基于神经网络集成的失业预警方法

提出采用神经网络集成技术对中国失业预警系统进行建模,以克服当前失业预警系统建模中存在的小样本、高维度、非线性、噪音数据等难题。采用BP神经网络回归模型对失业率进行预测;基于两种集成技术Bagging与AdaBoost对多个神经网络进行集成,以获得比单个预测模型更好的精度与稳定性;最后基于广东省的社会经济调查数据进行了实证分析,实验结果表明:在对失业率的预测上,Bagging集成方法的预测效果优于Adaboost集成方法,也优于单个最好的神经网络模型。     

Analyzing the financial distress of Chinese public companies using probabilistic neural networks and multivariate discriminate analysis

Many studies have applied backpropagation feedforward neural networks (BPNNs) as an alternative to multivariate discriminant analysis (MDA) in attempts to predict business distress using relatively small data sets. Although these studies have generally reported the superiority of BPNNs vs. MDA, they seem to ignore the fact that the former suffers from overfitting if the data set is too small compared to the free parameters of the network. We thus suggest an alternative approach that involves use of a probabilistic neural network (PNN). From our study of financially distressed Chinese public companies, we found that both the PNN and MDA algorithms provide good classifications. Relative to MDA, however, the PNN method provides better prediction, and, at the same time, does not require multivariate normality of the data. Our results appear to offer an improvement from those of earlier efforts that employ MDA, BPNN, and other models. In particular, PNN was here able to predict company distress with greater than 87.5% short-term accuracy, and 81.3% medium-term accuracy.     

Neural networks for GEFCom2017 probabilistic load forecasting

This report describes the forecasting model which was developed by team “4C” for the global energy forecasting competition 2017 (GEFCom2017), with some modifications added afterwards to improve its accuracy. The model is based on neural networks. Temperature scenarios obtained from historical data are used as inputs to the neural networks in order to create load scenarios, and these load scenarios are then transformed into quantiles. By using a feature selection approach that is based on a stepwise regression technique, a neural network based model is developed for each zone. Furthermore, a dynamic choice of the temperature scenarios is suggested. The feature selection and dynamic choice of the temperature scenarios can improve the quantile scores considerably, resulting in very accurate forecasts among the top teams.     

The Role of Information and Influence in Social Networks

A bstract .   This article examines two mechanisms through which social networks are related to job mobility: (1) access to diverse sources of information about job openings and (2) nonredundant sources of influence. Using data on job changing and social networks among television station managers, we assess the extent to which job information and influence variables mediate the relationship between social network structure and getting a better job. Results indicate that there is an association between job mobility and having nonredundant contacts, but our measures of the information and influence mechanisms are not significant mediators. We conclude by reexamining the network-resource model that forms the basis for much of the research on the relationship between social networks and job mobility.     

Kaggle forecasting competitions: An overlooked learning opportunity

We review the results of six forecasting competitions based on the online data science platform Kaggle, which have been largely overlooked by the forecasting community. In contrast to the M competitions, the competitions reviewed in this study feature daily and weekly time series with exogenous variables, business hierarchy information, or both. Furthermore, the Kaggle data sets all exhibit higher entropy than the M3 and M4 competitions, and they are intermittent.In this review, we confirm the conclusion of the M4 competition that ensemble models using cross-learning tend to outperform local time series models and that gradient boosted decision trees and neural networks are strong forecast methods. Moreover, we present insights regarding the use of external information and validation strategies, and discuss the impacts of data characteristics on the choice of statistics or machine learning methods. Based on these insights, we construct nine ex-ante hypotheses for the outcome of the M5 competition to allow empirical validation of our findings.     

基于非线性主成分分析的信用评估模型研究

传统的主成分分析（PCA）本质上是一种线性映射算法,无法有效处理非线性关系的数据。本文在分析自联想神经网络（AANN）的基础上,借鉴传统PCA方法中的序数主成分概念,提出了基于顺序自联想神经网络（SAANN）的非线性主成分分析法（NLPCA）。进一步,结合神经网络（NN）和Logisitic模型,以我国上市公司为研究对象,分别构建了基于NLPCA-NN和NLPCA-Logisitic的信用评估模型。实证结果及ROC曲线分析表明,本文构建的NLPCA相比传统的线性PCA方法能有效地实现数据的非线性特征提取与降维,提高模型预测性能。此外,实证结果还表明,在相同PCA方法处理数据的条件下,神经网络模型的信用评估效果要好于Logisitic模型。     

Forecasting and trading on the VIX futures market: A neural network approach based on open to close returns and coincident indicators

Previous work has highlighted the difficulty of obtaining accurate and economically significant predictions of VIX futures prices. We show that both low prediction errors and a significant amount of profitability can be obtained by using a neural network model to predict VIX futures returns. In particular, we focus on open-to-close returns (OTCRs) and consider intraday trading strategies, taking into account non-lagged exogenous variables that closely reflect the information possessed by traders at the time when they decide to invest. The neural network model with only the most recent exogenous variables (namely, the return on the Indian BSESN index) is superior to an unconstrained specification with ten lagged and coincident regressors, which is actually a form of weak efficiency involving markets of different countries. Moreover, the neural network turns out to be more profitable than either a logistic specification or heterogeneous autoregressive models.     

用局部和半局部转移函数实现RBF型神经网络的逼近和分类

增量神经网络(IncNet)的结构是由增长和剪枝控制,并且与训练数据的复杂性相匹配。用扩展卡尔曼滤波算法作为其学习算法。双径向转移函数比其它常用于人工神经网络的转移函数更具有灵活性。最新的改进是在多维空间中(具有N-1个参数)增加转移函数的旋转常数值。通过对逼近基准和心理分类问题的结果分析,清楚地表明比其他分类网络模型具有更强的泛化性。     

异因同果关联神经网络的研究

提出了一种异因同果关联神经网络模型,可以从不同角度分别建立不同的模型,并由其得到互不相同的模型预测值。异因同果关联神经网络模型将不同角度建立的模型有机结合起来,进而能够将多个神经网络模型进行综合考虑,得到一个综合的统一的模型预测结果。研究了新型模型的机理,结合实例进行仿真并与传统的神经网络模型的预测仿真结果比较,结果表明新型模型具有更高的预测精度。     

Automatic neural network modeling for univariate time series

Artificial neural networks (ANNs) are an information processing paradigm inspired by the way the brain processes information. Using neural networks requires the investigator to make decisions concerning the architecture or structure used. ANNs are known to be universal function approximators and are capable of exploiting nonlinear relationships between variables. This method, called Automated ANNs, is an attempt to develop an automatic procedure for selecting the architecture of an artificial neural network for forecasting purposes. It was entered into the M-3 Time Series Competition. Results show that ANNs compete well with the other methods investigated, but may produce poor results if used under certain conditions.     

On the shape of posterior densities and credible sets in instrumental variable regression models with reduced rank: An application of flexible sampling methods using neural networks

Likelihoods and posteriors of instrumental variable (IV) regression models with strong endogeneity and/or weak instruments may exhibit rather non-elliptical contours in the parameter space. This may seriously affect inference based on Bayesian credible sets. When approximating posterior probabilities and marginal densities using Monte Carlo integration methods like importance sampling or Markov chain Monte Carlo procedures the speed of the algorithm and the quality of the results greatly depend on the choice of the importance or candidate density. Such a density has to be ‘close’ to the target density in order to yield accurate results with numerically efficient sampling. For this purpose we introduce neural networks which seem to be natural importance or candidate densities, as they have a universal approximation property and are easy to sample from. A key step in the proposed class of methods is the construction of a neural network that approximates the target density. The methods are tested on a set of illustrative IV regression models. The results indicate the possible usefulness of the neural network approach.     

Forecasting exchange rates using feedforward and recurrent neural networks

In this paper we investigate the out-of-sample forecasting ability of feedforward and recurrent neural networks based on empirical foreign exchange rate data. A two-step procedure is proposed to construct suitable networks, in which networks are selected based on the predictive stochastic complexity (PSC) criterion, and the selected networks are estimated using both recursive Newton algorithms and the method of nonlinear least squares. Our results show that PSC is a sensible criterion for selecting networks and for certain exchange rate series, some selected network models have significant market timing ability and/or significantly lower out-of-sample mean squared prediction error relative to the random walk model.     

Forecasting economic time series using flexible versus fixed specification and linear versus nonlinear econometric models

Nine macroeconomic variables are forecast in a real-time scenario using a variety of flexible specification, fixed specification, linear, and nonlinear econometric models. All models are allowed to evolve through time, and our analysis focuses on model selection and performance. In the context of real-time forecasts, flexible specification models (including linear autoregressive models with exogenous variables and nonlinear artificial neural networks) appear to offer a useful and viable alternative to less flexible fixed specification linear models for a subset of the economic variables which we examine, particularly at forecast horizons greater than 1-step ahead. We speculate that one reason for this result is that the economy is evolving (rather slowly) over time. This feature cannot easily be captured by fixed specification linear models, however, and manifests itself in the form of evolving coefficient estimates. We also provide additional evidence supporting the claim that models which ‘win’ based on one model selection criterion (say a squared error measure) do not necessarily win when an alternative selection criterion is used (say a confusion rate measure), thus highlighting the importance of the particular cost function which is used by forecasters and ‘end-users’ to evaluate their models. A wide variety of different model selection criteria and statistical tests are used to illustrate our findings.     

Networks to Nerdistan: The Role of Labor Market Intermediaries in the Entry‐level IT Labor Market

Although networks have long governed economic relations, they assume even more importance in a knowledge‐based economy. Yet, some argue that because of the lack of social networks and human capital, some groups are permanently ‘switched off’ the networks of the global economy. Evidence presented in this article suggests that instead there is latent potential for access to the network, due to the rise of networked community‐based organizations and the increasing accessibility of technology. Based on surveys and in‐depth interviews with almost 700 workers and training providers, I show how the switched off are entering jobs in information technology through network ties and the acquisition of soft skills, or communication and interaction skills. Although community‐based training providers are best positioned to help disadvantaged jobseekers enter the network society, changes in the US workforce development system are reinforcing network exclusivity, rather than facilitating this upward mobility.     

Forecasting performances of three automated modelling techniques during the economic crisis 2007–2009

In this work we consider the forecasting of macroeconomic variables during an economic crisis. The focus is on a specific class of models, the so-called single hidden-layer feed-forward autoregressive neural network models. What makes these models interesting in the present context is the fact that they form a class of universal approximators and may be expected to work well during exceptional periods such as major economic crises. Neural network models are often difficult to estimate, and we follow the idea of White (2006) of transforming the specification and nonlinear estimation problem into a linear model selection and estimation problem. To this end, we employ three automatic modelling devices. One of them is White’s QuickNet, but we also consider Autometrics, which is well known to time series econometricians, and the Marginal Bridge Estimator, which is better known to statisticians. The performances of these three model selectors are compared by looking at the accuracy of the forecasts of the estimated neural network models. We apply the neural network model and the three modelling techniques to monthly industrial production and unemployment series from the G7 countries and the four Scandinavian ones, and focus on forecasting during the economic crisis 2007–2009. The forecast accuracy is measured using the root mean square forecast error. Hypothesis testing is also used to compare the performances of the different techniques.     

U-Convolutional model for spatio-temporal wind speed forecasting

The increasing penetration of intermittent renewable energy in power systems brings operational challenges. One way of supporting them is by enhancing the predictability of renewables through accurate forecasting. Convolutional Neural Networks (Convnets) provide a successful technique for processing space-structured multi-dimensional data. In our work, we propose the U-Convolutional model to predict hourly wind speeds for a single location using spatio-temporal data with multiple explanatory variables as an input. The U-Convolutional model is composed of a U-Net part, which synthesizes input information, and a Convnet part, which maps the synthesized data into a single-site wind prediction. We compare our approach with advanced Convnets, a fully connected neural network, and univariate models. We use time series from the Climate Forecast System Reanalysis as datasets and select temperature and u- and v-components of wind as explanatory variables. The proposed models are evaluated at multiple locations (totaling 181 target series) and multiple forecasting horizons. The results indicate that our proposal is promising for spatio-temporal wind speed prediction, with results that show competitive performance on both time horizons for all datasets.     

Geography of Italian student mobility: A network analysis approach

Understanding the phenomenon of intra- and international student mobility has become increasingly relevant to the organization of tertiary education systems. Using microdata information provided by the Italian National Student Archive on the cohorts of students enrolled at university in the academic years 2011–12 and 2014–15, we consider a network analysis approach to investigate the incoming and outgoing student flows between territories and universities. More specifically, the paper aims to shed light on the dynamics of Italian student mobility networks at both the bachelor's and master's degree levels by considering attractiveness indicators combined with network centrality measures, clustering techniques for network data and explanatory models. We analyze the partition of the global network structure by means of blockmodeling analysis and we explain the determinants of the differences among universities in attracting students adopting a quantile regression analysis.     

An integrated fuzzy mathematical model and principal component analysis algorithm for forecasting uncertain trends of electricity consumption

This paper introduces an integrated algorithm for forecasting electricity consumption (EL) based on fuzzy regression, time series and principal component analysis (PCA) in uncertain markets such as Iran. The algorithm is examined by mean absolute percentage error, analysis of variance (ANOVA) and Duncan Multiple Range Test. PCA is used to identify the input variables for the fuzzy regression and time series models. Monthly EL in Iran is used to show the superiority of the algorithm. Moreover, it is shown that the selected fuzzy regression model has better estimated values for total EL than time series. The algorithm provides as good results as intelligent methods. However, it is shown that the algorithm does not require utilization of preprocessing methods but genetic algorithm, artificial neural network and fuzzy inference system require preprocessing which could be a cumbersome task to deal with ambiguous data. The unique features of the proposed algorithm are three fold. First, two type of fuzzy regressions with and without preprocessed data are prescribed by the algorithm in order to minimize the bias. Second, it uses PCA approach instead of trial and error method for selecting the most important input variables. Third, ANOVA is used to statistically compare fuzzy regression and time series with actual data.     

边坡影响因素的自组织神经网络归类处理

文章根据自组织神经网络的基本原理,结合55个边坡实例,应用matlab进行编程,建立了边坡影响因素分类处理的神经网络模型,并运用该模型对不同的边坡进行了分类,分类结果提高了神经网络的边坡指标数据的学习效率,从而证明了自组织神经网络对提高用于预测边坡稳定性神经网络性能的有效性。