<![CDATA[The electronics industry, particularly solar cell manufacturing, demands production processes that are fast, precise, and supported by high data integrity. One critical component in the production flow is the chip embedded in the flower basket, which functions to store and transmit data through an RFID system. Damage to the chip can lead to information loss, tag reading failures, and disruptions in production efficiency and continuity. This study aims to predict chip status, classified as either normal or damaged, based on various process parameters, including immersion temperature, ambient humidity, process pressure, machine vibration, drying speed, heating and cooling duration, firing temperature, usage frequency, and RFID reading conditions. A feature engineering approach is applied to construct more representative derived features, while SMOTE is utilized to address class imbalance in the dataset. This study focuses on developing a predictive model using the Random Forest method to identify the most influential process variables related to chip damage risk. The data used in this study are obtained from historical production process records of a solar cell manufacturing plant. The results indicate that combinations of multiple process parameters significantly contribute to the potential risk of chip damage, and the Random Forest model demonstrates good predictive performance in classifying chip conditions. These findings suggest that the proposed model can serve as an early warning system to detect chip damage risks before they impact production processes. With proper implementation, the predictive model is expected to support preventive actions, enhance data integrity, and minimize disruptions in the solar cell manufacturing workflow.]]>
