Neural Networks for Financial Time Series Forecasting

Financial and economic time series forecasting has never been an easy task due to its sensibility to political, economic and social factors. For this reason, people who invest in financial markets and currency exchange are usually looking for robust models that can ensure them to maximize their profile and minimize their losses as much as possible. Fortunately, recently, various studies have speculated that a special type of Artificial Neural Networks (ANNs) called Recurrent Neural Networks (RNNs) could improve the predictive accuracy of the behavior of the financial data over time. This paper aims to forecast: (i) the closing price of eight stock market indexes; and (ii) the closing price of six currency exchange rates related to the USD, using the RNNs model and its variants: the Long Short-Term Memory (LSTM) and the Gated Recurrent Unit (GRU). The results show that the GRU gives the overall best results, especially for the univariate out-of-sample forecasting for the currency exchange rates and multivariate out-of-sample forecasting for the stock market indexes.     

KEADA: Identifying Key Classes in Software Systems Using Dynamic Analysis and Entropy-Based Metrics

Software maintenance is indispensable in the software development process. Developers need to spend a lot of time and energy to understand the software when maintaining the software, which increases the difficulty of software maintenance. It is a feasible method to understand the software through the key classes of the software. Identifying the key classes of the software can help developers understand the software more quickly. Existing techniques on key class identification mainly use static analysis techniques to extract software structure information. Such structure information may contain redundant relationships that may not exist when the software runs and ignores the actual interaction times between classes. In this paper, we propose an approach based on dynamic analysis and entropy-based metrics to identify key classes in the Java GUI software system, called KEADA (identifying KEy clAsses based on Dynamic Analysis and entropy-based metrics). First, KEADA extracts software structure information by recording the calling relationship between classes during the software running process; such structure information takes into account the actual interaction of classes. Second, KEADA represents the structure information as a weighted directed network and further calculates the importance of each node using an entropy-based metric OSE (One-order Structural Entropy). Third, KEADA ranks classes in descending order according to their OSE values and selects a small number of classes as the key class candidates. In order to verify the effectiveness of our approach, we conducted experiments on three Java GUI software systems and compared them with seven state-of-the-art approaches. We used the Friedman test to evaluate all approaches, and the results demonstrate that our approach performs best in all software systems.     

面向神经形态感知的人工脉冲神经元的研究进展

近年来,随着人工智能技术和脉冲神经网络(SNN)的迅猛发展,人工脉冲神经元的研究逐渐兴起。人工脉冲神经元的研究对于开发具有人类智能水平的机器人、实现自主学习和自适应控制等领域具有重要的应用前景。传统的电子器件由于缺乏神经元的非线性特性,需要复杂的电路结构和大量的器件才能模拟简单的生物神经元功能,同时功耗也较高。因此,最近研究者们借鉴生物神经元的工作机制,提出了多种基于忆阻器等新型器件的人工脉冲神经元方案。这些方案具有功耗低、结构简单、制备工艺成熟等优点,并且在模拟生物神经元的多种功能等方面取得了显著进展。文章将从人工脉冲神经元的基本原理出发,综述和分析目前已有的各种实现方案。具体来说,将分别介绍基于传统电子器件和基于新型器件的人工脉冲神经元的实现方案,并对其优缺点进行比较。此外,还将介绍不同类型的人工脉冲神经元在实现触觉、视觉、嗅觉、味觉、听觉和温度等神经形态感知方面的应用,并对未来的发展进行展望。希望能够为人工脉冲神经元的研究和应用提供有益的参考和启示。     

Investigation Driven by Network Pharmacology on Potential Components and Mechanism of DGS,a Natural Vasoprotective Combination,for the Phytotherapy of Coronary Artery Disease

Phytotherapy offers obvious advantages in the intervention of Coronary Artery Disease (CAD), but it is difficult to clarify the working mechanisms of the medicinal materials it uses. DGS is a natural vasoprotective combination that was screened out in our previous research, yet its potential components and mechanisms are unknown. Therefore, in this study, HPLC-MS and network pharmacology were employed to identify the active components and key signaling pathways of DGS. Transgenic zebrafish and HUVECs cell assays were used to evaluate the effectiveness of DGS. A total of 37 potentially active compounds were identified that interacted with 112 potential targets of CAD. Furthermore, PI3K-Akt, MAPK, relaxin, VEGF, and other signal pathways were determined to be the most promising DGS-mediated pathways. NO kit, ELISA, and Western blot results showed that DGS significantly promoted NO and VEGFA secretion via the upregulation of VEGFR2 expression and the phosphorylation of Akt, Erk1/2, and eNOS to cause angiogenesis and vasodilation. The result of dynamics molecular docking indicated that Salvianolic acid C may be a key active component of DGS in the treatment of CAD. In conclusion, this study has shed light on the network molecular mechanism of DGS for the intervention of CAD using a network pharmacology-driven strategy for the first time to aid in the intervention of CAD.     

改进的MEA-WNN图像复原方法

模糊图像复原是计算机视觉和图像处理领域的重要任务。针对思维进化算法(mind evolutionary algorithm,MEA)和小波神经网络(wavelet neural network,WNN)相结合的图像复原模型中,MEA的得分函数相对差别小、选优功能较弱等问题,提出了一种改进的MEA-WNN图像复原方法。该方法采用逻辑回归函数进行幂律变换,增加得分之间的差别,从而增强MEA的选优功能。将改进的模型与传统的基于WNN和MEA-WNN的图像复原模型进行对比,改进的模型把复原图像峰值信噪比(peak signal-to-noise ratio,PSNR)分别提高15%和6.5%、结构相似性(structural similarity,SSIM)提高了6.1%和5%,实验结果证明改进模型的有效性和优越性。     

一种基于CNN的混叠光谱解调算法及FPGA实现

为了解决光纤布拉格光栅(FBG)传感网络的光谱信号混叠问题,基于现场可编程门阵列(FPGA)提出了一种利用卷积神经网络(CNN)模型的混叠光谱信号解调算法,并对其进行硬件实现与加速。通过对模型参数进行定点数量化,压缩网络模型的存储空间,提高FPGA中DSP资源的利用率;利用循环展开和数组重排等硬件优化方法,提高了系统实时性,确定了算法的并行计算方案。研究结果表明,在100 MHz的时钟下,测试集解调精度为1.19 pm,推理速度为每帧14.96μs,光谱解调速率为60 kHz,对于FBG混叠光谱信号解调具有较高的精度和速率。     

Performance of cellular packet CDMA in an integrated voice/data network

This paper presents a simulation-based study of cellular packet CDMA systems operating in an integrated voice/data traffic scenario. Spread-spectrum CDMA provides a suitable framework for resource-shared packet transport capable of combining isochronous (voice, ISDN) and bursty data services. In this work, a general network model for cellular packet CDMA with mixed voice/data traffic is described and used to evaluate the capacity/performance impact of several key system parameters. First, the effect of spreading factor (N) and forward error correction (FEC) rate are studied, confirming earlier work indicating a weak dependence onN and a well-defined optimum code rate in the range of 0.5–0.7 (with BCH coding). Next, the effect of propagation loss coefficient () on network capacity is investigated over a range of possible assumptions for, including both constant and distance-dependent models. The results show that system capacity depends strongly on, varying by as much as a factor of 2 over the range of parameters considered. For a given distance-dependent assumption, performance results are also obtained for different cell sizes in order to understand the overall spatial reuse efficiency achievable in different cellular and microcellular scenarios. This is followed by an investigation of traffic source model effects: first the capacity improvement from voice activity detection VAD) is presented, showing the expected 21 gains. Results for varying proportions of voice and data traffic intensities indicate that the operating efficiency does not change significantly as the proportion of bursty data relative to voice is varied.     

BP网络过拟合现象满足的不确定关系新的改进式

类比信息传递过程中的一般测不准关系式 ,引进表征问题复杂性的函数复相关系数R和代表网络结构特性的隐节点数h ,揭示了BP网络过拟合现象出现时的网络学习能力与推广能力之间满足的不确定关系式 ;通过模拟了 12种不同类型复杂程度函数的过拟合数值试验 ,确定出关系式中的过拟合参数 p的取值范围已缩小为 1×10 -5～ 5× 10 -4;给出应用BP网络对给定样本集的训练过程中 ,判断出现过拟合现象的方法     

光链路检测系统在网管中的设计与实现

主要利用网管设计实现对集成了光时域反射仪(OTDR)板卡的OLT设备进行管理,达到监控管理光链路状况的目的。该方案具有自动检测和手动检测的功能,图形化的界面能实时显示测试与诊断结果。检测过程更简单,采集手段更完善,测试结论更准确,具有很强的实时性和实用性。     

VSAT的技术进展

本文从传输技术、多址接入方式、网络管理和网络体系结构等方面综述了甚小口径卫星地球站在技术上的最新进展。