Adaptive web mining of bilingual lexicon

A method, comprising:

• retrieving one or more bilingual web pages, each bilingual web page including a plurality of terms that comprises a search term and one or more additional terms in a first language;
• forming a plurality of candidate translation pairs by selecting one or more candidate translations for each of the plurality of terms and pairing one of the one or more candidate translations and one of the plurality of terms;
• extracting one or more translation layout patterns from the plurality of candidate translation pairs;
• computing one or more features based at least one the terms;
• deriving a term translation in a second language for the search term based on a hidden conditional random field (HCRF) model that includes the one or more candidate translations, the one or more translation layout patterns, and the one or more features;
• and
• storing the term translation in a lexicon repository.

A computer readable medium storing computer-executable instructions that, when executed, cause one or more processors to perform acts comprising:

• retrieving one or more bilingual web pages based at least on a search term, each bilingual web page including a plurality of terms that comprises the search term and one or more additional terms in a first language;
• forming a plurality of candidate translation pairs by selecting one or more candidate translations for each of the plurality of terms, and forming each candidate translation pair by including one of the one or more candidate translations and one of the plurality of terms;
• extracting one or more translation layout patterns from the plurality of candidate translation pairs;
• computing one or more features based at least one the terms;
• deriving a term translation in a second language for the search term based on a hidden conditional random field (HCRF) model that includes the one or more candidate translations, the one or more translation layout patterns, and the one or more features;
• and
• displaying the term translation and the search term to a user.