Search

Developing scientific and effective air quality early warning system has important scientific and practical significance for protecting people's physical health and promoting social harmony and stability. Therefore, the isolation forest (iForest) algorithm is employed to analyze the outlier data in original air quality index (AQI) data, and then an effective air quality early warning system is developed in this paper. The system is consists of four modules:data preprocessing module, optimization module, forecasting module and correction module, which combines time varying filtering based empirical mode decomposition (TVF-EMD), modified butterfly optimization algorithm (MBOA), outlier robust extreme learning machine (ORELM), and nonlinear correction strategy, and successfully achieves effective early warning for air quality. In order to verify the effectiveness of the developed air quality early warning system, five cities with different air pollution level in China are employed for empirical research. The results show that:1) Compared with empirical mode decomposition (EMD), TVF-EMD can be more effective to reduce the nonlinear and non-stationary features of the original data; 2) The developed MBOA-based error nonlinear correction strategy is superior to other error correction strategies, which can significantly improve the performance of the early warning system; 3) The performance of the developed system is superior to other compared methods, which can provide effective early warning for cities with different air pollution level. [ABSTRACT FROM AUTHOR]

Accurate prediction of social logistics demands is essential to government5s policy formulation for the logistics industry as well as to enterprise's logistics activity planning. In this paper, a logistics demand prediction model construction method based on the fuzzy cognitive map (FCM) is proposed. This method comprehensively considers the mutual influence between five economic elements (GDP, total import and export volume, etc.) and logistics demands, and acquires the mutual influence weight through machine learning of historical data. Finally, a logistics demand prediction model is built, which can realize accurate prediction of the future logistics demands. The experimental results provide solid evidence for high precision and favorable performance of the model in predicting logistics demands. [ABSTRACT FROM AUTHOR]

This paper aims to solve the problem of high throughput automated data analysis of brain microstructure reconstruction by using the solution of systems engineering that consists of hardware, software and algorithm. To solve the key issues of limit analysis efficiency such as registration of large-volume distorted image, dense neural structure recognition, ambiguous neural connections and multi-source data fusion etc, classification method in accordance with the biological characteristics of neural structure is proposed by the realization of the samples preparation, automatic cutting, microscopic imaging, 3D reconstruction and software platform building etc. Furthermore, it provides a theoretical basis and solution for building engineering platform for large volume neural structure reconstruction and meets the brain science research of high-throughput neural circuit network reconstruction demand. [ABSTRACT FROM AUTHOR]

Quantitative investment is an important application of system engineering in the financial investment area. Quantitative investment tries to automatically invest on securities using computational algorithms, and to obtain excess return. This paper proposes a novel quantitative trading algorithm based on machine learning and technical analysis, named "ML-TEA" (machine learning and technical analysis). ML-TEA predicts the stock's movements using the technical indicators calculated by prices and volumes. The empirical results show that firstly, three strategies can obtain an annual return of 25%, which outperforms the index's 10.60% and buy and hold's 3%, and the state of the art algorithms. The three algorithms also significantly outperform the benchmarks and the state of the art in terms of risk adjusted return, i.e., Sharpe ratio, Treynor ratio, and Jensen's alpha. Secondly, Ada-TEA and SVM-TEA can resist reasonable transaction costs that are much higher than the actual transaction costs. [ABSTRACT FROM AUTHOR]

Semi-supervised multi-class classification problem opens research focuses in machine learning and pattern recognition, currently it is decomposed into a set of binary classification problems. Two kinds of class label presentation methods that one was class label presentation method of unit disc and the other was that of binary string were proposed for fear that multiple binary classification problems were solved. Besides, local learning has the good feature in semi-supervised binary classification problem. On the basis of it, local learning semi-supervised multi-class classifier was presented in this paper. The effectiveness of the algorithms was confirmed with experiments on benchmark datasets compared to other related algorithms. [ABSTRACT FROM AUTHOR]

A credit scoring method for a large problem not only achieves a certain accuracy but also its speed, interpretability, simplicity and other performance are also very important. In this paper, a novel method called an improved SVM learning algorithm based on three-variable working set (ISVMTV) is presented. This algorithm is derived by solving a series of the QP problems with only three points and the corresponding relaxation subproblems are solved analytically so that the proposed method approaches to the optimal solution more quickly. The proposed method is introduced to credit scoring and three datasets from UCI machine learning datasets are selected to demonstrate the method's competitive performance. Moreover, ISVM-TV shows a superior performance in saving the computational cost and improving classification accuracy. [ABSTRACT FROM AUTHOR]