Computer database access system and method for categorizing by style ranking
In one embodiment, the present invention is directed to a method implemented on a computer system of database retrieval including organizing and suggesting search results by style, not just product category. In one embodiment, a computer implemented smoothing algorithm is used to adjust the ranking feature calculations from insufficient data. The use of smoothing improves the statistical confidence of the ranking features, creating a product ranking that is robust to small sample bias.
1. A computer implemented method for providing and ranking product search results, comprising:
receiving, at a server, search terms from a user over a network from a remote computing device, wherein the search terms are digitized, encrypted, and packetized;
classifying, by a processor at the server, the search terms as category search terms and style search terms;
applying, by the processor, the style search terms to a style table in a database, wherein applying the style search terms to the style table in a database comprises matching the style search terms to an index column, wherein the index column is associated with features and categories that correspond to each style search term, and wherein the features and categories that correspond to each style search term are used as additional style search terms;
obtaining the additional style search terms from the style table;
applying, by the processor, the category search terms, the style search terms, and the additional style search terms to a product database to obtain product results, wherein the product database comprises a hierarchy of database categories and subcategories including a root category and each database subcategory has a series of similar products and wherein each product has one or more features;
determining, by the processor, a ranking of the product results based on a degree of matching of the search terms and a number of searches, fetches, and user-selections;
applying a smoothing algorithm to the ranking of the product results such that a feature of a product in the product results is adjusted by the same feature calculated from a more general database category similar to the product in the product results, wherein the smoothing algorithm is applied iteratively starting from an immediate parent category having the same feature through each node of the hierarchy until information from the root category is incorporated and wherein the feature, for which the smoothing algorithm is applied, comprises one or more of seasonality, newness, and brand uniqueness;
transmitting the product results in ranked order from the server over the network to the remote computing device; and
displaying, on a display of the remote computing device, the product results in ranked order.
2. The computer implemented method of claim 1 further comprising:
weighting, by the processor, the additional style search terms from the style table.
3. The computer implemented method of claim 2 wherein:
the weights of the additional style search terms comprise both first weights assigned in the style table and second weights based on a user profile.
4. The computer implemented method of claim 3 further comprising:
applying, by the processor, additional weights based on similar user profiles.
5. The computer implemented method of claim 2 further comprising:
applying, by the processor, additional weights based on how many searches in a recent period of time.
6. The computer implemented method of claim 1 wherein:
the smoothing algorithm is applied to products having a limited search history.
7. The computer implemented method of claim 6 wherein:
the search results of a more general database category are weighted differently from the search results of a narrower database category.
8. A computer implemented method for providing and ranking product search results, comprising:
receiving, at a server, search terms from a user over a network from a remote computing device, wherein the search terms are digitized, encrypted, and packetized and wherein the search terms are classified as category search terms and style search terms, wherein the style search terms are applied to a style table in a database to obtain additional style search terms from the style table, and wherein applying the style search terms to the style table in a database comprises matching the style search terms to an index column, wherein the index column is associated with features and categories that correspond to each style search term, and wherein the features and categories that correspond to each style search term are used as additional style search terms;
applying the category search terms, style search terms, and additional style search terms to a product database to obtain product results, wherein the product database comprises a hierarchy of database categories and subcategories including a root category and each database subcategory has a series of similar products and wherein each product has one or more features;
ranking the product results according to at least one feature and according to a number of searches, fetches, and user-selections;
applying a smoothing algorithm to the ranking of the product results such that the at least one feature of a product in the product results is adjusted by the same feature calculated from a more general database category similar to the product in the product results, wherein the smoothing algorithm is applied iteratively starting from an immediate parent category having the same feature through each node of the hierarchy until information from the root category is incorporated and wherein the feature, for which the smoothing algorithm is applied, comprises one or more of seasonality, newness, and brand uniqueness; and
displaying the product results in ranked order.
9. The computer implemented method of claim 8 wherein:
the smoothing algorithm is applied to products having a limited search history.
10. The computer implemented method of claim 9 wherein:
the search results of a more general database category are weighted differently from search results of a narrower database category.
11. A system for providing and ranking product search results, the system comprising:
at least one database storing a plurality of documents; and
a server coupled to the at least one database, the server comprising:
a processor; and
a memory storing instructions that, when executed by the processor, cause the processor to perform operations comprising:
receiving search terms from a user over a network from a remote computing device, wherein the search terms are digitized, encrypted, and packetized;
classifying the search terms as category search terms and style search terms;
applying the style search terms to a style table in a database, wherein applying the style search terms to the style table in a database comprises matching the style search terms to an index column, wherein the index column is associated with features and categories that correspond to each style search term, and wherein the features and categories that correspond to each style search term are used as additional style search terms;
obtaining the additional style search terms from the style table;
applying the category search terms, the style search terms, and the additional style search terms to a product database to obtain product results, wherein the product database comprises a hierarchy of database categories and subcategories including a root category and each database subcategory has a series of similar products and wherein each product has one or more features;
determining a ranking of the product results based on a degree of matching of the search terms and a number of searches, fetches, and user-selections;
applying a smoothing algorithm to the ranking of the product results such that a feature of a product in the product results is adjusted by the same feature calculated from a more general database category similar to the product in the product results, wherein the smoothing algorithm is applied iteratively starting from an immediate parent category having the same feature through each node of the hierarchy until information from the root category is incorporated and wherein the feature, for which the smoothing algorithm is applied, comprises one or more of seasonality, newness, and brand uniqueness;
outputting the product results in ranked order for transmission over the network to the remote computing device for displaying, on a display of the remote computing device, the product results in ranked order.
12. The system of claim 11 wherein the instructions, when executed by the processor, cause the processor to perform operations further comprising:
weighting the additional style search terms from the style table.
13. The system of claim 12 wherein:
the weights of the additional style search terms comprise both first weights assigned in the style table and second weights based on a user profile.
14. The system of claim 13 wherein the instructions, when executed by the processor, cause the processor to perform operations further comprising:
applying additional weights based on similar user profiles.
15. The system of claim 12 wherein the instructions, when executed by the processor, cause the processor to perform operations further comprising:
applying additional weights based how many searches in a recent period of time.