Abstract Independent component analysis (ICA) is a broadly used method for blind source separation (BSS). The mature ICA model has a restriction that the number of sources must equal that of sensors used to collect data, which is hard to meet in practice. An overdetermined ICA method was proposed and successfully used in the analysis of human colonic pressure signals. Using principal component analysis (PCA), the method estimated the number of sources and reduced the dimension of observed signals so that it is the same with that of the sources; and then, FastICA was used to estimate all the sources. From 26 groups of colonic pressure recordings, colonic motility patterns were extracted, which not only proved the effectiveness of this method, but also would greatly facilitate further medical researches.
|
|
Received: 17 September 2009
Published: 30 November 2010
|
|
|
|
|
|
