Open matwey opened 2 months ago
Hi,
It would be great to have features extracted from parabolic (quadratic) fit similar to what we have from linear fit:
The model is $m_i = m_0 + g \left( t_i - t_0 \right)^2$
Equations:
$$ \begin{bmatrix} \sum_i t^4_i & \sum_i t^3_i & \sum_i t^2_i \ \sum_i t^3_i & \sum_i t^2_i & \sum_i t_i \ \sum_i t^2_i & \sum_i t_i & N \end{bmatrix} \begin{bmatrix} a \ b \ c \end{bmatrix} = \begin{bmatrix} \sum_i t^2_i m_i \ \sum_i t_i m_i \ \sum_i m_i \end{bmatrix} $$
$g = a$ $t_0 = -\frac{b}{2a}$ $m_0 = c - \frac{b^2}{4a}$
Hi,
It would be great to have features extracted from parabolic (quadratic) fit similar to what we have from linear fit:
The model is $m_i = m_0 + g \left( t_i - t_0 \right)^2$
Equations:
$$ \begin{bmatrix} \sum_i t^4_i & \sum_i t^3_i & \sum_i t^2_i \ \sum_i t^3_i & \sum_i t^2_i & \sum_i t_i \ \sum_i t^2_i & \sum_i t_i & N \end{bmatrix} \begin{bmatrix} a \ b \ c \end{bmatrix} = \begin{bmatrix} \sum_i t^2_i m_i \ \sum_i t_i m_i \ \sum_i m_i \end{bmatrix} $$
$g = a$ $t_0 = -\frac{b}{2a}$ $m_0 = c - \frac{b^2}{4a}$