Alzheimer’s Disease using BERT Text Classification

Abstract

The progression of Alzheimer's disease is relentless, leading to a worsening of mental faculties over time. Currently, there is no remedy for this illness. Accurate detection and prompt intervention are pivotal in mitigating the progression of the disease. Recently, researchers have been developing new methods for detecting Alzheimer at earlier stages, including genetic testing, blood tests for biomarkers, and cognitive assessments. Cognitive assessments involve a series of tests to measure memory, language, attention, and other brain functions. Although there is still no definitive test for Alzheimer, research is ongoing and new techniques are being developed. For disease detection, optimal performance necessitates enhanced accuracy coupled with efficient computational capabilities. Our proposition involves, after data augmentation of textual data from Kaggle , it will then be analyzed using a BERT-based deep learning model in an effort to take use of its advanced capabilities for improved feature extraction and text comprehension. Our model is able to accurately detect Alzheimer's disease from textual data. We conduct a thorough assessment of our proposed BERT-based deep learning model for text categorization using a dataset made up of patient-reported medical records in order to determine its efficacy. In our comparison analysis, we compare our model to cutting-edge machine learning techniques frequently used for text classification tasks, such as recurrent neural networks (RNNs), convolutional neural networks (CNNs), and long short-term memory (LSTM) networks. This thorough evaluation intends to evaluate the performance and superiority of our suggested model in terms of precision, computational effectiveness, and capability to successfully capture complex textual patterns in the medical domain. Our result showed that our BERT-based model outperforms previously implemented methods based on BERT-CNN and BERT-RCNN in terms of accuracy, precision, and recall. Additionally, we used ensemble state-of-art techniques that allowed us to leverage the collective intelligence of the ensemble and make highly accurate and reliable predictions on the dataset.