【摘要】 根据陕西解放以来的灾害损失资料，将各种灾害造成的成灾面积、人口死亡、房屋倒塌和水、旱、风雹、病虫害、霜冻以及其它灾害造成的农田成灾面积分成４和５个档次，用人工神经网络方法进行聚类，可分成５个不同的灾害损失组合年，其代表了不同损失特点，有助于对各年定量评价灾害年景。分析这些不同损失组合类型的降水特征，可通过对降水的预报来预测可能出现的灾害组合，从而为重点防御某种或某几种灾害损失指明方向。更多还原

【Abstract】 Several disaster constitution dividings of disaster losses such as disaster area, casualty and house collapse as well as farm land areas that suffer disasters including flood, drought, hail, pest and frost have been drawn from the loss data in Shaanxi province since foundation of China via artificial nerve network method.Its quantity dividing indices and the connection with rainfalls have been summed up.This sets a base for annual disaster situation evaluation and prediction quantitatively.更多还原