a sequence of feature vectors labeled with doubles
Expresses this training data in "SVMlight" format, which is <line> .=.
Expresses this training data in "SVMlight" format, which is <line> .=. <target> <feature>:<value> ... <feature>:<value> # <info> <target> .=. +1 | -1 | 0 | <float> <feature> .=. <integer> | "qid" <value> .=. <float> <info> .=. <string>
the signature to use for encoding feature names as integer
the training data in SVMlight format
Creates "positive" and "negative" feature vectors according to whether the feature
cost is greater than margin
or less than -margin
, respectively.
Creates "positive" and "negative" feature vectors according to whether the feature
cost is greater than margin
or less than -margin
, respectively.
Feature vectors that are within margin
of zero are filtered from the traing data.
the absolute threshold that determines whether a vector is kept
a TrainingData instance where all costs are -1 or 1
The set of feature names found in the training data.
The set of feature names found in the training data.
a sequence of feature vectors labeled with doubles
a sequence of feature vectors labeled with doubles
A subinstance of TrainingData whose labels are -1 or 1.