Elemental Analysis and Classification of Nicotine Pouches Using Machine Learning Assisted Laser Induced Breakdown Spectroscopy

The elemental analysis and classification of nicotine pouches using machine learning assisted Laser Induced Breakdown Spectroscopy (LIBS) is reported for the very first time. Nicotine pouches are widely recognized as non-tobacco products. These pouches have become popular among individuals. LIBS analysis has identified many elements present in the pouches such as Aluminium, Barium, Calcium, Chromium, Copper, Iron, Magnesium, Sodium, Scandium, Strontium, and Titanium. Twenty-nine machine-learning classification models from the classification learner app were utilized to classify the nicotine pouches. In the context of classification, our analysis was focused on two separate classes. In 1st Class, the flavors remain the same while the nicotine strength changes. Contrarily, in 2nd Class, flavors change while the nicotine strength remains the same. The application of supervised machine learning classification techniques yielded noteworthy outcomes. In 1st Class, the highest test accuracy achieved was 98%, while in 2nd Class, a multitude of models achieved a remarkable 100% test accuracy, emphasizing the precision achieved in this configuration. This shows machine learning models are really good in classification based on flavors of nicotine pouches as well as on nicotine strength.