design/tpl (Matches): Add light clarifying text
This is just some plain English to go along with and help rationalize the text. Further rationale will be provided in a dedicated section in the future; such information is vitally important to understand why the system evolved as it did.master
parent
972ea13623
commit
9b0e97d0b9
|
@ -391,7 +391,20 @@ For example,
|
|||
This is conceptually like a symbol table lookup in the compiler.}
|
||||
|
||||
|
||||
|
||||
\subsection{Matches}
|
||||
A classification consists of a set of binary predicates called
|
||||
\emph{matches}.
|
||||
Matches may reference any values,
|
||||
including the results of other classifications
|
||||
(as in \thmpref{class-compose}),
|
||||
allowing for the construction of complex abstractions over the data being
|
||||
classified.
|
||||
|
||||
Matches are intended to act intuitively across inputs of different ranks---%
|
||||
that is,
|
||||
one can match on any combination of matrix, vector, and scalar values.
|
||||
|
||||
\begin{axiom}[Match Input Translation]\axmlabel{match-input}
|
||||
Let $j$ and $k$ be free variables intended to be bound in the
|
||||
context of \axmref{class-pred}.
|
||||
|
@ -441,6 +454,14 @@ For example,
|
|||
\end{align*}
|
||||
\end{axiom}
|
||||
|
||||
|
||||
\index{package!core/match@\tamepkg{core/match}}
|
||||
Matches are represented by \xmlnode{match} nodes in \tame{}.
|
||||
Since the primitive is rather verbose,
|
||||
\tamepkg{core/match} also defines templates providing a more concise
|
||||
notation
|
||||
(\xmlnode{t:match-$\zeta$} below).
|
||||
|
||||
\index{classification!match@\xmlnode{match}}
|
||||
\begin{axiom}[Match Introduction]\axmlabel{match-intro}
|
||||
\begin{alignat*}{2}
|
||||
|
@ -593,6 +614,7 @@ We therefore establish a relationship to the notation of linear algebra
|
|||
|
||||
% This must be an axiom because it defines how the connectives operate; see
|
||||
% the remark.
|
||||
\index{classification!matrix notation}
|
||||
\begin{axiom}[Classification Matrix Notation]\axmlabel{class-mat-not}
|
||||
Let $\Gamma^2$ be defined by \axmref{class-yield}.
|
||||
Then,
|
||||
|
@ -614,9 +636,21 @@ We therefore establish a relationship to the notation of linear algebra
|
|||
combined.
|
||||
\end{remark}
|
||||
|
||||
\index{classification!intuition}
|
||||
\axmref{class-mat-not} makes it easy to visualize classification
|
||||
operations simply by drawing horizontal boxes across the predicates,
|
||||
as demonstrated by \spref{f:class-mat-boxes}.
|
||||
This visualization helps to show intuitively how the classification system
|
||||
is intended to function,
|
||||
with matrices serving as higher-resolution vectors.\footnote{%
|
||||
For example,
|
||||
with insurance,
|
||||
one may have a vector of data by risk location,
|
||||
and a matrix of chosen class codes by location.
|
||||
Consequently,
|
||||
we expect $M_j$ to be the set of class codes associated with
|
||||
location~$j$ so that it can be easily matched against location-level
|
||||
data~$v_j$.}
|
||||
|
||||
% NB: Give this formatting extra attention if the document's formatting is
|
||||
% substantially changed, since it's not exactly responsible with it's
|
||||
|
|
Loading…
Reference in New Issue