Smoothing Your Data with Polynomial Fitting: A Signal Processing Perspective
summary
The talk explains Savitzky–Golay polynomial smoothing and analyzes it from a frequency-domain / signal-processing viewpoint, showing that polynomial fitting equals convolution with a fixed kernel and how that affects noise attenuation. It demonstrates how to choose parameters (polynomial degree and window size) using frequency analysis and gives implementation examples using NumPy and SciPy (code and slides provided on GitHub).
AI-written summary · about
chapters (13)
- 0:00Polynomial smoothing intro
- 2:13Why polynomial fitting matters
- 4:21Polynomial fitting basics
- 6:49Sliding window approach
- 9:40Linear time-invariant systems
- 13:58Impulse response in filtering
- 15:46Frequency domain analysis
- 17:31Convolution theory
- 19:36Key smoothing parameters
- 21:00Power consumption filtering
- 22:46Ambient light sensing challenges
- 23:49Results after filtering
- 24:32Summary
related talks
34mDimensionally reducing data squeezing out the good stuff
31mUsing Simpson’s Paradox to Discover Interesting Patterns in...
33mfastplotlib-scientific discovery through data visualization
01h 38mBasic Python Packages for Science
28mA decade of gr-specest -- Free Spectral Estimation!
32mCutting Edge Football Analytics in Python
02h 23mEffective Data Visualization
04mData Sonification: Making Music from the yield curve