Jin Gyeong Son (Korea Research Institute of Standards and Science)

Jin Gyeong Son (Korea Research Institute of Standards and Science)

Jin Gyeong Son (Korea Research Institute of Standards and Science)

Jin Gyeong Son (Korea Research Institute of Standards and Science)

Jin Gyeong Son (Korea Research Institute of Standards and Science)

Jin Gyeong Son (Korea Research Institute of Standards and Science)

Hyun Kyong Shon (Korea Research Institute of Standards and Science)

In-Ho Lee (Korea Research Institute of Standards and Science)

Tae geol Lee (Korea Research Institute of Standards and Science)

In this study, time-of-flight secondary ion mass spectrometry (ToF-SIMS) data and machine learning techniques were employed for the purpose of categorising six distinct types of plastics. In order to address the characteristics of the measurement data, the local maxima were subjected to analysis during the pre-processing step. Subsequently, a variety of machine learning techniques were employed to develop a model with the capacity to effectively classify the plastics. To facilitate the visualisation of the data distribution, dimensionality reduction techniques, including Principal Component Analysis and t-Distributed Stochastic Neighbor Embedding, were employed. In order to distinguish between the six plastic types, an ensemble analysis was performed using four tree-based algorithms: Decision Tree, Random Forest, Gradient Boosting, and LightGBM. This methodology enables the identification of feature importance for plastic samples and facilitates the inference of the chemical properties of each plastic type. Therefore, ToF-SIMS data can be employed to accurately classify plastics and improve interpretability.

​