Evaluation of noise reduction techniques in the splice junction recognition problem

Abstract

The Human Genome Project has generated a large amount of sequence data. A number of works are currently concerned with analyzing these data. One of the analyses carried out is the identification of genes' structures on the sequences obtained. As such, one can search for particular signals associated with gene expression. Splice junctions represent a type of signal present on eukaryote genes. Many studies have applied Machine Learning techniques in the recognition of such regions. However, most of the genetic databases are characterized by the presence of noisy data, which can affect the performance of the learning techniques. This paper evaluates the effectiveness of five data pre-processing algorithms in the elimination of noisy instances from two splice junction recognition datasets. After the pre-processing phase, two learning techniques, Decision Trees and Support Vector Machines, are employed in the recognition process.

Computational discovery of regulatory DNA motifs using evolutionary computation

“…A properly designed data filtering mechanism is considered to favor the system performance and efficiency [17]. In this paper, we adopt a filtering approach proposed in [33] to increase the signal-to-noise ratio of the investigated dataset by discarding background k-mers without false dismissals.…”

“…Recently, the concept of negative data filtering (e.g. noise reduction) has been successfully applied for solving various problems in Bioinformatics [17]- [20]. Therefore, we believe a properly designed filtering mechanism by removing the biologically-irrelevant background sequences without touching any true motifs could increase the signal-to-noise ratio of the DNA sequences and potentially improve the predication accuracy.…”

Noise reduction for instance-based learning with a local maximal margin approach

“…It is generally accepted that inductive learning systems in the medical domain are dependent on the quality of the data (Pechenizkiy et al, 2006) and there has been significant research into data cleansing in bioinformatics (Malossini et al, 2006;Gamberger et al, 2000;Lorena and Carvalho, 2004;Tang and Chen, 2008a,b). Although instance based techniques such as k-NN are not generally used for classification in much of this research, noise reduction is an important element in the process as it can result in the simplification of the models created.…”

“…Although instance based techniques such as k-NN are not generally used for classification in much of this research, noise reduction is an important element in the process as it can result in the simplification of the models created. Lorena and Carvalho (2004), for example, found that preprocessing the training data to remove noise resulted in simplifications in induced SVM classifiers and higher comprehensiveness in induced decision tree classifiers.…”

“…Since case-based explanation will invoke individual cases as part of the explanation process it is important that noisy cases can be eliminated if possible. Even if noise reduction will not improve the classification accuracy of learning algorithms that have been developed to be noise tolerant, researchers have argued that noise reduction as a preprocessing step can simplify resulting models, an objective that is desirable in many circumstances (Lorena and Carvalho, 2004).…”