Identification of earthquake patterns for predictive modeling using decision tree classifiers to maintain sustainability
Industry 6.0: Technology, Practices, Challenges, and Applications, CRC Press, ss.248-268, 2024
- Yayın Türü: Kitapta Bölüm / Araştırma Kitabı
- Basım Tarihi: 2024
- Doi Numarası: 10.1201/9781003517993-11
- Yayınevi: CRC Press
- Sayfa Sayıları: ss.248-268
- Kırklareli Üniversitesi Adresli: Evet
Özet
Natural catastrophes like earthquakes can have disastrous effects on infrastructure and human life. Though seismology and earthquake monitoring technologies have advanced, precise prediction is still a difficult endeavor. Enhancing earthquake forecast accuracy and dependability will allow for more efficient and prompt mitigation plans. By combining machine learning methods with seismic data, the suggested model uses Decision Tree algorithms to analyze and categorize intricate patterns linked to the occurrence of earthquakes. Because decision trees are good at handling non-linear relationships and capturing complex interactions within the dataset, they are especially well-suited for this kind of work. A numerical dataset with a wide range of seismic parameters is used to validate the suggested model. The dataset contains statistics on depth, frequency, amplitude, and historical occurrences of earthquakes. The robustness of the model is increased by this large dataset, which guarantees that the decision tree classifier can learn and generalize patterns over diverse geological settings. Preprocessing the dataset for handling values that are missing, outliers, and feature standardization is part of the methodology. The dataset is then split into both training sets as well as testing sets to facilitate training and model evaluation. The underlying patterns and correlations in the seismic data are then taught to the decision tree classifier through the use of the training set. Accuracy, precision and few others are among the strict evaluation measures used to gauge the models performance. The improved performance of the model is attributed to its ability to recognize subtle patterns and interactions within the dataset. The work fills a vital need in public safety and catastrophe management by supporting continued efforts to create accurate and effective earthquake prediction systems. The findings of this chapter open the door to develop more reliable and efficient earthquake prediction systems in the future by providing insightful information to the fields of seismic research and disaster management.