Sha256: 1e57d9a6b70f714d8052c7669bb6adddc02188f5ab1afa0033c7ceb452659e53
Contents?: true
Size: 951 Bytes
Versions: 11
Compression:
Stored size: 951 Bytes
Contents
module TextRank
##
# Rank filters are post-process filters which can filter, enhance, or modify
# the results of the PageRank algorithm. A common use case is to collapse highly
# ranked tokens which are found to be adjacent in the original text. Other
# filters might modify the PageRank scores with some sort of external modifier.
# Another use might be to remove collapsed tokens which are not desired (since
# token filters only operate on a single, non-collapsed token).
#
# Rank filters are applied as a chain, so care should be taken to use them
# in the desired order.
##
module RankFilter
autoload :CollapseAdjacent, 'text_rank/rank_filter/collapse_adjacent'
autoload :NormalizeProbability, 'text_rank/rank_filter/normalize_probability'
autoload :NormalizeUnitVector, 'text_rank/rank_filter/normalize_unit_vector'
autoload :SortByValue, 'text_rank/rank_filter/sort_by_value'
end
end
Version data entries
11 entries across 11 versions & 1 rubygems