Basic Pitfalls in Waveform Analysis

Talk

Go to NumFOCUS academy page.

This talk focuses on the usual preparation steps of waveform analysis, digitizing, and converting to spectra. A simple python module DEEPS shows the errors of waves having frequencies lower than Nyquist frequencies, which is verified on an open dedicated dataset and YouTube sounds. The verification may improve domain-specific feature extractions of your analyses.

Speaker

Yukio Okuda

An Engineer on industrial hardware and software in 50 years

OSS Activities

• Python Acceleration, PyConJP Sep. 2018
• TensorFlow Eager, TFUG Jan. 2018
• HDD life prediction by R, Kaggle Jun. 2017
• CERN/Root bug, 2008

Industrial Data Analysis

• Photovoltaic degradation modeling from time-series data sets, IEEE-IRPS 2012, PVSEC-Frankfurt 2012 3DO.13.6
• A member of the technical program committee of IEEE-ITC(Inter. Test Conference), 2004 - 2008
• Estimating LSI path delay variations, IEEE-ITC 2007, Figures@SEMANTIC
• Modeling device mal-characteristics, IEEE-ITC 2003, Figures@SEMANTIC
• Obtaining a regularity of IDDQ currents, IEEE-VTS 2002
• Clustering LSIs by intrinsic leakage current, IEEE-ITC 2000, Figures@SEMANTIC
• Predicting the failure rate from the final test data sets, IEEE-ITC 1998