Computational Model for River Water Level Prediction
Research Verification
Official research membership confirmation:
Overview
This research focuses on developing a computational model capable of accurately predicting river water levels, with the goal of supporting early warning systems for potential flooding events. The primary objective is to design a reliable, custom-built model that effectively captures the unique hydrological characteristics of rivers in Indonesia and the Philippines. A deep learning model will also be implemented solely as a comparision.
Current Progress
We have completed data collection. Our team has established connections with two main river monitoring station.
Our data collection phase start from February 1, 2025 through March 30, 2025, focusing on two major Indonesian rivers:
Bengawan Solo River
Indonesia's longest river (548.53 km) flowing through Central and East Java.
Citarum River
West Java's primary river (269 km) with critical importance for agriculture, industry, and 25 million residents.
Notably, our research leader informed me that the core objective of this project is to develop a computational model tailored to the hydrological characteristics of Indonesian and Philipine river 😓. Our research leader and the Philippine research team during a discussion : 'Utamanya kita akan buat model komputasional sendiri' -That’s a whole different level of pain 😅. BUTT!! part of my role within the team, me and some another research member have been tasked with designing and implementing a neural network-based model to serve as a comparison?
My Contribution
As a research team member, I am primarily responsible for:
- Conducting literature reviews on recent advances in flood prediction models
- Build neural network-based model
- Data collection and developing data processing pipelines for handling time-series hydrological and meteorological data
Future Work
In the coming months, we plan to:
- Build computational model
- Experiment with different neural network architectures
- Intialize and developing fixed model based neural network as a comparison