当前位置:网站首页>Check the inverse relationship between the shift distance and the number of iterations
Check the inverse relationship between the shift distance and the number of iterations
2022-08-06 07:06:00 【black elm】
(A,B)---m*n*k---(1,0)(0,1)
移位距离和假设

用神经网络分类A和B,把参与分类的A和B中的数字看作是组成A和B的粒子,分类的过程就是让A和B中的粒子互相交换位置,寻找最短移位路径的过程.而熵HWith the shift of shortest distance andS成正比,迭代次数n与熵H成反比.
移位规则汇总
每个粒子移位一次,位置重合不移位,A single shift distance if the image in order to1为底等于1-元素数值若以0为底则为元素本身.

分类8116,A中只有一个0.8,B中有3个不为0的值.The number of iterations to estimate this group of pictures.
按照移位距离和假设,This group of pictures shift distance of4-0.8-0.1-0.1-0.6=2.4

8123The shift distance as2.6,The number of iterations the last experiment has been
8123 | |
δ | 迭代次数n |
5.00E-04 | 28525.63 |
4.00E-04 | 34343.1 |
3.00E-04 | 44964.48 |
2.00E-04 | 65561.12 |
1.00E-04 | 124472.5 |
s | 2.6 |
According to the shift distance and inverse relationship with the number of iterations2.4<2.6,Is the relationship between the number of iterations should be8116>8123
8116Number of iterations of the measured values for
8116 | |
δ | 迭代次数n |
5.00E-04 | 28917.729 |
4.00E-04 | 35497.623 |
3.00E-04 | 46176.704 |
2.00E-04 | 67179.523 |
1.00E-04 | 126701.39 |
s | 2.4 |
Analyzing the relationship between
8116 | 8123 | 8116-8123 | |
δ | 迭代次数n | 迭代次数n | |
5.00E-04 | 28917.729 | 28525.63 | 392.09864 |
4.00E-04 | 35497.623 | 34343.1 | 1154.5231 |
3.00E-04 | 46176.704 | 44964.48 | 1212.2235 |
2.00E-04 | 67179.523 | 65561.12 | 1618.4026 |
1.00E-04 | 126701.39 | 124472.5 | 2228.892 |
s | 2.4 | 2.6 |
8116>8123,The results conform to the assumption.
771 | 122 | |
δ | 迭代次数n | 迭代次数n |
5.00E-04 | 28114.19 | 25862.05 |
4.00E-04 | 34375.59 | 31524.1 |
3.00E-04 | 44667.95 | 41011.36 |
2.00E-04 | 64534.3 | 59270.11 |
1.00E-04 | 123420.9 | 112397.9 |
S | 2 | 3 |
The experiment has beens=2和3的迭代次数,粗略估算s=2.4的迭代次数为119011,实测值为126701,二者相差6.5%,Small forecasts.
According to the hypothesis as long ass=2.4The number of iterations should be all the same,继续实验

分类8116,8125,8134,8224,8233,这5组的s都等于2.4.The number of iterations for
8134 | 8116 | 8125 | 8224 | 8233 | |
δ | 迭代次数n | 迭代次数n | 迭代次数n | 迭代次数n | 迭代次数n |
5.00E-04 | 29615.58 | 28917.73 | 29552.14 | 29905.67 | 30245.28 |
4.00E-04 | 36047.14 | 35497.62 | 36133.77 | 36531.3 | 36726.87 |
3.00E-04 | 46776.77 | 46176.7 | 46885.2 | 46887.09 | 47950.08 |
2.00E-04 | 68050.53 | 67179.52 | 67496.68 | 68083.92 | 68372.68 |
1.00E-04 | 128497.9 | 126701.4 | 129570.6 | 130194.2 | 129414.1 |
s | 2.4 | 2.4 | 2.4 | 2.4 | 2.4 |
Shift distance and assumptions to estimate the image binarization of iterations will bring the error,But this group of data is very close to each other,Only between the maximum and the minimum2.7%,This conforms to the assumption.用130194去和119011More than forecast9.4%
Compared with the experimental data
981 | 871 | 971 | 881 | 861 | 771 | 122 | 961 | 875 | |
δ | 迭代次数n | 迭代次数n | 迭代次数n | 迭代次数n | 迭代次数n | 迭代次数n | 迭代次数n | 迭代次数n | 迭代次数n |
5.00E-04 | 34219.01 | 34553.02 | 28229.22 | 27843.89 | 28270.21 | 28114.19 | 25862.05 | 25523.99 | 25117.39 |
4.00E-04 | 41899.68 | 41568.85 | 34548.15 | 34511.79 | 34803.25 | 34375.59 | 31524.1 | 30958.15 | 31102.23 |
3.00E-04 | 53474.56 | 54287.27 | 44497.27 | 44407.41 | 45065.05 | 44667.95 | 41011.36 | 40262.78 | 40239.05 |
2.00E-04 | 77797.83 | 78173.77 | 64693.36 | 64832.42 | 64707.93 | 64534.3 | 59270.11 | 59096.97 | 58618.7 |
1.00E-04 | 148175 | 146473.4 | 123601.3 | 122874.1 | 123288 | 123420.9 | 112397.9 | 113446.1 | 111994.6 |
s | 1 | 1 | 2 | 2 | 2 | 2 | 3 | 3 | 3 |
851 | 951 | 866 | 777 | 941 | 854 | 931 | 921 | 911 | |
δ | 迭代次数n | 迭代次数n | 迭代次数n | 迭代次数n | 迭代次数n | 迭代次数n | 迭代次数n | 迭代次数n | 迭代次数n |
5.00E-04 | 25288.76 | 23904.73 | 23868.01 | 23579.5 | 22959.57 | 22738.92 | 22166.65 | 22094.36 | 23575.86 |
4.00E-04 | 31347.55 | 28978.03 | 28820.87 | 28965.92 | 27774.2 | 27653.1 | 27290.48 | 27253.6 | 28991.81 |
3.00E-04 | 40549.93 | 38252.24 | 37659.14 | 37460.82 | 35993.35 | 36072.61 | 35639.73 | 35340.25 | 37399.98 |
2.00E-04 | 59255.58 | 55426.48 | 54661.18 | 54804.67 | 52950.71 | 53116.99 | 51890.25 | 52155.43 | 54335.16 |
1.00E-04 | 114094.8 | 106880.8 | 106599.8 | 106112.6 | 101146.2 | 100986.2 | 100158 | 98502.28 | 102787.2 |
s | 3 | 4 | 4 | 4 | 5 | 5 | 6 | 7 | 8 |
8116 | 8123 | ||||||||
δ | 迭代次数n | 迭代次数n | |||||||
5.00E-04 | 28917.729 | 28525.63 | |||||||
4.00E-04 | 35497.623 | 34343.1 | |||||||
3.00E-04 | 46176.704 | 44964.48 | |||||||
2.00E-04 | 67179.523 | 65561.12 | |||||||
1.00E-04 | 126701.39 | 124472.5 | |||||||
s | 2.4 | 2.6 | |||||||
误差 | 6.5-9.4% | 5.9-6.7% |
边栏推荐
- 拦截器函数
- 【Harmony OS】【ARK UI】组件内转场api 基本使用
- 中英文互译在线翻译
- Unity 引导系统.点击目标物体后提示文字变色进入下一步
- ZLMediakit Exclusive Features Introduction
- 更换oracle nls,Oracle更改客户端语言字符集设置NLS_LANG的方法
- 推荐系统-排序层-2017:DCN(Deep&Cross)【对Wide&Deep模型的改进】【Wide部分需人工特征工程;而Cross部分可以进行特征的自动交叉, 避免了基于业务理解的人工特征组合】
- Qt 异常处理的Qt版及两处易坑的地方
- SAP-ABAP-RFC类型接口创建步骤
- pacman包 管理各种R包
猜你喜欢
随机推荐
高并发、多线程、分布式都不懂,你拿什么跳槽阿里、腾讯、京东?还不好好学习啊
推荐系统-排序层-2010:因子分解机(FM)【LR(逻辑回归)模型的改进版】【解决在稀疏数据的场景下模型参数难以训练的问题。考虑了特征的二阶交叉,弥补了LR表达能力差的缺陷】
明日立秋 autumn begins,天气渐凉
推荐系统-排序层-2018:xDeepFM模型【xDeepFM是对DCN(Deep&Cross)的改进,不是对DeepFM的改进】【中科大】
vscode配置typescript及简单使用
【Harmony OS】【ARK UI】组件内转场api 基本使用
C 基础语法2 —— 数组
动手学深度学习_Batch Normalization
C语言指针(大白话+图文聊指针)
reportROC 一行代码输出ROC曲线的各项统计数值及ROC曲线
echart折线/柱状虚线设置
No URLs will be polled as dynamic configuration sources warning handling
Cesium关于Entity中的parent、isShowing、entityCollection和监听事件的探讨
技能大赛训练:交换机端口带宽限速配置
不敲一个代码,10分钟做出数据可视化大屏,这个教程太赞了
ZLMediakit Exclusive Features Introduction
LeetCode50天刷题计划(Day 13—— 四树之和(8.50-10.00)
CrossOver22试用期到了如何免费使用?
卡片层叠Banner
推荐系统-排序层-2017:AFM模型【FM家族】【NFM的改进版】









