IoT-Based Classification Solution with SVM Machine Learning Method

In this paper, an IoT system is designed to identify the type of Osmancik and Cammeo rice. In the designed system, a data set containing the physical properties of Osmancik and Cammeo type rice was classified using SVM method. The actual feature weight values and the intercept value obtained by the SVM method are sent to the ESP32 development board via Wi-Fi, The program on the ESP32 development board creates an HTML-based user interface. Users connect to the IOT network with their computers or mobile devices and access this interface via the TCP/IP protocol. The rice classification process is performed by ESP32 using the rice physical property values entered in this interface. This process uses the feature weight values and the intercept value received from the computer via Wi-Fi. In the IoT network, a wireless modem was used to communicate between the computer, mobile devices, and the ESP32 development board. The user interface is displayed over the Wi-Fi network using the TCP/IP protocol with the web browsers of the devices. This interface is used to enter seven physical properties of the rice to be predicted. These seven properties are used by the ESP32 development board to perform the prediction process. According to the result of this prediction process, 2 roles are controlled by ESP32. With this experimental study, a machine learning-based applicable IoT system model is presented.