Package info.ephyra.answerselection.filters

Class Summary
AdaptiveNumberOfKeywordsFilter An adaptive version of the NumberOfKeywordsFilter that also takes occurrences of keywords in preceding text snippets into account.
AnswerPatternFilter The AnswerPatternFilter uses answer patterns to extract factoid answers from text passages and to rank them.
AnswerProjectionFilter The AnswerProjectionFilter projects answers from the Web onto a corpus.
AnswerTypeFilter The AnswerTypeFilter extracts factoid answers of the expected answer types from text passages.
CacheResultsFilter Caches the results along with the query string of the first result.
CombineScoresFilter Combines the extra scores of each Result and sets the main score to their sum.
CutKeywordsFilter Trims answer candidates for definitional questions to improve the precision by chopping of leading news agency acronyms and cities.
CutStatementProviderFilter Trims answer candidates for definitional questions to improve the precision by chopping of leading introductions of indirect speech.
DeserializationFilter The DeserializationFilter reads serialized results from one or more files.
DirectSpeechFilter Drops answer candidates for definitional questions that consist of direct speech.
DuplicateFilter The DuplicateFilter drops duplicate results.
DuplicateSnippetFilter A variant of the DuplicateFilter for answer candidates that are text snippets.
FactoidsFromPredicatesFilter The FactoidsFromPredicatesFilter extracts factoid answers from predicate-argument structures.
FactoidSubsetFilter The FactoidSubsetFilter checks a set of factoid answers for subset relations.
Filter A Filter is part of a pipeline for answer extraction and selection.
FunctionWordsFilter A filter that drops a result if the answer string contains only function words.
GoogleTermImportanceFilter A web term importance filter that counts term frequencies in text snippets retrieved with the Google search engine.
HitPositionComparator<T> The HitPositionComparator compares two Result objects by comparing their hit positions.
HitPositionFilter Scores the results according to their hit positions.
HitPositionSorterFilter The HitPositionSorterFilter sorts the results by their hit positions in ascending order.
NormalizedScoreSorterFilter The NormalizedScoreSorterFilter sorts the results by their normalized scores in descending order.
NounPhraseFilter Scores answer candidates for definitional questions according to the number of novel noun phrases they contain.
NuggetEvaluationFilter Automatically evaluates answer candidates for the 'other' questions in TREC 13-16.
NumberOfKeywordsFilter Filters results by the number of keywords.
OverlapAnalysisFilter Generates overlap analyses for different answer extraction strategies.
PredicateExtractionFilter Extracts predicates which are similiar to those in the question from documents.
PreferNamedEntitiesFilter The PreferNamedEntitiesFilter prefers factoid answers that are named entities over other factoid answers.
ProperNameFilter Drops answer candidates for definitional questions that are enumerations of proper names.
QueryScoreFilter This filter simply increments the score of each result by the score of the query that was used to obtain the result.
QuestionKeywordsFilter A filter that drops a result if the answer string contains a keyword from the question.
ResultLengthComparator<T> The ResultLengthComparator compares two Result objects by comparing the lengths of the answer strings.
ResultLengthFilter A filter that cuts off results if the total length of non-whitespace characters exceeds a threshold.
ResultLengthSorterFilter The ResultLengthSorterFilter sorts the results by the lengths of the answer strings in ascending order.
ReverseScoreSorterFilter The ReverseScoreSorterFilter sorts the results by their scores in ascending order.
ScoreCombinationFilter The ScoreCombinationFilter combines the normalized scores of an answer extracted with several techniques into a single score.
ScoreNormalizationFilter A filter that normalizes the scores of the answer candidates by applying a trained classifier.
ScoreResetterFilter Filter setting the score of each result to 0.
ScoreSorterFilter The ScoreSorterFilter sorts the results by their scores in descending order.
SentenceExtractionFilter Splits the answer strings of results that are not from knowledge annotators into sentences and creates a new result for each sentence.
SentenceSplitterFilter Splits text snippets into sentences in order to facilitate the detection of redundant information.
SerializationFilter The SerializationFilter serializes the results and writes them to a file.
StopwordFilter A filter that drops a result if the answer string contains only function words and single characters (except digits) contains an interrogative contains a single bracket or quotation mark is an adverb This filter is not applied to answer strings that have been extracted with a rule- or list-based NE tagger since these answers are assumed to be properly formatted.
SubclauseSplitterFilter Splits sentences into subclauses in order to facilitate the detection of redundant information.
TermFilter Drops answer candidates for the 'other' questions in TREC 13-16 that contain answers to previous factoid or list questions.
TermImportanceFilter Prefers answer candidates for definitional questions that contain common keywords.
TripletFilter Scores answer candidates for definitional questions according to the number of novel NP-VP-NP triplets they contain.
TruncationFilter A filter that truncates the answer strings.
UnnecessaryCharactersFilter Drops unnecessary characters from answer candidates for definitional questions to improve precision.
WebDocumentFetcher A thread that fetches a web document containing a given search engine snippet.
WebDocumentFetcherFilter A filter that fetches web documents that contain the given search engine snippets.
WebTermImportanceFilter A web reinforcement approach that ranks answer candidates for definitional questions.
WebTermImportanceFilter.CacheEntry  
WebTermImportanceFilter.TargetGeneratorTest  
WikipediaGoogleTermImportanceFilter A web term importance filter that counts term frequencies in a Wikipedia article on the target of the question.
WikipediaTermImportanceFilter A web term importance filter that counts term frequencies in a Wikipedia article on the target of the question.
YahooTermImportanceFilter A web term importance filter that counts term frequencies in text snippets retrieved with the Yahoo search engine.