Suicide Trend Analysis and Prediction in India using Facebook Prophet

Abstract

Suicide analysis is an area of vital importance to the National Institute of Mental Health and various other agencies working in the field of suicide prevention. Studying on this aspect helps to analyze the suicide pattern and trends that suicides follow over the years. This paper explores time-series data of the suicides that occurred in India to find whether there is a notable change in trend after a certain time point. A predictive approach is applied to forecast into the future of the suicide trend. The paper applies Facebook Prophet, a time-series prediction algorithm for drawing inferences and conclusions. The paper also suggests an inflection point algorithm that highlights the suicide trend between two points in time. Additionally, the model is also capable of predicting the trend for “n” number of years to come. We have used MAPE and SMAPE error techniques for accurate measurement. The mean absolute percentage error (MAPE) is a predictive accuracy measure while the symmetric mean absolute percentage error (SMAPE) is a percentage (or relative) error-dependent accuracy measure. The values of MAPE and SMAPE were found to be in the range of 0.1-0.2 and less than 12 respectively. The conclusion derived is that the result is an increasing nature in the current year and there is a need for utmost attention.