SISTEM MONITORING PENGENDALIAN PENGAIRAN SAWAH MENGGUNAKAN METODE DECISION TREE PADA PLATFORM THINGSPEAK BERBASIS INTERNET OF THINGS (IOT)

Abstract

This research aims to find out how to design IoT-based rice field watering monitoring and control systems using decision tree data processing. The process of irrigation of rice fields is a way of draining or bringing water to the rice field area or lading regularly. Irrigation can also be interpreted using water sources as something beneficial for plant life. If there is excessive water in lading, especially rice fields, it will interfere with the plant growth process. Irrigation is a major factor because the majority of rice production in Indonesia comes from irrigation fields.
 The design of the system uses hc-sr04 ultrasonic sensors to read irrigation water, soil moisture sensors to read soil moisture in rice fields, water level sensors to read water volumes in rice fields, and Arduino Uno as controlling and signaling to the cloud. Data from sensors sent to Arduino is then sent and displayed to the cloud with the platform thingspeak data from the platform in the process using data mining techniques with algorithm C. 45 on algorithm C.45 initial data in numerical form is transformed into categorical data according to standard standards. In C.45 the data is processed through the first calculated stage looking for the total entropy, entropy of each attribute, and looking for the highest gain according to the parameters for the class.
 Results of accuracy algorithm C.45 using the matric confusion method obtained an accuracy result of 99.50% in the next researchers expected action after receiving notification of field condition and irrigation both in the low or high state. So that monitoring the irrigation of these rice fields becomes more effective for the future.