Polynomial Regression: Difference between revisions
From Rice Wiki
m (Rice moved page Polynomial Regression Model to Polynomial Regression) |
No edit summary |
||
| (3 intermediate revisions by the same user not shown) | |||
| Line 1: | Line 1: | ||
'''Polynomial regression''' describes the relationship between ''x'' and ''y'' as an n<sup>th</sup> degree polynomial. | '''Polynomial regression''' is a model that describes the relationship between ''x'' and ''y'' as an n<sup>th</sup> degree polynomial. | ||
<math>f(x) = \omega_0 + \omega_1 x + \omega_2 x^2 + ... + \epsilon</math> | |||
Notice that polynomial regression is a ''form'' of [[Linear Regression]], except there are higher-degree attributes. | |||
To build a polynomial model involves determining the degree of polynomial and using normal linear regression methods. | |||
= Degree = | |||
A higher degree means a more complex model. While it increases the accuracy of the model, higher degree models more computationally expensive and an overly high degree will result in [[Curve Fitting#Overfitting|overfitting]]. [[RMS]] error can be used to determine what degree is sufficient; After a certain degree of polynomial, the RMS error on the test data would increase. | |||
[[Category:Machine Learning]] | |||
Latest revision as of 18:45, 15 April 2024
Polynomial regression is a model that describes the relationship between x and y as an nth degree polynomial.
Notice that polynomial regression is a form of Linear Regression, except there are higher-degree attributes.
To build a polynomial model involves determining the degree of polynomial and using normal linear regression methods.
Degree
A higher degree means a more complex model. While it increases the accuracy of the model, higher degree models more computationally expensive and an overly high degree will result in overfitting. RMS error can be used to determine what degree is sufficient; After a certain degree of polynomial, the RMS error on the test data would increase.
