Research Article

Seismic Activity Analysis in California: Patterns, Trends, and Predictive Modeling

Authors

  • Pravakar Debnath School of Business, Westcliff University Irvine, California, USA
  • Mitu Karmakar School of Business, International American University, Los Angeles, California, USA
  • MD Tushar Khan Masters of Science in Business Analytics, Trine University
  • MD Azam Khan School of Business, International American University, Los Angeles, California, USA
  • Abdullah Al Sayeed Masters of Business Administration in Project Management, Central Michigan University
  • Arifur Rahman School of Business, International American University, Los Angeles, California, USA
  • Md Fakhrul Islam Sumon School of Business, International American University, Los Angeles, California, USA

Abstract

For decades, California has been considered one of the most seismically active areas in the United States, if not the most, due to the frequent occurrence of earthquakes from tectonic activity, notably along the San Andreas Fault.  Understanding seismic activity is a matter not only of safety but also of major importance for urban planning. In the recent past, Machine learning algorithms (ML) have emerged as a promising invention for advancing earthquake prediction by facilitating the pinpointing of patterns in large datasets that may be hard to detect through traditional techniques. The primary objective of this research project is to assess historical seismic data to identify trends and patterns in California’s seismic activity. Equally important, the goals of this research focused on developing predictive models that can provide insight into the possibility of seismic events in the future. To assess seismic activity in California, data were gathered from a credible and reputable source, most notably, the United States Geological Survey (USGS) Earthquake Database, which provides detailed records of earthquakes spanning over six decades. This dataset included all recorded seismic events in California, capturing the key details about the place of occurrence in latitude and longitude, the magnitude of the seismic event, the depth at which it happened, and the time when the seismic activity took place. To devise and curate a reliable earthquake prediction algorithm for California, distinct machine learning models were considered, comprising, Random Forest, XG-Boost, and Logistic Regression. Overall, the Random Forest algorithm exemplified high accuracy. Optimizing this algorithm will lead to more reliable predictions, potentially aiding disaster preparedness and risk mitigation in California's earthquake-prone areas. The findings from seismic activity analysis have deep implications for urban planning and disaster preparedness in California. Knowledge of the pattern, place, and magnitude of earthquakes over time will help urban planners and policymakers structure communities that can remain resilient during such calamities. For instance, higher earthquake frequencies would automatically call for increased stringency in building codes, especially along fault lines, to ensure that houses situated on such lines are designed to withstand major seismic activity. This operation may even limit or reallocate urban development out of high-risk zones into less risky zones.

Article information

Journal

Journal of Computer Science and Technology Studies

Volume (Issue)

6 (5)

Pages

50-60

Published

2024-11-07

How to Cite

Seismic Activity Analysis in California: Patterns, Trends, and Predictive Modeling. (2024). Journal of Computer Science and Technology Studies, 6(5), 50-60. https://doi.org/10.32996/jcsts.2024.6.5.5

Downloads

Views

1

Downloads

1

Keywords:

Seismic activity, earthquake prediction, California, earthquake trends, urban planning, disaster preparedness, machine learning