Time Series Forecasting of Indian Coal Mines Fatal Accidents
A. Mohanty, D. S. Nimaje
Department of Mining Engineering, National Institute of Technology Rourkela, Odisha, India
Keywords: Accident forecasting, ARIMA, exponential smoothing, neural network, coal mine
Abstract
The present study analyzes the fatal accident occurrences of seventy years from 1951 to 2020 in Indian coal mines. The autoregressive integrated moving average (ARIMA) model, Brown’s double exponential smoothing method, Holt’s double exponential smoothing method, and neural network time series forecasting are used in this research to analyze fatal accidents and forecast future accident incidents. By analyzing various parameters of the applied models, the neural network model was found to be the most appropriate model for the collected data to forecast Indian coal mine accidents as it provides the least root mean squared error (RMSE) (17.62), and mean absolute error (MAE) (13.33) among all models. According to this study, the Neural Network model is the most suitable one to predict Indian coal mine fatality.
|
