quantile LINEAR regression by using python Deep neural networks packages

quantile LINEAR regression by using python Deep neural networks packages do it for all main Deep NN packages , but do not use hidden layers LINEAR regression MEANS NO hidden layers (upper + lower ) / 2 performance should be not worse than [login to view URL] should work not slower than [login to view URL] keras for example point to start [login to view URL] tensor flow [login to view URL] pytorch then it should be 4 solutions, all of them should provide good performance each solution in separate file data is mixture of categorical and continues features at least 3 categorical features at least 3 continues features at least 10000 rows you find needed data sets by yourself : several datasets ( more than 2) all data sets used for all models see primitive linear (or log-linear) model W \cdot \mathbf{x} + \mathbf{b} (where W is a weight matrix, \mathbf{x} = (x_1, x_2) is an input vector, and \mathbf{b} is a bias vector) without hidden layers, [login to view URL] better to find and use existing code from web, then you include web links for source to prove all done correctly: make prediction manually use only one dimension vector coefficients like [login to view URL] coef_ : ndarray of shape (n_features,) or (n_targets, n_features) Weight vector(s). you prove correctness by upper prediction = Data * coef_upper_ you prove correctness by lower prediction = Data * coef_lower_ coef_upper_ coef_lower_ like [login to view URL] coef_ : array of shape (n_features,) Estimated coefficients for the features YOU NEED TO EXTRACT coef_upper_ coef_lower_ FROM ML MODEL you baseline FOR PREDICTOIN [login to view URL] you baseline FOR UPPER AND LOWER INTERVALS [login to view URL] [login to view URL] [login to view URL] and of course hyperparameters search like optimizer adam , mrop and etc