In this paper a Bayesian regularized artificial neural network is proposed as a novel method to forecast financial market behavior Daily market prices and financial technical indicators are utilized as inputs to predict the one day future closing price of individual stocks The prediction of stock price movement is generally considered to be a challenging and important task for financial time series analysis The accurate prediction of stock price movements could play an important role in helping investors improve stock returns The complexity in predicting these trends lies in the inherent noise and volatility in daily stock price movement The Bayesian regularized network assigns a probabilistic nature to the network weights, allowing the network to automatically and optimally penalize excessively complex models The proposed technique reduces the potential for overfitting and overtraining, improving the prediction quality and generalization of the network Experiments were performed with Microsoft Corp and Goldman Sachs Group Inc stock to determine the effectiveness of the model The results indicate that the proposed model performs as well as the more advanced models without the need for preprocessing of data, seasonality testing, or cycle analysis

A Bayesian regularized artificial neural network for stock market forecasting

The paper proposed an improved bacterial chemotaxis optimization (IBCO), which is then integrated into the back propagation (BP) artificial neural network to develop an efficient forecasting model for prediction of various stock indices. Experiments show its better performance than other methods in learning ability and generalization.

Stock market prediction of S&P 500 via combination of improved BCO approach and BP neural network

Extensive calculations based on density functional theory have been carried out to understand the origin of magnetism in undoped ZnO thin films as found in recent experiments. The observed magnetism is confirmed to be due to Zn, instead of O, vacancy. The main source of the magnetic moment, however, arises from the unpaired 2p electrons at O sites surrounding the Zn vacancy with each nearest-neighbor O atom carrying a magnetic moment ranging from 0.490 to 0.740 B. Moreover, the study of vacancy-vacancy interactions indicates that in the ground state, the magnetic moments induced by Zn vacancies prefer to ferromagnetically couple with the antiferromagnetic state lying 44 meV higher in energy. Since this is larger than the thermal energy at room temperature, the ferromagnetic state can be stable against thermal fluctuations. Calculations and discussions are also extended to ZnO nanowires that have larger surface to volume ratio. Here, the Zn vacancies are found to lead to the ferromagnetic state too. The present theoretical study not only demonstrates that ZnO samples can be magnetic even without transition-metal doping but also suggests that introducing Zn vacancy is a natural and an effective way to fabricate magnetic ZnO nanostructures. In addition, vacancy mediated magnetic ZnO nanostructures may have certain advantages over transition-metal doped systems in biomedical applications.

https://absimage.aps.org/image/MAR08/MWS_MAR08-2007-004197.pdf

Zinc Vacancy induced magnetism in ZnO thin films and nanowires

The problem of acoustic noise is becoming increasingly serious with the growing use of industrial and medical equipment, appliances, and consumer electronics. Active noise control (ANC), based on the principle of superposition, was developed in the early 20th century to help reduce noise. However, ANC is still not widely used owing to the effectiveness of control algorithms, and to the physical and economical constraints of practical applications. In this paper, we briefly introduce some fundamental ANC algorithms and theoretical analyses, and focus on recent advances on signal processing algorithms, implementation techniques, challenges for innovative applications, and open issues for further research and development of ANC systems.

/pdf/recent-advances-on-active-noise-control-open-issues-and-1wg2ydx57i.pdf

Recent Advances on Active Noise Control: Open Issues and Innovative Applications

Vehicle system dynamics integration encompasses interdisciplinary challenges innovations in various aspects related to vehicle system/subsystems/components dynamic characteristics, modeling and validation, vehicle dynamics state measurement and estimation, vehicle/chassis control systems, coordination of power management and dynamics/stability control, etc [1-4]. These are becoming more critical due to the increasing concern and rapid development in electric and hybrid vehicles, which tend to exhibit different vehicle dynamics/stability and control characteristics when compared to conventional vehicles.

Mechanical systems and signal processing

Adaptive nonlinear active noise control algorithm for active headrest with moving error microphones

This paper proposes a simple and novel method of designing and developing LVDT based sensing system. Conventionally, precise adjustment of windings is made to enhance the linearity range of LVDT. The tedious job of pitch adjustment of windings of LVDT can be overcome by use of the proposed method. Functional link artificial neural network has been successfully used in the paper for nonlinear compensation of LVDT. The effectiveness of the proposed method is demonstrated through computer simulation with the experimental data of a simple LVDT. The complete algorithm with practical set-up for development of LVDT is presented in the paper.

A novel method of designing LVDT using artificial neural network

Ball mills are extensively used in the size reduction process of different ores and minerals. The fill level inside a ball mill is a crucial parameter which needs to be monitored regularly for optimal operation of the ball mill. In this paper, a vibration monitoring-based method is proposed and tested for estimating the fill level inside a laboratory-scale ball mill. A vibration signal is captured from the base of a laboratory-scale ball mill by using a ± 5 g accelerometer. Features are extracted from the vibration signal by using different transforms such as fast Fourier transform, discrete wavelet transform, wavelet packet decomposition, and empirical mode decomposition. These features are given as input to an artificial neural network which is used to predict the percentage fill level inside the ball mill. In this paper, the predicted fill level obtained by using different features are compared. It is found that the predicted fill level due to features obtained after fast Fourier transform outperforms other transforms.

Monitoring the fill level of a ball mill using vibration sensing and artificial neural network

This paper describes real-time implementation of a Lyapunov energy function based fuzzy adaptive step size filtered-x-LMS (FxLMS) algorithm for active control of noise in a single channel duct. The proposed method uses a Lyapunov energy function of the error for step size adjustment in the filtered-x-LMS (FxLMS) algorithm. The energy function and its rate of change are taken as the principal signals to generate the step size through a Fuzzy Inference System (FIS). Experimental results of the algorithm are compared with a recently proposed variable step size algorithm as well as the fixed step size FxLMS algorithm. The paper also reports a new method of implementation of the FIS by generating a nonlinear mapping through a feed-forward neural network to reduce the computational complexity. The CPU time of this method has been compared with the other existing methods to establish the feasibility of implementation with a 10 kHz sampling.

An Energy Function Based Fuzzy Variable Step Size FxLMS Algorithm for Active Noise Control

Atmospheric Ar:H2 plasma is an eco-friendly option for the reduction of metal oxides. For better reduction performance and safety concern, the hydrogen gas injected into the reactor should be monitored. A hydrogen flow rate estimation system is presented in this paper by using an artificial neural network (ANN) model fed with features of optical emission spectra of the plasma. ANN models are studied with two different sets of input, i.e. for the first case the inputs to the model are the three features of Hα line such as the peak intensity count, full-width half maximum and area under Hα line, while for the second case, the peak intensity count of a group of emission lines like Hα, Ar I, O I, K I, Na D lines are considered as the inputs. ANN model is developed for estimating four different sets of hydrogen flow rates 5, 8, 10 and 12 litres per minute (lpm) when the argon flow rate is constant at 10 lpm. For both the input features, the model performances are compared, and it is shown that improved estimation accuracy is observed from the second case, i.e. from peak intensity count of a group of emission lines instead of only hydrogen emission line.

Debi Prasad Das

Papers

Adaptive nonlinear active noise control algorithm for active headrest with moving error microphones

A novel method of designing LVDT using artificial neural network

Monitoring the fill level of a ball mill using vibration sensing and artificial neural network

An Energy Function Based Fuzzy Variable Step Size FxLMS Algorithm for Active Noise Control

Estimation of hydrogen flow rate in atmospheric Ar:H 2 plasma by using artificial neural network