多类支持向量机在语音识别中的应用

针对语音模式识别,引入多类（M-ary）支持向量机来进行分类。并详细的介绍了M-ary支持向量机的概念以及语音识别系统中的实现方式。通过仿真实验,与BP神经网络、SVM进行比较,验证了M-ary支持向量机具有识别率高,抗噪能力强等优越性。     

通用走航式海洋测量作业平台的设计与实现

扼要回顾了我国走航式海洋测量作业平台的研制历程,提出了通用走航式海洋测量作业平台研制的总体目标、设计原则、功能需求,介绍了该平台软件硬件研制实现及其主要性能、功能技术指标,以及平台的应用情况,最后对平台的未来发展进行了展望。     

Use of Gene Expression Programming in regionalization of flow duration curve

In this paper, a recently introduced artificial intelligence technique known as Gene Expression Programming (GEP) has been employed to perform symbolic regression for developing a parametric scheme of flow duration curve (FDC) regionalization, to relate selected FDC characteristics to catchment characteristics. Stream flow records of selected catchments located in the Auckland Region of New Zealand were used. FDCs of the selected catchments were normalised by dividing the ordinates by their median value. Input for the symbolic regression analysis using GEP was (a) selected characteristics of normalised FDCs; and (b) 26 catchment characteristics related to climate, morphology, soil properties and land cover properties obtained using the observed data and GIS analysis. Our study showed that application of this artificial intelligence technique expedites the selection of a set of the most relevant independent variables out of a large set, because these are automatically selected through the GEP process. Values of the FDC characteristics obtained from the developed relationships have high correlations with the observed values.     

Market driven management of climate change impacts in the Spanish mussel sector

Mussel farming is the largest aquaculture activity in Spain both in volume and value, being a key pillar for the development of the coastal areas in Galicia (NW Spain), where this production is concentrated. The available scientific knowledge suggests that in the long term the primary productivity of the Galician estuaries will be reduced due to climate change. Consequently, adaptive management will be required to face this challenge. The impact of the likely production decrease will depend, among other factors, on the availability of substitutes, being their identification the main object of this research. So, it was analysed whether this markets are integrated or not through cointegration tests, finding that only the French and the Spanish markets are partially integrated. This implies that since very little substitutes are available for the Galician fresh mussels, decreases in production will not necessarily affect producer's income, as prices increases will compensate the (eventual) reduction in volume. In terms of policy, this study suggests that market intelligence may contribute to an adaptive, pragmatic and flexible framework to face climate change impacts, avoiding too anticipated or blind responses that may result in economic, environmental and social costs.     

Real time earthquake forecasting in Italy

We have applied an earthquake clustering epidemic model to real time data at the Italian Earthquake Data Center operated by the Istituto Nazionale di Geofisica e Vulcanologia (INGV) for short-term forecasting of moderate and large earthquakes in Italy. In this epidemic-type model every earthquake is regarded, at the same time, as being triggered by previous events and triggering following earthquakes. The model uses earthquake data only, with no explicit use of tectonic, geologic, or geodetic information. The forecasts are displayed as time-dependent maps showing both the expected rate density of M_l ≥ 4.0 earthquakes and the probability of ground shaking exceeding Modified Mercalli Intensity VI (PGA ≥ 0.01 g) in an area of 100 × 100 km² around the zone of maximum expected rate density in the following 24 h. For testing purposes, the overall probability of occurrence of an M_l ≥ 4.5 earthquake in the same area of 100 × 100 km² is also estimated. The whole procedure is tested in real time, for internal use only, at the INGV Earthquake Data Center.Forecast verification procedures have been carried out in forward-retrospective way on the 2006–2007 INGV data set, making use of statistical tools as the Relative Operating Characteristics (ROC) diagrams. These procedures show that the clustering epidemic model performs up to several hundred times better than a simple random forecasting hypothesis. The seismic hazard modeling approach so developed, after a suitable period of testing and refinement, is expected to provide a useful contribution to real time earthquake hazard assessment, even with a possible practical application for decision making and public information.     

Focus-of-attention techniques in the automatic interpretation of seismograms

The focus-of-attention techniques implemented in SNA2, a knowledge-based system for seismogram interpretation, are presented. They consist of data compression of the input digital records, scanning of the compressed traces to detect candidate seismograms and extraction of seismogram features. A criterion is given to rate the clarity of seismograms; the clarity defines the order in which the system will consider them to build up the interpretation. The proposed techniques are simple and fast; they allow quick rejection of noise and focussing the attention of the system on the portions of traces containing relevant information.     

GPS在"数字地球"中的应用

从“数字地球”工程的引入开始，通过介绍GPS在“数字地球”工程几个领域中的应用，说明了GPS不可替代的重要地位。     

基于计算智能的土地适宜性评价模型

将计算智能理论引入土地评价领域，构建了一个全新的土地适宜性评价模型。首先基于模糊逻辑和人工神经网络构造了一个模糊神经网络模型，然后采用改进的遗传算法进行训练，能够快速收敛到最优解，对初始的规则库进行修正，形成了一个自学习、自适应的评价系统。     

人工神经网络与符号推理在江淮气旋预报中的应用研究

本文通过对符号推理及神经网络系统的剖析，实现了神经网络与符号推理的有机结合，并成功地运用到天气预报中去。提出了一种新的，充分利用预报经验，数值预报产品及计算机计算能力的预报方法。     

专家系统与智能系统

为使专家系统技术尽早在地学领域取得突破，应该研制智能系统。在人机交换系统中加入智能模块，管理调度各功能模块，即实现人机交互系统到智能系统的升级。将专家系统、骨架系统加以扩充，增加与数据库管理系统、科学计算等功能子系统的接口，即成为一个研制智能系统的工具—智能骨架系统。文末描述了一个智能骨架系统—ＳＧＥＳ的结构和用法