Exploiting the relationships among several binary classifiers via data transformation

The structural resemblance among several existing classifiers has motivated us to investigate their underlying relationships. By exploring into the mapping solutions of these classifiers, we found that they can be linked by simple feature data scaling. In other words, the key to these relationships lies upon how the replica of feature data are being scaled. This finding leads us directly to an exploration of novel classifiers beyond existing settings. Based on an extensive empirical evaluation, we show that the proposed formulation facilitates a tuning capability beyond existing settings for classifier generalization.     

Determination of concrete rebars characteristics by enhanced post-processing of GPR scan raw data

A method for the estimation of rebar radius by post-processing the raw data acquired by the ground penetrating radar as a B-scan radargram is proposed. Considering the hyperbola trace and the diffracted amplitudes an inverse problem consisting of two steps was stated. Using at first hyperbola fitting, the wave velocity and the coordinates of the hyperbola apex are identified. Then the rebar radius is retrieved as solution of a further optimization problem for which the cost function measures the misfit between the actual value of the maximum diffracted amplitudes and their theoretical predictions. The procedure was implemented in Matlab and tested in realistic situations. The obtained results showed improved accuracy.     

Perturbation sensitivity analysis of azimuth estimation of the acoustic source for the planar microphone array

Aiming at the requirements of quantitative analysis of the array scheme and positioning performance of a practical compact acoustic detection node system, the upper bound of the far-field azimuth estimation error for an arbitrary small-size planar array is derived and the four elements of the corresponding perturbation sensitivity factor with explicit physical significance are determined. It is proved that the perturbation sensitivity of the source direction-finding algorithm to the time difference of arrival (TDOA) estimates between microphones can be lowered obviously by means of increasing the number of distinct microphone pairs as well as employing spatially uniform distribution of array elements. Simulation results demonstrate that this study provides a quantitative approach to evaluating and comparing the azimuth estimation performance of different planar arrays.     

中国农村居民口粮消费研究

中国粮食安全尤其是口粮安全问题引起人们广泛的关注。为研究农村居民口粮消费与经济因素的关系,探讨家庭收入和价格变化对中国农村居民口粮消费的影响,以及价格变化对口粮间的消费替代研究,本文对全国2003—2009年的省级农村消费数据进行了整理,采用非参数估计方法,确定AIDS消费模型的类型,并得到了口粮消费的自价格弹性、互价格弹性、支出弹性等。结果发现,随着收入的上涨,居民倾向消费更多的是水稻;小麦和水稻之间有较强的替代性。农村居民收入的增长,将推动口粮消费的变化,尤其促进水稻消费增长,因此文章最后提出了,优先重视发展水稻生产的建议。     

Estimation of target station data using satellite data and deep learning algorithms

In this study, an innovative model has been developed for wind speed estimation through the Deep Learning method using hourly wind speed data from the measurement stations of the General Directorate of Meteorology in Van and Hakkari provinces in Turkey in conjunction with simultaneous satellite images from Eumetsat. Obtained satellite images were used during the introduction of the model, while wind speed data were used at the output stage. As a result of the findings, it was found that 85% accuracy performance could be achieved to provide sufficient insight for systems that are widely established worldwide. The model, developed as a result of the study, eliminates the need to install wind measuring stations for any region on earth within the satellite field in terms of determining wind potential. Since the field of view of the Meteosat 7 satellite covers the whole of Eastern Europe, it was determined that it could predict a high rate of up to 6 hours later by the method used in image analysis. The systems to be controlled with this method will be able to examine the weather events instantly at each point in the satellite field of view and make more accurate decisions. Also, companies will be able to perform a more detailed and rapid field scan compared to existing limited methods, and reduce initial investment costs and operating costs in terms of renewable energy resources investments.     

A novel evolutionary-negative correlated mixture of experts model in tourism demand estimation

Mixtures of experts (ME) model are widely used in many different areas as a recognized ensemble learning approach to account for nonlinearities and other complexities in the data, such as time series estimation. With the aim of developing an accurate tourism demand time series estimation model, a mixture of experts model called LSPME (Lag Space Projected ME) is presented by combining ideas from subspace projection methods and negative correlation learning (NCL). The LSPME uses a new cluster-based lag space projection (CLSP) method to automatically obtain input space to train each expert focused on the difficult instances at each step of the boosting approach. For training experts of the LSPME, a new NCL algorithm called Sequential Evolutionary NCL algorithm (SENCL) is proposed that uses a moving average for the correlation penalty term in the error function of each expert to measure the error correlation between it and its previous experts. The LSPME model was compared with other ensemble models using monthly tourist arrivals to Japan from four markets: The United States, United Kingdom, Hong Kong and Taiwan. The experimental results show that the estimation accuracy of the proposed LSPME model is significantly better than the other ensemble models and can be considered to be a promising alternative for time series estimation problems.     

Improved Kalman-based attitude estimation framework for UAVs via an antenna array

Accurate attitude estimation is crucial for Unmanned Aerial Vehicles (UAVs) in order to facilitate automated activities such as landing or trajectory tracking. Recently antenna array based communication systems have been installed in UAVs. This array structure can also be applied for attitude estimation by computing the line-of-sight (LOS) path between the base station and UAV. In this paper, we propose a complete framework for attitude estimation by exploiting 3D LOS vector obtained from the antenna array system. We present all the steps to incorporate the estimated LOS vector into the TRIaxial Attitude Determination (TRIAD), QUaternion ESTimator (QUEST) and Kalman algorithms. As an additional contribution, the error covariance matrix of the LOS vector is analytically calculated by first finding the phase shift mean squared error using the known perturbation model from Singular Value Decomposition and assuming that the antenna array measured data error can be modeled as a circularly symmetric white noise. We evaluate five array configurations via Monte Carlo simulations. We show that array configurations that provide orthogonal components of the LOS vector achieve a better performance. The usage of more than three pairs of antennas to improve the estimation of the LOS vector is also proposed for low and intermediate signal-to-noise ratio regimes.     

非对称成对载波多址信号的相位误差分析及幅度改进算法

针对由主站和小站信号同频混合而成的非对称成对载波多址（PCMA）信号解调问题，构建了一种实现此类信号解调的框架。参数估计是非对称PCMA通信系统在实现两路信号分离解调时不可或缺的环节，对于幅度参数估计精度问题，提出一种基于四次方法的搜索式幅度估计算法。首先建立非对称PCMA系统解调模型并作出基本假设，然后对不同假设下的相位误差进行对比并分析相位误差对幅度估计算法的影响，最后提出一种新的幅度估计算法。实验结果表明在相同信噪比（SNR）下，正态相位误差下的小站信号解调性能要劣于其均值条件下的解调性能。当误比特率（BER）在数量级为10^-4时，改进算法下小站信号的解调性能提升了1 dB，说明改进算法优于四次方法。     

Multilinear Weighted Regression (MWE) with Neural Networks for trend prediction

The ability to define accurate linear models to find patterns or relationships between variables is one of the most challenging fields in Computer Science. In particular, extrapolative applications are widely used to predict values in Biological, Behavioral and Social Sciences. Analysts usually focus on reducing the approximation error in order to ensure fairly reliable predictions. Using pattern detection models such as Multilayer Artificial Neural Networks can offer good results in predicting possible outcomes, however they might not be the optimal fit for predicting evolutionary trends due to their inherent overfitting characteristics. Regression models, on the other hand are more often used to predict trending behaviors and therefore they can be more helpful when studying the evolution of a given instrument, symbol or series. However, a simple regression model can be very inaccurate with an unacceptable prediction error rate. This paper shows an automatic method to find trends in known instruments, by using a massive linear regression technique combined with a conventional machine learning proposal that works on optimizing the weights of the linear regressive structures. In this paper the authors show an application of the model in finance (stocks market instruments), however the proposal is designed to work for to any discipline that studies the trends of any evolutionary object.     

Maximum likelihood estimation of noise covariance matrices for state estimation of autonomous hybrid systems

A critical aspect of developing Bayesian state estimators for hybrid systems, that involve a combination of continuous and discrete state variables, is to have a reasonably accurate characterization of the stochastic disturbances affecting their dynamics. Recently, Bavdekar et al. (2011) have proposed a maximum likelihood (ML) based framework for estimation of the noise covariance matrices from operating input–output data when an EKF is used for state estimation. In this work, the ML framework is extended to estimation of the noise covariance matrices associated with autonomous hybrid systems, and, to a wider class of recursive Bayesian filters. Under the assumption that the innovations generated by an estimator form a white noise sequence, the proposed ML framework computes the noise covariance matrices such that they maximize the log-likelihood function of the estimator innovations. The efficacy of the proposed scheme is demonstrated through the simulation and experimental studies on the benchmark three-tank system.