Compare values in two matrix with different length by time in first row

Hello,

I have one aim to do comparison of values in two matrices without using for loops (array are huge).

I show you example with two small arrays:

Matrix_R =

9 x 2 double

[time value]

0.020 77

0.030 98

0.040 79

0.070 95

0.080 96

0.090 99

0.100 100

0.100 200

0.120 44

Matrix_G =

13 x 2 double

[time value]

0.010 75

0.020 75

0.030 75

0.040 77

0.050 78

0.060 96

0.070 99

0.080 100

0.090 100

0.100 110

0.110 100

0.120 100

0.130 110

Matrix_R (measure) has not homogenous time values

Matrix_G (ground truth) has homogenous time values

I want compare values between two matrices (with tolerance) from second column by same time. Comparison should be work in both directions.

Example (value tolerance 10):

at time 0.010 in Matrix_G is value 75, in Matrix_R there are no value -> False negative

at time 0.020 in Matrix_G is value 75, in Matrix_R is value 77 -> True positive

at time 0.030 in Matrix_G is value 75, in Matrix_R is value 98 -> False negative and False positive

Question:

• Can I use ismembertool by matrix with different lengths?
• Maybe it makes sense to change time rows of Matrix_R to homogenous and take 0 like values?

Matrix_R_mod =

9 x 2 double

[time value]

0.010 0

0.020 77

0.030 98

0.040 79

0.050 0

0.060 0

0.070 95

0.080 96

0.090 99

0.100 100

0.100 200

0.110 0

0.120 44

what can I make with double answers of time?

I will be happy to hear some ideas from you! Thank you!

Answer 1

[Abhishek Gupta]

Hi,

You can perform the task mentioned above as follows: -

clc; clear;

%% Input variables
Matrix_R = [0.020 77; 0.030 98; 0.040 79; 0.070 95; 0.080 96; ...
    0.090 99; 0.100 100; 0.100 200; 0.120 44];

Matrix_G = [0.010 75; 0.020 75; 0.030 75; 0.040 77; 0.050 78; 0.060 96; ...
    0.070 99; 0.080 100; 0.090 100; 0.100 110; 0.110 100; 0.120 100; 0.130 110];

tol = 10;

%% Initialize output matrix, where
% column 1 -- all the elements in the Matrix_G with appropriate repetition
% colum 2 -- output column, where -1 -> false negative; 1 -> true positive; 0 -> False negative and False positive

freq = [Matrix_G(:,1), histc(Matrix_R(:,1),Matrix_G(:,1))]; % get frequncy of each element of Matrix_G in Matrix_R
repetitionNumber = freq(freq(:,2) > 1, 1);  % get the element which repeats more than once
repetitionTimes = freq(freq(:,2) > 1, 2);   % get number of time element repeats more than once
outputTime = sort([freq(:,1); repelem(repetitionNumber, repetitionTimes - 1)]); % sort w.r.t time
output = [outputTime -1*ones(size(outputTime,1), 1)];   % initialize output matrix with all time rows == -1

%% 
% get common elements between Matrix_R and Matrix_G
[Lia,Locb] = ismember(Matrix_R(:,1), Matrix_G(:,1));

A = Matrix_R(Lia, 2);
B = Matrix_G(Locb, 2);

toleranceCheck = abs(A-B) <= 10;   % find elements which are within the tolerance

output(ismember(output(:,1), Matrix_R(:,1)), 2) = toleranceCheck;

[Nik Rocky]

Wow, thank you very much! Is a great solution of my case!