Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
1999-03-04
2002-05-28
Alam, Hosain T. (Department: 2172)
Data processing: database and file management or data structures
Database design
Data structure types
C707S793000, C705S007380
Reexamination Certificate
active
06397212
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Technical Field
The invention relates to search engines. More particularly, the invention relates to a self-learning and self-personalizing knowledge search engine that delivers holistic results.
2. Description of the Prior Art
With the Internet, information, i.e. “. . . a collection of facts and data” (The American Heritage Dictionary of English Language) is becoming a commodity, such that nearly any fact can be found. Knowledge, i.e. a set of facts or information which are tied together by some logic module relationship, on the other hand is of high value.
Modern search technology is focused on surfacing facts. However, if a user desires knowledge rather than facts, the results of a search on the Internet using today's search engines or product wizards are often of little use. Such searches are typically key word based, using one or two words linked by Boolean logic module. The search engines typically provide a long list of headlines in accordance with these key words. These lists can contain tens of thousands of results.
The better search engines attempt to rank the results, using various methodologies such as frequency of a site visited. Thus, such search engines assume that a higher frequency of visitation is indication of value of information. Alternatively, such search engines may use some internal manual/automatic ranking system. The methods employed are not user specific. Each result typically provides a URL link to the Web page or site containing the information. The URLs must be activated for one site at a time. The user must then scan the Web page pointed to by the URL to determine if the desired information is found on that Web page. If it is not, then the user must select the next item in the list and continue this process until the desired result is found, or until the user simply gives up in frustration.
In typical search engine technology, the user is not able to provide the search engine with any feedback directly or indirectly that might help the search engine find a better result. Other than the key words, the user can not be more specific or multi-dimensional in describing the information for which he is searching.
One type of search engine is referred to as “Product wizards” or product finders, which are commonly used within Web sites of product vendors and merchants. These search engines are either:
Key word based search engines having a pull down window that shows all product names and/or product groups/categories. An alternative to such pull down windows are icons and product pictures, or lists with links. These methods are usually only useful and practical for small product selections;
Question based search engines for more complex products, in which the user must answer a set of questions before the system makes a recommendation. Frequently, as in the case with technical products, it is difficult for the user to answer a number of these questions; or
Modular search engines, where the user must select a base model to which a number of options can then be added by using pull down windows. Based on the selection, a final price is then calculated. DELL computers has such as Web site. Unfortunately, when assembling a personal computer using such search engine, the user can not move between base models, e.g. between a desk top solution and a laptop, or even between one specific laptop series and another laptop series. Nor can the user lock an important feature, such as display size and/or price range, and seamlessly and easily compare solutions which satisfy those requirements.
The shortcomings of today's search engines are numerous:
All of the above search technologies require a keyboard/mouse interface and their use therefore is not available or limited in applications that provide a television remote control or other non-personal computer Internet devices.
Such other human interfaces are expected to proliferate rapidly over the next few years.
The user needs usually determine the search terms (i.e. key words). The user is not presented with solutions to which he can respond from among major characteristics given to him.
The above search technologies do not allow the user to provide any feedback, which would help the search engine to provide better searches in the next iteration. They are not iterative in the sense of moving from a first, coarse solution to a subsequent, fine solution.
The above search technologies do not learn about the searches of all users and the specific user to provide better assistance. That is, they are not adaptive.
None of the above search technologies allow for a simultaneous multi-dimensional search, which delivers a complete result, even if not all characteristics or parameters have been specified.
None of the above search technologies allow a user to vary simultaneously a number of major and minor characteristics, and at the same time lock in other characteristics.
None of the search engines allow the users to make easily trade off's. For example if a user selects a flight at a travel site such as Expedia, the search engine is able to find the lowest cost flight. However, the user can not easily explore—as would any experienced travel agent—other alternatives or “neighboring solutions” such as if the customer would return a day earlier, or would depart from the airport across town he could reduce the cost of the flight substantially.
None of the search technologies allow to summarize and present a number of searches for the same or similar product/service in a tabular form for user friendly comparison.
The shortcomings of today's methods result in a great level of frustration for the user and are neither natural nor intuitive.
A natural approach is usually more holistic and iterative. For example, a potential buyer comes to a used car lot where she finds many different models, and alternatives. The sales person, might not begin the sales process by asking the potential buyer a large number of questions to learn which car might be of interest to the potential customer. Instead, assuming the user provides a little guidance (verbal or non-verbal), the sales person takes a best guess regarding the needs of the potential customer. The sales person “sizes up” the potential buyer and her needs, i.e. he personalizes the search, comparing the few facts known/observed about the user, such as gender, age, and dress, against experience (i.e. a personal, historic data base of all users), leads the person on the lot, and shows her a car (.e. a holistic solution). The buyer reacts about most important characteristics, e.g. “would prefer sports utility,” “too expensive,” “mileage ok,” “like the brand,” “a little expensive.” Based on this information, the sales person picks a better fitting car (e.g. a next iteration), which again is a complete and holistic result. Again the buyer confirms some of the characteristics and might add a few more of the less important specifications (e.g. characteristics). “Yes like this one but do you have it in a different color, with a sunroof.” Another car is presented. This car might not fulfill all specifications, but might be a good compromise considering the relative importance of the characteristics.
It would be advantageous to provide a search engine that was self-learning and self-personalizing, just as with the example of the car salesman above, such that a user was delivered holistic results.
SUMMARY OF THE INVENTION
The invention provides a search engine that allows for intelligent multi-dimensional searches, in which the search engine always presents a complete, holistic result, and in which the search engine presents knowledge (i.e. linked facts) and not just information (i.e. facts). The system is adaptive, such that the search results improve over time as the system learns about the user and develops a user profile. Thus, the search engine is self personalizing, i.e. it collects and analyzes the user history, e.g. it has the user react to solutions and learns from such user reactions. The search engine generates profiles, e.g. it learns from
Alam Hosain T.
Biffar Peter
Glenn Michael A.
Wong Kirk D.
LandOfFree
Self-learning and self-personalizing knowledge search engine... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Self-learning and self-personalizing knowledge search engine..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Self-learning and self-personalizing knowledge search engine... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2878123