Sha256: 097857bd27d99bd70f003512d35b79d50f82903a4d071dca19efaeea7f75e811
Contents?: true
Size: 1.87 KB
Versions: 2
Compression:
Stored size: 1.87 KB
Contents
module Ankusa module Classifier attr_reader :classnames def initialize(storage) @storage = storage @storage.init_tables @classnames = @storage.classnames end # text can be either an array of strings or a string # klass is a symbol def train(klass, text) th = TextHash.new(text) th.each { |word, count| @storage.incr_word_count klass, word, count yield word, count if block_given? } @storage.incr_total_word_count klass, th.word_count doccount = (text.kind_of? Array) ? text.length : 1 @storage.incr_doc_count klass, doccount @classnames << klass if not @classnames.include? klass # cache is now dirty of these vars @doc_count_totals = nil @vocab_sizes = nil th end # text can be either an array of strings or a string # klass is a symbol def untrain(klass, text) th = TextHash.new(text) th.each { |word, count| @storage.incr_word_count klass, word, -count yield word, count if block_given? } @storage.incr_total_word_count klass, -th.word_count doccount = (text.kind_of? Array) ? text.length : 1 @storage.incr_doc_count klass, -doccount # cache is now dirty of these vars @doc_count_totals = nil @vocab_sizes = nil th end protected def get_word_probs(word, classnames) probs = Hash.new 0 @storage.get_word_counts(word).each { |k,v| probs[k] = v if classnames.include? k } vs = vocab_sizes classnames.each { |cn| # use a laplacian smoother probs[cn] = (probs[cn] + 1).to_f / (@storage.get_total_word_count(cn) + vs[cn]).to_f } probs end def doc_count_totals @doc_count_totals ||= @storage.doc_count_totals end def vocab_sizes @vocab_sizes ||= @storage.get_vocabulary_sizes end end end
Version data entries
2 entries across 2 versions & 1 rubygems
Version | Path |
---|---|
ankusa-0.0.8 | lib/ankusa/classifier.rb |
ankusa-0.0.7 | lib/ankusa/classifier.rb |