HackMD
1/26 Study papers for water project
Real time series
Machine learning for real-time aggregated prediction of hospital admission for emergency patients
They use sequential training
- It helps generate more data from a given limited time series and then thorough training of the ANN.
- It makes the data heterogeneous so that the overfitting issue of the ANN can be reduced; and
- It facilitates the learning of patterns of the data not only for the entire time series but also for short segments of sequential data.
They apply this approach to a diverse set of financial time series data and show that it can be used to make accurate predictions across different markets.
Cross validation
A Consolidated Cross-Validation Algorithm for Support Vector Machines via Data Reduction
https://neptune.ai/blog/cross-validation-in-machine-learning-how-to-do-it-right
https://medium.com/@poudelsushmita878/cross-validation-in-time-series-forecasting-db2bc7601875
LSTM with water quality
Short-term water quality variable prediction using a hybrid CNN–LSTM deep learning model
Water Quality Predictions Based on Grey Relation Analysis Enhanced LSTM Algorithms
Water Quality Prediction Method Based on LSTM Neural Network
hyp-tuning
Algorithms for Hyperparameter Tuning of LSTMs for Time Series Forecasting
Optimizing LSTM for time series prediction in Indian stock market
LSTM with air quality
An LSTM-based aggregated model for air pollution forecasting
LSTM + attention head
Improving time series forecasting using LSTM and attention models
