Help understand likelyhood function notation?

32 Views Asked by At

I'm reading An Introduction to Statistical Learning book and I've been struggling to understand this product notation. The usual form was $\prod_{i=1}^nx_i$.

What I don't understand is the subscript $i:y_{i}=1$. Is $i$ a variable or $y_{i}$? or both. What are they ranges? $$ l(B_0, B_1) = \prod_{i:y_{i}=1} p(x_i) \prod_{j:y_{j}=0} (1 - p(x_j)) $$

1

There are 1 best solutions below

0
On

I presume the intention was to make the product over the $i$-s (resp. $j$-s) such that $y_i=1$ (resp. $y_{j}=0$). Therefore, those might be mistypings of LaTeX for $$\prod_{\{i\,:\, y_i=1\}}p(x_i)\quad\text{and}\quad\prod_{\{j\,:\, y_j=0\}}(1-p(x_j))$$ where the parenthesis did not appear because the person who wrote it has typed { and } instead of \{ and \}. However, even leaving just "$i:y_i=1$" and "$j:y_j=0$" might be a conscious choice of notation to the same end.