Knowledge Mining for Defining Systemic Engineering Practices

Abstract

This paper addresses the problem on generating knowledge graphs from research papers related to wildfires, and their impacts on the electrical grid infrastructure. A novel framework based on part-of-speech tagging, word frequency statistics, and document clustering is proposed for extracting information and generating knowledge graphs from a selfcurated corpus of wildfire-related research work abstracts. The proposed method is capable of capturing a wide range of insightful information from the self-generated domain-specific corpus. Systemic engineering practices, such as strategies for wildfire mitigation by electric utilities has been included as a case study in this paper. An application of the research is implemented for simulating the management and mitigation of wildfires by electric utilities in California.