2025-02-282025-02-282022-02-21MOREIRA, Igor Matheus Souza. On reducing the dimensionality of small molecule data for visual-exploratory analysis in human intestinal absorption prediction. Orientador: Claudomiro de Souza de Sales Júnior 20222. 137 f. Trabalho de Conclusão de Curso (Bacharelado em Ciência da Computação) – Faculdade de Computação, Instituto de Ciências Exatas e Naturais, Universidade Federal do Pará, Belém, 2022. Disponível em:. Acesso em:.https://bdm.ufpa.br/jspui/handle/prefix/7754Oral bioavailability is a desirable property in drug development. Virtual screening of compounds according to their properties with computational intelligence can accelerate the prediction of their human intestinal absorption (HIA). Despite the existence of studies aimed at predicting HIA of compounds, dimensionality reduction (DR) techniques that extract features are seldom employed to enable visual-exploratory analyses and pre-process data for machine learning (ML) algorithms. This work applies six DR projectors (ivis, KPCA, PCA, PCS, TSVD, and UMAP) to produce two- and three-dimensional projections alongside four ML classifiers (KNN, MLP, RF, and SVM) in predicting HIA of small molecules, an effort that encompassed the analysis of fifty-two pipelines. Results demonstrate that, despite reducing the dimensionality by more than 98%, DR-encompassing pipelines still delivered competitive results while also facilitating visualization, demonstrating the viability and potential of DR via feature extraction as an automated pre-processing step.Acesso AbertoQuimioinformáticaFarmacêutica computacionalRedução de dimensionalidadeDescoberta e desenvolvimento de drogasExtração de característicasAbsorção intestinal humanaAprendizado de máquinaChemoinformaticsComputational pharmaceuticsDimensionality reductionDrug discovery and developmentFeature extractionHuman intestinal absorptionMachine learningCNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAOTrabalho de Curso - Graduação - Monografia