Deep Learning and Time Series-to-Image Encoding for Financial Forecasting

In the last decade,market financial forecasting has attracted high interests amongst the researchers in pattern recognition.Usually,the data used for analysing the market,and then gamble on its future trend,are provided as time series;this aspect,along with the high fluctuation of this kind of data,cuts out the use of very efficient classification tools,very popular in the state of the art,like the well known convolutional neural networks(CNNs)models such as Inception,Res Net,Alex Net,and so on.This forces the researchers to train new tools from scratch.Such operations could be very time consuming.This paper exploits an ensemble of CNNs,trained over Gramian angular fields(GAF)images,generated from time series related to the Standard&Poor's 500 index future;the aim is the prediction of the future trend of the U.S.market.A multi-resolution imaging approach is used to feed each CNN,enabling the analysis of different time intervals for a single observation.A simple trading system based on the ensemble forecaster is used to evaluate the quality of the proposed approach.Our method outperforms the buyand-hold(B&H)strategy in a time frame where the latter provides excellent returns.Both quantitative and qualitative results are provided.

Performance evaluation of series and parallel strategies for financial time series forecasting

Background:Improving financial time series forecasting is one of the most challenging and vital issues facing numerous financial analysts and decision makers.Given its direct impact on related decisions,various attempts have been made to achieve more accurate and reliable forecasting results,of which the combining of individual models remains a widely applied approach.In general,individual models are combined under two main strategies:series and parallel.While it has been proven that these strategies can improve overall forecasting accuracy,the literature on time series forecasting remains vague on the choice of an appropriate strategy to generate a more accurate hybrid model.Methods:Therefore,this study’s key aim is to evaluate the performance of series and parallel strategies to determine a more accurate one.Results:Accordingly,the predictive capabilities of five hybrid models are constructed on the basis of series and parallel strategies compared with each other and with their base models to forecast stock price.To do so,autoregressive integrated moving average(ARIMA)and multilayer perceptrons(MLPs)are used to construct two series hybrid models,ARIMA-MLP and MLP-ARIMA,and three parallel hybrid models,simple average,linear regression,and genetic algorithm models.Conclusion:The empirical forecasting results for two benchmark datasets,that is,the closing of the Shenzhen Integrated Index(SZII)and that of Standard and Poor’s 500(S&P 500),indicate that although all hybrid models perform better than at least one of their individual components,the series combination strategy produces more accurate hybrid models for financial time series forecasting.

Short-Term Financial Time Series Forecasting Integrating Principal Component Analysis and Independent Component Analysis with Support Vector Regression

Financial time series forecasting could be beneficial for individual as well as institutional investors. But, the high noise and complexity residing in the financial data make this job extremely challenging. Over the years, many researchers have used support vector regression (SVR) quite successfully to conquer this challenge. In this paper, an SVR based forecasting model is proposed which first uses the principal component analysis (PCA) to extract the low-dimensional and efficient feature information, and then uses the independent component analysis (ICA) to preprocess the extracted features to nullify the influence of noise in the features. Experiments were carried out based on 16 years’ historical data of three prominent stocks from three different sectors listed in Dhaka Stock Exchange (DSE), Bangladesh. The predictions were made for 1 to 4 days in advance targeting the short term prediction. For comparison, the integration of PCA with SVR (PCA-SVR), ICA with SVR (ICA-SVR) and single SVR approaches were applied to evaluate the prediction accuracy of the proposed approach. Experimental results show that the proposed model (PCA-ICA-SVR) outperforms the PCA-SVR, ICA-SVR and single SVR methods.

Improvements onthe Earnings Forecast Model——Based on Correlation between Financial Ratio, Auditor Opinion and Future Earnings

This paper aims to find evidence for the improvements on the present earnings forecast models through analyzing the correlation among financial ratios, auditor opinion of listed companies and their future earnings. This paper uses two statistical regression methods including Logistic model and Linear model to examine the inner interaction between financial ratios and future earnings from qualitative and quantitative perspectives respectively. Empirical tests find that financial ratios, especially ROE, can help to predict future earnings. Then we add auditor opinion variable into Logistic model to test whether going concern opinion in the auditor reports can be helpful for earnings forecast. Result shows the degree of optimistic statement of going concern opinion is significantly correlated with future earnings but with the disturbance of earnings management.

Feature Selection with Optimal Variational Auto Encoder for Financial Crisis Prediction

Financial crisis prediction(FCP)received significant attention in the financial sector for decision-making.Proper forecasting of the number of firms possible to fail is important to determine the growth index and strength of a nation’s economy.Conventionally,numerous approaches have been developed in the design of accurate FCP processes.At the same time,classifier efficacy and predictive accuracy are inadequate for real-time applications.In addition,several established techniques carry out well to any of the specific datasets but are not adjustable to distinct datasets.Thus,there is a necessity for developing an effectual prediction technique for optimum classifier performance and adjustable to various datasets.This paper presents a novel multi-vs.optimization(MVO)based feature selection(FS)with an optimal variational auto encoder(OVAE)model for FCP.The proposed multi-vs.optimization based feature selection with optimal variational auto encoder(MVOFS-OVAE)model mainly aims to accomplish forecasting the financial crisis.For achieving this,the proposed MVOFS-OVAE model primarily pre-processes the financial data using min-max normalization.In addition,the MVOFS-OVAE model designs a feature subset selection process using the MVOFS approach.Followed by,the variational auto encoder(VAE)model is applied for the categorization of financial data into financial crisis or non-financial crisis.Finally,the differential evolution(DE)algorithm is utilized for the parameter tuning of the VAE model.A series of simulations on the benchmark dataset reported the betterment of the MVOFS-OVAE approach over the recent state of art approaches.

Forecasting cryptocurrency returns and volume using search engines

In the context of the debate on the role of cryptocurrencies in the economy as well as their dynamics and forecasting,this brief study analyzes the predictability of Bitcoin volume and returns using Google search values.We employed a rich set of established empirical approaches,including a VAR framework,a copulas approach,and non-parametric drawings,to capture a dependence structure.Using a weekly dataset from 2013 to 2017,our key results suggest that the frequency of Google searches leads to positive returns and a surge in Bitcoin trading volume.Shocks to search values have a positive effect,which persisted for at least a week.Our findings contribute to the debate on cryptocurrencies/Bitcoins and have profound implications in terms of understanding their dynamics,which are of special interest to investors and economic policymakers.

Financial Level of Czech and Slovak Employees

This paper deals with the development of wage distribution by gender in the Czech and Slovak Republics in the years of 2005-2012. Special attention is given to changes in the behavior of wage distribution in relation to the onset of the global economic recession. The different behavior of the wage distribution of Czech and Slovak employees during the period is the subject of research. The article discusses the differences in the wage level between men and women in the Czech and Slovak Republics. There are the total wage distributions of men and women together, both in the Czech Republic and in the Slovak Republic on one hand, and wage distributions according to the gender separately for men and women on the other hand. Special attention was paid to the development of Gini coefficient of the concentration in both countries according to the gender in the period under review, too.

Riemann Hypothesis, Catholic Information and Potential of Events with New Techniques for Financial and Other Applications

In this research we are going to define two new concepts: a) “The Potential of Events” (EP) and b) “The Catholic Information” (CI). The term CI derives from the ancient Greek language and declares all the Catholic (general) Logical Propositions () which will true for every element of a set A. We will study the Riemann Hypothesis in two stages: a) By using the EP we will prove that the distribution of events e (even) and o (odd) of Square Free Numbers (SFN) on the axis Ax(N) of naturals is Heads-Tails (H-T) type. b) By using the CI we will explain the way that the distribution of prime numbers can be correlated with the non-trivial zeros of the function ζ(s) of Riemann. The Introduction and the Chapter 2 are necessary for understanding the solution. In the Chapter 3 we will present a simple method of forecasting in many very useful applications (e.g. financial, technological, medical, social, etc) developing a generalization of this new, proven here, theory which we finally apply to the solution of RH. The following Introduction as well the Results with the Discussion at the end shed light about the possibility of the proof of all the above. The article consists of 9 chapters that are numbered by 1, 2, …, 9.

Economic Effect of China's Rural Financial Market Growth during 1952-2013

Through study,it is found that since 1952,there has been a long-run equilibrium relationship between China's rural financial market growth and rural economic growth,the government-led rural financial market growth has effectively supported rural economic growth,and increasing the farmers' financing ratio has always helped to boost long-term growth of the rural economy.However,dominated by market mechanism from 1978,there is only one-way support relationship:rural economic growth brings about quantitative growth of rural financial market.

Replacing the Annual Budget with Business Intelligence Driver-Based Forecasts

The fixed annual budget process can be a cumbersome and static process, often failing to deliver intended benefits. Typically detached from business operations and strategic planning goals, the annual budget suffers from inherent weaknesses caused by a lack of business intelligence regarding its underlying assumptions. This weakness is well documented in existing literature and there is ample evidence of improved alternatives to static corporate financial planning. One such alternative utilizes business intelligence as an essential component in the annual budget process, along with rolling forecasts as a critical tool. Utilizing business intelligence supported, driver-based rolling forecasting can align an organization’s budget process with strategic objectives and can further the operational and financial strength of an organization, as well as maximize shareholder value. In order to fully explore this topic, this article will present a review of the conventional annual budget process and the manner in which an approach that bases financial forecasts on business intelligence drivers can align operations with strategic objectives and add value to an organization. An assessment of intelligence-supported, driver-based rolling forecasting will also be presented, demonstrating an im- proved approach to the traditional annual budgeting process.

Evaluation of forecasting methods from selected stock market returns

Forecasting stock market returns is one of the most effective tools for risk management and portfolio diversification.There are several forecasting techniques in the literature for obtaining accurate forecasts for investment decision making.Numerous empirical studies have employed such methods to investigate the returns of different individual stock indices.However,there have been very few studies of groups of stock markets or indices.The findings of previous studies indicate that there is no single method that can be applied uniformly to all markets.In this context,this study aimed to examine the predictive performance of linear,nonlinear,artificial intelligence,frequency domain,and hybrid models to find an appropriate model to forecast the stock returns of developed,emerging,and frontier markets.We considered the daily stock market returns of selected indices from developed,emerging,and frontier markets for the period 2000–2018 to evaluate the predictive performance of the above models.The results showed that no single model out of the five models could be applied uniformly to all markets.However,traditional linear and nonlinear models outperformed artificial intelligence and frequency domain models in providing accurate forecasts.

An integrated new threshold FCMs Markov chain based forecasting model for analyzing the power of stock trading trend

This paper explores the power of stock trading trend using an integrated New ThresholdFuzzy Cognitive Maps(NTFCMs)Markov chain model.This new model captures thepositive as well as the negative jumps and predicts the trend for different stocks over 4months which follow an uptrend,downtrend and a mixed trend.The mean absolute percent error(MAPE)tolerance limits,the root mean square error(RMSE)tolerance limits aredetermined for various stock indices over a multi-timeframe period and observed for theexisting methods lying within the defined limits.The results show for every‘n’number ofpredictions made,the predicted close value of the day’s stock price was within tolerancelimit with 0%error and with 100%accuracy in predicting the future trend.

An Artificial Neural Network Model to Forecast Exchange Rates

For the purposes of this research, the optimal MLP neural network topology has been designed and tested by means the specific genetic algorithm multi-objective Pareto-Based. The objective of the research is to predict the trend of the ex-change rate Euro/USD up to three days ahead of last data available. The variable of output of the ANN designed is then the daily exchange rate Euro/Dollar and the frequency of data collection of variables of input and the output is daily. By the analysis of the data it is possible to conclude that the ANN model developed can largely predict the trend to three days of exchange rate Euro/USD.

Methodological Approaches to Study Internet Financial Communication

The aim of this paper is to show how qualitative and quantitative approaches can be complementary to study internet financial communication in a thesis by papers and how grounded theory （GT） can be the link among the different papers of the thesis. The study context of our thesis was the unregulated markets of New York Stock Exchange （NYSE） Euronext Brussels and the problematic rose from this context： What is the voluntary effort of communication when there is no obligation of internet financial communication？ Four papers tried to answer this central question and other following research questions. To answer those research questions, several methodological approaches were used： content analysis of websites and scoring, linear regression, paired sample, and interviews. At the end of our thesis by papers, we discovered that GT was the general methodological travel among the papers： Every article had for vocation to try to answer the questions raised by the previous article.

Business and financial information integration and voluntary management earnings forecasts

In this study,the impact of business and financial information integration(BFⅡ)on the voluntary management earnings forecasts(VMEFs)of listed firms in China between 2008 and 2018 is investigated.Drawing on litigation cost and ability signaling theories,we find that the adoption of BFⅡencourages top managers to disclose VMEFs.BFⅡfirms are identified through the textual analysis of management discussion and analysis(MD&A)reports,and the empirical results indicate that BFⅡfirms have a higher probability and frequency of issuing VMEFs than non-BFⅡfirms.The results remain robust after we identify causality by applying a propensity score matching and difference-in-differences(PSM-DID)test and use an alternate measure of BFⅡ.Further tests show that BFⅡfirms issue more accurate VMEFs and are able to issue them at an earlier stage.We also find that the positive relationship between BFⅡand VMEFs is weakened if the media expresses concern about the uncertainty of BFⅡadoption.

A MULTISCALE MODELING APPROACH INCORPORATING ARIMA AND ANNS FOR FINANCIAL MARKET VOLATILITY FORECASTING

The financial market volatility forecasting is regarded as a challenging task because of irreg ularity, high fluctuation, and noise. In this study, a multiscale ensemble forecasting model is proposed. The original financial series are decomposed firstly different scale components （i.e., approximation and details） using the maximum overlap discrete wavelet transform （MODWT）. The approximation is pre- dicted by a hybrid forecasting model that combines autoregressive integrated moving average （ARIMA） with feedforward neural network （FNN）. ARIMA model is used to generate a linear forecast, and then FNN is developed as a tool for nonlinear pattern recognition to correct the estimation error in ARIMA forecast. Moreover, details are predicted by Elman neural networks. Three weekly exchange rates data are collected to establish and validate the forecasting model. Empirical results demonstrate consistent better performance of the proposed approach.

基于XGBoost模型的中小企业财务危机风险预测方法

为提高中小企业财务危机风险预测精度,提高预测性能,引入XGBoost模型,开展中小企业财务危机风险预测方法设计研究。建立中小企业财务危机预警指标体系,包含盈利能力、营运能力等一级指标和营业利润率、净资产收益率等二级指标。针对指标变量相关性和重叠性的问题,引入主成分分析方法,对数据处理。利用XGBoost,构建财务危机预测模型,实现风险预测。通过对比实验证明,新的预测方法预测结果ROC曲线更趋近于左上,说明预测性能得到显著提升。

我国金融形势指数的构建与混频预测

近年来,我国黄金和成品油的市场需求有所升温,本文尝试探讨是否应将黄金价格和西德克萨斯轻质(WTI)原油价格纳入构建我国金融形势指数。研究表明,纳入黄金价格的我国金融形势指数能够更好地拟合居民消费价格指数(CPI)的演进路径,且在时序上领先CPI三个季度;但纳入WTI原油价格不仅加剧了金融形势指数的波动,还弱化了其与CPI的关系。通过构建16变量混频贝叶斯向量自回归(MF-BVAR)模型,并综合运用点预测、区间预测、密度预测三种方法,对比混频贝叶斯向量自回归模型与季频贝叶斯向量自回归(QF-BVAR)模型、季频向量自回归(QF-VAR)模型对金融形势指数的预测能力。实证结果显示,MF-BVAR模型对于金融形势指数的预测效果优于QF-BVAR和QF-VAR模型,且该结论不受预测期的影响。季度内不同月份的信息差异对金融形势指数预测产生显著影响,且随着预测期的增加,不同组别之间金融形势指数的预测能力逐渐趋同。本研究对于加强金融形势预测和防范金融市场风险具有理论价值与实践意义。

基于改进深度学习的金融时间序列波动率研究

线性模型和传统神经网络模型是常用的传统金融时间序列预测方法,但在非线性、非平稳的金融时间序列预测中存在一定的局限性。对此,文章提出了一种改进的深度学习模型。该模型结合了卷积神经网络和长短时记忆网络,可以有效捕捉金融时间序列中的非线性特征。通过真实数据与预测结果对比发现,文章提出的算法模型有助于捕捉金融时间序列中的非线性特征,可提高波动率预测的准确性。

财务报告问询函对业绩预告质量治理效应研究

以沪深证券交易所和上市公司为博弈双方构建不完全信息静态博弈模型,通过均衡分析揭示财务报告问询函对业绩预告质量治理效应,并以2015—2019年沪深A股上市公司为研究样本进行实证检验。研究发现:财务报告问询函改善了前瞻性和后视性业绩预告质量,表明财务报告问询函对双重业绩预告质量具有真实治理效应;深圳证券交易所财务报告问询函对双重业绩预告质量的治理效应更显著;财务报告问询函治理效应在公司收函后第二年依然显著,但在收函后第三年不复存在,表明财务报告问询函对业绩预告质量的治理效应具有短暂性。