| Abstract Submission No. | ABS-3.1-0112 |
| Title of Abstract | Intensity Prediction in the Himalayan Region using Statistical and Artificial Neural Network Technique |
| Authors | PUNIT PAURUSH*, Kapil Mohan, OP Mishra |
| Organisation | National Centre for Seismology |
| Address | Mausam Bhawan Delhi, Delhi, India Pincode: 110003 E-mail: punit.vickey@gmail.com |
| Country | India |
| Presentation | Poster |
| Abstract | A limited intensity information of the pre-instrumental and early instrumental era earthquakes is a hindrance to understand the effect of these earthquakes on construction elements/buildings and precise assessment of their magnitudes. An effort has been made to predict intensity using hypocentral distance and magnitude of the pre-instrumental and early instrumental era earthquakes utilising 24 earthquakes of magnitude 6.1 to 8.2, hypocentral distance 10 km to 2128 km having sample strength of 2183 recent earthquake intensity data. The study has been conducted on NW Himalayas, NE Himalayas and Central Himalayan region using Multi Variate Linear Regression technique by validating almost 75% (1743 samples) of the datasets with the 25% (440 samples) unused data. In the present study, the relation has been derived to predict the intensity based on magnitudes at different hypocentral distances using the multi-variate linear regression technique. The multi-layer perceptron (ANN) technique has been used on the derived prediction equation to verify the authenticity of the prediction equation for the intensity. The ANN technique established the verification of the prediction equation with R2 value of 0.67 (for Central Himalayan region), 0.58 (for NW Himalayan region) and 0.69 (for NE Himalayan region). The validation process strengthens the results and assurance as well as robustness of the developed equation for predicting the intensity of the pre-instrumental and early instrumental era earthquakes with root mean square error of 0.9 (for NW Himalayan region), 1.19 (for NE Himalayan region) and 0.7 (for Central Himalayan region). |
| Keywords | Intensity, MLR, Magnitude, Hypocentral distance, ANN |