International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 10 Issue: 07 | Jul 2023
p-ISSN: 2395-0072
www.irjet.net
Predicting Flood Impacts: Analyzing Flood Dataset using Machine Learning Algorithms Naga Ravindra Babu M1, B Naresh2, A Satya Vamsi Kumar3, G Ganga Bhavani4, A Sai Ram5, G Chakradhara Rao6 1Assoc. Prof., Dept. of Computer Science, B V Raju College, Bhimavaram, AP, India 2Assoc. Prof., Dept. of Computer Science, B V Raju College, Bhimavaram, AP, India 3Asst. Prof., Dept. of Computer Science, B V Raju College, Bhimavaram, AP, India
4Asst. Prof., Dept. of Computer Science, B V Raju College, Bhimavaram, AP, India 5Asst. Prof., Dept. of Computer Science, B V Raju College, Bhimavaram, AP, India 6Asst. Prof., Dept. of Computer Science, B V Raju College, Bhimavaram, AP, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Floods are one of the most destructive and
in non-flood proof structures, lack access to warning systems, or are unaware of the risk of flooding [1].
challenging to anticipate natural catastrophes. The frequency or severity of floods has grown in recent years due to climate change and urbanization, as have the associated fatalities and financial losses. Machine learning-based flood forecasting models have slowly started to appear as a result of the fast expansion of computing power. Rich information is included in these models since they were trained on historical data, which is advantageous for data analysis and use. Machine-learningbased models are more effective than the conventional physical flood forecasting model in achieving satisfying results. This study provides a summary of contemporary machine learning-based flood prediction techniques to illustrate recent developments in flood forecasting.
75% of those who pass away in flood catastrophes drown. Disasters caused by flooding are occurring increasingly often, and this tendency is predicted to continue. Flooding increases the danger of drowning, especially in low- and middle-income nations where residents live in flood-prone locations and flood warning, evacuation, and protection systems are still underdeveloped or insufficient [2].
2. LITERATURE SURVEY The author of this research analyzed several machine learning-based flood forecasting systems, including linear regression, decision trees, and SVM-based approaches, as well as deep learning-based algorithms like BP and LSTM models. The study shows that the applicability of various approaches varies. Furthermore, since the most recent algorithms are largely influenced by the sophisticated models in deep learning, this paper's conclusion from the current research is that the advancement of deep learning technology has a significant impact on further improving the accuracy of flood prediction performance [3].
We list a range of current works in flood prediction and construct the model based on several methodologies. For flood warnings, flood reduction, or flood prevention, machine learning (ML) models are useful. Machine-learning (ML) addresses have become more well-liked in this regard because to their minimal computing demands and predominance of observational data. Key Words: Flood, human injure, urbanization, Random Forest, Support Vector, Neural Network etc.
Based on historical rainfall datasets spanning 33 years, the goal of this project was to develop a machine learning model that can forecast floods in Kebbi state so that it may be applied to other Nigerian states with high flood risk. In this study, three machine learning algorithms—Decision Tree, Logistic Regression, and Support Vector Classification (SVR)—were assessed and their Accuracy, Recall, and Receiver Operating Characteristics (ROC) scores were compared. When compared to the other two methods, logistic regression yields more accurate findings and offers excellent performance accuracy and recall. The Decision Tree fared better than the Support Vector Classifier as well. Due to Decision Tree's above-average accuracy and below-average recall ratings, it did pretty well [4].
1.INTRODUCTION The number of natural and man-made disasters has grown globally in recent years. The most common natural catastrophe is a flood, which happens when an excess of water submerges normally dry ground. Floods are typically caused by protracted periods of intense rain, rapid snowmelt, storm surges from tropical cyclones, or tsunamis in coastal areas. Floods may wreak havoc across a large area, causing fatalities as well as damage to private property and vital public health facilities. Worldwide, more than 2 billion people were impacted by floods between 1998 and 2017. Most at risk from floods are those who reside in floodplains,
© 2023, IRJET
|
Impact Factor value: 8.226
This study introduces an innovative method for using the ensemble model to estimate water level in relation to flood
|
ISO 9001:2008 Certified Journal
|
Page 50