Abstract: Techniques are described for automatically mining query dimensions from web pages resulting from execution of a search query. Lists of items such as words, terms, or phrases are extracted from the web pages based on the recognition of free text, metadata tag, or repeated region patterns within the web page text. Extracted item lists are weighted according to document matching and/or inverse document frequency, and item lists are clustered based on shared or similar items within the lists to generate query dimensions. The generated query dimensions, and the items within each query dimension, are ranked according to quality, and high-quality query dimensions are provided for display alongside top search results.

Abstract: Disclosed herein are techniques and systems for building “information sensors,” which are programmable “focused crawlers” that periodically discover, extract, analyze and aggregate structured information around a topic from the Web. A platform for building an information sensor allows a user to specify one or more data elements within a data source that the user desires to monitor, and an update frequency at which the data elements are to be extracted. Code may be generated based on the user specifications for creation and submission of the information sensor for storage in a database with metadata containing the code and update frequency. Once created, information sensors are scanned to check if running conditions are met, and if met, they may be executed by retrieving the metadata using a sensor identifier (ID). The code is executed to locate a data source, and periodically extract specified data elements therefrom to output structured time-series data.

Abstract: A content application employs page links to merge pages of articles. The content application retrieves an initial page of an article. An article such as a web article spread into multiple pages is retrieved for analysis. A page link of a following page of the article is detected within the initial page. The page link is a top choice among candidates sorted based on a weight score. The following page is retrieved using the page link and appended into the initial page to form an aggregate article. The aggregate article is presented for consumption.

Abstract: Techniques are described for automatically mining query dimensions from web pages resulting from execution of a search query. Lists of items such as words, terms, or phrases are extracted from the web pages based on the recognition of free text, metadata tag, or repeated region patterns within the web page text. Extracted item lists are weighted according to document matching and/or inverse document frequency, and item lists are clustered based on shared or similar items within the lists to generate query dimensions. The generated query dimensions, and the items within each query dimension, are ranked according to quality, and high-quality query dimensions are provided for display alongside top search results.

Abstract: This document describes tools for adjusting anchor text weight to provide more relevant search engine results. Specifically, these tools take advantage of a site-relationship model to consider relationships not only between an anchor text source site and a destination page but also relationships between multiple anchor text source sites to improve web searches. Consideration of these relationships aids in determining a new an anchor text weight, which in turn results in more relevant search results.

Abstract: This document describes tools for adjusting anchor text weight to provide more relevant search engine results. Specifically, these tools take advantage of a site-relationship model to consider relationships not only between an anchor text source site and a destination page but also relationships between multiple anchor text source sites to improve web searches. Consideration of these relationships aids in determining a new an anchor text weight, which in turn results in more relevant search results.

Abstract: Described is a data-centric web search engine technology/architecture, in which document metadata, including offline-extracted metadata, is used as part of a search indexing and ranking pipeline. A web data management component receives crawled documents and extracts document metadata from the documents. An indexing component uses the document metadata to build an index for the documents. A serving component uses the index and the document metadata to serve content, e.g., search results. Also described is the use of query metadata extracted from queries of a query log for use in the pipeline.

Abstract: Described is the running of search-related experiments on a full (or partial) offline snapshot copy of the search engine documents of an actual production system. A snapshot experimentation subsystem runs experimental code related to web searches on the offline data, including to run experimental index building code to build an experimental index (e.g., to test a new document feature), and/or to run experimental search-related code, such as to rank search results according to experimental ranking code, to implement an experimental search strategy, and/or to generate experimental captions.