SVM train and Classify

10 vues (au cours des 30 derniers jours)
Rajith
Rajith le 5 Août 2014
Commenté : Image Analyst le 22 Août 2014
Hi,
I am using SVM to train and clasify the audio files based upon extracting the MFCC and PLP features for 10 audio files And store it in a .mat file. I have to classify depending upon the high pass filter, low pass filter,etc...
I am new to SVM I have to train some of the audio files and test the data (different) later on so I am bit confused what to do next? I have to store the features data in .mat file and input that as the training matrix?
Thanks in advance
  2 commentaires
Dasharath Gulvady
Dasharath Gulvady le 5 Août 2014
  • What have you tried so far?
  • what is the format of the audio files?
  • If you are looking for MATLAB functions for train and test, "svmtrain" and "svmclassify" are the functions.
  • If you are looking for functions related to audio, refer to this link
Rajith
Rajith le 5 Août 2014
Modifié(e) : Rajith le 5 Août 2014
@Dasharath Gulvady Till now i have extracted MFCC and PLP features of all 10 audio file and saved in .mat file........ format of the audio file is .wav............... I want to know what input I should give to train in SVM?......... And how to create a class to classify? I want the class to be as Highpass, Low pass, etc..
Thanks

Connectez-vous pour commenter.

Réponse acceptée

Dasharath Gulvady
Dasharath Gulvady le 5 Août 2014
Modifié(e) : Dasharath Gulvady le 5 Août 2014
Rajith,
For training, use "svmtrain" function. Here is the documentation page for "svmtrain". Note that in the documentation, input argument "Group" is where you can mention the class labels - high pass filter, low pass filter,etc and "Training" is your training data.
You can find few examples on SVM here
  5 commentaires
Dasharath Gulvady
Dasharath Gulvady le 6 Août 2014
Modifié(e) : Dasharath Gulvady le 6 Août 2014
I have modified the above link(in my previous comment). Please try it now.
Image Analyst
Image Analyst le 22 Août 2014
You must have logged in with an Answers account that is not associated with your license. You should be able to find the same documentation as is online in your local help from MATLAB itself.

Connectez-vous pour commenter.

Plus de réponses (0)

Catégories

En savoir plus sur Introduction to Installation and Licensing dans Help Center et File Exchange

Community Treasure Hunt

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

Start Hunting!

Translated by