Chlorophyll Prediction System with Machine Learning Algorithms in Lake Titicaca (Peruvian Sector)

Arroyo Paz A.; Ponce Y. · Communications in Computer and Information Science (2023) · DOI: 10.1007/978-3-031-35641-4_12

Resumen

Water is a fundamental resource, eutrophication is an important factor in determining water quality, and chlorophyll-a is an indicator of this degree. Artificial intelligence offers important tools to help improve classification and prediction for inland water studies such as lakes. This study evaluated the water quality in the Lake Titicaca basin (Peruvian sector), using the monitoring results of the Instituto del Mar del Perú (IMARPE), between 2011 and 2021. The dataset used in the study included 579 samples and their meta-data collected over eleven years. A program was designed to compare the performance of four machine learning models as follows, linear regression model, decision tree, random forest, and XGBoost. The result of the research showed that the XGBoost model performed better, with an r2 of 0.5455, MAPE of 25.8168, and RMSE of 0.6509, compared to the other algorithms. The findings strengthen the argument that ML models, especially XGBoost, can be used for chlorophyll prediction in lakes.

Instituciones: Universidad Nacional del Altiplano; Universidad Nacional de Juliaca