基于机器学习LSTM&US模型的消费者信心指数预测研究

消费者信心指数等宏观经济指标具有时间上的滞后效应和动态变化的多维性,不易精确预测。本文基于机器学习长短时间记忆(Long Short-Term Memory,LSTM)神经网络模型,结合大数据技术挖掘消费者信心指数相关网络搜索数据(User Search,US),进而构建一种LSTM&US预测模型,并将其应用于对我国消费者信心指数的长期、中期与短期的预测研究,同时引入多个基准预测模型进行了对比分析。结果发现:引入网络搜索数据能够提高LSTM神经网络模型的预测性能与预测精度;LSTM&US预测模型具有较好的泛化能力,对不同期限的预测效果均较稳定,其预测性能与预测精度均优于其他六种基准预测模型(LSTM、SVR&US、RFR&US、BP&US、XGB&US和LGB&US);预测结果显示本文提出的LSTM&US预测模型具有一定的实用价值,该预测方法为消费者信心指数的预测与预判提供了一种新的研究思路,丰富了机器学习方法在宏观经济指标预测领域中的理论研究。     

Estimating time-varying directed neural networks

Reconstructing the functional network of a neuron cluster is a fundamental step to reveal the complex interactions among neural systems of the brain. Current approaches to reconstruct a network of neurons or neural systems focus on establishing a static network by assuming the neural network structure does not change over time. To the best of our knowledge, this is the first attempt to build a time-varying directed network of neurons by using an ordinary differential equation model, which allows us to describe the underlying dynamical mechanism of network connections. The proposed method is demonstrated by estimating a network of wide dynamic range neurons located in the dorsal horn of the rats’ spinal cord in response to pain stimuli applied to the Zusanli acupoint on the right leg. The finite sample performance of the proposed method is also investigated with a simulation study.

     

Markov switching component GARCH model: Stability and forecasting

ABSTRACT

This paper introduces an extension of the Markov switching GARCH model where the volatility in each state is a convex combination of two different GARCH components with time varying weights. This model has the dynamic behavior to capture the variants of shocks. The asymptotic behavior of the second moment is investigated and an appropriate upper bound for it is evaluated. Using the Bayesian method via Gibbs sampling algorithm, a dynamic method for the estimation of the parameters is proposed. Finally, we illustrate the efficiency of the model by simulation and also by considering two different set of empirical financial data. We show that this model provides much better forecasts of the volatility than the Markov switching GARCH model.     

基于首末位质量因子的BP神经网络财务风险预警模型

上市公司往往存在粉饰财务数据来美化企业经营状况的动机,这会降低财务风险预警模型预测的准确性。文章利用Benford律和Myer指数两种数据质量评估方法,构建Benford和Myer质量因子,引入BP神经网络模型,构造BM-BP神经网络财务风险预警模型;并进一步利用2000—2019年中国A股上市公司数据,评价数据质量因子对财务风险预警模型预测准确性的影响,分析新模型预测准确性的稳定性。实证分析结果显示:Benford和Myer质量因子提高了BP神经网络财务风险预警模型预测的准确性;在不同质量因子的比较结果中,包含评选指标Benford和Myer质量因子的BP神经网络财务风险预警模型具有较高的预测准确率和较低的二类误判率,稳定性良好;利用决策树算法筛选指标有效提高了新模型的预测准确性。     

Sensitivity analysis of reliability functions of the exponential power series lifetime distribution

ABSTRACT

Hazard rate functions are often used in modeling of lifetime data. The Exponential Power Series (EPS) family has a monotone hazard rate function. In this article, the influence of input factors such as time and parameters on the variability of hazard rate function is assessed by local and global sensitivity analysis. Two different indices based on local and global sensitivity indices are presented. The simulation results for two datasets show that the hazard rate functions of the EPS family are sensitive to input parameters. The results also show that the hazard rate function of the EPS family is more sensitive to the exponential distribution than power series distributions.     

互联网企业广告收入预测研究——基于低频数据的神经网络和时间序列组合模型

神经网络模型对大样本时间序列的拟合效果优于传统时间序列模型,但对于年度、月度、日度等低频时间序列的预测则难以发挥其优势。鉴于此,本文应用传统时间序列模型和神经网络模型,建立Holtwinters-BP组合模型,利用Holtwinters模型分别拟合各解释变量序列,利用BP模型拟合解释变量和自变量的非线性关系,基于某社交新闻类APP的日广告收入数据进行互联网企业广告收入预测研究。通过与循环神经网络(RNN)模型、长短期记忆神经网络(LSTM)模型等预测结果的对比发现:Holtwinters-BP组合模型的预测精度和稳定性更高;证明多维变量对于广告收入的显著影响,多变量模型的预测准确性高于单变量模型;构建的Holtwinters-BP组合模型对于低频数据预测有较好的有效性和适用性。     

基于深度学习LSTM神经网络的全球股票指数预测研究

作为深度学习技术的经典模型之一，长短期记忆（LSTM）神经网络在挖掘序列数据长期依赖关系中极具优势。基于深度神经网络优化技术，本文构造了一个深层LSTM神经网络并将其应用于全球30个股票指数三种不同期限的预测研究，结果发现：①LSTM神经网络具有很强的泛化能力，对全部指数不同期限的预测效果均很稳定；②LSTM神经网络具有优秀的预测精度，相比三种对照模型（SVR，MLP和ARIMA），其对全部指数的平均预测精度在不同期限上均有提升；③LSTM神经网络能够有效控制误差波动，其对全部指数的平均预测稳定度相比三种对照模型在不同期限上亦均有提高。鉴于LSTM神经网络在预测精度和稳定度两方面的优势，其未来在金融预测中将有广阔的应用前景。     

递归预测器神经网络在中国金融市场的实证研究

金融时间序列预测是金融理论领域的研究热点之一。以金融市场中普遍存在的弱混沌为基础,对递归预测器神经网络在中国金融市场的预测应用进行实证研究。在网络训练上,提出用遗传算法优化网络的阈值、权值以及激发函数的幅值和斜率。对国内股票、期货和黄金市场中几个有代表性的品种进行实证检验,计算了预测均方根误差（RMSE）和预测精度（PA）,并和两种典型的神经网络预测模型——BP神经网络、径向基函数神经网络做了比较,结果表明该模型有较好的预测效果。     

Standardization of Process Norms in Baker's Yeast Fermentation through Statistical Models in Comparison with Neural Networks

Achieving consistency of growth pattern for commercial yeast fermentation over batches through addition of water, molasses and other chemicals is often very complex in nature due to its bio-chemical reactions in operation. Regression models in statistical methods play a very important role in modeling the underlying mechanism, provided it is known. On the contrary, artificial neural networks provide a wide class of general-purpose, flexible non-linear architectures to explain any complex industrial processes. In this paper, an attempt has been made to find a robust control system for a time varying yeast fermentation process through statistical means, and in comparison to non-parametric neural network techniques. The data used in this context are obtained from an industry producing baker's yeast through a fed-batch fermentation process. The model accuracy for predicting the growth pattern of commercial yeast, when compared among the various techniques used, reveals the best performance capability with the backpropagation neural network. The statistical model used through projection pursuit regression also shows higher prediction accuracy. The models, thus developed, would also help to find an optimum combination of parameters for minimizing the variability of yeast production.     

A continuous-time Markov chain approach with the analytic likelihood in studies of behavioral changes

Abstract

We develop an analytic likelihood approach for a four-state CTMC by solving the backwards Kolmogorov differential equations, reducing this bias in transition rate estimates. A simulation study is performed to assess the performance of this new method and confirms that it achieves good coverage probabilities with low bias and standard errors. Finally, we analyzed data from Project SUCCESS to estimate the study each participant’s transitions among behavioral stage, consisting of risky drinking and possible ineffective using of contraception, which comprise the primary endpoint of risk of an alcohol-exposed pregnancy.     

An analytical approximation method for pricing barrier options under the double Heston model

Abstract

The purpose of the paper is to provide an efficient pricing method for single barrier options under the double Heston model. By rewriting the model as a singular and regular perturbed BS model, the double Heston model can separately mimic a fast time-scale and a slow time-scale. With the singular and regular perturbation techniques, we analytically derive the first-order asymptotic expansion of the solution to a barrier option pricing partial differential equation. The convergence and efficiency of the approximate method is verified by Monte Carlo simulation. Numerical results show that the presented asymptotic pricing method is fast and accurate.     

Spatial prediction of crystalline defects observed in molecular dynamic simulations of plastic damage

Molecular dynamic computer simulation is an essential tool in materials science to study atomic properties of materials in extreme environments and guide development of new materials. We propose a statistical analysis to emulate simulation output with the ultimate goal of efficiently approximating the computationally intensive simulation. We compare several spatial regression approaches including conditional autoregression (CAR), discrete wavelets transform (DWT), and principle components analysis (PCA). The methods are applied to simulation of copper atoms with twin wall and dislocation loop defects, under varying tilt tension angles. We find that CAR and DWT yield accurate results but fail to capture extreme defects, yet PCA better captures defect structure.     

Sampling plans based on balanced incomplete block designs for evaluating the importance of computer model inputs

The importance of individual inputs of a computer model is sometimes assessed using indices that reflect the amount of output variation that can be attributed to random variation in each input. We review two such indices, and consider input sampling plans that support estimation of one of them, the variance of conditional expectation or VCE (Mckay, 1995. Los Alamos National Laboratory Report NUREG/CR-6311, LA-12915-MS). Sampling plans suggested by Sobol’, Saltelli, and McKay, are examined and compared to a new sampling plan based on balanced incomplete block designs. The new design offers better sampling efficiency for the VCE than those of Sobol’ and Saltelli, and supports unbiased estimation of the index associated with each input.     

Results concerning the generalized partially linear single-index model

The central topic of this article is the estimation of parameters of the generalized partially linear single-index model (GPLSIM). Two numerical optimization procedures are presented and an S-plus program based on these procedures is compared to a program by Wand in a simulation setting. The results from these simulations indicate that the estimates for the new procedures are as good, if not better, than Wand's. Also, this program is much more flexible than Wand's since it can handle more general models. Other simulations are also conducted. The first compares the effects of using linear interpolation versus spline interpolation in an optimization procedure. The results indicate that by using spline interpolation one gets more stable estimates at a cost of increased computational time. A second simulation was conducted to assess the performance of a method for estimating the variance of alpha. A third set of simulations is carried out to determine the best criterion for testing that one of the elements of alpha is equal to zero. The GPLSIM is applied to a water quality data set and the results indicate an interesting relationship between gastrointestinal illness and turbidity (cloudiness) of drinking water.     

Robust estimation and variable selection for varying-coefficient single-index models based on modal regression

ABSTRACT

In this paper, we propose a new efficient and robust penalized estimating procedure for varying-coefficient single-index models based on modal regression and basis function approximations. The proposed procedure simultaneously solves two types of problems: separation of varying and constant effects and selection of variables with non zero coefficients for both non parametric and index components using three smoothly clipped absolute deviation (SCAD) penalties. With appropriate selection of the tuning parameters, the new method possesses the consistency in variable selection and the separation of varying and constant coefficients. In addition, the estimators of varying coefficients possess the optimal convergence rate and the estimators of constant coefficients and index parameters have the oracle property. Finally, we investigate the finite sample performance of the proposed method through a simulation study and real data analysis.     

Variable Selection via Regression Trees in the Presence of Irrelevant Variables

Many tree algorithms have been developed for regression problems. Although they are regarded as good algorithms, most of them suffer from loss of prediction accuracy when there are many irrelevant variables and the number of predictors exceeds the number of observations. We propose the multistep regression tree with adaptive variable selection to handle this problem. The variable selection step and the fitting step comprise the multistep method.

The multistep generalized unbiased interaction detection and estimation (GUIDE) with adaptive forward selection (fg) algorithm, as a variable selection tool, performs better than some of the well-known variable selection algorithms such as efficacy adaptive regression tube hunting (EARTH), FSR (false selection rate), LSCV (least squares cross-validation), and LASSO (least absolute shrinkage and selection operator) for the regression problem. The results based on simulation study show that fg outperforms other algorithms in terms of selection result and computation time. It generally selects the important variables correctly with relatively few irrelevant variables, which gives good prediction accuracy with less computation time.     

BP神经网络在基本养老保险基金支付风险预警中的应用

尝试将BP神经网络引入基本养老保险基金的风险预警中,以期为其提供一种新的预警工具和方法。首先,建立了一套基本养老保险基金风险预警指标体系,构建了基于BP网络的基本养老保险基金风险预警模型;其次,采集1996年至2008年间的上海市各年度基本养老保险基金的历史数据对该模型进行了反复训练和学习,取得了误差率仅为3.86%的预测结果,说明该模型有很好的拟合度;最后,依据国际国内经验对基本养老保险基金的警情指标设置了五个警度输出区间。     

半参数变系数回归模型的空间相关性检验

对半参数变系数回归模型,构造了新的空间相关性检验统计量,利用三阶矩 逼近方法导出了其检验 值的近似计算公式,蒙特卡罗模拟结果表明该统计量在检测空间相关性方面具有较高的准确性和可靠性。同时考察了误差项服从不同分布时的检验功效,体现出该检验方法的稳健性。进一步,我们还给出了检验统计量的Bootstrap方法以及检验水平的模拟效果。     

Logistic and neural network models for predicting a hospital admission

Feedforward neural networks are often used in a similar manner as logistic regression models; that is, to estimate the probability of the occurrence of an event. In this paper, a probabilistic model is developed for the purpose of estimating the probability that a patient who has been admitted to the hospital with a medical back diagnosis will be released after only a short stay or will remain hospitalized for a longer period of time. As the purpose of the analysis is to determine if hospital characteristics influence the decision to retain a patient, the inputs to this model are a set of demographic variables that describe the various hospitals. The output is the probability of either a short or long term hospital stay. In order to compare the ability of each method to model the data, a hypothesis test is performed to test for an improvement resulting from the use of the neural network model.     

A new metamodel-based method for solving semi-expensive simulation optimization problems

In this article, a new algorithm for rather expensive simulation problems is presented, which consists of two phases. In the first phase, as a model-based algorithm, the simulation output is used directly in the optimization stage. In the second phase, the simulation model is replaced by a valid metamodel. In addition, a new optimization algorithm is presented. To evaluate the performance of the proposed algorithm, it is applied to the (s,S) inventory problem as well as to five test functions. Numerical results show that the proposed algorithm leads to better solutions with less computational time than the corresponding metamodel-based algorithm.