Patents are a valuable source of technical intelligence, offering a wealth of knowledge that remains underutilized in academia, business, and industry—fields where such information could yield significant benefits. Expanding access to patent databases can provide cutting-edge insights across various technological domains. In the field of computational linguistics, the ontology of the Theory of Inventive Problem Solving (TRIZ), derived from extensive patent analyses, presents a significant opportunity to advance text mining for patents written in Portuguese. This study aims to develop an ontology based on concepts extracted from engineering and scientific knowledge, encompassing pure and applied physical effects, which can suggest potential generic solutions. The proposed semantic network of terms, originating from the patent domain, seeks to enhance the retrieval of technical intelligence embedded in patent documents. Classified as applied research of an exploratory nature, this work introduces an ontology with hierarchically organized and interrelated terms in Portuguese. It establishes a lexical and semantic bridge to enable the application of computational tools for extracting technical intelligence from Portuguese-language patents, an area that remains underexplored in text mining.

--

• Ciências da Computação e da Informação - Ciências Naturais
• Engenharia Civil - Engenharia e Tecnologia