Chapter 4 : Data Processing Flashcards
Where does data come from?
Agent’s sensors
What does higher sampling frequency mean?
Less information missing
What is aliasing?
When sampling, signals become aliases of each other and can’t reconstruct original signal anymore
What is noise?
Obscures features in data
What is artificat?
Makes it appear as a feature exists when it does not
is DFT created on idea of in time any signal can be seen as a sum of sine functions
True
What is PSD (Power Spectrum Density)?
Shows the power at each point of frequency. Co-related to squared amplitude of DFT.
What is an modified Periodogram?
When the window size is not a vector of 1 x number of data points
What is welch method?
- Divide signal into segments
- Take periodogram of each of the segments
- Average all periodograms
Purpose of welch method?
Smoother PSD with reduced variance
What domain are filters applied?
Frequency domain
What is frequency response?
Explains how a filter or system effects signals in frequency domain in terms of amplitude response and phase response
What is FIR filter?
Impulse Response is finite because no feedback loop
What type of filter is moving average filter?
Low pass filter
Why normalize filter?
So the size of the output does not depend on size of input
What is the FIR Output Filter Delay?
Output delay of symmetric FIR filters is
(Window size - 1) / 2
Why do we need other filters other then moving average filter?
- Moving average filter needs to be large to remove a lot of noise
- A large filter causes more delay
- Transition band of moving average filter can be very large
What is a matched filter?
FIR filter which tries to extract features from a known spatial signal
What are problems with matched filter?
- Very sensitive to signal change
- Very fine-tuned for signals
What type of output does median filter give?
Constant output (box version)
What is the purpose of feature selection?
- Achieves faster training
- Select ML models with less complexities which leads to a less chance of overfit and easier to interpret
- Better generalization can be achieved, accuracy can be improved
What does variance-based feature selection do?
Trying to extract features with most information in them
What does correlation-based feature selection do?
Tries to remove features that are very similar to other features
What does univariate-based feature selection do?
Evaluate each feature individually with respect to output to see which ones are most important