Prediction of Rainfall Based on Deep Learning and Internet of Things to Prevent Landslide

Rainfall data is very important for the disaster that is, people and their property lose. Myanmar has two hilly regions, there are the Western region and the Eastern region, they have more landslide and also Chin state areas and jade mine areas in Kachin state always you can see a lot of landslides. This model with the Internet of Things system can predict the landslides. This monitoring system has a soil moisture sensor, weather sensor, raspberry pi 3 B, and send the message and data to rural areas, mountainous areas, government, and non-government organizations. Soil moisture and precipitation data are mainly for a landslide event, this information may change dynamically from the soil layer, and due to the heavy rain. Moisture sensors are deploying in the soil and scale monitoring of moisture. Landslide is due to heavy rainfall, the prediction is depending on the soil moisture, precipitation, rarely depend on a slope. Thus, we have to know the precipitation how much and rainfall data. This system forecasts a univariate time series, approach with baseline method, Recurrent Neural Network, Long Short-Term Memory model from deep learning using the daily rainfall data. This system is a low-cost landslide prediction system.