing a search result comprising one or more references to information.
The present invention also relates to an apparatus for searching and presenting electronic information from one or more information sources, said apparatus comprising:
presenting means adapted to present one or more choices to a user,
registering means adapted to register one or more user specified choices,
retrieving means adapted to retrieve one or more references to information from said information sources using said one or more user specified choices, and
means adapted to provide a search result comprising one or more references to information.
BACKGROUND AND PROBLEM
Searching in large collections of electronic information, e.g. the Internet, large Intranets, etc., is normally done by software programs called search engines, which typically have an interface with one or more text boxes so that a user may type text usually in the form of keywords describing what he wants to find information about. The search engine supplies links to information sources or the sources themselves containing the relevant information by searching through indexes or taxonomies, i.e. a hierarchy of related contexts and/or topics.
Several search engines are available on the Internet today, the best known engines being Yahoo, Alta-Vista, Lycos, etc.
Searching can generally be described as a transaction oriented type of searching.
In a transaction oriented type of searching a search is executed as fast as possible and in one transaction only. There are no feedback possibilities and therefore no possibility for a search engine to learn what a specific user of the search engine prefers or is generally interested in for later use. If the user is not satisfied with the result of the search, e.g. because it contained too many or too few hits, the user has to input a new search criterion and start all over again.
In a typical transaction oriented type of searching the user specifies one or more keywords e.g. combined with Boolean operators like AND, OR and other criteria like closeness of words, use of synonyms, use of phonetics, etc. The search engine uses one optimal sorting algorithm to find the relevant documents/pieces of information that contain the specified keyword(s), and generally presents this information to the user as a long list of links sorted with regard to how well the information/document contains the keyword(s), i.e. typically sorted according to the highest frequency of the appearance of the specified keyword(s) in the documents in accordance with Boolean rules, phonetics, etc., if any.
All users of the search engine are normally presented with the same standard user interface, i.e. a static user interface, with a text box to specify the keyword(s) and possible Boolean operators and in some cases check boxes representing other possibilities like closeness of words, use of synonyms, use of phonetics or the like. The result of the search is presented in another interface as a list of links to documents containing the specified keyword(s) e.g. sorted as mentioned above. In this way two very different users specifying the same keywords, e.g. safety of cars, will receive exactly the same list of links regardless of whether e.g. one is a potential buyer of a car and the other is a scientist in safety of cars. This gives an often irritating great amount of uninteresting links for both of them.
After being presented with the result, the user generally only has the option of selecting one of the links to the found documents, thereby leaving the search engine or initiating a new search.
There is no possibility of returning to the search engine after reviewing the selected document with information about how relevant the user found that particular document or piece of information and why. So it is impossible for the search engine to make a better choice of presenting the search result the next time.
Another problem of transaction oriented type of searching is that there is no way of determining what the information value of a result is for a user, since the best search result only depends on the best match of the keywords with e.g. a document. For example, if the user specifies the same keyword(s) at two different times and the result is a document saying that the price of oil is increasing, then this document will only have an information value for the user the first time and none the second time.
Yet another drawback of transaction oriented type of searching is the lacking reference to time. If e.g. the keyword ‘car’ is specified, the corresponding result will always be the same regardless of which prior keywords the user has specified. The user could e.g. have typed Volvo as the last keyword or have shown the greatest interest in Volvo in 75% of all searches related to cars. This knowledge is not available in a search system without a time reference and individuality of the users.
One way of obtaining individuality in searching in the prior art is by the use of user profiles. An individual user profile is obtained by monitoring and storing the keywords which a user specifies over a period of time and tries to determine one or more trends for that period, thereby trying to be able to predict what a user is interested in the future. Additionally a user profile may be updated to reflect how interesting the user found a particular document, etc.
However, such a user profile is not very dynamic or flexible, and unless a user very often searches for the same things and uses the same group of keywords again and again, this kind of user profile does not give any advantages, e.g. if a user one day wants to find information about a certain subject and the next day about a completely different subject, e.g. because he has just learned about it or often does not have an actual goal but just wants to be entertained. Also a keyword is specified at two different times, the user profile generally will change meanwhile, so that two different lists would be presented.
Generally speaking, prior art search engines and user profiles are good at trying to find or re-find documents, which are about the same as the user has shown an interest in before, but they match a query against an already decided and objective topic category in a hierarchy of related categories (sports is parent of ice-hockey, football, soccer, etc.). So the information value of a document in a given topic category is determined by others.
SOLUTIONS
The object of the present invention is to provide a method which enables a user to perform a search for electronic information from one or more information sources, the method enabling a representation of relations between different types of information and different users.
This object is achieved by a method of the type mentioned above, said method further comprising
a context representation for one or more references, and
a context representation for one or more users, where
each context representation contains one or more indications representing other contexts with which each is related, and in that the method further comprises the steps of:
presenting the search result to the user in a number of different ways at the same time based on one or more context representations, and
presenting additional choices relating to the search result based on one or more context representations.
Hereby, a method is provided where each object, i.e. references to text document, picture, user, etc., has a representation, i.e. context representation, describing the relations between that particular object and other objects. In
Fruensgaard Finn Ove
Kjærsgaard Jesper
Channavajjala Srirama
Finnegan Henderson Farabow Garrett & Dunner L.L.P.
Mindpass A/S
LandOfFree
Method and apparatus for searching and presenting search... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for searching and presenting search..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for searching and presenting search... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3067545