快速自顶向下优化神经网络结构的方法

引荐了一种优化神经网络结构的改进方法。近年来，各种神经网络结构优化方法相继被提出，但素负盛名的还是神经中枢手术优化算法(OBS)。提出一种与修剪技术相关的快速方法，这种方法起始于适当复杂的初始网络，然后剔除冗余的神经元。研究说明，这种快速方法与以往的常规神经中枢手术优化算法相比，具有更好的归纳性能、更简单的网络结构和更快的学习速度。     

An algorithm to find agreement subtrees

Given two binary trees, a largest subtree contained in both of the original trees that has been obtained by pruning vertices is called an agreement subtree. An exact algorithm for finding an agreement subtree is presented.Research of F.R.M. supported by grant number N00014-89-J-1643 from the Office of Naval Research. The authors would like to thank the referees, the Editor, and William H. E. Day for many valuable suggestions.     

“米良1号”生长结果习性与夏季修剪技术

“米良1号”猕猴桃新梢生长强旺，容易形成长枝和徒长枝；一年生枝发梢部位高，结果枝组及村体重心外移。夏季修剪的主要任务是控制新消长势和培养更新枝组。文章最后阐述了具体的夏季修剪技术。     

湿地松林抚育间伐与修枝技术

对立地、密度、林龄基本一致的湿地松林分采取不同强度的抚育间伐和修枝试验，确定了湿地松林第1次间伐和修枝年龄为7～8a；林分标准郁闭度达0．85以上时，胸径连年生长量已下降，株数间伐强度以不超过30%为宜。第2次间伐应根据培育目的和经营管理水平确定强度，一般林分间伐强度不超过35％；间伐间隔期为5～6a。以培育大径级建筑用材林的林分，应进行第3次间伐，每公顷保留600～680株为宜。间伐后应进行修枝，修枝强度第1次以树高的1／3为宜，第2次以树高的1／2为宜，间隔期为3～5a，时间以深秋至早春为好。     

决策树挖掘技术在手术诊断上的应用研究

在解决分类问题的各种方法中,决策树是比较常用的一种方法。基于决策树理论,在 ID3算法基础上提出基于悲观错误剪枝的后剪枝算法,并将其运用于医疗系统手术诊断的数据挖掘分析过程中,所得实验结果与专家诊断结果基本吻合,取得了较好的实际应用效果。     

原空间中最小二乘支持向量机的新算法

为了解决原空间中最小二乘支持向量机的解缺乏稀疏性的缺点,提出了Pruning法、MFCV法和IMFCV法并对BDFS法进行了修改和运用。对一个不含有奇异点的系统而言,Pruning法、BDFS法和MFCV法在一定程度上都能实现原空间中最小二乘支持向量机解的稀疏性。BDFS法无论是训练时间还是预测时间都比Pruning法短;和MFCV法比起来,虽然BDFS法的训练时间短,但比MFCV的预测时间长。对一个含有奇异点的系统而言,Pruning法几乎失去了效用;虽然BDFS和MFCV法的训练时间都比IMFCV法的训练时间短,但IMFCV法能成功抑制奇异点从而缩短预测时间。     

修枝强度和季节对杨树人工林生长的影响

以江苏省泗阳农场８ｈｍ＾２的试验林为基础,评价了修枝强度和修枝季节对Ⅰ－６９杨人工林生长及萌枝发生数量的影响。该试验采用裂区设计,主处理为修枝强度,分为３种水平（１／３Ｈ、１／２Ｈ和ＣＫ）;副处理为修枝季节,分２个水平（生长季和非生长季）。研究结果表明,修枝处理后当年,不同修枝强度的林分叶面积指数存在显著差异,中度修枝（１／２Ｈ）林分的叶面积指数仅为对照处理的４９％,弱度修枝林分叶面积指数为对照的８４％;不同修枝季节对杨树人工林的生长影响不显著,但对萌枝发生数量影响显著,非生长季修枝林分的萌枝生数量比生长季修枝林分高出７９．８％;与对照相比,修枝处理４ａ后,其平均胸径和平均单株材积,都略有下降,但均未达差异显著水平;而不同修枝强度的林分枝下高存在显著差异,中度修枝林分枝下高达６ｍ,弱度修枝的林分为４．５ｍ左右,     

Application of Partially Connected Neural Network

This paper focuses mainly on application of Partially Connected Backpropagation Neural Network （PCBP） instead of typical Fully Connected Neural Network （FCBP）. The initial neural network is fully connected, after training with sample data using cross-entropy as error function, a clustering method is employed to cluster weights between inputs to hidden layer and from hidden to output layer, and connections that are relatively unnecessary are deleted, thus the initial network becomes a PCBP network. Then PCBP can be used in prediction or data mining by training PCBP with data that comes from database. At the end of this paper, several experiments are conducted to illustrate the effects of PCBP using Iris data set.     

A rough sets based pruning method for bagging ensemble

Ensemble techniques train a set of component classifiers and then combine their predictions to classify new patterns. Bagging is one of the most popular ensemble techniques for improving weak classifiers. However, it is hard to deploy in many real applications because of the large memory requirement and high computation cost to store and vote the predictions of component classifiers. Rough set theory is a formal mathematical tool to deal with incomplete or imprecise information, which has attracted a lot of attention from theory and application fields. In this paper, a novel rough sets based method is proposed to prune the classifiers obtained from bagging ensemble and select a subset of the component classifiers for aggregation. Experiment results show that the proposed method not only decreases the number of component classifiers but also obtains acceptable performance.