Investigating the occurrence of flooded areas and developing a predictive model for identifying regions at risk of flooding
This project focuses on investigating the occurrence of flooded areas and developing a predictive model for identifying regions at risk of flooding. The study utilizes various data mining techniques to explore correlations between flooding and environmental factors. The dataset, obtained from publicly available sources, includes features such as land cover, elevation, and temperature data. Through random sampling and feature extraction, the study derives key features relevant to flood prediction. Data preprocessing involves frequency encoding of categorical variables, followed by training a machine learning model using XGBoost. The results indicate high accuracy in predicting flood occurrences, with elevation and land cover emerging as significant predictors. Temperature-related features, while less influential, still contribute to the predictive model.
Application Language(s): N/A
Programming Languages and Technologies: Python, GeoPandas, Shapely, OWSLib, XGBoost
Member(s): Pari Naz Tabari, Taha Rostami