Code covered by the BSD License  

Highlights from
Log Binning of Data

Log Binning of Data

by

 

06 Apr 2010 (Updated )

Takes in data, and then produces data for a log binned histogram

[midpts Freq eFreq]=lnbin(x,BinNum)
% This function take the input of a data vector x, which is to be binned;
% it also takes in the amount bins one would like the data binned into. The
% output is two vectors, one containing the normalised frequency of each bin 
% (Freq), the other, the midpoint of each bin (midpts).
% Added and error to the binned frequency: eFreq (As of June 30 2010). If this
% option is not required, just call the function without including the third out
% put; i.e.: [midpts Freq]=lnbin(x,BinNum). 

function [midpts Freq eFreq]=lnbin(x,BinNum)
x=sort(x);
i=1;
while x(i)<=0;
   i=i+1; 
end
str = num2str((length(x)-i)/length(x)*100);
stuff='Percentage of input vec binned ';
disp([stuff str])

FPT=x(i:length(x));
LFPT=log(FPT);
max1=log( ceil(max(FPT)) );
min1=log(floor(min(FPT)));
% min1=1;

LFreq=zeros(BinNum,1);
LTime=zeros(BinNum,1);
Lends=zeros(BinNum,2);

step=(max1-min1)/BinNum;


% ------------ LOG Binning Data ------------------------
for i=1:length(FPT)
    for k=1:BinNum
       if( (k-1)*step+min1 <= LFPT(i) && LFPT(i) < k*step+min1)
           LFreq(k)=LFreq(k)+1;
       end
        LTime(k)=k*step-(0.5*step)+min1;
        Lends(k,1)=(k-1)*step+min1;
        Lends(k,2)=(k)*step+min1;
    end  
end 
ends=exp(Lends);

widths=ends(1:length(ends),2)-ends(1:length(ends),1);
Freq=LFreq./widths./length(x);
eFreq=1./sqrt(LFreq).*Freq;
midpts = exp(LTime);

Contact us