Automatic Determination of Stellar Atmospheric Parameters Using Neural Networks and Instance-Based Machine Learning

In this article we show how machine learning methods can beeffectively applied to the problem of automatically predictingstellar atmospheric parameters from spectral information, a veryimportant problem in stellar astronomy. We apply feedforwardneural networks, Kohonen's self-organizing maps andlocally-weighted regression to predict the stellar atmosphericparameters effective temperature, surface gravity and metallicityfrom spectral indices. Our experimental results show that thethree methods are capable of predicting the parameters with verygood accuracy. Locally weighted regression gives slightly betterresults than the other methods using the original dataset asinput, while self-organizing maps outperform the other methods when significant amounts of noise are added. We also implemented a heterogeneous ensemble of predictors, combining the results given by the three algorithms. This ensemble yields better results than any of the three algorithms alone, using both the original and the noisy data.     

基于卷积神经网络的快速射电暴候选体分类

针对目前从海量的快速射电暴(Fast Radio Burst, FRB)候选体中人工筛选FRB事件难以为继的现状,提出了一种基于卷积神经网络(Convolutional Neural Networks, CNN)的FRB候选体分类方法.首先,通过真实的观测数据和仿真FRB组成训练和测试样本集.其次,建立了二输入的深度卷积神经网络模型,并对其进行训练、测试和优化,获取FRB候选体分类器.最后,利用来自脉冲星的单脉冲数据对该分类器的有效性和性能进行了验证.实验结果表明,该方法可以快速从大量候选体中准确识别出单脉冲事件,极大地提高了FRB候选体的处理速率和效率.     

Parameterizing Stellar Spectra Using Deep Neural Networks

Large-scale sky surveys are observing massive amounts of stellar spectra.The large number of stellar spectra makes it necessary to automatically parameterize spectral data,which in turn helps in statistically exploring properties related to the atmospheric parameters.This work focuses on designing an automatic scheme to estimate effective temperature(T_(eff)),surface gravity(log g) and metallicity[Fe/H] from stellar spectra.A scheme based on three deep neural networks(DNNs) is proposed.This scheme consists of the following three procedures:first,the configuration of a DNN is initialized using a series of autoencoder neural networks;second,the DNN is fine-tuned using a gradient descent scheme;third,three atmospheric parameters T_(eff),log g and [Fe/H] are estimated using the computed DNNs.The constructed DNN is a neural network with six layers(one input layer,one output layer and four hidden layers),for which the number of nodes in the six layers are 3821,1000,500,100,30 and 1,respectively.This proposed scheme was tested on both real spectra and theoretical spectra from Kurucz's new opacity distribution function models.Test errors are measured with mean absolute errors(MAEs).The errors on real spectra from the Sloan Digital Sky Survey(SDSS) are 0.1477,0.0048 and 0.1129 dex for log g,log T_(eff) and [Fe/H](64.85 K for T_(eff)),respectively.Regarding theoretical spectra from Kurucz's new opacity distribution function models,the MAE of the test errors are 0.0182,0.0011 and 0.0112 dex for log g,log T_(eff) and [Fe/H](14.90 K for T_(eff)),respectively.     

Automatic Solar Flare Detection Using Neural Network Techniques

We present a new method for automatic detection of flare events from images in the optical range. The method uses neural networks for pattern recognition and is conceived to be applied to full-disk Himages. Images are analyzed in real time, which allows for the design of automatic patrol processes able to detect and record flare events with the best time resolution available without human assistance. We use a neural network consisting of two layers, a hidden layer of nonlinear neurodes and an output layer of one linear neurode. The network was trained using a back-propagation algorithm and a set of full-disk solar images obtained by HASTA (HSolar Telescope for Argentina), which is located at the Estación de Altura Ulrico Cesco of OAFA (Observatorio Astronómico Félix Aguilar), El Leoncito, San Juan, Argentina. This method is appropriate for the detection of solar flares in the complete optical classification, being portable to any Hinstrument and providing unique criteria for flare detection independent of the observer.     

Radio Emission of Meteor Bodies in the Extremely Low Frequency Range

Solar System Research - A model for generating electromagnetic fields in the ELF/VLF range accompanying the final stage of the evolution of meteorites in the Earth’s atmosphere is presented....     

Detecting HI Galaxies with Deep Neural Networks in the Presence of Radio Frequency Interference

In the neutral hydrogen(H I) galaxy survey, a significant challenge is to identify and extract the H I galaxy signal from the observational data contaminated by radio frequency interference(RFI). For a drift-scan survey, or more generally a survey of a spatially continuous region, in the time-ordered spectral data, the H I galaxies and RFI all appear as regions that extend an area in the time-frequency waterfall plot, so the extraction of the H I galaxies and RFI from such data can be regarded a...     

Radio and Meteor Science Outcomes From Comparisons of Meteor Radar Observations at AMISR Poker Flat, Sondrestrom, and Arecibo

Radio science and meteor physics issues regarding meteor “head-echo” observations with high power, large aperture (HPLA) radars, include the frequency and latitude dependency of the observed meteor altitude, speed, and deceleration distributions. We address these issues via the first ever use and analysis of meteor observations from the Poker Flat AMISR (PFISR: 449.3 MHz), Sondrestrom (SRF: 1,290 MHz), and Arecibo (AO: 430 MHz) radars. The PFISR and SRF radars are located near the Arctic Circle while AO is in the tropics. The meteors observed at each radar were detected and analyzed using the same automated FFT periodic micrometeor searching algorithm. Meteor parameters (event altitude, velocity, and deceleration distributions) from all three facilities are compared revealing a clearly defined altitude “ceiling effect” in the 1,290 MHz results relative to the 430/449.3 MHz results. This effect is even more striking in that the Arecibo and PFISR distributions are similar even though the two radars are over 2,000 times different in sensitivity and at very different latitudes, thus providing the first statistical evidence that HPLA meteor radar observations are dominated by the incident wavelength, regardless of the other radar parameters. We also offer insights into the meteoroid fragmentation and “terminal” process.     

Annual and Diurnal Variation of Meteor Rates by the Forward-Scatter Radio Observation

Forward-scatter radio meteor observations have been made at Japan since 1996 using inexpensive and low-end equipment. The activity of some major meteor showers and the seasonal variability of sporadic meteors in 2006 are presented.     

Automatic Video System for Continues Monitoring of the Meteor Activity

In this paper we present current progress in development of new observational instruments for the double station video experiment. The Meteor Automatic Imager and Analyser (MAIA) system is based on digital monochrome camera JAI CM-040 and well proved image intensifier XX1332. Both the observations as well as the data processing will be fully automatic. We are expecting the recorded data of better quality and both spatial and time resolution in comparison with currently used analogue system. The main goal of the MAIA project is to monitor activity of the meteor showers and sporadic meteor each night for the period of at least 3 years. First version of the system was already assembled and has been intensively tested in the optical laboratory. Optical properties were measured and the result confirmed our expectations according to image quality and resolution. First night sky observation was already carried out.     

基于独立成分分析的射频干扰信号消除方法

射电天文已成为人类研究宇宙的重要途径。但随着人类生产、生活的发展,射频干扰信号对射电天文观测的影响越来越严重,观测数据的好坏关系到科学成果的质量甚至结论的真伪。目前广泛采用基于阈值判断射频干扰,对干扰信号直接舍弃部分观测数据的方法。此类方法存在阈值确定困难、观测带宽和时间被缩减等问题。针对脉冲星观测射电信号中,各干扰信号及射电信号统计独立以及呈现出的非高斯性,利用独立成分分析对混合信号进行分解,并根据观测信号中脉冲星信号和干扰信号的分布特点识别脉冲星信号,实现干扰信号消除。使用该方法对云南天文台40 m射电望远镜接收到的脉冲星观测信号进行独立成分分析,分解出独立的射频干扰信号和脉冲星信号,消除射频干扰信号。独立成分分析法在干扰信号消除、射电信号保留及信噪比方面均取得良好效果。     

Observation and Research on Strong Meteor Showers and Their Catastrophic Space Weather Events

During the first international joint observation of the Leo strong meteor shower, multidisciplinary and multi-media synthetic observation of Leo and Draco strong meteor showers and their catastrophic space weather events were carried out. The comprehensive analysis of the observed and related data of the Leo, Perseus and Draco strong meteor showers obtained for near half a century (from 1957 to 2003) fully verifies that the non-sporadic periodic strong meteor shower may lead to catastrophic space weather events. Preliminary identification is made of the following: the mechanism of formation of strong meteor showers, the law of occurrence of the f_bE_s abnormal peak and serious safety hazards for astronavigation. Also discussed in this paper are the evolutionary process of cometary dust, the law of occurrence and loss of cosmic dust storm and the mechanism of maintaining cosmic dust in the mid-latitude E_s layer.     

Meteor Orbit Determinations with Multistatic Receivers Using the MU Radar

The MU radar of RISH (Research Institute for Sustainable Humanosphere, Kyoto University), which is a MST radar (46.5 MHz, 1 MW peak power), has been successfully applied to meteor studies by using its very high versatility. The system has recently renewed with 25 channel digital receivers which significantly improved the sensitivity and precision of interferometer used in meteor observation. The transmission is now synchronized to GPS signals, and two external receiving sites with a ranging capability has additionally been operated in order to determine the trajectories and speeds of meteoroids.     

用陕西天文台流星雷达监测人为空间碎片可能性的分析

本文分析了利用中国科学院陕西天文台的流星雷达进行人为空间碎片监测的可能性．详细计算了到达接收机的回波的信噪比Ｓ／Ｎ依赖于目标散射横截面。和高度距离Ｒ的关系。理论计算表明，利用陕西天文台的流星雷达完全有可能监测在２００ｋｍ至１２００ｋｍ的高度范围内，半径大于０．５ｍ的人为空间碎片．     

Nowcasting Solar Energetic Particle Events Using Principal Component Analysis

We perform a principal component analysis (PCA) on a set of six solar variables (i.e. width/size (\(s\)) and velocity (\(u\)) of a coronal mass ejection, logarithm of the solar flare (SF) magnitude (\(\log\mathit{SXRs}\)), SF longitude (\(\mathit{lon}\)), duration (\(\mathit{DT}\)), and rise time (\(\mathit{RT}\))). We classify the solar energetic particle (SEP) event radiation impact (in terms of the National Oceanic and Atmospheric Administration scales) with respect to the characteristics of their parent solar events. We further attempt to infer the possible prediction of SEP events. In our analysis, we use 126 SEP events with complete solar information, from 1997 to 2013. Each SEP event is a vector in six dimensions (corresponding to the six solar variables used in this work). The PCA transforms the input vectors into a set of orthogonal components. By mapping the characteristics of the parent solar events, a new base defined by these components led to the classification of the SEP events. We furthermore applied logistic regression analysis with single, as well as multiple explanatory variables, in order to develop a new index (\(I\)) for the nowcasting (short-term forecasting) of SEP events. We tested several different schemes for \(I\) and validated our findings with the implementation of categorical scores (probability of detection (POD) and false-alarm rate (FAR)). We present and interpret the obtained scores, and discuss the strengths and weaknesses of the different implementations. We show that \(I\) holds prognosis potential for SEP events. The maximum POD achieved is 77.78% and the relative FAR is 40.96%.     

日冕物质抛射及其伴生的射电辐射观测特征

统计分析了太阳活动周下降段(2003～2005年)发生的76个共生CME的射电爆发事件.射电爆发资料来自国家天文台和Culgoora的微波和米波频谱仪.在76个事件中有50个快速CME和26个慢速CME.从中发现,快速CME和慢速CME的产率分别随着太阳活动周的降低而下降和上升,这可能说明CME的产率与太阳活动周中日冕磁结构的位形和位置变化有关.同时也发现,射电爆发的类型和寿命有一个变化规律,即随着频率的降低射电爆发的寿命变长,此特征支持了伴生CME的Ⅱ型爆发统一模型的思想.另外还发现在厘米一分米波范围,CME开始前后,容易发生射电Ⅲ型爆发或快速精细结构.这说明射电辐射的精细结构可能是CME的前兆现象或CME早期发展阶段由于磁重联引发的低日冕小尺度磁扰动的结果.     

Stellar Spectral Classification using Principal Component Analysis and Artificial Neural Networks

A fast and robust method of classifying a library of optical stellar spectra for O to M type stars is presented. The method employs, as tools: (1) principal component analysis (PCA) for reducing the dimensionality of the data and (2) multilayer back propagation network (MBPN) based artificial neural network (ANN) scheme to automate the process of classification. We are able to reduce the dimensionality of the original spectral data to very few components by using PCA and are able to successfully reconstruct the original spectra. A number of NN architectures are used to classify the library of test spectra. Performance of ANN with this reduced dimension shows that the library can be classified to accuracies similar to those achieved by Gulati et al. but with less computational load. Furthermore, the data compression is so efficient that the NN scheme successfully classifies to the desired accuracy for a wide range of architectures. The procedure will greatly improve our capabilities in handling and analysing large spectral data bases of the future.     

来自6538活动区的太阳射电米波事件

利用云南天文台声光频谱仪在1991年3月记录到的太阳射电米波辐射事件、光学活动及相关事件作了分析，得到来自6538活动区太阳射电米波事件的一些基本特性。     

Feedback Stabilization of Burn Conditions in a Thermonuclear Reactor Using Artificial Neural Networks: Preliminary Results

In this work we present some preliminary results concerning the use of Artificial Neural Networks (ANN) for the feedback stabilization of a thermonuclear reactor at nearly ignited burn conditions. An algorithm to train ANN for this purpose, involving the method of backpropagation through time and the conjugate gradients technique, is discussed. The training results show that ANN can be used successfully for the stabilization of burn conditions in thermonuclear reactors which are unstables to thermal excursions.     

发生在米波段的太阳射电辐射

综述云南天文台在太阳活动22周峰年期间观测到的米波射电频谱资料,处理资料时发现230~300MHz频段有一些有趣的观测现象。如米波Ⅲ型爆发;“blips”;毫秒级Spike与Ⅲ型爆发共存事件;快速脉动等。根据这些事件的现象,讨论了它们的产生机制。