On the Application of Topological Data Analysis and Machine Learning to Flood Incidents, and Decision Making

Flooding annually destroys properties, building structures, farmland, brought about the loss of life, displacement of people, economic and financial losses. In this study, the researchers proposed a synthetic clustering technique that integrated the functions of Topological data analysis, K-means, a machine learning process for information extraction and decision making. The technique automatically chooses the threshold value. It has joint topological and geometric attributes; it applies both properties to obtain the relevant features (shape) of datasets. The aim is to provide information from datasets for decision making in solving flood problems by identifying the feature patterns of floods. Three regions containing seven states in Nigeria where selected: the relationships in the feature patterns of our datasets were obtained. After our model's training process, the best set was obtained at the point where the highest Silhouette coefficient (SI) produced an efficiency of 80%. Finally, the outcome on flooding and no flooding areas provided information that inundate the risk managers to make the best decision to avert flood disasters. This study conducted the validity test to authenticate its findings