This is the readme file for the supplementary matlab code provided with `On Acoustic Emotion Recognition: Compensating for Covariate Shift' submitted to IEEE Transactions on Audio, Speech and Language Processing.

Here I give the description of each folder and the useful function.

./Data
----------
	This folder should contain your mat files for the extracted features.  I have used OpenEAR toolkit to extract features.  I am expected following variables in each file.
	Data 		-- contains extracted features, each row is a observation vector
	feat_name 	-- name of each feature
	Lbl		-- class label
	wavname		-- name of each observation vector, this will be required for SI-cv. OpeanEAR toolkit gives you the name of each observation vector

./KLIEP
----------
	This folder contains the script for calculating the Importance weights using KLIEP method given in Y. Tsuboi, H. Kashima, S. Hido, S. Bickel, and M. Sugiyama, ``Direct density ratio estimation for large-scale covariate shift adaptation'', in Proceedings of the 2008 SIAM International Conference on Data Mining, Atlanta, Georgia, 2008, pp. 443454.

./KMM
----------
This folder contains the script for calculating the Importance weights using KMM method given in A. Gretton, A. Smola, J. Huang, M. Schmittfull, K. Borgwardt, and B. Schoelkopf, Dataset Shift in Machine Learning. Cambridge, MA: MIT Press, 2009, ch. 8. Covariate Shift and Local Learning by Distribution Matching, pp. 131160.

./uLSIF
----------
This folder contains the script for calculating the Importance weights using KLIEP method given in T. Kanamori, S. Hido, and M. Sugiyama, A least-squares approach to direct importance estimation, Journal of Machine Learning Research, vol. 10, pp. 13911445, 2009.

./svmSteeveGunnMod
----------
Contains the modified version of S.R.Gunn's code for matlab.  This code has been modified so that it can accept IW calcluted by the above methods. Compiled mex files for the code are available in ./svmSteeveGunnMod/Optimiser.  They have been compiled for windows 32 & 64 bit edition and Linux 64 bit. Copy the required file in the ./svmSteeveGunnMod folder.

./function
----------
This folder contains useful functions that have been useful in my work.

./scripts
----------
Contains the test scripts for running the SI-CV and SD-CV on the data using OvR importance weighted SVMs.  These scripts can be used directly/modified as required.  

It will be nice if you refer to my paper.