Yazar "Elnaggar, Ashraf Y." seçeneğine göre listele

Listeleniyor 1 - 4 / 4
  • [ X ]
    Öğe
    A fast and efficient machine learning assisted prediction of urea and its derivatives to screen crystal propensity with experimental validation
    (Elsevier Ltd, 2025) Güleryüz, Cihat; Sumrra, Sajjad H.; Hassan, Abrar U.; Mohyuddin, Ayesha; Noreen, Sadaf; Elnaggar, Ashraf Y.
    Predicting crystal propensity is crucial yet challenging in various industries where it significantly influences product stability, performance, and efficacy. Predicting a crystal propensity identifies their optimal chemical structures for desired properties including solubility, bioavailability, shelf-life stability etc. Herein, A machine learning (ML) assisted analysis is performed to predict their crystal propensity by collecting a dataset of 6000 non-crystalline and over 200 crystalline urea and its derivatives. The data is trained by employing a Support Vector Machine (SVM) with its Radial Basis Function (RBF) and linear kernels along with Random Forest regression analysis. The trained data is compared with four other ML models, including Linear Regression, Gradient Boosting, Random Forest and Decision Tree Regressions to predict their crystal propensity. It yields an accuracy of 79 % for identifying their non-crystalline compounds and 59 % in predicting crystallization failure. Their dimensionality reduction via t-SNE reveals their distinct clustering patterns to underscore their complex interplay between molecular structure and crystal propensity. Their experimental validation also corroborates the current findings to demonstrate their efficacy to streamline their crystal engineering for pharmaceutical formulation-based workflows. Notably, the number of rotatable bonds and molecular connectivity index (χov) emerges as pivotal descriptors for enabling their accurate classification with minimal input features. This study elucidates its quantitative structure-crystallinity relationship to provide a valuable tool for crystal design and optimization.
  • [ X ]
    Öğe
    A graph neural network assisted reverse polymers engineering to design low bandgap benzothiophene polymers for light harvesting applications
    (Elsevier Ltd, 2025) Hassan, Abrar U.; Güleryüz, Cihat; El Azab, Islam H.; Elnaggar, Ashraf Y.; Mahmoud, Mohamed H.H.
    In this study, we present a novel approach to reverse polymer engineering utilizing a Graph Neural Network (GNN) framework to design low bandgap benzothiophene (BT) polymers for light harvesting applications. We have curated an extensive dataset comprising 57,556 structure-property pairs of BT-based compounds, leveraging expert knowledge to enhance the quality and relevance of the data. Our Transformer-Assisted Oriented pretrained model for on-demand polymer generation (TAO) demonstrates exceptional performance, achieving a chemical validity rate of 99.27 % in top-1 generation mode across a test set of 6000 generated polymers, marking the highest success rate reported among polymer generative models to date. Throughout the training process, the loss steadily decreased with each epoch, indicating that the model was learning effectively from the data. The model predictive accuracy is further validated by an impressive average R2 value of 0.96 for 15 defined properties, highlighting the TAO with its robust capabilities in polymer design. The newly designed polymers exhibit a bandgap range of 1.5–3.40 eV, making them promising candidates for light harvesting applications. Additionally, their highest Synthetic Accessibility Likelihood Index (SALI) scores reach up to 17 and also indicates that the majority of these polymers are amenable to synthesis. This work not only advances the field of polymer design but also provides a powerful tool for the targeted development of materials with specific electronic properties.
  • [ X ]
    Öğe
    A machine learning analysis to predict the stability driven structural correlations of selenium-based compounds as surface enhanced materials
    (Elsevier Ltd, 2025) Güleryüz, Cihat; Sumrra, Sajjad H.; Hassan, Abrar U.; Mohyuddin, Ayesha; Elnaggar, Ashraf Y.; Noreen, Sadaf
    The selenium-based compounds are gaining significance for their surface-enhanced properties. In order to accelerate their discovery, a machine learning (ML) approach has been employed to predict their structural correlations. For this a dataset of 618 compounds is collected from literature and is trained by using Support Vector Machine (SVM) with its Linear Kernal. Among ten ML evaluated models, three top-performing models are selected to make predictions for their stability energy. A Convex Hull Distribution (CHD) is constructed to elucidate the relationship for their stability and structural correlations. The main finding of this study reveals its strong correlation between stability and its related structural descriptors, particularly Bertz Branching Index" corrected for the number of Terminal atoms (BertzCT), Partial Equalization of Orbital Electronegativities-Van der Waals Surface Area with 14 bins (PEOE_VSA14), and First-Order Connectivity Index (chi 1). The analysis demon strates that the current ML models can effectively predict the stability of such materials to enable their rapid screening. Their calculations can provide a framework to understand their complex relationships between their material properties, structure, and stability.
  • [ X ]
    Öğe
    Exploring the structural basis of crystals that affect nonlinear optical responses: An experimental and machine learning quest
    (Elsevier B.V., 2025) Hassan, Abrar U.; Güleryüz, Cihat; El Azab, Islam H.; Elnaggar, Ashraf Y.; Mahmoud, Mohamed H.H.
    Machine learning can enable a computational framework to learn from data, thereby enhancing decision-making for targeted properties. Based on the significance of nonconjugated crystals as effective switches, an ML based approach has been applied to evaluate driving forces behind their polarizability/hyperpolarizability related hyper-Rayleigh Scattering (βHRS). For this, a dataset of relevant 1,3,5-triazine-2,4,6-triamine related structures in collected from peer reviewed literature to design its molecular descriptors. The designed dataset is trained on different regression models along with their cross-validation techniques include K-Fold and Leave One Group Out. It shows that Random Forest Regression can predict their polarizabilities with a fair accuracy (R2 = 0.83). Additionally, it shows its energy gaps (Egaps) ranging from 4.62 to 4.89 eV, with the smallest gap observed in ethanol. Understanding both these theoretical and experimental calculations can significantly help in selecting materials for targeted purposes, including sensors, electronic devices, and catalysis. Furthermore, insights into nucleophilic tendencies and charge distributions aids in designing new materials with tailored properties, expanding their use in various applications across chemistry, materials science, and other fields. The ML techniques prove its effectiveness to predict polarizabilities in response to its computational realm due to feature design, regression models with their cross-validations.