: This thesis applies neural network feature selection techniques to multivariate time series data to improve prediction of a target time series. Two approaches to feature selection are used. First, a subset enumeration method is used to determine which financial indicators are most useful for aiding in prediction of the S&P 500 futures daily price. The candidate indicators evaluated include RSI, Stochastics and several moving averages. Results indicate that the Stochastics and RSI indicators result in better prediction results than the moving averages. The second approach to feature selection is calculation of individual saliency metrics. A new decision boundary-based individual saliency metric, and a classifier independent saliency metric are developed and tested. Ruck's saliency metric, the decision boundary based saliency metric, and the classifier independent saliency metric are compared for a data set consisting of the RSI and Stochastics indicators as well as delayed closing price values. The decision based metric and the Ruck metric results are similar, but the classifier independent metric agrees with neither of the other metrics. The nine most salient features, determined by the decision boundary based metric, are used to train a neural network and the results are presented and compared to other published results. (AN)

/pdf/nonlinear-time-series-analysis-1j3a35n4y3.pdf

Nonlinear Time Series Analysis.

(1990). ENERGY FUNCTION ANALYSIS FOR POWER SYSTEM STABILITY. Electric Machines & Power Systems: Vol. 18, No. 2, pp. 209-210.

Energy function analysis for power system stability

ライブラリー Annual Energy Outlook 2000

Elementary Fixed Point Theorems * Theorem of Borsuk and Topological Transversality * Homology and Fixed Points * Leray-Schauder Degree and Fixed Point Index * The Lefschetz-Hopf Theory * Selected Topics * Index

Fixed Point Theory

Thank you for reading stochastic processes and filtering theory. Maybe you have knowledge that, people have look numerous times for their favorite novels like this stochastic processes and filtering theory, but end up in harmful downloads. Rather than reading a good book with a cup of tea in the afternoon, instead they are facing with some infectious bugs inside their computer. stochastic processes and filtering theory is available in our digital library an online access to it is set as public so you can download it instantly. Our digital library saves in multiple locations, allowing you to get the most less latency time to download any of our books like this one. Merely said, the stochastic processes and filtering theory is universally compatible with any devices to read.

Stochastic Processes And Filtering Theory

Growing model complexities in load modeling have created high dimensionality in parameter estimations, and thereby substantially increasing associated computational costs. In this letter, a tensor-based method is proposed for identifying composite load modeling (CLM) parameters and for conducting a global sensitivity analysis. Tensor format and Fokker-Planck equations are used to estimate the power output response of CLM in the context of simultaneously varying parameters under their full parameter distribution ranges. The proposed tensor structure is shown as effective for tackling high-dimensional parameter estimation and for improving computational performances in load modeling through global sensitivity analysis.

Global Sensitivity Analysis in Load Modeling via Low-Rank Tensor

This letter proposes a novel and highly efficient distribution system state estimation (DSSE) method with nonlinear measurements from supervisory control and data acquisition (SCADA) systems. Conventional DSSE based on the weighted least square (WLS) criterion requires multiple Monte Carlo simulations for overall accuracy evaluation and high calculation cost due to a nonlinear iterative process. The proposed method uses the Taylor series of voltages for constructing a linear DSSE model in the interval form and then solves this model by interval arithmetic. This method obtains accurate and robust estimates via a single random sampling of measurements and is computationally efficient. The comparative analysis in the IEEE 34-bus distribution system points to improved estimation results relative to the nonlinear WLS-based method.

Towards Highly Efficient State Estimation With Nonlinear Measurements in Distribution Systems

This paper presents a new solution for reconstructing missing data in power system measurements. An Enhanced Denoising Autoencoder (EDAE) is proposed to reconstruct the missing data through the input vector space reconstruction based on the neighbor values correlation and Long Short-Term Memory (LSTM) networks. The proposed LSTM-EDAE is able to remove the noise, extract principle features of the dataset, and reconstruct the missing information for new inputs. The paper shows that the utilization of neighbor correlation can perform better in missing data reconstruction. Trained with LSTM networks, the EDAE is more effective in coping with big data in power systems and obtains a better performance than the neural network in conventional Denoising Autoencoder. A random data sequence and the simulated Phasor Measurement Unit (PMU) data of power system are utilized to verify the effectiveness of the proposed LSTM-EDAE.

Reconstruction of Power System Measurements Based on Enhanced Denoising Autoencoder

The risk of cascading blackouts greatly relies on failure probabilities of individual components in power grids. To quantify how component failure probability (CFP) influences blackout risk (BR), this paper proposes a sample-induced semianalytic approach to characterize the relationship between CFP and BR. To this end, we first give a generic component failure probability function (CoFPF) to describe CFP with varying parameters or forms. Then, the exact relationship between BR and CoFPFs is built on the abstract Markov-sequence model of cascading outages. Leveraging a set of samples generated by blackout simulations, we further establish a sample-induced semianalytic mapping between the unbiased estimation of BR and CoFPFs. Finally, we derive an efficient algorithm that can directly calculate the unbiased estimation of BR when the CoFPFs change. Since no additional simulations are required, the algorithm is computationally scalable and efficient. Numerical experiments well confirm the theory and the algorithm.

Quantifying the Influence of Component Failure Probability on Cascading Blackout Risk

A solution to the power flow problem is imperative for many power system applications and several iterative approaches are employed to achieve this objective. However, the chance of finding a solution is dependent on the choice of the initial point because of the non-convex feasibility region of this problem. In this paper, a non-iterative approach that leverages a convexified relaxed power flow problem is employed to verify the existence of a feasible solution. To ensure the scalability of the proposed convex relaxation, the problem is formulated as a sparse semi-definite programming problem. The variables associated with each maximal clique within the network form several positive semidefinite matrices. Perturbation and network reconfiguration schemes are employed to improve the tightness of the proposed convex relaxation in order to validate the existence of a feasible solution for the original non-convex problem. Multiple case studies including an ill-conditioned power flow problem are examined to show the effectiveness of the proposed approach to find a feasible solution.

/pdf/a-convex-relaxation-approach-for-power-flow-problem-41qhmf8p6b.pdf

Jianhui Wang

Papers

Global Sensitivity Analysis in Load Modeling via Low-Rank Tensor

Towards Highly Efficient State Estimation With Nonlinear Measurements in Distribution Systems

Reconstruction of Power System Measurements Based on Enhanced Denoising Autoencoder

Quantifying the Influence of Component Failure Probability on Cascading Blackout Risk

A convex relaxation approach for power flow problem