Uniform Prior Weighted Sampling Stimulus Generation

This is a stimulus generation class in which the number of filled bins is selected from a uniform distribution on [min_bins, max_bins], but which bins are filled is determined by a non-uniform distribution.

Unique Properties

This stimulus generation class has two properties in addition to those inhereted from the Abstract and Abstract Binned classes. Default:

- bin_probs = [] % Assigned via `set_bin_probs()`
- alpha_ = 1

generate_stimulus

[stim, Fs, spect, binned_repr, frequency_vector] = generate_stimulus(self)

Generates a stimulus by generating a frequency spectrum with self.unfilled_dB and self.filled_dB dB amplitudes. The number of filled bins is selected from a uniform distribution on [self.min_bins, self.max_bins], but which bins are filled is determined from a non-uniform distribution.

OUTPUTS:

stim: self.nfft + 1 x 1 numerical vector, the stimulus waveform,

Fs: 1x1 numerical scalar, the sample rate in Hz.

spect: self.nfft / 2 x 1 numerical vector, the half-spectrum, in dB.

binned_repr: self.n_bins x 1 numerical vector, the binned representation.

frequency_vector: self.nfft / 2 x 1 numerical vector, the frequencies associated with the spectrum, in Hz.

Class Properties Used:

- n_bins

See Also

• UniformPriorWeightedSamplingStimulusGeneration.sample

---

UniformPriorWeightedSamplingStimulusGeneration

class constructor

---

get_bin_occupancy

bin_occupancy = self.get_bin_occupancy();

Compute the bin occupancy, which is a self.n_bins x 1 vector which counts the number of unique frequencies in each bin. This bin occupancy quantity is not related to which bins are "filled".

OUTPUTS

• bin_occupancy: self.n_bins x 1 representing the bin occupancy quantity, e.g. bin_occupancy(1) is the occupancy for the first bin.

See Also

• AbstractBinnedStimulusGenerationMethod.get_freq_bins

---

set_bin_probs

self.set_bin_probs()
self.set_bin_probs(1.3)

Sets self.bin_probs equal to the bin occupancy, exponentiated by alpha_. If alpha_ is empty, uses the existing self.alpha_ value. Otherwise, self.alpha_ is set as well, and that value is used.

ARGUMENTS

• self: the object
• alpha_: 1x1 nonnegative scalar

See Also

• UniformPriorWeightedSamplingStimulusGeneration.get_bin_occupancy

---

sample

filled_bins = self.sample(weights, values)

Get a vector of indices referred to bins that should be filled, by taking successive weighted samples without replacement from a list of values with associated weights.

ARGUMENTS - n_bins_to_fill: 1x1 integral scalar indicating how many bins to fill

OUTPUTS - filled_bins: n_bins_to_fill x 1 vector of bin indices