An Automated Learner for Extracting New Ontology Relations

Recently, the NLP community has shown a renewed interest in automatic recognition of semantic relations between pairs of words in text which called lexical semantics. This approach to semantics is concerned with psychological facts associated with the meaning of words. Lexical semantics is an important task with many potential applications including but not limited to, Information Retrieval, Information Extraction, Text Summarization, and Language Modeling. As this task "automatic recognition of semantic relations between pairs of words in text" can be used in many NLP applications, its implementation are demanding and may include many potential methodologies. And as it includes semantic processing, the results produced still need enhancements and the outcome was always limited in terms of domain or coverage. In this research we developed a buffered system that handle the whole process of extracting causation relations in general domain ontologies. The main achievement of this work is the heavy analysis of statistical and semantic information of causation relation context to generate the learner. The system also builds relation resources that made it possible to learn from itself, were each time it runs the resources incremented with new relations information recording all the statistics of such relation, making its performance enhanced each time it runs. Also we present a novel approach of learning based on the best lexical patterns extracted, besides two new algorithms the CIA and PS that provide the final set of rules for mining causation to enrich ontologies.