Implementation of Machine Learning Classification Technique for Detecting the Thermo-mechanical Properties of Chalogenide Glass Datasets

308 International Journal of Advance and Applied Research www.ijaar.co.in ISSN – 2347-7075 Impact Factor – 8.141 Peer Reviewed Bi-Monthly Vol. 6 No. 38 September - October - 2025 Implementation of Machine Learning Classification Technique for Detecting the Thermo-mechanical Properties of Chalogenide Glass Datasets Mrs. Swati Mule Dr. D Y Patil Arts, Commerce and Science College Akurdi, Pune-44 Corresponding Author –Mrs. Swati Mule DOI - 10.5281/zenodo.17315909 Abstract: Chalcogenide glasses are based on the chalcogen elements S, Se, and Te. These glasses are formed by the addition of other elements such as Ge, As, Sb, Ga, etc. These glasses are low-phononenergy materials and are generally transparent from the visible up to infrared. Chalcogenide glasses can be doped by rare-earth elements, such as Er, Nd, Pr, etc., and hence numerous applications of active optical devices have been proposed. These glasses are optically highly non-linear and could therefore be useful for all-optical switching. Chalcogenide glasses are sensitive to the absorption of electromagnetic radiation and show a variety of photoinduced effects as a result of illumination. The proposed paper presents an artificial intelligence approach in determining the thermo-mechanical properties of Chalogenide glass datasets. K-Nearest Neighbor (KNN) machine learning classification technique is used for the prediction purpose. The RMSE of values 13.88 and R2 Score is predicted as 0.81. Keywords: KNN Classification, RMSE, Chalogenide Glass, Machine Learning, Photon-Energy Material Introduction: Chalcogenides, despite their versatile functionality, share a notably similar local structure in their amorphous states. Particularly in electronic phase-change memory applications, distinguishing these glasses from neighboring compositions that do not possess memory capabilities is inherently difficult when employing traditional analytical methods. This has led to a dilemma in materials design since an atomistic view of the arrangement in the amorphous state is the key to understanding and optimizing the functionality of these glasses [3][4]. To tackle this challenge, we present a machine learning (ML) approach to separate electronic phase-change materials (ePCMs) from other chalcogenides, based upon subtle differences in the short-range order inside the glassy phase [1][2]. Leveraging the established structure–property relations in chalcogenide glasses, we select suitable features to train accurate machine learning models, even with a modestly sized dataset [5]. The trained model accurately discerns the critical transition point between glass compositions suitable for use as ePCMs and those that are not, particularly for both GeTe– GeSe and Sb2Te3–Sb2Se3 materials, in line with experiments [12-14]. Properties of Chalogenide Glasses: ● Infrared Transmission: Unlike oxide glasses, chalcogenide glasses are transparent over a wide range, including midand far-infrared wavelengths (800 IJAAR Vol. 6 No. 38 ISSN – 2347-7075 Mrs. Swati Mule 309 nm to 16 µm or more), making them ideal for IR imaging, sensing and communication [6]. ● Nonlinearity: They possess high optical nonlinear coefficients, which allows for applications in all-optical switching. ● Semiconductivity & Photosensitivity: Their small band gaps lead to semiconducting properties and a sensitivity to electromagnetic radiation, enabling various photo-induced effects used in optical devices like diffractive structures and waveguides [7]. ● Chemical Stability: They are chemically stable and resistant to water and acids, ensuring durability in harsh environmental conditions. ● Thermal Properties: They have low phonon energies and low temperature coefficients of refractive index, which means they maintain their focal length better over temperature changes compared to materials like germanium [8]. ● Visco-Plasticity: Chalcogenide glasses can be easily shaped into complex forms like lenses and fibers at relatively low temperatures, just above their glass transition temperature. Methodology: The proposed implementation consists of implementing a KNN machine learning algorithm. A synthetic Chalogenide glass dataset is used to predict thermo-mechanical properties of the given dataset using KNN. the process is as follow: ● Dataset Preparation: Glass composition ratios (e.g., %Ge, %As, %Se, %Sb). With other descriptors such as density, molar volume, atomic mass averages, etc. Output targets:Thermo-mechanical properties ● Preprocessing stage: Normalize or standardize input features ● KNN Model Implementation ● Training & Validation: Split dataset into train and test or use k-fold cross-validation ● Evaluate with metrics like RMSE for regression tasks Results: Fig. 1 Prediction and Distribution error observed for given Chalogenide glass dataset IJAAR Vol. 6 No. 38 ISSN – 2347-7075 Mrs. Swati Mule 310 Fig. 2 KNN Classification analysis for predicted and observed values Conclusion: The proposed implementation of KNN classification technique to predict the thermomechanical properties of chalogenide glass datasets has been studied. For the implementation purpose synthetic chalogenide dataset is used and results are generated. It is found that the KNN models works satisfactory to evaluate the thermo-mechanical properties as the R2 score obtained as 0.81 which shows the errors in predicting the property values are less. Acknowledgment: I would like to express my gratitude to Dr. Mohan Waman Principal, Dr. D Y Patil Arts, Commerce and Science College Akurdi Pune for valuable guidance. References: 1. Yunlai Zha, Maike Waldmann, and Craig B. Arnold; “A review on solution processing of chalcogenide glasses for optical components”; Vol. 3, Issue 9, pp. 1259-1272, 2013 2. Singh, P. K. and Dwivedi, D. K; “Chalcogenide glass: Fabrication techniques, properties and applications”; Ferroelectrics, 520(1), 256–273; 2017 3. Sayam Singla, Sajid Mannan; “Accelerated design of chalcogenide glasses through interpretable machine learning for composition–property relationships”; J. Phys. Mater., 2023 4. Bekir Karasu, Ali Ozan Yanar; “ Chalcogenide Glasses”; El-Cezerî Journal of Science and Engineering Volume: 6, No: 3, 2019 5. Sonam Tripathi, Brijesh Kumar; “Chalcogenide Glasses: Thin Film Deposition Techniques and Applications in the Field of Electronics”; IEEE 2018 6. Tamara Petkova, Vania Ilcheva, P. Ilchev, P. Petkov; “Ge-Chalcogenide Glasses – Properties and Application as Optical Material”; Key Engineering Materials, Volume 538, January 2013 7. D. Lezal , J. Pedlikova, J. Zavadil; “Chalcogenide Glasses for Optical and Photonics Applications”; Chalcogenide Letters Vol. 1, No. 1, January 2004 8. M. Hariharan, R.D. Eithiraj; “Machine learning model used to predict DFTcomputed thermoelectric parameters of chalcogenide halides CuHgSCl and CuHgSI”; Elsevier, Volume 43, 31 January 2025 9. Qundao Xu, Meng Xu, et.al.; “Enhancing elemental phase-change chalcogenide glass through tailored alloying”, Acta Materialia, Elsevier, Volume 289, 1 May 2025 10. Qundao Xu, Meng Xu, Siqi Tang, Shaojie Yuan; “Machine learning for IJAAR Vol. 6 No. 38 ISSN – 2347-7075 Mrs. Swati Mule 311 discrimination of phase-change chalcogenide glasses”, Willy; 4 February 2025 11. Saulo Martiello Mastelini , Daniel R. Cassar; “Machine learning unveils composition-property relationships in chalcogenide glasses”; Elsevier, Acta Materialia, Volume 240, November 2022 12. Rola Houhou, Thomas Bocklitz; “Trends in artificial intelligence, machine learning, and chemometrics applied to chemical data”; Volume2, Issue3-4, Recent Advances in Analytical Sciences, April 2021 13. Mohammad Yazdani-Asrami, Alireza Sadeghi, Wenjuan Song; “Artificial intelligence methods for applied superconductivity: material, design, manufacturing, testing, operation and condition monitoring”, 14 October 2022 14. Rola Houhou, Thomas Bocklitz; “Trends in artificial intelligence, machine learning, and chemometrics applied to chemical data”; Anal Sci Adv. 2021