Knowledge Extraction

From Natural Language Text to Graph Database.

Extract information from big volumes of English language text, process it and store the results in a graph database for easy-to-do computation. Knowledge is represented as triplets of the form subject-action-object.

Abstract

Definition and analysis of a sample approach to the subject of knowledge extraction. The goal is to extract knowledge from large volumes of English language text, process it and finally store the results in a graph database for future usage. The basic and atomic element of knowledge is represented as a triplet of the form subject-predicate-object. We propose some approaches, solutions, and relative tools for each of the two major aspects (i.e. NLP and graph database storage/processing), while addressing the more specific scenario of open-ended ontology extraction. The experiments are specifically conducted in order to verify the feasibility of the solution with big amount of natural language text. Analysis and comparisons, in terms of performances and quality, of all the approaches encountered are then presented.

License

Released under version 2.0 of the Apache License.