Independent Component Analysis Using Random Projection For Data Pre-Processing
Adeel Ansari, Afza Bt, B Abas, Bandar Seri Iskandar · 2012
There is an inherent difficulty of finding out latent structures within high dimensional data repository centers. It is assumed that this data is generated by these unknown latent variables and with the relationship and interaction between each of them. The task is to find these latent variables and the way they interact, given the observed data only. It is assumed that the latent variables do not depend on each other but act independently. A popular method for counteracting with the above stated problem scenario is independent component analysis (ICA). An ICA algorithm for analyzing complex valued signals is given; and an ICA-type algorithm is used for analyzing the topics in dynamically changing text data. Experimental results are given on all of the presented methods. Another, partially overlapping problem considered in this paper is dimensionality reduction. Empirical validation is given on a computationally simple method called random projection: it does not introduce severe distortions in the data. It is also proposed that random projection could be used as a preprocessing method prior to ICA, and experimental results are shown to support this claim.