Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
2000-10-25
2003-08-26
Shah, Sanjiv (Department: 2172)
Data processing: database and file management or data structures
Database design
Data structure types
C707S793000, C707S793000
Reexamination Certificate
active
06611842
ABSTRACT:
FIELD OF THE INVENTION
The present application relates to a method and apparatus for generating profile data indicative of individuals' personal preferences. In particular, embodiments of the present invention relate to the generation of profile data that indicate the underlying qualities of items selected by a user as derived from a user history.
BACKGROUND OF THE INVENTION
When people buy, rent or use products, they rely on their personal preferences for choosing the products which they buy, rent or use. Where previous purchases, rentals or usage are recorded, purchase or rental histories can be used to determine the personal preferences on which those users rely on for selecting products or services. If a user's personal preferences can be accurately classified, suitable additional products or services can be automatically recommended or advertised to the user which correspond to the user's underlying personal preferences.
Furthermore, to the extent that a users selections are a reflection of the users individual preferences, the profile data also provides a profile of that user in terms of the core values held by that user. The profile data can therefore be used to identify not only products of the same type as those identified in a purchase or rental history but any product or service which can be classified in a similar way reflecting similar core values.
One area in which purchase and rental histories are available is in relation to user selections of entertainment media such as the hiring of videos. A known strategy for automatically recommending videos to users is to classify all videos available for rental in terms of categories such as drama, action, romance, comedy or children's videos and to identify the categories which a users previous selections fall within. When these categories have been identified, other films falling within such categories can be automatically recommended to a user.
A problem with attempting to classify user personal preferences in terms of the broad categories of products within a users purchase or rental history is that such broad categories do not capture individual differences very well. Thus, for example, a video rental history for a given individual may indicate that the individual has selected action films and also comedies and dramas. The personal preferences of such an individual classified in terms of the broad categories of action films, comedies and dramas therefore only serves to reduce slightly the total number of possible products which might be recommended to that individual.
An alternative method of identifying and classifying user personal preferences from user purchase or rental histories is to associate all available products with data indicative of assessments of the products in a range of different categories and to then determine an average profile based on previous selections. This average profile is then taken to be representative of the core values relied upon when selecting products.
However, although such profile data accurately reflects an individual's personal preferences if an individual is consistent about how they select products, frequently individuals will adopt different criteria for selecting products on different occasions with the individual criteria relied upon at times being inconsistent. The effect of averaging these inconsistencies to generate profile data results in the generated profile data not accurately reflecting the true values being relied upon by the individual to make their selections.
Thus, for example, in the case of an individual who at times selects fast paced films and at other times selects slow paced films, a determined average of data indicative of the pace of film selected might indicate that the individual had no particular preference with respect to the pace of films whereas in fact all of the user's selections were in fact made specifically for either extremely fast paced or extremely slow paced films.
There is therefore a need for an alternative method of establishing categorization data classifying individuals personal preferences for products utilizing purchase and rental or similar user selection histories which more accurately classifies the personal preferences of individuals. In particular, there is a need for generating and recording data indicative of a more accurate classification of personal preferences in a way which can be utilized to select further products or services based on the underlying core values of preferences held by an individual.
SUMMARY OF THE INVENTION
In accordance with one aspect of the present invention there is provided a computer apparatus for generating data indicative of underlying qualities of selections made by a user comprising:
means for receiving user history data identifying a plurality of items selected by a user;
a selection profile database for storing records associating selectable items with a plurality of category values, each indicative of qualities of said items in respect of a plurality of categories; and
processing means for processing said profile data associated with items identified within user history data received by said receiving means,
wherein said processing means generates user profile data comprising data indicative of groups of one or more of said categories, based on a determination of the number of category values associated with said products within said user history determined to fall within predetermined ranges.
In this way, by providing a database that associates selectable items with data indicative of the underlying values embodied by the selectable items, items associated with values within a high or low range can be identified. Identification of values falling within extreme ranges for different categories for items within a user selection history can then be used to generate data indicative of a user's preferences in terms of the number of times high or low values are associated with categories or groups of categories.
The applicants have appreciated that since individuals are not always consistent about the selections they make identifying selections with high or low category ratings provides data which better reflects individual preferences.
Furthermore, the applicants have appreciated that where user selections are associated with a group of different category ratings all indicative of high or low assessments of content in respect of different categories, this combination of categories provides a good representation of the underlying qualities embodied in a selection since the interdependence of different qualities is indicated by such category groups.
In particular, where certain combinations of high and low values for groups of categories appear frequently in data associated with selections within a user history, the most frequently appearing groups provide a more accurate assessment of the personal preferences of a user.
In addition to providing means by which users may be classified by generated profile data, embodiments of the present invention relate to a system for selecting and outputting advertisement data associated with products having profile data similar to the generated profile data. In this way, profile data indicative of values relied upon by a user for selecting previous products can be utilized to direct targeted advertising to users so that the products or services advertised correspond to-products and services which are most likely to appeal to that particular user.
Embodiments of the present invention also relate to a system for selecting products and services on the basis of generated profile data to mimic the selections adopted. For example, a recording apparatus may be provided for automatically recording received television signals on the basis of a comparison of profile data associated with received signals and user profile data generated utilizing user histories of viewed programs, where the profile data comprises groups of categories indicative of underlying qualities of the content of previously viewed programs. Thus, a system can be pro
Applied Psychology Research Limited
Frishauf Holtz Goodman & Chick P.C.
Shah Sanjiv
LandOfFree
Method and apparatus for generating profile data 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 generating profile data, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for generating profile data will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3114162