[Sequence Alignment Methods] Cross-Recurrent Plot (CRP)

A recurrence plot (RP) is a straightforward way to visualize characteristics of similar system states attained at different times (Eckmann et al., 1987). 

,即RP可识别在时间上伸缩的状态对。Cross-Recurrent Plot,形成一个二维坐标轴,坐标系里的黑色代表相似的状态对,实际上在这个坐标系里,大部分的点都不是黑色的。在对角线上的路径代表了两个序列的相似子序列。

明显,上图(a)比(b)序列对更有相似性。然而这只是人眼的判别,虽然事实也是如此,但肯定还是需要定量的判别的吧。比如做一个图片二分类的分类器,判别两副图哪个更匹配,但问题在于这样一个稀疏的黑白两色图,提取哪个特征适合用来做分类呢?基于这个考虑(事实上是基于大量的实验表明该分类目的的特征不好提取),研究者们还是自己定了一些计算方法,如DP:

将路径斜率限定为1太死了,但是如果改成0到无穷大又太大了,所以采取折衷的策略:

折衷策略的路径斜率为[1/2, 2],跟DTW的策略很像有没有,其实类似应用到动态规划的问题都有这个问题出现和类似这个的解决方案。总之,现在路径的大方向把握好了,接下来解决细节问题。这个问题就是如果两个连续路径中间由于极个别的点没匹配上,却错过了形成optimal路径的机会,这太可惜了。所以为了解决这个问题,引入了衰减变量,使得如果有个别点没匹配上,不会让路径直接被斩断,而是给路径打的分会降低:

衰减力度不能太小,至少要比1大吧。OK,这样就很合适了

时间: 2024-10-05 21:50:06

[Sequence Alignment Methods] Cross-Recurrent Plot (CRP)的相关文章

[Sequence Alignment Methods] Dynamic time warping (DTW)

本系列介绍几种序列对齐方法,包括Dynamic time warping (DTW),Smith–Waterman algorithm,Cross-recurrence plot Dynamic time warping (DTW) is a well-known technique to find an optimal alignment between two given (time-dependent) sequences under certain restrictions. ——Mei

[Sequence Alignment Methods] Smith–Waterman algorithm

Smith–Waterman algorithm 首先需要澄清一个事实,Smith–Waterman algorithm是求两个序列的最佳subsequence匹配,与之对应的算法但是求两个序列整体匹配的算法是Needleman-Wusch algorithm,即 Smith–Waterman algorithm:Local Needleman-Wusch algorithm: Global Needleman-Wusch algorithm与longest common subsequence

[转]Sequence of methods in form and table in AX

转自:http://axvuongbao.blogspot.jp/2013/09/sequence-of-methods-in-form-and-table.html Form: Sequence of Methods calls while opening the Form Form — init () Form — Datasource — init () Form — run () Form — Datasource — execute Query () Form — Datasource

AX 2012 Sequence of methods in form and table in AX

Form: Sequence of Methods calls while opening the Form Form - init () Form - Datasource - init () Form - run () Form - Datasource - execute Query () Form - Datasource - active () Sequence of Methods calls while closing the Form Form - canClose () For

2018.3.5-6 knapsack problem, sequence alignment and optimal binary search trees

这周继续dynamic programming,这三个算法都是dynamic programming的. knapsack problem有一种greedy的解法,虽然简单但是不保证正确,这里光头哥讲的是dynamic的解法.其实和上次那个max weight independent set的算法差不多,同样是每个物件都判断一遍有这个物件和没这个物件两种情况,用bottom-up的方法来解,然后得到一个最大的value值,这时因为没有得到具体的选择方案,所以最后还需要一部重构的步骤得到具体方案.

吴恩达《深度学习》-课后测验-第五门课 序列模型(Sequence Models)-Week 1: Recurrent Neural Networks(第一周测验:循环神经网络)

Week 1 Quiz: Recurrent Neural Networks(第一周测验:循环神经网络) \1. Suppose your training examples are sentences (sequences of words). Which of the following refers to the jth word in the ith training example?( 假设你的训练样本是句子(单词序列),下 面哪个选项指的是第??个训练样本中的第??个词?) [ ]

[Stanford Algorithms: Design and Analysis, Part 2] c28 Sequence Alignment Optimal Substructure

原文地址:https://www.cnblogs.com/ecoflex/p/10658398.html

pandas.DataFrame.plot

pandas.DataFrame.plot¶ DataFrame.plot(x=None, y=None, kind='line', ax=None, subplots=False, sharex=None, sharey=False, layout=None, figsize=None, use_index=True, title=None, grid=None, legend=True, style=None, logx=False, logy=False, loglog=False, xt

{ICIP2014}{收录论文列表}

This article come from HEREARS-L1: Learning Tuesday 10:30–12:30; Oral Session; Room: Leonard de Vinci 10:30  ARS-L1.1—GROUP STRUCTURED DIRTY DICTIONARY LEARNING FOR CLASSIFICATION Yuanming Suo, Minh Dao, Trac Tran, Johns Hopkins University, USA; Hojj