| Title |
Forest-Based Underwater Temperature Prediction Model For Iot Devices |
| ID_Doc |
26891 |
| Authors |
Mondal S.; Roy S.; Jeevaraj R.; Mary Jasmine E.; Ghosh M.; Nag A. |
| Year |
2023 |
| Published |
2023 IEEE 3rd Mysore Sub Section International Conference, MysuruCon 2023 |
| DOI |
http://dx.doi.org/10.1109/MysuruCon59703.2023.10396911 |
| Abstract |
IoT is a booming domain in the field of technology. The main reason for its advancements in recent years has been its applications in various domains such as the environment, healthcare, smart cities, and agriculture. Real-Time data collection and fast exchange of data allow for effective functioning. Underwater temperature detection can be more fruitful and effective when coupled with IoT. This study focuses on building a predictive model that may help Automated Underwater Vehicles (AUVs) detect the ocean bed's temperature in real-Time. The AUV embedded with IoT onboard sensors enables the exploration of submarines in an underwater environment and lives. For this purpose, six machine learning regression models are considered to be trained over a dataset containing underwater temperature records at the corresponding depths near the islands in Brazil. The standard performance metrics were utilized for the evaluation of the performance of the models. Tree-based ensemble models such as Random Forest Regressor and Decision Tree Regressor produced the best predictive results, with their R2 values being 96.13% and 94.21%, respectively. The results clearly indicate that accurate underwater temperature prediction can be achieved by implementing tree-based models. Application of this technology can help AUVs get more insight into marine life and make inferences of any anomalies regarding this whatsoever. © 2023 IEEE. |
| Author Keywords |
Coefficient of Determination (R2); Mean Absolute Error; Prediction; Random Forest Regression; Regression Model; Underwater temperature |
