Feature engineering for energy systems

• Main Author: Liew, Wee Yeong
• Other Authors: Hung Dinh Nguyen
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2020
• Subjects: Engineering::Electrical and electronic engineering
• Online Access: https://hdl.handle.net/10356/140224
• Institution: Nanyang Technological University

Features play an important role in machine learning applications. This report explores several fundamental elements of feature engineering as well as automated feature Specifically, we deploy the state-of-the-art feature engineering tool known as FeatureTools in order to obtain relevant features. We also compare directly such features to corresponding manually calculated ones using the same data obtained from Centrum Enet. To ensure that the experiment is as unbiased and functional as possible, we apply several measures such as repeated iterations for up to 25 times, 5-fold cross-validation techniques, and Matthews Correlation Coefficient. This project is a pioneer work that proposes to leverage automated feature engineering in a power engineering context, in particular, partial discharge analysis for condition monitoring.