Storm Surge Prediction Using Different Machine Learning Methods

Description

• Storm surge prediction using ML.ipynb - This file includes:

  • Data preprocessing
    • Data cleaning, feature selection and creating time lagging predictors data
    • Data preparation (predictor and predictand data)
    • Splitting the data into training and testing datasets
    • Standardizing the training and testing datasets
  • Autocorrelation in time series
  • Multi Layer Perceptron (MLP)
    • Building a MLP sequential model, Training the model, Model evaluation and Plotting the results
    • MLP Hyperparameter Tuning
  • Long Short-Term Memory Networks (LSTM)
    • Building LSTM model, Data preparation, Training the model, Model evaluation and Plotting the results
  • Auto Regressive Integrated Moving Average (ARIMA)
    • Model building and analysis
    • Forecasting
    • Model evaluation
    • Results visualization
  • Convolutional Neural Networks (CNN)

• projectML.py - This file includes

  • Data preprocessing
  • Random Forest Regression (RFR)
    • Data preprocessing
    • Training the model
    • Model evaluation
    • Results visualization
  • Support Vector Regression (SVR)
    • Data preprocessing
    • Training the model
    • Linear, RBF and Polynomial kernels for SVM
    • Model evaluation
    • SVR hyper parameters tuning
    • Results visualization
    • Improvement of the SVR method by changing temporal resolution to "daily max surge" instead of hourly