Searching products catalogs

Data processing: database and file management or data structures – Database design – Data structure types

Reexamination Certificate

Rate now

  [ 0.00 ] – not rated yet Voters 0   Comments 0

Details

Reexamination Certificate

active

06728706

ABSTRACT:

FIELD OF THE INVENTION
The invention relates to searching product catalogs and relates particularly, though not exclusively, to improved methods suitable for allowing shoppers to effectively search online product catalogs of items able to purchased from a retailer.
BACKGROUND OF THE INVENTION
With the growing popularity of e-commerce, searching online catalogs for products has become an important problem for both online retailing for business to consumer (B2C) as well as business to business (B2B) trading portals. These Web sites are virtual shops that have searchable catalogs providing product details.
A shopper may either browse through the various products for sale, or select products by specific product identification numbers. When using conventional product catalogs, the shopper may either (a) know the appropriate product identification number, for example, when ordering a replacement item; (b) be looking for items similar to some item in mind, for example, when looking for a shirt; or (c) be browsing with no specific product in mind but hoping to make up her mind as she looks through the range of products available for sale, for example, when shopping for a gift.
Prior-art commerce systems like WebSphere™ Commerce Suite from IBM, Open Market™ from Open Market Inc, and Broad Vision™ from Broad Vision Inc save product information as numeric and nominal values in a database. These product attributes form a high-dimensional space where each product occupies a unique point. Exact, partial and range queries may be performed using vectors from this space to retrieve products, and present them to the shopper.
Typically, the shopper wants to search for products by description, that is, product attributes rather than by product identification numbers since these may differ from store to store. Further, she may want to browse through the catalog and explore the various possibilities to decide what suits her requirements. It is often not practical or convenient to do exact searches on catalog databases due to several reasons. As a consequence, similarity searching has been proposed to enable “vague” or “fuzzy” querying of the product catalog. However, one limitation of such systems is that different shoppers have different notions of similarity from one another. Hence, two products may be similar to each other in one shopper's perception but another shopper may find them quite different. The use of fixed similarity metrics to evaluate closeness of two products has limitations since it cannot capture the subjectivity of shoppers' product requirements. For example, a shopper may find a Toyota Camry motor vehicle similar to a Volkswagen Wagon since their engines have similar horsepower, but another shopper find a Honda Accord more similar to the Camry since they are both sedans.
Another key limitation of existing commerce systems is their inability to capture the rich content available in pictures of the actual product. Typically, details about visual properties of products such as appearance, colour distribution, texture and so on, are represented by keywords and stored in a multi-attribute product catalog. However, there are cases in which keywords cannot adequately describe product characteristics such as floral prints, upholstery patterns, jewellery designs and so on. Further, describing visual characteristics is often a subjective process and hence it is not possible to appropriately associate unique keywords from pictures of products to enable a search based on a product's visual appearance. Thus, if a shopper wants to search upholstery that looks similar to a pattern that a shopper has with her, then an annotation-based catalog does not work well, as the shopper may not be able to describe the pattern in words. She may have to browse through the catalog for her requirements even though she has a pattern in her mind.
U.S. Pat. No. 4,996,642 (Hey), titled “System and method for recommending items”, discloses a system in which a user is recommended a product, such as a movie title in a video store, based on the similarity of the current user and other users. This system is built on the hypothesis that if two users have generally bought similar things in the past, then they will tend to buy similar products in future.
However, any given shopper typically has different product requirements at different times and hence, such recommendations cannot effectively be made depending solely on her and/or other shoppers' previous responses. Such systems are based on the shoppers' histories and profiles, and are thus “state-less”, that is, the shopper's interaction with the system during a shopping session is not used to recommend items or find items closer to her requirements.
U.S. Pat. No. 6,041,311 (Chislenko et al), titled “Method and apparatus for item recommendation using automated collaborative filtering”, discloses a collaborative filtering approach to product recommendation. It is claimed that content-based systems do not work well and hence products must be matched only at the higher level. Though feature selection and extraction remains a difficult problem, partial solutions do exist for specific domains. In particular, extensive research in image and video analysis has resulted in acceptable automatic feature extraction methods. These methods work even better when assisted by an expert.
U.S. Pat. No. 5,666,442 (Wheeler), titled “Comparison system for identifying the degree of similarity between objects by rendering a numeric measure of closeness, the system including all available information complete with errors and inaccuracies”, discloses a method of similarity searching where attributes can be weighted by the user and where the all items may not have all the feature attributes. This work has been incorporated in the “Similarity Search Engine” from Infoglide Inc, a search engine that returns a rank-order list of items weighted according to the preferences given by the user at query time. The results are ranked with the items most closely matching the search criteria at the top of the list. The system allows the results of multiple databases searches to be brought together. In such a system, the user has to explicitly give the weights for each search criteria. This can make the system difficult to use, especially if the features used for searching can not be adequately understood by the user. Further, this system is better suited to cases in which a definite object is being searched (for example, a criminal record from a police database). However, while shopping, a shopper may start with only an approximate query and arrive at decision only after exploring the product offerings from the store, such as when selecting a gift.
Ivo Vollrath, Wolfgang Wilke, and Ralph Bergmann present a case-based reasoning approach to product selection using intelligent online catalogs in
Intelligent Electronic Catalogs for Sales Support: Introducing Case
-
Based Reasoning Techniques to On
-
Line Product Selection Applications
, R. Roy, T. Furuhashi, P. K. Chawddhry, (Eds.): Advances in Soft Computing—Engineering Design and Manufacturing, Springer-Verlag, London, 1999. The described system, however, expects the user to weigh the relative importance of different attributes during the search process.
None of the above product search approaches outlined above are rigorously optimal or universally applicable for use in presenting product information to shoppers. Accordingly, it is an object of the invention to at least attempt to address these and other limitations associated with the prior art. In particular, it is an object of the invention to generally improve the results provided to those searching online databases, such as retail product catalogs.
SUMMARY OF THE INVENTION
The inventive concept resides in a recognition that the searching of online catalog databases by shoppers is advantageously improved by performing similarity searching on searches performed by the shopper, in conjunction with adjusting the similarity metric used during the search to interactively imp

LandOfFree

Say what you really think

Search LandOfFree.com for the USA inventors and patents. Rate them and share your experience with other people.

Rating

Searching products catalogs does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Searching products catalogs, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Searching products catalogs will most certainly appreciate the feedback.

Rate now

     

Profile ID: LFUS-PAI-O-3240826

  Search
All data on this website is collected from public sources. Our data reflects the most accurate information available at the time of publication.