Identifying Types of Waste as Efforts in Plastic Waste Management Based on Deep Learning

Purpose: This research aims at designing a computer algorithm for automatic waste sorting.Design/methodology/apprach: This research is quantitative and uses secondary data, specifically images of various types of waste. The images will be classified into organic and inorganic waste types with the assistance of a deep learning model. In this research, we propose the EfficientNet method for Waste Type Identification as an Effort in Plastic Waste Management. Experiments were conducted on a secondary dataset from Kaggle.com, which involved classifying various types of waste into "Plastic" and "Non-Plastic" categories, showing the effectiveness of the proposed method.Findings/result: The measurement is performed to compute the accuracy of the designed deep learning model in classifying waste images into the appropriate waste types. Based on the research results, our system achieved the highest accuracy of 97% during testing.Originality/value/state of the art: The designed method can perform fast and automatic waste sorting, which is useful in reducing the increasing amount of waste accumulating each year.