A hybrid Bayesian back‐propagation neural network approach to multivariate modelling

There is growing interest in the use of back‐propagation neural networks to model non‐linear multivariate problems in geotehnical engineering. To overcome the shortcomings of the conventional back‐propagation neural network, such as overfitting, where the neural network learns the spurious details and noise in the training examples, a hybrid back‐propagation algorithm has been developed. The method utilizes the genetic algorithms search technique and the Bayesian neural network methodology. The genetic algorithms enhance the stochastic search to locate the global minima for the neural network model. The Bayesian inference procedures essentially provide better generalization and a statistical approach to deal with data uncertainty in comparison with the conventional back‐propagation. The uncertainty of data can be indicated using error bars. Two examples are presented to demonstrate the convergence and generalization capabilities of this hybrid algorithm. Copyright © 2003 John Wiley & Sons, Ltd.     

Application of a neural network model in establishing a stage–discharge relationship for a tidal river

This paper presents the applicability of neural network (NN) modelling for forecasting and filtering problems. The multilayer feedforward (MLFF) network was first constructed to forecast the tidal‐level variations at the mouth of the River Chao Phraya in Thailand. Unlike the well‐known conventional harmonic analysis, the NN model uses a set of previous data for learning and then forecasting directly the time‐series of tidal levels. It was found that lead time of 1 to 24 hourly tidal levels can be predicted successfully using only a short‐time hourly learning data. The MLFF network was further used to establish a stage–discharge relationship for the tidal river. The results show a considerably better performance of the NN model over the conventional models. In addition, the stage–discharge relationship obtained by the NN model can indicate reasonably well the important behaviour of the tidal influences. Copyright © 2003 John Wiley & Sons, Ltd.     

福建省地壳形变观测网络建设

介绍了福建省地壳形变观测网络的建设。福建省地壳形变观测网络，以GPS观测技术为主，结合精密水准、精密重力和流动地磁测量，将3者重合到GPS观测基准站上获取观测信息，初步构成一个时空相结合，点线面与长中短兼顾的高精度、高时空分辨率的地壳形变观测网络。该网络可动态监测福建省东南沿海地区地壳运动变化，为地震预测预报提供背景信息，将取得显著的社会效益和经济效益。     

Primary proton spectrum around the knee deduced from the emulsion-chamber data obtained at Mts. Fuji and Kanbala

We have done extensive Monte Carlo simulations using the new simulation codes of CORSIKA and COSMOS to compare with the gamma-family data obtained at Mts. Fuji (3750 m above sea level) and Kanbala (5500 m above sea level). Then, we estimated the primary proton and helium spectra around the knee energy region using a multiple-layered feed-forward neural network as a classifier of primary particle kind. The selection efficiency of proton-induced family events is estimated to be 82%. The flux value of protons at 2×10¹⁵ eV is (5.5±1.5)×10⁻¹⁴ (m⁻² s⁻¹ sr⁻¹ GeV⁻¹). The result suggests heavy-enriched primary composition around the knee region.     

REVIEWS OF GEOGRAPHIC SOFTWARE

Geographic Software Reviewed in this article: THE DIFFUSION GAME . Christopher J. Lovelock and Charles B. Weinberg . MAPIT . Raymond J. Kalush , Jr . SIMODEL . Peter A. Williams and A. Stewart Fotheringham . WEATHER FRONTS . P. C. Moyer . WORLD DYNAMICS . Mark Lewis Baldwin     

SOFTWARE REVIEWS

software reviews are in this article BMDP/PC CONCEPTS COMPUTERIZED ATLAS EVALUATION OF ALTERNATE PROPOSALS (EAP) EXTRA: EXPRESS TRANSIT ANALYSIS MATC CAD MOTORS TRANSPORTATION PACKAGE PCMAP: Thematic Mapping Software for the IBM Personal Computer, Version 2.3 PLACE TRANPLAN . US-ATLAS     

RBF神经网络的汛期旱涝预报方法研究

分析广州2003年1～6月的SO2、NO2和PM10指数,发现(1)SO2、NO2和PM10具有非常好的同步变化关系,且以NO2和PM10的相关最大。(2)从月平均值来看,PM10是整个时段的主要污染物,SO2次之,NO2最小(1月除外);PM10和NO2呈现出逐月减小趋势,SO2变化平稳。主要讨论PM10指数的变化。(3)从月方差值来看,SO2、NO2和PM10的最大值均在1月,次大值在2月。(4)给出了几种地面气压型,结论显示主要清洁时段出现在锋区强风型,主要污染时段出现在脊内回流型。在脊内回流型下,若由前期的静风转吹东南(从东到南)风时,指数会增大。     

计算机模拟引擎及其在铀成矿过程模拟中的应用研究

本文介绍了计算机模拟引擎的设计与开发 ,铀成矿过程模拟的实施步骤 ,并对其在伊犁盆地南缘的应用作了探讨。     

Pyroclastic Flow Hazard at Volcán Citlaltépetl

Volcán Citlaltépetl (Pico de Orizaba) with an elevation of 5,675 m is the highest volcano in North America. Its most recent catastrophic events involved the production of pyroclastic flows that erupted approximately 4,000, 8,500, and 13,000 years ago. The distribution of mapped deposits from these eruptions gives an approximate guide to the extent of products from potential future eruptions. Because the topography of this volcano is constantly changing computer simulations were made on the present topography using three computer algorithms: energy cone, FLOW2D, and FLOW3D. The Heim Coefficient (), used as a code parameter for frictional sliding in all our algorithms, is the ratio of the assumed drop in elevation (H) divided by the lateral extent of the mapped deposits (L). The viscosity parameter for the FLOW2D and FLOW3D codes was adjusted so that the paths of the flows mimicked those inferred from the mapped deposits. We modeled two categories of pyroclastic flows modeled for the level I and level II events. Level I pyroclastic flows correspond to small but more frequent block-and-ash flows that remain on the main cone. Level II flows correspond to more widespread flows from catastrophic eruptions with an approximate 4,000-year repose period. We developed hazard maps from simulations based on a National Imagery and Mapping Agency (NIMA) DTED-1 DEM with a 90 m grid and a vertical accuracy of ±30 m. Because realistic visualization is an important aid to understanding the risks related to volcanic hazards we present the DEM as modeled by FLOW3D. The model shows that the pyroclastic flows extend for much greater distances to the east of the volcano summit where the topographic relief is nearly 4,300 m. This study was used to plot hazard zones for pyroclastic flows in the official hazard map that was published recently.