genome.models.linear

A simple Linear regression model

class genome.models.linear.LinearRegression(input, n_vars, W=None, b=None)

Bases: genome.models.base.BaseModel

Initializes a Linear Regression model

This class creates a generic linear regression model: y = Wx + b.

Parameters:

• input – A symbolic Tensor input
• n_vars (int) – Number of input variables
• W – Slope parameters - A TensorSharedVariable (optional)
• b – Intercept parameter - A TensorSharedVariable (optional)

mean_squared_error(y)

see: mean_squared_error()