Artificial Intelligence Class

EarthquakePrediction

The earthquake is the seismic fluctuations that occur as a result of the energy that happens spontaneously in the crust of the earth and the event of these waves shaking the earth. One of the main events in nature that causes both irretrievable financial and physical damage is the occurrence of earthquakes. Predicting the earthquake is very important in terms of preventing loss of life in order to protect human life. In the event of an earthquake that may occur in the future, warning people beforehand and taking the necessary precautions can prevent major damage and loss of life. Artificial intelligence helps to detect seismic events meaningfully, to predict earthquake activities more accurately, to determine the unknown characteristics of the natural disaster. In this project, a system that will estimate the magnitude of the earthquake using AI algorithms including machine learning and evaluate the estimation results of these algorithms, consider the algorithm that gives the best performance is proposed

Dataset

The dataset was created by merging data from the public Kaggle source, divided by months until January 2019 and March 2020. The purpose of this dataset is to examine the date, latitude, longitude and depth features of the historical data in order to estimate the earthquake magnitude in Turkey

Datasets available on: https://www.kaggle.com/mielek/datasets?search=turkey+earthquake

Methods

Regression Techniques:

1. Multiple Linear Regression
2. Polynomial Linear Regression
3. Decision Tree
4. Random Forest
5. Support Vector Regression

Multivariate Time Series

1. Measure of Stationary
2. Making a Time Series Stationary
3. Vector Auto Regression (VAR)
4. Long Short-term Memory (LSTM)

Measuring the Performance of Regression Models:

1. Root Mean Squared Error (RMSE)
2. R-Square Score
3. Mean Squared Error (MSE)
4. Mean Absolute Error (MAE)