Find Several values closest to zero

Question

Manuel Steinbrecher 2017-5-30

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/342493-find-several-values-closest-to-zero

评论： Anthony Snow 2020-4-15

Hey all,

I have a simple vector containing my data and wanna find the index of its value closest to zero. At the moment, I am just doing:

zeroIX=dsearchn(mydata,0);

However, this only gives me the very first value. If I have for example a vector like this:

mydata= [1;2;5;0.1;0.1;0.1;2;3]

I omit all the other two values, which are exactly as far away from 0. But in this case for example, I need the index of the middle one. I read through several ideas but haven't figured out a way to find similar values close to zero. Can anyone of you help me?

Thanks in advance and best, Manuel

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

KSSV 2017-5-30

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/342493-find-several-values-closest-to-zero#answer_268897

在 MATLAB Online 中打开

Define a small fraction..say eps of your required order...

eps = 10^-2 ;
idx = mydata<=eps ;

2 个评论
显示无隐藏无

Stephen23 2017-5-30

编辑：Stephen23 2017-5-30

@KSSV: do NOT use eps as a variable name: it has a specific meaning (the machine epsilon), so re-defining it is as confusing as re-defining pi. And because it is such an important value in computer science, it is already defined as the very important MATLAB function eps.

What you used in your example is a tolerance, but it is certainly not the machine epsilon.

Anthony Snow 2020-4-15

What would you input if you only want numbers in a matrix on the order of lets say 1e-2 (to keep it constant with your previous reply) set equal to zero but NOT the numbers that are negative ( ~mydata<0)

请先登录，再进行评论。

Answer 2

Manuel Steinbrecher 2017-5-30

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/342493-find-several-values-closest-to-zero#answer_268902

Well, I tend to forget how easy Matlab can be. Thanks for your reply.

Even though this doesn't fully do the job for me, since my problem is slightly different actually, this gave me the idea how to do it. So my actual problem is: I can have data, with several actual zeros in it (they are consecutive then!) or I can have data where I always slightly miss the zero. So I will now just do an if-check and check whether idx=mydata==0 gives any result, and if not, I just do idx = dsearchn(mydata,0).

If there is any smarter way, let me know :) Thx KSSV