00000ctm a22000004i 4500 UP-99796217613097275 Buklod 20230215105418.0 m |o d | ta 190917s xx d r |||| u| (iLib)UPD-00413166515 DENGII rda eng DMLUC LG 995 2019 E653 E84 Esguerra, Givette Kristine Y. author. Development of an artificial neural network (ANN)-based prediction model for the occurrence of algal blooms in Laguna de Bay based on in situ data and remotely sensed data thesis by Givette Kristine Y. Esguerra ; Florencio C. Ballesteros Jr/, thesis adviser. Quezon City College of Engineering, University of the Philippines Diliman 2019. xiv, 124 leaves, 19 unnumbered leaves illustrations (some color) 28 cm text rdacontent unmediated rdamedia volume rdacarrier Thesis (Master of Science in Environmental Engineering)--University of the Philippines Diliman June 2019. Available to the general public. Algal blooms pertain to an undesirable formation of unicellular freely-floating algal scum caused by the rapid growth of phytoplankton, which can become a hazard for the water body ecosystem. Laguna de Bay serves as both a source of livelihood and water supply for the residents in the region and the risk of algal blooms should be detected for safe and efficient management. The research presents a method for predicting the amount of phytoplankton to alert the monitoring agencies of incidences of high phytoplankton as a scalable and inexpensive early-warning tool. The study primarily focuses on the development of a prediction model based on the following water quality parameters measured by the Laguna Lake Development Authority(LLDA) from 2008 to 2018: nitrate, orthophosphate, water temperature, turbidity chlorophyll-a, and phytoplankton counts. Gaps in the data were augmented with the data from Moderate Resolution Imaging Spectroradiometer (MODIS) satellite images for sea surface temperature (SST), found to be in good agreement to field values. The system predicts the phytoplankton counts of the next month using three months of previous values of the water quality parameters, modeled through the multilayer perceptron neural network method. The research uses a walk-forward validation method to obtain the root-mean-square-error (RMSE) of the model. Each station was modeled separately and the results present some stations (Stations 2 and 4) as having statistically less RMSE, whereas other stations have statistically no significant difference with the zero rule algorithm baseline model and the ordinary least square regression. Algal blooms. Neural networks (Computer science) Ballesteros, Florencio C. Jr. adviser. Thesis FI UP UPD DENG-II LG 995 2019 E653 E84 Thesis