Feature extraction method should use

I have generated 1-D array from my walking pattern. I'm trying to use some feature extraction method to tell the difference in walking fast, normal, and run between male subjects. I have did cross correlation between normal and fast walking, but I need to go further to some machine learning or some algorithms. I can tell the difference between subject motion by their average weight or from the peak at the heel contact, so that is my parameter I can diagnose. But still which method I should go through so that the MATLAB can illustrated some other strong explanation between walking patterns. Im attaching my txt files for the normal walk and fast walk. it is [1 X some-Numbers].

close all;
clear all;
normal_w = load('normal_w.txt');
fast_w = load('fast_w.txt');

title('normal walking');

   figure(1);
   plot(fast_w(1:600),'r');hold on;
   plot(normal_w(1:600),'b');
   Grid;
   title('Normal and fast walking');
   legend('fast','normal');
   xlabel('time in seconds X 100');
   ylabel('KPa');

   %%%%%Finding cross correlation between two files.%%%%%%%
    nor_ = normal_w(1:600);
    fst_ = fast_w(1:600);
    [corr_seq lags] = xcorr(nor_,fst_); %return the lag time
   figure(2);
   plot(lags, corr_seq);
   xlabel('lag');
   ylabel('Correlation measurment');

   %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

   %%%Find autocorrelation for normal_walking%%%
   nor_auto = xcorr(normal_w);
   figure(3);
   plot(nor_auto);hold on;
  %%%Find autocorrelation for fast_walking%%%
   f_auto = xcorr(fast_w);
   plot(f_auto,'r');
  title('Autocorrelation in normal and fast walking individually');
  xlabel('lags');
  ylabel('correlation seq');

% %%%Correlation cofficicent %%%
  nor_walk = normal_w(1:900);
  fst_walk = fast_w(1:900);
%  a= corrcoef(nor_walk,fst_walk);
%  figure(4);
%   plot(a);
% %%%%%%%%

r_cpr = corr2(nor_walk,fst_walk); 

s3 = load('s3.txt');
s5 = load('s5.txt');
corr_sensor_3_5 =corr2(s3,s5);

s6 = load('s6.txt');
s7 = load('s7.txt');
corr_sensor_6_7 = corr2(s6,s7);
s8 = load('s8.txt');
corr_sensor_7_8 = corr2(s7,s8);

figure(4);
C = cov(nor_walk, fst_walk);
plot(C);

[Image Analyst]

There's nothing built in to do this at a high level. You'll have to construct it from lower level functions, or sponsor a university group to solve it for you. I can't work with you on developing this project because it could take years or research.

Answer 1

[Ahmed Mostfa]

hello, anyone could assist me?