Adaptive Information Disclosure (AID)

Description
learnRelation is a web service for supervised learning of relations in text. It requires annotated text as input (in the format specified below), and produces a ‘model’ as output. This model can subsequently be used for RelationExtraction.

Output
modelFile (log file modelFile.log is stored in the same directory as modelFile)

Input
inputData has to be annotated as follows:all tokens/words are separated by newlinesentences are separated by the newline. The first line right after the last word in a sentence contains *all* available/detected named entities, the second line lists all relation mentions. Annotations are done assuming the position of the first word in a sentence to be 1.

Example input:

This

is

PROTEIN1

interacting

with

PROTEIN2

but

not

interacting

with

PROTEIN3

.

protein(3) protein(6) protein(11)

relation_mention(3,6)

This

is

a

test

.

From javadocs

learnRelation

public String learnRelation(String inputData,                            String modelFile)

Creates a model modelFile based on the inputData and returns the results of the 10-fold cross-validation.

Parameters:

inputData - annotated training set

Returns:

modelFile (log file modelFile.log is stored in the same directory as modelFile) Note that this is a supervised learner which means that the inputData has to be annotated as follows:all tokens/words are separated by newlinesentences are separated by the newlineThe first line right after the last word in a sentence contains *all* available/detected named entities, the second line lists all relation mentions. Annotations are done assuming the position of the first word in a sentence to be 1. Examples:ThisisPROTEIN1

interacting

with

PROTEIN2

but

not

interacting

with PROTEIN3

.

protein(3) protein(6) protein(11)

relation_mention(3,6)

This

is

a

test

.