Module qsarmodelingpy.calculate_parameters

Define algorithms for some statistics, such as SSY, PRESS, R², MAE, RMSE, R…

Functions

def calcMAE(yreal, ypred) ‑> float

Calculates the Mean Absolute Error.

$$MAE = \dfrac{\sum\limits_{i=1}^n \left\vert y_i - \hat y_i \right\vert}{n}$$

where $y$ is the predicted value, $\hat y$ is the true value and $n$ is the number of data points.

Args

yreal : DataFrame,list,array

The real data.

ypred : DataFrame,list,array

The predicted data.

Returns

float

MAE

def calcPress(yreal, ypred) ‑> float

Calculates predicted residual error sum of squares of ypred regarding yreal.

$$PRESS = \sum\limits_{i} (y_i - \hat y_{i})^2$$

Args

yreal : DataFrame,list,array

The real data.

ypred : DataFrame,list,array

The predicted data.

Returns

float

PRESS

def calcR(yreal, ypred) ‑> float

Calculates $R$.

Args

yreal : DataFrame,list,array

The real data.

ypred : DataFrame,list,array

The predicted data.

Returns

float

R

def calcR2(yreal, ypred, mean_y=None) ‑> float

Calculates the coefficient of determination ( $R^2$ ).

$$R^2 = 1-\dfrac{PRESS}{SSY}$$

Args

yreal : DataFrame,list,array

The real data.

ypred : DataFrame,list,array

The predicted data.

mean_y : float, optional

Mean of yreal. If None, it'll be calculated. Defaults to None.

Returns

float

$R^2$

See also:

• calcPress()
• ssy()

def calcRMSE(yreal, ypred) ‑> float

Calculates the Root Mean Square Error.

$$RMSE = \sqrt{\dfrac{PRESS}{n}}$$

Args

yreal : DataFrame,list,array

The real data.

ypred : DataFrame,list,array

The predicted data.

Returns

float

RMSE

See also:

• calcPress()

def ssy(y, mean_y=None) ‑> float

Calculates Total Sum of Squares of the dependent variable.

$$SSY = \sum\limits_{i} (y_i - \bar y)^2$$

Args

y : DataFrame, list, array

The vector to calculate SSY.

mean_y : float, optional

Mean of y. If None, it'll be calculated. Defaults to None.

Returns

float

SSY