This further imposes two self-inflicting limitations on any proposed model: (i) processing all the data is not always feasible and (ii) processing only a subset of data may result in loss of information, overfitting, and local maxima. The ratio of number of genes (features) to the number of patients (samples) is very skewed which results in the well-known curse-of-dimensionality problem. Microarray experiments produce a huge amount of gene-expression data from a single sample. The proposed approach addresses a data scarcity problem, is flexible to the choice of heterogeneous base classifiers, and is able to produce HAHR models comparable to the established MAQC-II results. The successful soldiers are combined through heterogeneous ensembles for optimal results. Every soldier of the ITO army is a base model with its own independently chosen Subset selection method, pre-processing, and validation methods and classifier. The algorithm produces results in two phases: (i) Lightweight Infantry Group (LIG) converges quickly to find non-local maxima and produces comparable results (i.e., 70 to 88% accuracy) (ii) Followup Team (FT) uses advanced tuning to enhance the baseline performance (i.e., 75 to 99%). The proposed ITO algorithm combines parameter-free and parameter-based classifiers to produce a high-accuracy-high-reliability (HAHR) binary classifier.
This paper presents a warzone inspired “infiltration tactics” based optimization algorithm (ITO)-not to be confused with the ITO algorithm based on the Itõ Process in the field of Stochastic calculus. Thus, generalized optimization is still a challenge open for further research. However, higher accuracy may not necessarily mean higher reliability of the model. The former are quick but may result in local maxima while the latter use dataset-specific parameter-tuning for higher accuracy. A number of different feature selection and classification techniques have been proposed in literature including parameter-free and parameter-based algorithms.
0 kommentar(er)
