Datasets
Standard Dataset
Dataset for MTS4WaterR: Predicting Gate Operation in Open Canal Control with Multi-Task Sequential Model
- Citation Author(s):
- Submitted by:
- Shiyuan Guo
- Last updated:
- Sun, 12/31/2023 - 02:33
- DOI:
- 10.21227/h22a-xj82
- License:
- Categories:
- Keywords:
Abstract
This is the dataset used in the paper MTS4WaterR: Predicting Gate Operation in Open Canal Control with Multi-Task Sequential Model, consisting of 2 main parts, used to train the evaluator and the learner neural networks, respectively.
Each part contains several files:
- in_x.txt: The input of the neural network of canal segment No. x. Each row is one piece of data. The first column is the initial total water discharge. For the evaluator, other columns represent the gate opening size of each canal gate. For the learner, other columns represent the given water discharge requirements of each canal gate.
- out_x.txt: The output of the neural network of canal segment No. x. Each row is one piece of data. Each column represents the target water discharge of each canal gate.
- data_x.txt: The normalized static feature values of each canal gate of segment No. x, processed by min-max normalization method. Each row represents a canal gate.
- index_x.txt: The static feature values of each canal gate of segment No. x after feature discretization. Each row represents a canal gate.
- prev_x.txt: Only exists in the learner part. Each row is one piece of data, represents the gate opening size of the last timestamp of each canal gate of segment No. x.
You may load the dataset using numpy.loadtxt, design the proper neural network model using frameworks like PyTorch, set other attributes like the maximum gate opening height, the maximum water discharge, evaluate metrics and other hyperparameters, etc. Then you can train and evaluate the model using the given dataset.
Comments
None