Map bool to int for cython

Create ``Dataset`` abstraction for sparse and dense inputs.

Map learning rate string to int for cython

Map task string to int for cython

Maps labels to [-1, 1] space

Validate input params.

Fit factorization machine using Stochastic Gradient Descent with Adaptive Regularization. Parameters ---------- X : {array-li

Predict using the factorization machine Parameters ---------- X : sparse matrix, shape = [n_samples, n_features] or