Posterior probability math to code

2 ビュー (過去 30 日間)
Zakaria Djebbara
Zakaria Djebbara 2021 年 7 月 21 日
コメント済み: Zakaria Djebbara 2021 年 7 月 21 日
I'm unsure whether I "translated" the math correctly into the code. I can't seem to make sense of the summation sign here. I know the values of x, but does the math say "sum of x from index 1 to whatever t I'm at?
sum(x(1:t))
Here's what I've come up with instead. Obviously the summation is missing.
x = [-0.46, 0.83, -3.26, -0.14, -0.68, -2.31, 0.57, 1.34, 4.58, 3.77];
T = length(x); % 10 numbers
sigma = 1;
posterior = [];
for t = 1:T
posterior = [posterior; exp((2/sigma^2) .* x(t))]; % This part
end
plot(posterior);
Could someone please have a look and direct me to some existing forum, but please not to https://se.mathworks.com/help/matlab/ref/sum.html because this doesn't help me translating the math to code.
Thanks a lot

採用された回答

Yazan
Yazan 2021 年 7 月 21 日
編集済み: Yazan 2021 年 7 月 21 日
First of all, you have to pay attention to the fact that there is no equality in the relation you presented, but rather a proportionality. Meaning that the posterior is given by this relation up to a nonzero constant. Now, imagine that you have a random variable , the posterior at t is given by summing the values from to , where , multiplying the result by a constant then taking e to the power of the result. Obviously, some constraints should be imposed on the definition of X such that p becomes a proper probability function.
Assuming that you have defined X, T, and sigma properly in your code, you can use this to compute the posterior
p = arrayfun(@(t) exp(sum(X(t:T))*2./sigma^2), t);
  1 件のコメント
Zakaria Djebbara
Zakaria Djebbara 2021 年 7 月 21 日
Thanks for a proper answer. The key thing I didn't understand, which I do now given your explanation, was that the posterior is given by the summing of vector x from x1 to xT. So my code simply needed this change:
x = [-0.46, 0.83, -3.26, -0.14, -0.68, -2.31, 0.57, 1.34, 4.58, 3.77];
T = length(x); % 10 numbers
sigma = 1;
posterior = [];
for t = 1:T
posterior = [posterior; exp((2/sigma^2) .* sum(x(t:T)))]; % This part
end
plot(posterior);
It works beautifully now. Thanks again, Yazan.

サインインしてコメントする。

その他の回答 (0 件)

カテゴリ

Help Center および File ExchangeMathematics についてさらに検索

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by