Incremental Machine Learning Predictive Analysis for Agricultural Applications Using A Modular Weather Monitoring Sensor in Infanta, Quezon

The Philippines considers agriculture as an important part of the country's economic security, so when it comes to agriculture, there is always a demand when it comes to management. Specifically, Infanta, Quezon, a city in Southern Luzon, is often plagued with typhoons and landslides that affect their agriculture. This is why weather sensors play a big role in their city. These weather sensors play a big role in their agricultural areas as they help with irrigation control and flood mapping, which are often paired with predictive analysis using machine learning to help evaluate critical weather conditions. The study's objective is to develop an incremental machine learning algorithm using pre-existing machine learning techniques as a baseline to classify data from the modified weather sensor system to determine the status of the agricultural fields. The researcher tested LSTM, SVM, and Random Forest as its baseline model for the incremental model, and Random Forest was used for the baseline. Through thorough exploration of the parameters and models, the researchers were able to utilise the best model for predictions in the agricultural aspect. The incremental models for soil moisture prediction and crop yield prediction, which work hand-in-hand, are 98% and 98.3 %, respectively.