Using Machine Learning Algorithms to Detect Dysplasia in Barrett's Esophagus

Abstract

Machine learning is a division or branch of mathematical models that are used to learn data to generate computerized algorithms. This concept could be a predictive model. Machine learning is part of the learning key function that can predict other noisy data. One possible method is to separate the two types according to the features of the measurement content. These models can be used to cause people associated with the disease. In our research, different algorithms are applied to the data set to correctly predict dysplasia. This opens the door to standardized training and qualitative analysis steps for staff performing endoscopy in Barrett's esophagus. In this study, Five classification techniques were used in the implementation of principal component analysis, e.g. K-Nearest Neighbors, SVM (Support Vector Machines). The basic goal is to measure the exact value of information about output and feasibility, with each evaluation being for accuracy, recall, and specificity. The exploratory output shows the support of vector analysis with PCA. The single score for the K neighbor (0.97) and the SVM value is 0.91.