Module org.elasticsearch.xcore
Package org.elasticsearch.xpack.core.ml.inference.preprocessing.customwordembedding
package org.elasticsearch.xpack.core.ml.inference.preprocessing.customwordembedding
-
ClassDescriptionExtracted feature that has a continuous weight valueDiscrete extracted value with a static weight of 1.0A collection of messy feature extractorsExtracted feature values from the textThis provides an array of
FeatureValuefor the given nGram size and dimensionId Each feature value contains the average occurrence of an nGram and its "id".Derived from: https://github.com/google/cld3/blob/master/src/relevant_script_feature.cc This extracts an array ofFeatureValuefrom the given text.These are the custom script codes that match up to the appropriate id row for the CLD3 weights and quantiles See https://github.com/google/cld3/blob/master/src/script_span/generated_ulscript.hDerived from https://github.com/google/cld3/blob/master/src/script_detector.h We take advantage of Java codepoints to determine the specific script value we care aboutThis is the lone discrete feature.