Replication data for "NQF-RNN: Probabilistic Forecasting via Neural Quantile Function-based Recurrent Neural Networks"

Replication data for "NQF-RNN: Probabilistic Forecasting via Neural Quantile Function-based Recurrent Neural Networks". The synthetic dataset creates six different groups of synthetic time series (Group 1 to Group 6), each group containing seven probability distributions with different parameters. the complete details of synthetic dataset are here: sample_x : input covariates in training dataset. sample_x_vali : input covariates in validation dataset. sample_x_test : input covariates in test dataset. sample : simple scaled target observations in training dataset. sample_vali : simple scaled target observations in validation dataset. sample_test : target values in test dataset. sample_v : a scale factor in training dataset. sample_v_vali : a scale factor in validation dataset. sample_v_test : a scale factor in test dataset. the complete details of real-world datasets are available here: (electricity) https://archive.ics.uci.edu/dataset/321/electricityloaddiagrams20112014 [1]. (traffic) https://archive.ics.uci.edu/dataset/204/pems+sf [2]. (solar) http://www.nrel.gov/grid/solar-power-data [3]. (M4-hourly) https://github.com/Mcompetitions/M4-methods/tree/master [4]. (tourism-monthly, tourism-quarterly) https://robjhyndman.com/publications/the-tourism-forecasting-competition/ [5].