Aijing Lin

APPLICATION OF EMPIRICAL MODE DECOMPOSITION COMBINED WITH k-NEAREST NEIGHBORS APPROACH IN FINANCIAL TIME SERIES FORECASTING

The purpose of this paper is to forecast the daily closing prices of stock markets based on the past sequences. In this paper, keeping in mind the recent trends and the limitations of previous researches, we proposed a new technique, called empirical mode decomposition combined with k-nearest neighbors (EMD–KNN) method, in forecasting the stock index. EMD–KNN takes the advantages of the KNN and EMD. To demonstrate that our EMD–KNN method is robust, we used the new technique to forecast four stock index time series at a specific time. Detailed experiments are implemented for both of the proposed forecasting models, in which EMD–KNN, KNN method and ARIMA are compared. The results demonstrate that the proposed EMD–KNN model is more successful than KNN method and ARIMA in predicting the stock closing prices.

Distribution of eigenvalues of detrended cross-correlation matrix

This letter is devoted to the cross-correlation analysis of non-stationary multivariate data, in which the detrended cross-correlation matrix based on the detrended cross-correlation coefficient is studied. The relationship between Pearsons cross-correlation coefficient and the detrended cross-correlation coefficient is analyzed. As a special case of random matrix theory, the distribution of the eigenvalues of the detrended cross-correlation matrix for purely random variables is derived.

MINIMIZING PERIODIC TRENDS BY APPLYING LAPLACE TRANSFORM

Rescaled range analysis (R/S analysis), detrended fluctuation analysis (DFA) and detrended moving average (DMA) are widely-used methods for detection of long-range correlations in time series. Detrended cross-correlation analysis (DCCA) is a recently developed method to quantify the cross-correlations of two non-stationary time series. Another method for studying auto-correlations and cross-correlations was presented by Sergio Arianos and Anna Carbone in 2009. Recent studies have reported the susceptibility of this methods to periodic trends, which can result in spurious crossovers. In this paper, we propose the modified methods base on Laplace transform to minimizing the effect of periodic trends. The effectiveness of our techniques are demonstrated on stock data corrupted with periodic trends.

THE ORTHOGONAL V-SYSTEM DETRENDED FLUCTUATION ANALYSIS

The Detrended Fluctuation Analysis (DFA) and its extensions (MF-DFA) have been proposed as robust techniques to determine possible long-range correlations in self-affine signals. However, many studies have reported the susceptibility of DFA to trends which give rise to spurious crossovers and prevent reliable estimations of the scaling exponents. Lately, several modifications of the DFA method have been reported with many different techniques for eliminating the monotonous and periodic trends. In this study, a smoothing algorithm based on the Orthogonal V-system (OVS) is proposed to minimize the effect of power-law trends, periodic trends, assembled trends and piecewise function trends. The effectiveness of the new method is demonstrated on monofractal data and multifractal data corrupted with different trends.

Effects of Exponential Trends on Correlations of Stock Markets

Detrended fluctuation analysis (DFA) is a scaling analysis method used to estimate long-range power-law correlation exponents in time series. In this paper, DFA is employed to discuss the long-range correlations of stock market. The effects of exponential trends on correlations of Hang Seng Index (HSI) are investigated with emphasis. We find that the long-range correlations and the positions of the crossovers of lower order DFA appear to have no immunity to the additive exponential trends. Further, our analysis suggests that an increase in the DFA order increases the efficiency of eliminating on exponential trends. In addition, the empirical study shows that the correlations and crossovers are associated with DFA order and magnitude of exponential trends.

Multiscale Symbolic Phase Transfer Entropy in Financial Time Series Classification

We address the challenge of classifying financial time series via a newly proposed multiscale symbolic phase transfer entropy (MSPTE). Using MSPTE method, we succeed to quantify the strength and direction of information flow between financial systems and classify financial time series, which are the stock indices from Europe, America and China during the period from 2006 to 2016 and the stocks of banking, aviation industry and pharmacy during the period from 2007 to 2016, simultaneously. The MSPTE analysis shows that the value of symbolic phase transfer entropy (SPTE) among stocks decreases with the increasing scale factor. It is demonstrated that MSPTE method can well divide stocks into groups by areas and industries. In addition, it can be concluded that the MSPTE analysis quantify the similarity among the stock markets. The symbolic phase transfer entropy (SPTE) between the two stocks from the same area is far less than the SPTE between stocks from different areas. The results also indicate that four stocks from America and Europe have relatively high degree of similarity and the stocks of banking and pharmaceutical industry have higher similarity for CA. It is worth mentioning that the pharmaceutical industry has weaker particular market mechanism than banking and aviation industry.