Help with math terminology in tutorial

  • Thread starter Thread starter hotvette
  • Start date Start date
Click For Summary

Discussion Overview

The discussion focuses on the clarity and potential ambiguity of mathematical terminology used in tutorials, particularly regarding the notation for functions and their parameters. Participants explore various notational conventions for representing function parameters and iterates in optimization contexts.

Discussion Character

  • Exploratory
  • Technical explanation
  • Conceptual clarification

Main Points Raised

  • One participant expresses concern about using ##\beta_k## to denote both a component of a vector and an iterate, suggesting that this dual meaning could lead to confusion.
  • Another participant proposes using vector notation, specifically ##\vec \beta## for iterations, while keeping ##\beta_k## for components, as a way to differentiate the meanings.
  • A different proposal introduces an n+1 dimensional vector notation, ##\mathbf{x}=(x_0,x_1,x_2,...,x_n)##, where components correspond to original variables, and suggests using upper suffixes to denote values at different iterations.
  • Participants discuss the implications of these notational choices on clarity and connection between the function and its parameters.

Areas of Agreement / Disagreement

There is no consensus on the best notation to use, as participants present multiple competing views and suggestions regarding how to clearly represent the concepts without ambiguity.

Contextual Notes

Limitations include the potential for confusion arising from the dual use of symbols and the need for clear differentiation between components and iterates, which remains unresolved.

hotvette
Homework Helper
Messages
1,001
Reaction score
11
I write math related tutorials and would appreciate comments/advice about terminology that may be potentially ambiguous or confusing. As an example, I use ##f(x;\beta)## to define a generic continuous function with independent variable ##x## and function parameters ##\beta = (\beta_1, \beta_2, \dots, \beta_n)## that are also considered variables (to be determined by an iterative optimization process). The partial derivative of ##f(x;\beta)## with respect to the ##k##th component of ##\beta##, or ##\beta_k## would be:
$$\frac{\partial f(x;\beta)}{\partial \beta_k} = .....$$
However, later on, ##\beta_k## is referred to as the ##k##th iterate of the values of the vector ##\beta##, for example: ##\beta_{k+1} = \beta_k + \Delta \beta##. Thus, ##\beta_k## has a different interpretation based on context. I've seen use of the following to denote iterates, which I don't like (seems too complicated):
$$\beta^{(k+1)} = \beta^{(k)} + \Delta \beta \quad\text{or}\quad \beta_{(k+1)} = \beta_{(k)} + \Delta \beta$$
I've also thought about using a different symbol for the vector, so that:
$$\phi = (\beta_1, \beta_2, \dots, \beta_n)
\qquad
\frac{\partial f(x;\phi)}{\partial \beta_k} = .....
\qquad
\phi_{k+1} = \phi_k + \Delta \phi$$
which seems confusing to me because there appears to be no connection between ##f## and ##\beta_k##. Lastly, I've thought about using a capital greek letter to refer to the vector, something like:
$$\Phi = (\phi_1, \phi_2, \dots, \phi_n)
\qquad
\frac{\partial f(x;\Phi)}{\partial \phi_k} = .....
\qquad
\Phi_{k+1} = \Phi_k + \Delta \Phi$$
I think the last choice is probably the most clear, but I don't like the bold symbols sprinkled around in the tutorial. What I really want to do is have two different meanings of ##\beta_k##, one as vector component and the other as a vector iterate. Is this bad form?
 
Science news on Phys.org
hotvette said:
the vector β
How about the vector notation, ##\vec \beta##?
Accordingly, ##\beta_k## for components and ##\vec \beta_k## for iterations.
 
Thanks for the suggestion!
 
A proposal for simple and clear notation.

Intorducing n+1 dimension vector
\mathbf{x}=(x_0,x_1,x_2,...,x_n)
components of which correspond to the original variables as
x=x_0, \beta_1=x_1,\beta_2=x_2,...,\beta_n=x_n
the function is expressed as
f(\mathbf{x})=f(x_0,x_1,x_2,...,x_n)

Values of variables are identified by upper suffix
\mathbf{x}^{(i)}=(x_0^{(i)},x_1^{(i)},x_2^{(i)},...,x_n^{(i)})
VAlues of partial derivatives are
\frac{\partial f}{\partial x_k}(x_0^{(i)},x_1^{(i)},x_2^{(i)},...,x_n^{(i)})=\frac{\partial f}{\partial x_k}(\mathbf{x}^{(i)})
 
Thanks, I'll chew on it.
 

Similar threads

  • · Replies 12 ·
Replies
12
Views
3K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 3 ·
Replies
3
Views
1K
  • · Replies 0 ·
Replies
0
Views
2K
  • · Replies 1 ·
Replies
1
Views
1K
Replies
5
Views
3K
  • · Replies 0 ·
Replies
0
Views
248
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 1 ·
Replies
1
Views
1K