An efficient feature selection method for mobile devices with application to activity recognition

This paper presents a feature selection method for data classification, which combines a model-based variable selection technique and a fast two-stage subset selection algorithm. The relationship between a specified (and complete) set of candidate features and the class label is modelled using a non-linear full regression model which is linear-in-the-parameters. The performance of a sub-model measured by the sum of the squared-errors (SSE) is used to score the informativeness of the subset of features involved in the sub-model. The two-stage subset selection algorithm approaches a solution sub-model with the SSE being locally minimized. The features involved in the solution sub-model are selected as inputs to support vector machines (SVMs) for classification. The memory requirement of this algorithm is independent of the number of training patterns. This property makes this method suitable for applications executed in mobile devices where physical RAM memory is very limited. An application was developed for activity recognition, which implements the proposed feature selection algorithm and an SVM training procedure. Experiments are carried out with the application running on a PDA for human activity recognition using accelerometer data. A comparison with an information gain based feature selection method demonstrates the effectiveness and efficiency of the proposed algorithm.

Non Classical Design in MOSFETs for Improving OTA gain-bandwidth trade off

In this paper, gain-bandwidth (GB) trade-off associated with analog device/circuit design due to conflicting requirements for enhancing gain and cutoff frequency is examined. It is demonstrated that the use of a nonclassical source/drain (S/D) profile (also known as underlap channel) can alleviate the GB trade-off associated with analog design. Operational transconductance amplifier (OTA) with 60 nm underlap S/D MOSFETs achieve 15 dB higher open loop voltage gain along with three times higher cutoff frequency as compared to OTA with classical nonunderlap S/D regions. Underlap design provides a methodology for scaling analog devices into the sub-100 nm regime and is advantageous for high temperature applications with OTA, preserving functionality up to 540 K. Advantages of underlap architecture over graded channel (GC) or laterally asymmetric channel (LAC) design in terms of GB behavior are demonstrated. Impact of transistor structural parameters on the performance of OTA is also analyzed. Results show that underlap OTAs designed with spacer-to-straggle ratio of 3.2 and operated below a bias current of 80 microamps demonstrate optimum performance. The present work provides new opportunities for realizing future ultra wide band OTA design with underlap DG MOSFETs in silicon-on-insulator (SOI) technology. Index Terms—Analog/RF, double gate, gain-bandwidth product, .

Improved Nonlinear PCA for Process Monitoring Using Support Vector Data Description

Nonlinear principal component analysis (PCA) based on neural networks has drawn significant attention as a monitoring tool for complex nonlinear processes, but there remains a difficulty with determining the optimal network topology. This paper exploits the advantages of the Fast Recursive Algorithm, where the number of nodes, the location of centres, and the weights between the hidden layer and the output layer can be identified simultaneously for the radial basis function (RBF) networks. The topology problem for the nonlinear PCA based on neural networks can thus be solved. Another problem with nonlinear PCA is that the derived nonlinear scores may not be statistically independent or follow a simple parametric distribution. This hinders its applications in process monitoring since the simplicity of applying predetermined probability distribution functions is lost. This paper proposes the use of a support vector data description and shows that transforming the nonlinear principal components into a feature space allows a simple statistical inference. Results from both simulated and industrial data confirm the efficacy of the proposed method for solving nonlinear principal component problems, compared with linear PCA and kernel PCA.

Cardiopulmonary function in two human disorders of the hypoxia-inducible factor (HIF) pathway: von Hippel-Lindau disease and HIF-2 alpha gain-of-function mutation

The hypoxia-inducible factors (HIFs; isoforms HIF-1 alpha, HIF-2 alpha, HIF-3 alpha) mediate many responses to hypoxia. Their regulation is principally by oxygen-dependent degradation, which is initiated by hydroxylation of specific proline residues followed by binding of von Hippel-Lindau (VHL) protein. Chuvash polycythemia is a disorder with elevated HIF. It arises through germline homozygosity for hypomorphic VHL alleles and has a phenotype of hematological, cardiopulmonary, and metabolic abnormalities. This study explores the phenotype of two other HIF pathway diseases: classic VHL disease and HIF-2 alpha gain-of-function mutation. No cardiopulmonary abnormalities were detected in classic VHL disease. HIF-2 alpha gain-of-function mutations were associated with pulmonary hypertension, increased cardiac output, increased heart rate, and increased pulmonary ventilation relative to metabolism. Comparison of the HIF-2 alpha gain-of-function responses with data from studies of Chuvash polycythemia suggested that other aspects of the Chuvash phenotype were diminished or absent. In classic VHL disease, patients are germline heterozygous for mutations in VHL, and the present results suggest that a single wild-type allele for VHL is sufficient to maintain normal cardiopulmonary function. The HIF-2 alpha gain-of-function phenotype may be more limited than the Chuvash phenotype either because HIF-1 alpha is not elevated in the former condition, or because other HIF-independent functions of VHL are perturbed in Chuvash polycythemia.-Formenti, F., Beer, P. A., Croft, Q. P. P., Dorrington, K. L., Gale, D. P., Lappin, T. R. J., Lucas, G. S., Maher, E. R., Maxwell, P. H., McMullin, M. F., O'Connor, D. F., Percy, M. J., Pugh, C. W., Ratcliffe, P. J., Smith, T. G., Talbot, N. P., Robbins, P. A. Cardiopulmonary function in two human disorders of the hypoxia-inducible factor (HIF) pathway: von Hippel-Lindau disease and HIF-2 alpha gain-of-function mutation. FASEB J. 25, 2001-2011 (2011). www.fasebj.org