Support vector machine via nonlinear rescaling method

In this paper we construct the linear support vector machine (SVM) based on the nonlinear rescaling (NR) methodology (see [Polyak in Math Program 54:177–222, 1992; Polyak in Math Program Ser A 92:197–235, 2002; Polyak and Teboulle in Math Program 76:265–284, 1997] and references therein). The formulation of the linear SVM based on the NR method leads to an algorithm which reduces the number of support vectors without compromising the classification performance compared to the linear soft-margin SVM formulation. The NR algorithm computes both the primal and the dual approximation at each step. The dual variables associated with the given data-set provide important information about each data point and play the key role in selecting the set of support vectors. Experimental results on ten benchmark classification problems show that the NR formulation is feasible. The quality of discrimination, in most instances, is comparable to the linear soft-margin SVM while the number of support vectors in several instances were substantially reduced.     

Fractional Order Disturbance Observer for Robust Vibration Suppression

For the first time, the fractional order disturbance observer (FO-DOB) is proposed for vibration suppression applications such as hard disk drive servo control. It has been discovered in a recently published US patent application (US20010036026) that there is a tradeoff between phase margin loss and strength of the low frequency vibration suppression. Given the required cutoff frequency of the low pass filter, also known as the Q-filter, it turns out that the relative degree of the Q-filter is the major tuning knob for this tradeoff. The solution in US20010036026 was based on an integer order Q-filter with a variable relative degree. This actually motivated the use of a fractional order Q-filter. The fractional order disturbance observer is based on the fractional order Q-filter. The implementation issue is also discussed. The nice point of this paper is that the traditional DOB is extended to the fractional order DOB with the advantage that the FO-DOB design is now no longer conservative nor aggressive, i.e., given the cutoff frequency and the desired phase margin, we can uniquely determine the fractional order of the low pass filter.     

A comparison between optically active CdZnSe/ZnSe and CdZnSe/ZnBeSe self-assembled quantum dots: effect of beryllium

The composition and size of optically active Cd_xZn_1−xSe/ZnSe quantum dots are estimated with a previously developed method. The results are then compared with those obtained for Cd_xZn_1−xSe/Zn_0.97Be_0.03Se QDs. We show that introducing Be into the barrier material enhances both Cd composition and quantum size effect of optically active quantum dots.     

Dual Homogeneous Forms and Varieties of Power Sums

We review the classical definition of the dual homogeneous form of arbitrary even degree which generalizes the well-known notion of the dual quadratic form. Following the ideas of S. Mukai we apply this construction to the study of the varieties parametrizing representations of a homogeneous polynomial as a sum of powers of linear forms.Research supported in part by NSF Grant DMS 0245203.Lecture held in the Seminario Matematico e Fisico on October 15, 2003Received: April, 2004     

A differential calculus on the quantum groupSU q (2) comparible with the gauge-transformation structure

We consider a three-dimensional bicovariant differential calculus on the quantum group SU_q(2) that is compatible with the structure of gauge transformations. In the approach based on the global covariance with respect to the stability subgroup U(1), we show that the gauge-covariant exterior differential algebra on SU_q(2) is uniquely determined. We construct explicit representations of q-deformed Lie algebras in terms of differential operators. Translated from Teoreticheskaya i Matematicheskaya Fizika, Vol. 116, No. 2, pp. 265–276, August. 1998.     

Microwave frequency shifting using photon acceleration

Using a Bostick Gun as a plasma source, we have been investigating the effects of a moving plasma front on the frequency of an incident R.F. signal. We have been able to measure frequency shifts of 10 MHz frequency with a 2.6 GHz signal. The density is such that the plasma frequency is well above the transmitter frequency 1.3 meters from the plasma source. No magnetic field is used.     

Minimum Disparity Estimation in Linear Regression Models: Distribution and Efficiency

This paper deals with the minimum disparity estimation in linear regression models. The estimators are defined as statistical quantities which minimize the blended weight Hellinger distance between a weighted kernel density estimator of errors and a smoothed model density of errors. It is shown that the estimators of the regression parameters are asymptotic normally distributed and efficient at the model if the weights of the density estimators are appropriately chosen.