Abstract: The present invention leverages query-related information based on a query and/or a search intention to provide a systematic means to facilitate a user in locating desired information despite lacking exact search parameters. This allows users to find information without first formulating an optimum search query. The query graph provides a navigable, graphical notion of the query-related information via nodes representative of the query-related information and edges representative of the associations between the information. In one instance of the present invention, the query graph is derived from an information source such as a query log, a query list, and/or a search engine and the like. Additional instances of the present invention utilize visual and/or audible indicators employed with the query graph to facilitate in relaying the query-related information to the user, including, but not limited to, overlays, icons, colors, and dimensional variances and the like.

Abstract: Systems and methods for predicting a target page associated with a search query are disclosed. Generally, a predictive user click model module defines a set of sessions and builds a model to predict whether a webpage is a target page associated with a search query based on a number of times, over the set of sessions, that a user as defined in the session clicks within the same session on a given webpage associated with a given search query versus a number of sessions in the set of sessions.

Abstract: Methods of providing a document relevance score to a document on a network are disclosed. Computer readable medium having stored thereon computer-executable instructions for performing a method of providing a document relevance score to a document on a network are also disclosed. Further, computing systems containing at least one application module, wherein the at least one application module comprises application code for performing methods of providing a document relevance score to a document on a network are disclosed.

Abstract: Techniques for generating features that are used to rank documents in a search results page are provided. A query is received and may be modified before being compared to queries in a query log of previously-issued queries. The comparisons may be made in a variety of ways. The comparisons may allow query terms to be ordered and terms to be inserted. Relevance features are generated from the results of the comparisons. The documents that are referenced in a search results page (generated as a result of the query) are ranked based on the generated relevance features.

Abstract: Methods of providing a document relevance score to a document on a network are disclosed. Computer readable medium having stored thereon computer-executable instructions for performing a method of providing a document relevance score to a document on a network are also disclosed. Further, computing systems containing at least one application module, wherein the at least one application module comprises application code for performing methods of providing a document relevance score to a document on a network are disclosed.

Abstract: Methods of providing a document relevance score to a document on a network are disclosed. Computer readable medium having stored thereon computer-executable instructions for performing a method of providing a document relevance score to a document on a network are also disclosed. Further, computing systems containing at least one application module, wherein the at least one application module comprises application code for performing methods of providing a document relevance score to a document on a network are disclosed.

Abstract: Search results of a search query on a network are ranked according to an additional ranking function for the prior probability of relevance of a document based on document property. The ranking function can be adjusted based on a comparison of the language that a document is written in and the language that is associated with a search query. Both query-independent values and query-dependent values can be used to rank the document.

Abstract: Search results of a search query on a network are ranked according to an additional ranking function for the prior probability of relevance of a document based on document property. The document property may be the document's file type, the file size, the document language, or another query-independent property of the document. The query-independent values for each document property may be weighted according to relevance measurements of the document based on the document property. As more relevance measurements of the documents may be obtained, the query-independent values for each document property may be updated to reflect the new measurements.

Abstract: A modified large margin perceptron learning algorithm (LMPLA) uses asymmetric margin variables for relevant training documents (i.e., referred to as “positive examples”) and non-relevant training documents (i.e., referred to as “negative examples”) to accommodate biased training sets. In addition, positive examples are initialized to force at least one update to the initial weighting vector. A noise parameter is also introduced to force convergence of the algorithm.

Abstract: Search results of a search query on a network are ranked according to a scoring function that incorporates anchor text as a term. The scoring function is adjusted so that a target document of anchor text reflect the use of terms in the anchor text in the target document's ranking. Initially, the properties associated with the anchor text are collected during a crawl of the network. A separate index is generated that includes an inverted list of the documents and the terms in the anchor text. The index is then consulted in response to a query to calculate a document's score. The score is then used to rank the documents and produce the query results.

Abstract: Search results of a search query on a network are ranked according to an additional click distance property associated with each of the documents on the network. The click distance is measurement of the number clicks or user navigations from a page or pages on the network designated as highest authority or root pages on the network. The precision of the results is increased by the addition of the click distance term when the site or intranet where the search query takes place is hierarchically structured.

Abstract: The present invention leverages query-related information based on a query and/or a search intention to provide a systematic means to facilitate a user in locating desired information despite lacking exact search parameters. This allows users to find information without first formulating an optimum search query. The query graph provides a navigable, graphical notion of the query-related information via nodes representative of the query-related information and edges representative of the associations between the information. In one instance of the present invention, the query graph is derived from an information source such as a query log, a query list, and/or a search engine and the like. Additional instances of the present invention utilize visual and/or audible indicators employed with the query graph to facilitate in relaying the query-related information to the user, including, but not limited to, overlays, icons, colors, and dimensional variances and the like.

Abstract: A field-weighted search combines statistical information for each term across document fields in a suitably weighted fashion. Both field-specific term frequencies and field and document lengths are considered to obtain a field-weighted document weight for each query term. Each field-weighted document weight can then be combined in order to generate a field-weighted document score that is responsive to the overall query.

Abstract: A modified large margin perceptron learning algorithm (LMPLA) uses asymmetric margin variables for relevant training documents (i.e., referred to as “positive examples”) and non-relevant training documents (i.e., referred to as “negative examples”) to accommodate biased training sets. In addition, positive examples are initialized to force at least one update to the initial weighting vector. A noise parameter is also introduced to force convergence of the algorithm.