Sap Se

Sandeep Chakravarty

Sudhir Bhojwani

Sudha Lakshman

Quan Zhang

Tu Truong

Richa Namballa

Fuming Wu

Lin Dong

Yue Li

Yuan Tung

Sanish Mondkar

Steven Weigand

Vengateswaran Chandrasekaran

Joseph Fox

Matthew Cromar

Grace Zhang

Rajanikanth Jayaseelan

Provided is a method and system for normalizing catalog item data to create higher quality search results. In one example, the method may include receiving a record comprising an unstructured description of an object, identifying a type of the object from among a plurality of object types and identifying a predefined attribute of the identified type of object, extracting a value from the unstructured description corresponding to the predefined attribute and modifying the extracted value to generate a normalized attribute value, and storing a structured record of the object in a structured format comprising a plurality of values of a plurality of attributes of the object from the unstructured description including the normalized attribute value for the predefined attribute of the object.

Normalization of unstructured catalog data

A search engine may detect a user selecting an object associated with a first category subsequent to inputting a first search phrase including a keyword. In response, the search engine may update a learning model by at least incrementing a relevance score for an association between the keyword and the first category. The search engine may suggest keywords for completing a second search phrase based on the updated learning model. The search engine may further respond to the second search phrase by determining, based on the updated learning model, that the first category is more relevant to the first user than a second category. A search result of the second search phrase may be refined by eliminating, from objects matching the second search phrase, objects associated with the second category but not the first category.

Learning model based search engine

The embodiments in this disclosure include a system for receiving search requests for one or more items from a client device having access to a local database that stores primary catalogs containing items of contracted entities. If the items are not found in the primary catalogs, the system may be configured to execute an API for processing the search request by routing the search request to a proxy vendor computer in a cloud network. The cloud network may be configured to store vendor master records for one or more non-contracted entities. The proxy vendor computer can conduct a text search of secondary catalogs associated with non-contracted entities, and can transmit search results to the client device for the requested items. The items may correspond to one or more vendor master records associated with a non-contracted entity and may have a matching entry in one or more of the secondary catalogs.

Methods and apparatus for storing and accessing vendor master data records in a cloud network via a proxy vendor system

Growing community of inventors

San Francisco, CA, United States of America

Sandeep Chakravarty