transformers

Media transformation functions for Marketing Mix Models.

Functions

batched_convolution(x, w, *, dim, kernel_dim)	Apply a 1D convolution in a vectorized way across multiple batch dimensions.
binomial_adstock(x[, alpha, l_max, ...])	Binomial adstock transformation.
delayed_adstock(x[, alpha, theta, l_max, ...])	Delayed adstock transformation.
geometric_adstock(x[, alpha, l_max, ...])	Geometric adstock transformation.
hill_function(x, slope, kappa)	Hill Function.
hill_saturation_sigmoid(x, sigma, beta, lam)	Hill Saturation Sigmoid Function.
inverse_scaled_logistic_saturation(x[, lam, eps])	Inverse scaled logistic saturation transformation.
logistic_saturation(x[, lam])	Logistic saturation transformation.
michaelis_menten(x, alpha, lam)	Evaluate the Michaelis-Menten function for given values of x, alpha, and lambda.
root_saturation(x, alpha)	Root saturation transformation.
tanh_saturation(x[, b, c])	Tanh saturation transformation.
tanh_saturation_baselined(x, x0[, gain, r])	Baselined Tanh Saturation.
weibull_adstock(x[, lam, k, l_max, mode, ...])	Weibull Adstocking Transformation.

Classes

ConvMode(*values)	Convolution mode for the convolution.
TanhSaturationBaselinedParameters(x0, gain, r)	Representation of tanh saturation parameters in baselined form.
TanhSaturationParameters(b, c)	Container for tanh saturation parameters.
WeibullType(*values)	Weibull type for the Weibull adstock.