Thermoluminescence-based simplified criteria for the detection of irradiated sesame seeds using artificial intelligence methods

Abstract

The practical application of unsupervised Artificial Intelligence (AI) numerical methods for analysing unexamined data has gained popularity for solving scientific and technological problems. This paper reports on the implementation of numerical algorithms set based on unsupervised AI methods to discriminate between irradiated and non-irradiated Mexican sesame sample by searching for behaviour patterns in the thermoluminescence (TL) response of polymineral samples adhered to the seeds. Two algorithms were tested, which were able to discriminate between irradiated and non-irradiated samples regardless of whether the whole or initial rise of the TL glow curve was considered. The use of AI algorithms can greatly increase the analytical process by using appropriate models to large datasets. Moreover, free software tools are now available for developers to implement these AI methods in their data analysis code, with Python being the primary language of choice.