ABSTRACT
Normalized variants of logarithmic algorithms including normalized LMLS (NLMLS) and normalized LLAD (NLLAD) along with their error-normalized variants are presented in this chapter to further increase the stability and convergence performance, lower the computational complexity, and improve the exon tracking ability. So as to further lessen the difficulty to perform calculations for an adaptive exon predictor (AEP), all these adaptive algorithms are combined with sign algorithms and their maximum versions are deliberated. In this chapter, we present distinct AEPs developed using logarithmic-based realization of various adaptive filtering techniques for exon position identification in DNA sequences. By considering this, we developed several AEPs for accurate prediction of exons in DNA signals using MATLAB. Towards analyzing the performance of various AEPs developed, simulation studies are done on ten real genomic datasets related to Homo sapiens taken from the National Center for Biotechnology Information (NCBI) genome databank. These techniques are further extended to sign-based realizations to decrease computational complexity. In addition to the adaptive filtering techniques, this chapter also provides performance analysis depending upon various metrics including specificity, precision, sensitivity, computational complexity issues, and convergence analysis. Also, a detailed discussion on various normalized logarithmic-based algorithms considered for genomic sequence analysis is presented.
