You can overfit anything with just one parameter

  • Context: Graduate 
  • Thread starter Thread starter BWV
  • Start date Start date
  • Tags Tags
    Parameter
Click For Summary
SUMMARY

The discussion centers on the paper by Piantadosi (2018), which demonstrates that a closed-form solution to the logistic map, defined as m(z) = 4z(1−z), can be manipulated using a finely tuned parameter θ to fit virtually any two-dimensional shape. The specific formulation mk(θ) = sin²[2k arcsin √θ] allows for extensive precision, extending to hundreds of decimal places. This technique parallels Tupper's Self-Referential Formula, showcasing the potential for overfitting in mathematical modeling.

PREREQUISITES
  • Understanding of logistic maps and their mathematical properties
  • Familiarity with closed-form solutions in mathematical modeling
  • Knowledge of Tupper's Self-Referential Formula and its implications
  • Basic proficiency in trigonometric functions and their applications
NEXT STEPS
  • Research the implications of overfitting in mathematical models
  • Explore advanced techniques in parameter tuning for logistic maps
  • Study the applications of Tupper's Self-Referential Formula in computational mathematics
  • Investigate the use of closed-form solutions in other mathematical contexts
USEFUL FOR

Mathematicians, data scientists, and researchers interested in mathematical modeling, particularly those exploring the concepts of overfitting and parameter optimization.

BWV
Messages
1,669
Reaction score
2,017
Interesting paper here:

https://colala.bcs.rochester.edu/papers/piantadosi2018one.pdf

which using a closed-form solution to the logistic map m(z) = 4z(1−z) of
mk (θ) = sin2 [ 2k arcsin √ θ]

with a finely tuned parameter θ (extending to hundreds of decimal places) can fit just about any 2 dimensional shape such as:

Elephant.png
 

Attachments

  • Elephant.png
    Elephant.png
    38.7 KB · Views: 660
  • Like
Likes   Reactions: Stephen Tashi, scottdave, BvU and 1 other person
Mathematics news on Phys.org

Similar threads

  • · Replies 4 ·
Replies
4
Views
3K
  • · Replies 62 ·
3
Replies
62
Views
12K
  • · Replies 21 ·
Replies
21
Views
7K
  • · Replies 9 ·
Replies
9
Views
2K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 6 ·
Replies
6
Views
5K
  • · Replies 2 ·
Replies
2
Views
8K
  • · Replies 1 ·
Replies
1
Views
4K
  • · Replies 1 ·
Replies
1
Views
4K
  • · Replies 2 ·
Replies
2
Views
3K