MATLAB Answers

Binary classification using SVM or ANN?

14 views (last 30 days)
Mario Grgic
Mario Grgic on 28 Feb 2020
Answered: Dinesh Yadav on 2 Mar 2020
Hello
I have a project in which i have to create a model using either SVM or ANN metode that predicts if the symbol is zero (0) or one (1).
My input data are complex numbers. Each symbol is detetment by column of 20 complex numbers.
For instance:
1 0
-30.5946 +45.6142i -9.5504 -52.0076i
-28.2553 +41.0503i -9.9506 -47.9197i
-27.5315 +44.5150i -8.1468 -51.1290i
-31.2449 +42.1544i -9.6489 -44.7405i
-26.4084 +42.9527i -4.9537 -49.0714i
-29.0869 +42.3309i -9.2641 -44.3664i
-26.8713 +41.1549i -5.1648 -50.1355i
-25.4910 +41.5865i -12.3554 -47.2671i
-28.8041 +43.6515i -8.9379 -49.0792i
-26.9008 +39.5717i -11.2396 -54.5616i
-28.1639 +42.3321i -6.1479 -55.7859i
-34.1447 +40.4079i -6.2652 -51.2927i
-25.9209 +42.8146i -14.1721 -53.1520i
-21.8625 +41.7201i -11.8595 -56.1244i
-25.8175 +42.4621i -12.2793 -54.7263i
-29.2388 +37.0193i -7.2035 -52.2509i
-26.9880 +38.4291i -4.3028 -55.7246i
-30.9404 +40.1483i -9.4492 -50.1819i
-22.5961 +41.8203i -6.3496 -54.7031i
-25.0664 +44.7902i -7.1984 -47.8074i
If anybody can help me how to start my project or can give me a example similar to my.
Thank you

  0 Comments

Sign in to comment.

Answers (1)

Dinesh Yadav
Dinesh Yadav on 2 Mar 2020
For Binary classification use SVM as it will be more efficient computationally. Try the following commands to train and classify new data.
load data.mat
SVMModel = fitcsvm(x,y);
[label,score] = predict(SVMModel, newxval);
Refer to the following links to understand more about SVM using MATLAB.
As of now MATLAB does not support complex datatype for SVM classification. Go through the following link

  0 Comments

Sign in to comment.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!