A Comprehensive Evaluation of a Novel Approach to Probabilistic Information Extraction from Large Unstructured Datasets

Abstract

In this paper, we discuss the evaluation of the probabilistic extraction as introduced in [1], by considering three different datasets introduced in [1] - [3]. the results show the potential of the approach, as well as its reliability and efficiency when analyzing datasets with different properties and structures. This is part of ongoing research aiming to provide a tool to extract, assess and visualize intelligence extracted from large unstructured datasets.