Fundamental factors versus herding in the 2000–2005 US stock market and prediction

We present a general methodology to incorporate fundamental economic factors to the theory of herding developed in our group to describe bubbles and antibubbles. We start from the strong form of rational expectation and derive the general method to incorporate factors in addition to the log-periodic power law (LPPL) signature of herding developed in ours and others’ works. These factors include interest rate, interest spread, historical volatility, implied volatility and exchange rates. Standard statistical AIC and Wilks tests allow us to compare the explanatory power of the different proposed factor models. We find that the historical volatility played the key role before August of 2002. Around October 2002, the interest rate dominated. In the first six months of 2003, the foreign exchange rate became the key factor. Since the end of 2003, all factors have played an increasingly large role. However, the most surprising result is that the best model is the second-order LPPL without any factor. We thus present a scenario for the future evolution of the US stock market based on the extrapolation of the fit of the second-order LPPL formula, which suggests that herding is still the dominating force and that the unraveling of the US stock market antibubble since 2000 is still qualitatively similar to (but quantitatively different from) the Japanese Nikkei case after 1990.     

The sensitivity analysis of propagator for path independent quantum finance model

Quantum finance successfully implements the imperfectly correlated fluctuation of forward interest rates at different maturities, by replacing the Wiener process with a two-dimensional quantum field. Interest rate derivatives can be priced at a more realistic value under this new framework. The quantum finance model requires three main ingredients for pricing: the initial forward interest rates, the volatility of forward interest rates, and the correlation of forward interest rates at different maturities. However, the hedging strategy only focused on fluctuation of forward interest rates. This hedging method is based on the assumption that the propagator, the covariance of forward interest rates, has an ergodic property. Since inserting the propagator is the main characteristic that distinguishes quantum finance from the Libor market model (LMM) and the Heath, Jarrow and Morton (HJM) model, understanding the impact of propagator dynamics on the price of interest rate derivatives is crucial. This research is the first step in developing a hedge strategy with respect to the evolution of the propagator. We analyze the dynamics of the propagator from Libor futures data and the integrated propagator from zero-coupon bond rate data. Then we study the sensitivity of the implied volatility of caplets and swaptions according to the three dominant dynamics of the propagator, and the change of the zero-coupon bond option price according to the two dominant dynamics of the integrated propagator.     

Extreme-value dependence: An application to exchange rate markets

Extreme value theory (EVT) focuses on modeling the tail behavior of a loss distribution using only extreme values rather than the whole data set. For a sample of 10 countries with dirty/free float regimes, we investigate whether paired currencies exhibit a pattern of asymptotic dependence. That is, whether an extremely large appreciation or depreciation in the nominal exchange rate of one country might transmit to another. In general, after controlling for volatility clustering and inertia in returns, we do not find evidence of extreme-value dependence between paired exchange rates. However, for asymptotic-independent paired returns, we find that tail dependency of exchange rates is stronger under large appreciations than under large depreciations.     

The long-range dependence behavior of the term structure of interest rates in Japan

This paper presents an empirical evidence suggesting that Japanese interest rates for different maturities possess long-range dependence in both mean and volatility. For long-term bonds, predictability in the term structure of interest rates increases with maturity, suggesting that there exists a term premium. Furthermore, the dynamics of short-term interest rates (6 months) is very different from longer term bonds, as the former are anti-persistent, which implies that the zero-interest rate policy is perceived to be temporary.     

基于小波分解和因子分析的白酒香型和年份鉴定的研究

提出了一种基于小波分解和因子模型分析白酒荧光光谱,对白酒香型进行分类和年份预测的方法。白酒的三维荧光光谱包含了其所含荧光物质信息,对其进行小波分解,其分解系数与特征峰的强度相关。选取高斯小波对三维荧光光谱进行分解,可以避免对二维荧光光谱进行分解时需要选取特定激发波长的问题。对样品的三维荧光光谱进行小波分解后,选取第4层近似系数构建正交因子模型,通过因子载荷系数对白酒进行鉴别。结果指出,贡献率较小的因子蕴含着样品的独特信息,在相似样品的比较中,不容忽视。在对10个品牌的白酒进行香型分类时,先将样品的三维荧光光谱进行高斯小波分解,使用第4层近似系数进行因子分析,得到贡献率由大到小的多个因子。根据因子的载荷系数,对样品进行聚类分析。结果表明,加入贡献率较小的因子可以将正确率提高至90%。通过对因子载荷系数与年份的相关性分析得出,贡献率排在前六位的因子和白酒年份关系较大,而排在后面的因子和白酒年份的相关性较小,因此可以选取前六位的因子建立白酒年份预测模型。通过选取不同贡献率的因子对白酒年份进行预测,其平均误差可降低至0.9年。     

Empirical investigation of a field theory formula and Black's formula for the price of an interest-rate caplet

The industry standard for pricing an interest-rate caplet is Black's formula. Another distinct price of the same caplet can be derived using a quantum field theory model of the forward interest rates. An empirical study is carried out to compare the two caplet pricing formulae. Historical volatility and correlation of forward interest rates are used to generate the field theory caplet price; another approach is to fit a parametric formula for the effective volatility using market caplet price. The study shows that the field theory model generates the price of a caplet and cap fairly accurately. Black's formula for a caplet is compared with field theory pricing formula. It is seen that the field theory formula for caplet price has many advantages over Black's formula.     

国债远期利率的量子场理论模型构建

随着我国利率市场化改革的全面推进和利率衍生品数量的增加,如何对远期利率进行精确与合理建模,就显得十分重要和紧迫.本文利用金融物理学中可有效纳入日历时间和到期时间两个维度上的国债远期利率之间不完全相关性的量子场论方法,对2011年1月4日到2016年12月30日的国债瞬时远期利率的实际市场演化进行建模,并将其结果与传统金融只能考虑日历时间方向上的相关性的主流两因子Heath-Jarrow-Morton (HJM)模型的实证结果进行比较.研究结果表明,考虑心理感知剩余时间变量后的量子场理论模型,提供了对实际的国债远期利率的92.67%的拟合优度,优于经典的最优两因子HJM模型69.02%的拟合精度.此外,分别将估计所得的最优参数代入最优量子场理论模型和两因子HJM模型下的远期利率更新方程,对2017年1月3日到2017年12月30日的100个期限的瞬时远期利率的250个瞬时远期利率的期限结构进行回测检验,从平均瞬时远期利率、均方根误差和Theil不等系数三个方面的结果均显示出量子场理论模型对国债远期利率建模的优越性.这些结果对将量子场理论引入到以国债为标的各种金融产品的定价和相关的利率风险管理、银行和金融公司的量化分析以及固定收益证券领域的实践者们均具有重要意义.     

An agent-based approach to financial stylized facts

An important challenge of the financial theory in recent years is to construct more sophisticated models which have consistencies with as many financial stylized facts that cannot be explained by traditional models. Recently, psychological studies on decision making under uncertainty which originate in Kahneman and Tversky's research attract a lot of interest as key factors which figure out the financial stylized facts. These psychological results have been applied to the theory of investor's decision making and financial equilibrium modeling. This paper, following these behavioral financial studies, would like to propose an agent-based equilibrium model with prospect theoretical features of investors. Our goal is to point out a possibility that loss-averse feature of investors explains vast number of financial stylized facts and plays a crucial role in price formations of financial markets. Price process which is endogenously generated through our model has consistencies with, not only the equity premium puzzle and the volatility puzzle, but great kurtosis, asymmetry of return distribution, auto-correlation of return volatility, cross-correlation between return volatility and trading volume. Moreover, by using agent-based simulations, the paper also provides a rigorous explanation from the viewpoint of a lack of market liquidity to the size effect, which means that small-sized stocks enjoy excess returns compared to large-sized stocks.     

Statistical regularities in the return intervals of volatility

We discuss recent results concerning statistical regularities in the return intervals of volatility in financial markets. In particular, we show how the analysis of volatility return intervals, defined as the time between two volatilities larger than a given threshold, can help to get a better understanding of the behavior of financial time series. We find scaling in the distribution of return intervals for thresholds ranging over a factor of 25, from 0.6 to 15 standard deviations, and also for various time windows from one minute up to 390 min (an entire trading day). Moreover, these results are universal for different stocks, commodities, interest rates as well as currencies. We also analyze the memory in the return intervals which relates to the memory in the volatility and find two scaling regimes, ℓ<ℓ^* with α₁=0.64±0.02 and ℓ> ℓ^* with α₂=0.92±0.04; these exponent values are similar to results of Liu et al. for the volatility. As an application, we use the scaling and memory properties of the return intervals to suggest a possibly useful method for estimating risk.     

Evaluation of generalized cross-correlation methods for direction of arrival estimation using two microphones in real environments

The localization of sound sources, and particularly speech, has a numerous number of applications to the industry. This has motivated a continuous effort in developing robust direction-of-arrival detection algorithms, in order to overcome the limitations imposed by real scenarios, such as multiple reflections and undesirable noise sources. Time difference of arrival-based methods, and particularly, generalized cross-correlation approaches have been widely investigated in acoustic signal processing, but there is considerable lack in the technical literature about their evaluation in real environments when only two microphones are used. In this work, four generalized cross-correlation methods for localization of speech sources with two microphones have been analyzed in different real scenarios with a stationary noise source. Furthermore, these scenarios have been acoustically characterized, in order to relate the behavior of these cross-correlation methods with the acoustic properties of noisy scenarios. The scope of this study is not only to assess the accuracy and reliability of a set of well-known localization algorithms, but also to determine how the different acoustic properties of the room under analysis have a determinant influence in the final results, by incorporating in the analysis additional factors to the reverberation time and signal-to-noise ratio. Results of this study have outlined the influence of the acoustic properties analysed in the performance of these methods.     

Controlling the false positive rate in fuzzy clustering using randomization: application to fMRI activation detection

Despite its potential advantages for fMRI analysis, fuzzy C-means (FCM) clustering suffers from limitations such as the need for a priori knowledge of the number of clusters, and unknown statistical significance and instability of the results. We propose a randomization-based method to control the false-positive rate and estimate statistical significance of the FCM results. Using this novel approach, we develop an fMRI activation detection method. The ability of the method in controlling the false-positive rate is shown by analysis of false positives in activation maps of resting-state fMRI data. Controlling the false-positive rate in FCM allows comparison of different fuzzy clustering methods, using different feature spaces, to other fMRI detection methods. In this article, using simulation and real fMRI data, we compare a novel feature space that takes the variability of the hemodynamic response function into account (HRF-based feature space) to the conventional cross-correlation analysis and FCM using the cross-correlation feature space. In both cases, the HRF-based feature space provides a greater sensitivity compared to the cross-correlation feature space and conventional cross-correlation analysis. Application of the proposed method to finger-tapping fMRI data, using HRF-based feature space, detected activation in sub-cortical regions, whereas both of the FCM with cross-correlation feature space and the conventional cross-correlation method failed to detect them.     

Investigation of interface roughness cross-correlation properties of optical thin films from total scattering losses

The interface roughness and interface roughness cross-correlation properties affect the scattering losses of high-quality optical thin films. In this paper, the theoretical models of light scattering induced by surface and interface roughness of optical thin films are concisely presented. Furthermore, influence of interface roughness cross-correlation properties to light scattering is analyzed by total scattering losses. Moreover, single-layer TiO₂ thin film thickness, substrate roughness of K9 glass and ion beam assisted deposition (IBAD) technique effect on interface roughness cross-correlation properties are studied by experiments, respectively. A 17-layer dielectric quarter-wave high reflection multilayer is analyzed by total scattering losses. The results show that the interface roughness cross-correlation properties depend on TiO₂ thin film thickness, substrate roughness and deposition technique. The interface roughness cross-correlation properties decrease with the increase of film thickness or the decrease of substrates roughness. Furthermore, ion beam assisted deposition technique can increase the interface roughness cross-correlation properties of optical thin films. The measured total scattering losses of 17-layer dielectric quarter-wave high reflection multilayer deposited with IBAD indicate that completely correlated interface model can be observed, when substrate roughness is about 2.84 nm.     

Random matrix theory analysis of cross-correlations in the US stock market: Evidence from Pearson’s correlation coefficient and detrended cross-correlation coefficient

In this study, we first build two empirical cross-correlation matrices in the US stock market by two different methods, namely the Pearson’s correlation coefficient and the detrended cross-correlation coefficient (DCCA coefficient). Then, combining the two matrices with the method of random matrix theory (RMT), we mainly investigate the statistical properties of cross-correlations in the US stock market. We choose the daily closing prices of 462 constituent stocks of S&P 500 index as the research objects and select the sample data from January 3, 2005 to August 31, 2012. In the empirical analysis, we examine the statistical properties of cross-correlation coefficients, the distribution of eigenvalues, the distribution of eigenvector components, and the inverse participation ratio. From the two methods, we find some new results of the cross-correlations in the US stock market in our study, which are different from the conclusions reached by previous studies. The empirical cross-correlation matrices constructed by the DCCA coefficient show several interesting properties at different time scales in the US stock market, which are useful to the risk management and optimal portfolio selection, especially to the diversity of the asset portfolio. It will be an interesting and meaningful work to find the theoretical eigenvalue distribution of a completely random matrix R for the DCCA coefficient because it does not obey the Mar?enko–Pastur distribution.     

非线性光电延时反馈混沌同步复用通信系统研究

提出了一种基于电光调制器的非线性光电延时反馈超混沌复用通信系统.与传统混沌通信系统不同,其混沌波形不是由激光器产生,而是由电光调制器产生,该系统具有非线性维数高、易于再生和精确控制的优势.介绍了信号调制反馈延时的编码方法和相关检测解码方法,数值仿真了三条支路的高速复用与解复用,分析了误码产生的主要原因,并进一步研究了光...     

模型与特征混合补偿法及其在耳语说话人识别中的应用

为了提高信道差异下短时耳语说话人的识别率,提出了一种在模型域和特征域进行混合补偿的方法。该方法首先在模型训练阶段以联合因子分析法为基础,通过估计训练语音的说话人空间和信道空间,提取出说话人因子,消除信道因子,其次在测试阶段,将测试语音的信道因子映射到特征空间,实施特征补偿,从而在模型和特征两方面去除信道信息,提高识别率。实验结果显示,在三种不同的信道训练环境下,混合补偿法都取得了相似的识别率,且新方法对短时耳语音的测试效果要优于联合因子分析法。      

Whispered speaker identification based on feature and model hybrid compensation

In order to increase short time whispered speaker recognition rate in variable channel conditions,the hybrid compensation in model and feature domains was proposed.This method is based on joint factor analysis in training model stage.It extracts speaker factor and eliminates channel factor by estimating training speech speaker and channel spaces.Then in the test stage,the test speech channel factor is projected into feature space to engage in feature compensation,so it can remove channel information both in model and feature domains in order to improve recognition rate.The experiment result shows that the hybrid compensation can obtain the similar recognition rate in the three different training channel conditions and this method is more effective than joint factor analysis in the test of short whispered speech.     

A multivariate long memory stochastic volatility model

This paper develops a multivariate long-memory stochastic volatility model which allows the multi-asset long-range dependence in the volatility process. The motivation is from the fact that both autocorrelations and cross-correlations of some proxies of exchange rate volatility exhibit strong evidence of long-memory behavior. The statistical properties of the new stochastic volatility model provide theoretical explanation to the common findings that long memory volatility properties are more apparent if we use absolute return as a volatility proxy than squared return. Results of the real data application show that our model outperforms an existing multivariate stochastic volatility model.     

The expectation hypothesis of interest rates and network theory: The case of Brazil

This paper investigates the topological properties of the Brazilian term structure of interest rates network. We build the minimum spanning tree (MST), which is based on the concept of ultrametricity, using the correlation matrix for interest rates of different maturities. We show that the short-term interest rate is the most important within the interest rates network, which is in line with the Expectation Hypothesis of interest rates. Furthermore, we find that the Brazilian interest rates network forms clusters by maturity.     

利用自相关函数和双耳自相关函数分析对交通噪声特征参量的提取~

对实地双通道测量获得的道路交通噪声和铁路噪声信号样本进行了自相关函数和双耳自相关函数(Interaural CrossCorrelation Function)的分析。进而通过对噪声样本时间因子和空间因子的相关性分析、主成分分析和主观评价实验,得到了3个铁路噪声源特征参量物理因子和4个道路交通噪声源特征参量物理因子。发现与传统的声压级测量相比,表征声音信号时间特性和空间特性的这7个物理量可以更全面、准确地表征交通噪声的特性。在对道路噪声进行测量或分析时,掌握与声源视觉宽度和音调感相对应的物理因子以及双耳时延和初始能量,就可获悉与人的主观评价相一致的道路交通噪声特征信息;对铁路噪声而言,掌握与声源视觉宽度相对应的物理因子以及双耳时延和声音的重复性特征,就可以得到与入主观评价相一致的铁路噪声特征信息。综合道路噪声特征参量和铁路噪声特征参量可以发现,双耳时延和与声源视觉宽度相对应的物理因子是与人的主观反应最为一致的主成分指标,说明噪声中决定人的评价的最主要的因素是代表空间特征的信号因子。