In this paper study, we develop a Bayesian Network model selection principle that address addresses the incommensurability of network features obtained from incongruous datasets and overcomes performance irregularities of the Minimum Description Length model selection principle. This is achieved (i) by approaching model evaluation as a classification problem, (ii) by estimating the effect that sampling error has on the satisfiability of conditional independence criterion, as reflected by Mutual Information, and (iii) by utilizing this error estimate to penalize uncertainty in the Minimum Uncertainty (MU) model selection principle. We validate our findings numerically and demonstrate the performance advantages of the MU criterion. Finally, we illustrate the advantages of the new model evaluation framework on a tRNA structural biology example.

A novel quasi-twisted miniaturized wideband Branch Line Coupler (BLC) is proposed. The design is based on bisecting the conventional microstrip line BLC transversely and folding bisected sections on double-layered substrates with a common ground plane in between. The pair of quarter wavelength horizontal parallel arms is converted to a Z-shape meandered microstrip line in the designed structure. On the other hand, the pair of quarter wavelength vertical arms are cut by half into two lines and converted to a periodically loaded slow wave structure. The bisected parts of BLC are placed on the opposite side of the doubled-layer substrate and connected through four vias passing through the common ground plane. This technique enabled a compact BLC size of 6.4×18 mm2, which corresponds to a surface area miniaturization by ~50% as compared to the classical BLC size of 10×23 mm2 at 6GHz, Moreover, the attained relative bandwidth is 73.9% (4.6-10 GHz) for S11, and 75.8% (4.5-10 GHz) for S33 as compared to 41% for the ordinary BLC at the same resonant frequency. The circuit is constructed on a double-layered low-cost FR4 substrate with a relative permittivity of 4.3, and a loss tangent of 0.025. An isolation of -13dB was realized in both S13 and S31 demonstrating an excellent performance. The transmission coefficients between input/output ports S21, S41, S23, and S43 are between -3.1dB to -3.5dB at the frequency of 6GHz. Finally, the proposed BLC provides phase differences between output ports of 90.5o and 94.8o at the frequency of 6GHz when the input ports 1 and 3 are excited, respectively. The presented design offers the potential of being utilized as a unit cell for building a Butler-matrix (BM) for sub-6GHz 5G beamforming networks.

This paper addresses the issues of how we can quantify structural information for nonparametric distributions and how we can detect its changes. Structural information refers to an index for a global understanding of a data distribution. When we consider the problem of clustering using a parametric model such as a Gaussian mixture model, the number of mixture components (clusters) can be thought of as structural information in the model. However, there does not exist any notion of structural information for nonparametric modeling of data. In this paper we introduce a novel notion of {\em kernel complexity} (KC) as structural information in the nonparametric setting. The key idea of KC is to combine the information bias inspired by the Gini index with the information quantity measured in terms of the normalized maximum likelihood (NML) code length. We empirically show that KC has a property similar to the number of clusters in a parametric model. We further propose a framework for structural change detection with KC in nonparametric distributions. With synthetic and real data sets we empirically demonstrate that our framework enables us to detect structural changes underlying the data and their early warning signals.