Search - 1 to 4 u

Search - 1 to 4 u - List

[AI-NN-PR] Substituter.java DL : 0: 代入法的启发示搜索我的代码实现是：按照自然语言各字母出现频率的大小从高到低（已经有人作国统计分析了）先生成一张字母出现频率统计表（A）--------(e),(t,a,o,i,n,s,h,r),(d,l),(c,u,m,w,f,g,y,p,b),(v,k,j,x,q,z) ，再对密文字母计算频率，并按频率从高到低生成一张输入密文字母的统计表（B），通过两张表的对应关系，不断用A中的字母去替换B中的字母，搜索不成功时就回退，在这里回朔是一个关键。 -generation into a search of inspiration I said a code is : According to the Natural Language alphabet frequency of the size of precedence (has been for the State Statistical Analysis), Mr. into an alphabet frequency tables (A )--------( e), (t, a, o, i, n, s, h r), (d, l), (c, u, m, w, f, g, y, p, b), (v, k, j, x, q, z), again close to the mother language calculated frequency and frequency input precedence generate a secret letter to the mother TAB (B), Table 2 by the corresponding relations, use of the letters A to B to replace the letters of the alphabet, when unsuccessful search regression, Here is a retrospective key.
Date : 2026-01-09 Size : 4kb User : rtshen
[AI-NN-PR] BP DL : 1: 基于BP神经网络的参数自学习控制（1）确定BP网络的结构，即确定输入层节点数M和隐含层节点数Q，并给出各层加权系数的初值和，选定学习速率和惯性系数，此时k=1；（2）采样得到rin(k)和yout(k)，计算该时刻误差error(k)=rin(k)-yout(k)；（3）计算神经网络NN各层神经元的输入、输出，NN输出层的输出即为PID控制器的三个可调参数，，；（4）根据(3.34)计算PID控制器的输出u(k)；（5）进行神经网络学习，在线调整加权系数和，实现PID控制参数的自适应调整；（6）置k=k+1，返回（1）。 -Based on the parameters of BP neural network self-learning control (1) to determine the structure of BP network, that is, determine the input layer nodes M and hidden layer nodes Q, and gives all levels of the initial value and the weighted coefficient, the selected learning rate and inertia coefficient, when k = 1 (2) sample has been rin (k) and the yout (k), calculate the moment of error error (k) = rin (k)-yout (k) (3) calculation of neural network NN all floors of the neurons in input and output, NN output layer is the output of PID controller for the three adjustable parameters,, (4) According to (3.34) Calculation of PID controller output u (k) (5) to carry out neural network learning, on-line adjustment of the weighted coefficient and, realize the adaptive PID control parameters adjust (6) purchase k = k+ 1, return (1).
Date : 2026-01-09 Size : 1kb User : dake
[AI-NN-PR] hunhechengxu DL : 0: 基于MATLAB完成的神经网络源程序大家-Based on the MATLAB neural network to complete the U.S. source
Date : 2026-01-09 Size : 1.06mb User : liyuanxia
[AI-NN-PR] SVM DL : 0: 多篇关于支持向量机参考文献，大家可以参考-Many articles about support vector machine references can refer to U.S.
Date : 2026-01-09 Size : 3.94mb User : 晓明
[AI-NN-PR] Matlabeg DL : 0: 基于 Ma t l a b语言的遗传算法工具箱支持二进制和浮点数编码方式，并且提供了多种选择、交叉、变异的方法。通过具体实例对 Ma t l a b的遗传算法工具箱的用法进行了说明介绍.-The Ge ne t i c Al g or it h m To o l b ox ba s e d on Ma t l a b s u ppo ~s t h e b i na r y a nd f lo a t ， a n d t he r e a r e t he e x c el l e nl o pe r at o r s o f s el e c t i on ，c r os s o v e r a nd mut a t i on i n t he To o l bo x ，t wo e xa mpl e s a b o ut ho w t o us e t h e To o l bo x a r e i n t r o du c ec i n t h i s pa p er。
Date : 2026-01-09 Size : 98kb User : 阿铁
[AI-NN-PR] RLS DL : 0: 仿真对象如下：其中， v( k )为服从N (0,1) 分布的白噪声。输入信号u ( k) 采用M 序列，幅度为 1。M 序列由 9 级移位寄存器产生，x(i)=x(i-4)⊕x(i-9)。选择如下辨识模型：加权阵取Λ = I。衰减因子β = 0.98，数据长度 L = 402。辨识结果与理论值比较，基本相同。辨识结果可信 -he simulation object is as follows: among them, v (k) to obey N (0, 1) distribution of white noise. The input signal u (k) using M sequence, amplitude is 1. M sequence by 9 level shift register generation, x (I) = x (I- 4) ⊕ x (I- 9). Choose the identification model: weighted array take Λ = I. Attenuation factor β = 0.98, the data length L = 402. Identify the theoretical calculation and comparison, basically the same. Identification results are reliable
Date : 2026-01-09 Size : 1kb User : 张鹏
[AI-NN-PR] kmeans1 DL : 0: 　　K-means算法，算法步骤如下： Step1.利用式(2)计算距离矩阵D=()，其中=dist[i, j] ()； Step2.扫描坐标距离矩阵D，寻找距离的最大值和最小值，用式(3)计算limit； Step3.扫描坐标距离矩阵D，寻找矩阵中距离最小的2个数据a，b，将数据a，b加入集合，={a,b}，同时将数据a，b从U中删除，更新距离矩阵D； Step4.利用 (4)式在U中寻找距离集合最近的数据样本t，如果小于limit，则将t加入集合，同时将t从集合U中删除，更新距离矩阵D，重复Step5，否则停止； Step5.若i<k，i=i+1，重复步骤Step3、Step4，直至k个集合完成； Step6.取集合中数据的算术平均值记作数据中心，并计算得到的坐标值，完成k个数据中心的选取。-Algorithm steps are as follows: Step1. Type (2) is used to calculate the distance matrix D = (), including = dist [I, j] () Step2. Scan coordinate distance matrix D, looking for the maximum and the minimum distance, use type (3) calculate the limit Step3. Scan coordinate distance matrix D, looking for matrix minimum distance of two data a, b, and the data to a, b to join the collection, = {a, b}, at the same time the data a, b is removed from the U, update the distance matrix D Step4. Using (4) in the U find closest to the collection of data samples t, if less than the limit, then t join collection, at the same time t is removed from the set U, update the distance matrix D, repeat Step5, otherwise stop Step5. If I < k, I = I+ 1, repeat steps Step3, Step4, until k collection is complete Step6. Take the arithmetic mean of the collection of data for the data center, and to calculate the coordinates, to complete the selection of k data center. The above steps distribution cu
Date : 2026-01-09 Size : 125kb User : ming
[AI-NN-PR] PARTICLE-FILTER-ISSUES DL : 0: 针对基于贝叶斯原理的序贯蒙特卡罗粒子滤波器出现退化现象的原因, 以无敏粒子滤波(U PF)、辅助粒子滤波 (A S IR) 及采样重要再采样(S IR) 等改进的粒子滤波算法为例, 对消除该缺陷的关键技术(优化重要密度函数及再采样) 进行了分析研究。说明通过提高重要密度函数的似然度、引进当前测量值、预增和复制大权值粒子等方式, 可以有效改善算法性能。最后通过对一无源探测定位问题进行仿真, 验证了运用该关键技术后, 算法的收敛精度和鲁棒性得到进一步增强。-　Abstract:W e analyze the degeneracy phenomenon of sequen t ialMon te Carlo part icle f ilters based on bayesian theo rem , pu t focu s on the key techn iques ( good cho ice of impo rtance den sity and u se of resamp ling ) to reduce it s effect s. Several imp roving schemes such as the U n scen ted Part icle F ilters (U PF) , the A ux iliary Samp ling Impo rtance Resamp ling (A S IR ) and the Samp ling Impo rtance Resamp ling (S IR ) algo rithm s are in t roduced to illu st rate th rough increasing the likelihood of the impo rtance den sity o r inco rpo rat ing new measu remen t, o r rep licat ing part icles w ith large w eigh t s w ith in the generic f rame of part icle f ilters, the convergence accu racy and robu stness behavio rs of the algo rithm can be effect ively imp roved. A typ ical passive detect ion and locat ion p rob lem is simu lated to p rove above conclu sion s.
Date : 2026-01-09 Size : 291kb User : Haiser
[AI-NN-PR] cod8 DL : 0: this code help u to buld a better simulation option for speed control of conventional motro
Date : 2026-01-09 Size : 276kb User : ram kapoor