cee master project
by jiacheng qu
2017.1.1
to execute: run .m file named 0zero to 0sixth in order
Updates:
Here is a pdf file that I used for presentations
This browser does not support PDFs. Please click here to view the PDF: View PDF.
zero_basic_plots:
select .wav file and plot its features (could be randomized using randi() function)
choosedialog_art.m: Choose a filename
(figure 1) Spectrogram using short-time Fourier transform
(figure 2) inputdlg(): Enter Endpoints for selected freq
(figure 3) pwelch(): Weltch Power Density using
(figure 4 & 5) loudness_1991.m:
plot the Partial Loudness sone/Bark (fig 5)
inside the loudness_1991.m, call filter_third_octaves_downsample.m: plot 1/3-octave
spectrum using bar() (fig 4)
first_welch_power_density:
this is the very first script that saves/plots welch_power_density of a particular category of noise
choosedialog_art_all.m: Choose a category
choosedialog_plot_power_all_avg_dba.m: let the user choose from 'save and plot all figures on one graph' or 'pull and save the data only'
second_avg_welch_power_density:
plot avg of welch_power_densit
use mean() and std() to calculate the mean and stddiv
call plotRstyleUncert.m to print the avg wpd for each category
third_pull_fraction_of_power:
pull the data?fraction_of_power
choosedialog_art_all.m: choose a noise categroy
use inputdlg() to get the freq ranges and save them as seven_octave_freq_band.mat
for e.g.
type: '22','44','88','176','353','707','1414'
for range sets: 0-22 23-44 45-88 89-176 177-353
354-707 708-1414 1415-3000 (Hz)
fractionPower.m: calculate the fraction of power according to the selected range set
and then, name them by 'fractionofpower_' CATEGORY '.mat'
fourth_notboxplot_fraction_of_power:
plot 'notboxplots' of the power in different frequency bins
load seven_octave_freq_band.mat: get the freq ranges
default freq bands are: {'0-22Hz','22-44Hz','44-88Hz','88-176Hz','176-353Hz','353-707Hz','707-1414Hz','1414-3000Hz'} ;
use choosedialog_art.m to choose a category
load 'fractionofpower_' CATEGORY '.mat' and plot the
'notboxplots' by calling notBoxPlot(fractionofpower_category)
fifth_boxplot_of_fraction_of_power_side_by_side:
plot all the boxplots of fraction_of_power side by side
load seven_octave_freq_band.mat: get the freq ranges
default freq bands are: {'0-22Hz','22-44Hz','44-88Hz','88-176Hz','176-353Hz','353-707Hz','707-1414Hz','1414-3000Hz'} ;
load 'fractionofpower_' CATEGORY '.mat'
plot the boxplots of 'fraction of power' side by side
rail_fracpower = 'g' = green
truck_fracpower = 'm' = magenta
aircraft_fracpower = 'b' = blue
sixth_gscatter_and_prediction:
plot gscatter labeled by group together with the classification process
load variance_all.mat, xmeans_all.mat, ALL_Xmean.mat;
call gscatter() to plot Xmeans(dB) v Variance and label
them by group
callchoosedialog_classifier.m to choose a classifier
'Naive Bayes'
'Discriminant Analysis'
'Classification Tree'
'K - Nearest Neighbor'
'3D Classification Probability'
plot the predicted class region