Enhanced Accuracy for Lung Adenocarcinoma (LUAD) Prediction based UMAP Feature Using Artificial Neural Network

The Lung adenocarcinoma (LUAD), a common histopathological manifestation of carcinoma of the lungs as well as a variant, grouped as Non-Small Cell Lung Cancer (NSCLC), makes up 45-5% of every instance of cancer of the lungs. Different variables, notably environmental exposures and genetic makeup have been discovered to contribute to the onset and advancement of LUAD. It has been shown that combining the expression of genes with other information can help detect lung cancer patients. It offers several more perspectives that improve the categorization of cancers. Based on the results, it is believed that identifying the genes that have an extensive expression in malignant cells as opposed to typical ones is a difficulty, which calls for the application of computational tools. Applied computing techniques encountered more problems with this data set because of its elevated resolution and small sample size. Many supervised and unsupervised educational strategies have been developed for the classification of GED cancer. Since the most important traits remain unidentified, ML techniques have not been very successful in reducing dimensionality or classifying malignancy in GED patients. This research aims to address these challenges by leveraging the power of Uniform Manifold Approximation and Projection (UMAP) to enhance the feature space representation. This paper offers a unique Artificial Neural Network (ANN) model to predict, particularly LUAD types of cancer among the other gene expression data such as BRCA, COAD, KIRC, and PRAD cancer. The results of the proposed UMAP with ANN model 3 demonstrate for the detection of LUAD data when evaluated using performance measures has the highest accuracy of 99.53%.