Attribute-Value Extraction: The case of a Real Estate Observatory

Abstract

Structured information is a valuable resource in information systems construction. The process of structuring unstructured data can be automated, but since machines can’t directly process natural language texts, NLP techniques are required. This work aims to evaluate different approaches to perform attribute-value extraction in real estate descriptions, in the context of the construction of a real estate observatory for the Province of Buenos Aires. The performance of each model is measured using precision, recall and F1-score with a partial matching approach.