共查询到20条相似文献,搜索用时 62 毫秒
1.
2.
3.
4.
5.
6.
介绍了一种进化式模糊分类系统.首先,介绍系统的基本特征及结构框架.然后,介绍了一种动态聚类算法,并运用动态聚类算法对输入的训练模式进行动态聚类,每一簇创建一条模糊规则.规则所对应的区域为类椭圆形区域.规则调整的策略是连续改变模糊分类规则的一个参数,使得分类系统对训练模式识别率不能再提高,对不能达到要求的调整,采用遗传算法进行调整.分析了规则调整的方法,给出了调整算法,也介绍了规则的插入和聚合策略.用两个典型的数据集来评测研究的系统,研究的分类系统在识别率与多层神经网络分类器相当,但训练时间远少于多层神经网络分类器的训练时间. 相似文献
7.
可进化的入侵检测系统的模糊分类器研究 总被引:2,自引:0,他引:2
周国良 《计算机工程与应用》2004,40(30):157-159
由于计算机网络中的正常行为和异常行为难以很好界定,所以许多入侵检测系统经常产生误报警。使用模糊逻辑推理方法,入侵检测系统的误报率则会明显降低,可以在入侵检测系统中,使用一套模糊规则和作用在该集合上的模糊推理算法,来判断是否发生了入侵事件。这种方法面临的主要问题是要有一个针对入侵检测的好的模糊算法。该文提出了一种使用遗传算法产生模糊分类器,以检测误用和入侵事件。主要思想是生成两个进化规则子集合,一个用于描述正常行为,一个用于描述异常行为。其中,正常行为规则进化信息来自正常使用时的操作行为,异常行为规则进化信息来自计算机网络受到入侵时的操作行为。 相似文献
8.
9.
10.
11.
本文介绍一个基于模糊神经网络的数据逼近和泛化建模方法,定义了一种模糊系统动态调节神经网络的学习率,给出了用迭代自组织数据分析算法确定神经网络结构、初始化神经网络参数的方法.在雷达天线罩视线误差建模中的应用表明,这种方法加快了网络的收敛速度,避免了局部极值,具有较高的数据逼近和泛化能力. 相似文献
12.
Adil Baykasoğlu 《控制论与系统》2013,44(6):475-489
There are many approaches in the literature to model and quantify manufacturing flexibility. Most of these models were developed to quantify several aspects of manufacturing flexibility like machine flexibility, routing flexibility, mix flexibility, volume flexibility, etc. This is mainly due to the fact that developing a generic model, which can be used to measure different types of flexibilities, is not straightforward. Recently, a generic flexibility measure, which is based on digraphs and permanent index, was proposed by the author. The main difficulty with that model like in all other flexibility models is the inability to collect precise data for computing the flexibility. In order to overcome this difficulty, a practical fuzzy linguistic approach is incorporated into the previous digraph model in this article. The extended fuzzy digraph model is explained in detail through an example in the present article. 相似文献
13.
新的模糊度量与模糊似然函数 总被引:3,自引:0,他引:3
闫德勤 《模式识别与人工智能》2001,14(1)
本文分析了Kosko等人提出的关于模糊子集的度量方法,提出其局限性,从而给出了一种新的模糊子集度量方法与模糊似然函数.由于这种方法能够更好地刻划模糊集合间的子集度与似然性,从而在模式识别、聚类分析、图像及信息处理中有着重要实际应用意义.同时文章也给出了两种新的模糊熵表示方法. 相似文献
14.
ABSTRACT In this work, we develop a neural model to solve causal reasoning problems (also called abduction) in the open, independent, and incompatibility classes. We model the reasoning process by a single and global energy function using cooperative and competitive neural computation. The update rules of the distinct connections of the network are derived from its energy function, using gradient descent techniques. Simulation results reveal a good performance of the model. 相似文献
15.
Yi-Chung Hu 《Applied Artificial Intelligence》2013,27(6):601-619
ABSTRACT A fuzzy if-then rule whose consequent part is a real number is referred to as a simplified fuzzy rule. Since no defuzzification is required for this rule type, it has been widely used in function approximation problems. Furthermore, data mining can be used to discover useful information by exploring and analyzing data. Therefore, this paper proposes a fuzzy data mining approach to discover simplified fuzzy if-then rules from numerical data in order to approximate an unknown mapping from input to output. Since several pre-specified parameters for deriving fuzzy rules are not easily specified, they are automatically determined by the genetic algorithm with binary chromosomes. To evaluate performance of the proposed method, computer simulations are performed on various numerical data sets, showing that the fitting ability and the generalization ability of the proposed method are comparable to the known fuzzy rule-based methods. 相似文献
16.
17.
In this paper, a new approach to designing fuzzy‐learning fuzzy controllers for a system plant without an exact mathematical model is presented. The cost function is defined as the square of the sliding function to alleviate the difficulty of overshoot when on‐line learning is conducted. The learning mechanism of a fuzzy controller is constructed so as to minimize the cost function with a set of linguistic rules. Moreover, to reduce the complexity of the fuzzy‐learning fuzzy controller, the fuzzy mechanism used for learning and the fuzzy mechanism contained in the fuzzy controller are designed so as to have the identical structures. Finally, simulations are included to show the effectiveness of the fuzzy‐learning fuzzy controllers. 相似文献
18.
提出了一种利用支持向量机(SVM)学习算法提取模糊规则进而实现施肥预测的方法.对于农业施肥中常用的正交实验法,由于其数据均衡分散使得曲线拟合的回归预测方法效果不佳.提出了一种利用SVM学习样本数据,再利用隶属度来提取模糊规则,通过阈值和可信度来控制规则的激活和准确性的预测方法,这一方法不仅避免了回归预测所产生的误差,并且模糊规则更具有实际意义,从而大大提高了知识获取的能力. 相似文献
19.
模糊控制推理ASIC 总被引:3,自引:0,他引:3
模糊逻辑控制(FLC)对于复杂的难定义的控制过程,具有简单而有效等许多优越性。近年来已越来越多也得到工业界的重视。本文提出一种模糊控制推理的ASIC方法,具有简单,灵活性和可扩充性特点,然后介绍一个基于FPGA的实验系统FCIS原型。 相似文献