Among all the tools one can find in descriptive statistics, principal component analysis (PCA) is likely to be of much importance to phonetics for it provides an automatic process to extract features. If we represent a set of speech sounds as a cloud of points in a multi-dimensional space, CPA determines what are, for this cloud, the axes according to which distances between points are maximized. Moreover, such axes permit us to consider the weighted cloud only through its shape, while its absolute position in the space is left out. A formal definition of its structure can be based in this sense upon them.
Cite as: Nguyen-Trong, N. (1989) A recent advance in factorial analysis, related to phonetic feature extraction. Proc. First European Conference on Speech Communication and Technology (Eurospeech 1989), 1369
@inproceedings{nguyentrong89_eurospeech,
author={Noel Nguyen-Trong},
title={{A recent advance in factorial analysis, related to phonetic feature extraction}},
year=1989,
booktitle={Proc. First European Conference on Speech Communication and Technology (Eurospeech 1989)},
pages={1369}
}