AAEO-MLP-GWL

Cite our code

@misc{nguyen_van_thieu_2022_6480834,
  author       = {Nguyen Van Thieu},
  title        = {AAEO-MLP-GWL},
  month        = {april},
  year         = {2022},
  publisher    = {Zenodo},
  version      = {v1.0.0},
  doi          = {10.5281/zenodo.6480834},
  url          = {https://doi.org/10.5281/zenodo.6480834}
}

Data Statement:

1. The gridded rainfall and temperature data are available freely at

https://www.imdpune.gov.in/Clim_Pred_LRF_New/Grided_Data_Download.html

2. The tidal height data are available at

https://www.psmsl.org/data/

3. The groundwater level data were provided on purchase by the the Department of Mines and Geology, Government of Karnataka state, India. So, the data is confidential and could not be shared publicly.

Groundwater Level Modeling using Augmented Artificial Ecosystem Optimization

Compared Algorithms

1. Genetic Algorithm (GA)
2. Differential Evolution (DE)
3. Particle Swarm Optimization (PSO)
4. Harris Hawks Optimization (HHO)
5. Hunger Games Search (HGS)
6. Sparrow Search Algorithm (SSA)
7. Multi-Verse Optimizer (MVO)
8. Equilibrium Optimizer (EO)
9. Electromagnetic Field Optimization (EFO)

10. Forensic-Based Investigation Optimization (FBIO)
11. Coronavirus Herd Immunity Optimization (CHIO)

12. Slime Mould Algorithm (SMA)
13. Chaos Game Optimization (CGO)
14. Artificial Ecosystem-based Optimization (AEO)
15. Improved AEO
16. Modified AEO
17. Enhanced AEO
18. Adaptive AEO (Our proposed model)

How to run models

1. All the models inside the script: script_mha_mlp.py
2. All the configuration inside the script: config.py
3. Calculate the statistics using the script: get_summary_statistics.py
4. For tradtional MLP run the script: script_mlp.py
5. Base class for all models is defined in models/based_mlp.py 
6. All the MHA-MLP model is defined in models/mha_mlp.py
7. All helper functions is located in utils
8. Results are located in data/input_data/results_paper
9. All figures are located in paper
10. The AAEO-MLP model figure used draw.io website to design

Environment

1. If using conda

+ Should create a new environment:
   conda create -n new ml python==3.7.5
   conda activate ml
   
   conda install -c conda-forge numpy
   conda install -c conda-forge pandas
   conda install -c conda-forge scikit-learn
   conda install -c conda-forge matplotlib
   conda install -c conda-forge tensorflow==2.1.0
   conda install -c conda-forge keras==2.3.1
   
   pip uninstall mealpy
   pip uninstall permetrics
   pip install mealpy==2.4.0
   pip install permetrics==1.2.2

2. If using pip

+ Should create a new environment:
   python -m venv ml
   ml\Scripts\activate.bat
   pip install -r requirements.txt

Export environment

pip list (pip auto-downloaded in python)

pip list --local (only packages installed from current environment)


1. Create a blank environment 

python -m venv env_name (create a blank environment named "env_name")

env_name\Scripts\activate.bat (activate env)

pip freeze > requirements.txt 


2. Create an inheritance environment 

python -m venv india --system-site-packages (create india environment inherits packages from based-system)

india\Scripts\activate.bat  (activate india env)

pip install numpy (install package to current env)

pip freeze --local > requirements.txt (Create requirements file that installed additional packages)

rmdir india /s 	(remove all environment)


3. Remove pip from python 

python -m pip uninstall pip

4. Export environments 

pip list --format=freeze > requirements.txt 

pip freeze --local > requirements.txt

Helper

https://pythontic.com/modules/pickle/dumps https://medium.com/fintechexplained/how-to-save-trained-machine-learning-models-649c3ad1c018 keras-team/keras#14180 https://ai-pool.com/d/how-to-get-the-weights-of-keras-model-

https://stackoverflow.com/questions/1894269/how-to-convert-string-representation-of-list-to-a-list

import json 
x = "[0.7587068025868327, 1000.0, 125.3177189672638, 150, 1.0, 4, 0.1, 10.0]"
solution = json.loads(x)
print(solution)