結果:
What is a rough number? What can they be used for? Today I'll take you down a journey into the land of prime numbers (in MATLAB). But remember that a journey is not always about your destination, but about what you learn along the way. And so, while this will be all about primes, and specifically large primes, before we get there we need some background. That will start with rough numbers.
Rough numbers are what I would describe as wannabe primes. Almost primes, and even sometimes prime, but often not prime. They could've been prime, but may not quite make it to the top. (If you are thinking of Marlon Brando here, telling us he "could've been a contender", you are on the right track.)
Mathematically, we could call a number k-rough if it is evenly divisible by no prime smaller than k. (Some authors will use the term k-rough to denote a number where the smallest prime factor is GREATER than k. The difference here is a minor one, and inconsequential for my purposes.) And there are also smooth numbers, numerical antagonists to the rough ones, those numbers with only small prime factors. They are not relevant to the topic today, even though smooth numbers are terribly valuable tools in mathematics. Please forward my apologies to the smooth numbers.
Have you seen rough numbers in use before? Probably so, at least if you ever learned about the sieve of Eratosthenes for prime numbers, though probably the concept of roughness was never explicitly discussed at the time. The sieve is simple. Suppose you wanted a list of all primes less than 100? (Without using the primes function itself.)
% simple sieve of Eratosthenes
Nmax = 100;
N = true(1,Nmax); % A boolean vector which when done, will indicate primes
N(1) = false; % 1 is not a prime by definition
nextP = find(N,1,'first'); % the first prime is 2
while nextP <= sqrt(Nmax)
% flag multiples of nextP as not prime
N(nextP*nextP:nextP:end) = false;
% find the first element after nextP that remains true
nextP = nextP + find(N(nextP+1:end),1,'first');
end
primeList = find(N)
Indeed, that is the set of all 25 primes not exceeding 100. If you think about how the sieve worked, it first found 2 is prime. Then it discarded all integer multiples of 2. The first element after 2 that remains as true is 3. 3 is of course the second prime. At each pass through the loop, the true elements that remain correspond to numbers which are becoming more and more rough. By the time we have eliminated all multiples of 2, 3, 5, and finally 7, everything else that remains below 100 must be prime! The next prime on the list we would find is 11, but we have already removed all multiples of 11 that do not exceed 100, since 11^2=121. For example, 77 is 11*7, but we already removed it, because 77 is a multiple of 7.
Such a simple sieve to find primes is great for small primes. However is not remotely useful in terms of finding primes with many thousands or even millions of decimal digits. And that is where I want to go, eventually. So how might we use roughness in a useful way? You can think of roughness as a way to increase the relative density of primes. That is, all primes are rough numbers. In fact, they are maximally rough. But not all rough numbers are primes. We might think of roughness as a necessary, but not sufficient condition to be prime.
How many primes lie in the interval [1e6,2e6]?
numel(primes(2e6)) - numel(primes(1e6))
There are 70435 primes greater than 1e6, but less than 2e6. Given there are 1 million natural numbers in that set, roughly 7% of those numbers were prime. Next, how many 100-rough numbers lie in that same interval?
N = (1e6:2e6)';
roughInd = all(mod(N,primes(100)) > 0,2);
sum(roughInd)
That is, there are 120571 100-rough numbers in that interval, but all those 70435 primes form a subset of the 100-rough numbers. What does this tell us? Of the 1 million numbers in that interval, approximately 12% of them were 100-rough, but 58% of the rough set were prime.
The point being, if we can efficiently identify a number as being rough, then we can substantially increase the chance it is also prime. Roughness in this sense is a prime densifier. (Is that even a word? It is now.) If we can reduce the number of times we need to perform an explicit isprime test, that will gain greatly because a direct test for primality is often quite costly in CPU time, at least on really large numbers.
In my next post, I'll show some ways we can employ rough numbers to look for some large primes.
tiledlayout(4,1);
% Plot "L" (y = 1/(x+1), for x > -1)
x = linspace(-0.9, 2, 100); % Avoid x = -1 (undefined)
y =1 ./ (x+1) ;
nexttile;
plot(x, y, 'r', 'LineWidth', 2);
xlim([-10,10])
% Plot "O" (x^2 + y^2 = 9)
theta = linspace(0, 2*pi, 100);
x = 3 * cos(theta);
y = 3 * sin(theta);
nexttile;
plot(x, y, 'r', 'LineWidth', 2);
axis equal;
% Plot "V" (y = -2|x|)
x = linspace(-1, 1, 100);
y = 2 * abs(x);
nexttile;
plot(x, y, 'r', 'LineWidth', 2);
axis equal;
% Plot "E" (x = -3 |sin(y)|)
y = linspace(-pi, pi, 100);
x = -3 * abs(sin(y));
nexttile;
plot(x, y, 'r', 'LineWidth', 2);
axis equal;
I've been trying this problem a lot of time and i don't understand why my solution doesnt't work.
In 4 tests i get the error Assertion failed but when i run the code myself i get the diag and antidiag correctly.
function [diag_elements, antidg_elements] = your_fcn_name(x)
[m, n] = size(x);
% Inicializar los vectores de la diagonal y la anti-diagonal
diag_elements = zeros(1, min(m, n));
antidg_elements = zeros(1, min(m, n));
% Extraer los elementos de la diagonal
for i = 1:min(m, n)
diag_elements(i) = x(i, i);
end
% Extraer los elementos de la anti-diagonal
for i = 1:min(m, n)
antidg_elements(i) = x(m-i+1, i);
end
end
Check out the result of "emoji matrix" multiplication below.
- vector multiply vector:
a = ["😁","😁","😁"]
b = ["😂";
"😂"
"😂"]
c = a*b
d = b*a
- matrix multiply matrix:
matrix1 = [
"😀", "😃";
"😄", "😁"]
matrix2 = [
"😆", "😅";
"😂", "🤣"]
resutl = matrix1*matrix2
enjoy yourself!
Creating data visualizations
79%
Interpreting data visualizations
21%
28 票
For Valentine's day this year I tried to do something a little more than just the usual 'Here's some MATLAB code that draws a picture of a heart' and focus on how to share MATLAB code. TL;DR, here's my advice
- Put the code on GitHub. (Allows people to access and collaborate on your code)
- Set up 'Open in MATLAB Online' in your GitHub repo (Allows people to easily run it)
I used code by @Zhaoxu Liu / slandarer and others to demonstrate. I think that those two steps are the most impactful in that they get you from zero to one but If I were to offer some more advice for research code it would be
3. Connect the GitHub repo to File Exchange (Allows MATLAB users to easily find it in-product).
4. Get a Digitial Object Identifier (DOI) using something like Zenodo. (Allows people to more easily cite your code)
There is still a lot more you can do of course but if everyone did this for any MATLAB code relating to a research paper, we'd be in a better place I think.
Here's the article: On love and research software: Sharing code with your Valentine » The MATLAB Blog - MATLAB & Simulink
What do you think?
Have you ever wanted to search for a community member but didn't know where to start? Or perhaps you knew where to search but couldn't find enough information from the results? You're not alone. Many community users have shared this frustration with us. That's why the community team is excited to introduce the new ‘People’ page to address this need.
What Does the ‘People’ Page Offer?
- Comprehensive User Search: Search for users across different applications seamlessly.
- Detailed User Information: View a list of community members along with additional details such as their join date, rankings, and total contributions.
- Sorting Options: Use the ‘sort by’ filter located below the search bar to organize the list according to your preferences.
- Easy Navigation: Access the Answers, File Exchange, and Cody Leaderboard by clicking the ‘Leaderboards’ button in the upper right corner.
In summary, the ‘People’ page provides a gateway to search for individuals and gain deeper insights into the community.
How Can You Access It?
Navigate to the global menu, click on the ‘More’ link, and you’ll find the ‘People’ option.
Now you know where to go if you want to search for a user. We encourage you to give it a try and share your feedback with us.
Simulink has been an essential tool for modeling and simulating dynamic systems in MATLAB. With the continuous advancements in AI, automation, and real-time simulation, I’m curious about what the future holds for Simulink.
What improvements or new features do you think Simulink will have in the coming years? Will AI-driven modeling, cloud-based simulation, or improved hardware integration shape the next generation of Simulink?
I got thoroughly nerd-sniped by this xkcd, leading me to wonder if you can use MATLAB to figure out the dice roll for any given (rational) probability. Well, obviously you can. The question is how. Answer: lots of permutation calculations and convolutions.
In the original xkcd, the situation described by the player has a probability of 2/9. Looking up the plot, row 2 column 9, shows that you need 16 or greater on (from the legend) 1d4+3d6, just as claimed.
If you missed the bit about convolutions, this is a super-neat trick
[v,c] = dicedist([4 6 6 6]);
bar(v,c)
% Probability distribution of dice given by d
function [vals,counts] = dicedist(d)
% d is a vector of number of sides
n = numel(d); % number of dice
% Use convolution to count the number of ways to get each roll value
counts = 1;
for k = 1:n
counts = conv(counts,ones(1,d(k)));
end
% Possible values range from n to sum(d)
maxtot = sum(d);
vals = n:maxtot;
end
You've probably heard about the DeepSeek AI models by now. Did you know you can run them on your own machine (assuming its powerful enough) and interact with them on MATLAB?
In my latest blog post, I install and run one of the smaller models and start playing with it using MATLAB.
Larger models wouldn't be any different to use assuming you have a big enough machine...and for the largest models you'll need a HUGE machine!
Even tiny models, like the 1.5 billion parameter one I demonstrate in the blog post, can be used to demonstrate and teach things about LLM-based technologies.
Have a play. Let me know what you think.
My following code works running Matlab 2024b for all test cases. However, 3 of 7 tests fail (#1, #4, & #5) the QWERTY Shift Encoder problem. Any ideas what I am missing?
Thanks in advance.
keyboardMap1 = {'qwertyuiop[;'; 'asdfghjkl;'; 'zxcvbnm,'};
keyboardMap2 = {'QWERTYUIOP{'; 'ASDFGHJKL:'; 'ZXCVBNM<'};
if length(s) == 0
se = s;
end
for i = 1:length(s)
if double(s(i)) >= 65 && s(i) <= 90
row = 1;
col = 1;
while ~strcmp(s(i), keyboardMap2{row}(col))
if col < length(keyboardMap2{row})
col = col + 1;
else
row = row + 1;
col = 1;
end
end
se(i) = keyboardMap2{row}(col + 1);
elseif double(s(i)) >= 97 && s(i) <= 122
row = 1;
col = 1;
while ~strcmp(s(i), keyboardMap1{row}(col))
if col < length(keyboardMap1{row})
col = col + 1;
else
row = row + 1;
col = 1;
end
end
se(i) = keyboardMap1{row}(col + 1);
else
se(i) = s(i);
end
% if ~(s(i) = 65 && s(i) <= 90) && ~(s(i) >= 97 && s(i) <= 122)
% se(i) = s(i);
% end
end
私の場合、前の会社が音楽認識アプリの会社で、アルゴリズム開発でFFTが使われていたことがきっかけでした。でも、MATLABのすごさが分かったのは、機械学習のオンライン講座で、Andrew Ngが、線型代数を使うと、数式と非常に近い構文のコードで問題が処理できることを学んだ時でした。
Let's celebrate what made 2024 memorable! Together, we made big impacts, hosted exciting events, and built new apps.
Resource links:
Attaching the Photoshop file if you want to modify the caption.
Toolbox 全部入りの MATLAB ライセンス
67%
まだ持っていない Toolbox (下記にコメントください）
0%
MATLAB T シャツ
17%
MATLAB ルービックキューブ
0%
MATLAB 靴下
6%
MathWorks オフィス訪問チケット
11%
18 票
この場は MATLAB や Simulink を使っている皆さんが、気軽に質問や情報交換ができる場所として作られました。日本語でも気軽に投稿ができるように今回日本語チャネルを解説します。
ユーザーの皆様とのやり取りを通じて、みんなで知識や経験を共有し、一緒にスキルアップしていきましょう。 どうぞお気軽にご参加ください。
そして日本語チャネル開設にあたってコメントくださった皆様、ありがとうございます！
We’d like to announce a change on the Machine Translation feature on MATLAB Answers.
When users are visiting our international domains (e.g. China or Japan), Answers provides the option to translate the content. Recently, we identified several security threats involving high-volume requests from certain IP addresses targeting our translation service.
As one of the countermeasures, we have now placed the Machine Translation feature behind a login requirement. While non-logged-in users will still see the 'Translate' button, it will be inactive (greyed out) until they log in.
We are actively collaborating with adjacent teams to develop solutions to better detect and block malicious requests.
Please let us know if you have any questions or concerns.
What better way to add a little holiday magic than the L-shaped membrane atop your evergreen? My colleagues output the shape and then added some thickness and an interior cylinder in Blender. Then, the shape was exported to STL and 3D printed (in several pieces). Then glued, sanded, primed, sanded again and painted. If you like, the STL file is attached. Thank you to https://blogs.mathworks.com/community/2013/06/20/paul-prints-the-l-shaped-membrane/ and a tip of the hat to MATLAB Ornament. Happy Holidays!
If you have a folder with an enormous number of files and want to use the uigetfile function to select specific files, you may have noticed a significant delay in displaying the file list.
Thanks to the assistance from MathWorks support, an interesting behavior was observed.
For example, if a folder such as Z:\Folder1\Folder2\data contains approximately 2 million files, and you attempt to use uigetfile to access files with a specific extension (e.g., *.ext), the following behavior occurs:
Method 1: This takes minutes to show me the list of all files
[FileName, PathName] = uigetfile('Z:\Folder1\Folder2\data\*.ext', 'File selection');
Method 2: This takes less than a second to display all files.
[FileName, PathName] = uigetfile('*.ext', 'File selection','Z:\Folder1\Folder2\data');
Method 3: This method also takes minutes to display the file list. What is intertesting is that this method is the same as Method 2, except that a file seperator "\" is added at the end of the folder string.
[FileName, PathName] = uigetfile('*.ext', 'File selection','Z:\Folder1\Folder2\data\');
I was informed that the Mathworks development team has been informed of this strange behaviour.
I am using 2023a, but think this should be the same for newer versions.
This post is more of a "tips and tricks" guide than a question.
If you have a folder with an enormous number of files and want to use the uigetfile function to select specific files, you may have noticed a significant delay in displaying the file list.
Thanks to the assistance from MathWorks support, an interesting behavior was observed.
For example, if a folder such as Z:\Folder1\Folder2\data contains approximately 2 million files, and you attempt to use uigetfile to access files with a specific extension (e.g., *.ext), the following behavior occurs:
Method 1: This takes minutes to show me the list of all files
[FileName, PathName] = uigetfile('Z:\Folder1\Folder2\data\*.ext', 'File selection');
Method 2: This takes less than a second to display all files.
[FileName, PathName] = uigetfile('*.ext', 'File selection','Z:\Folder1\Folder2\data');
Method 3: This method also takes minutes to display the file list. What is intertesting is that this method is the same as Method 2, except that a file seperator "\" is added at the end of the folder string.
[FileName, PathName] = uigetfile('*.ext', 'File selection','Z:\Folder1\Folder2\data\');
I was informed that the Mathworks development team has been informed of this strange behaviour.
I am using 2023a, but think this should be the same for newer versions.
