Framework for Thought to Text Classification
1R. Manasa, Suchita Ghose, R. Ragasudha and P. Vijayakumar
People with neurological disorders are unable to communicate their basic requirements because they lost the ability to speak. Designing a brain-computer interface that could convey their basic needs would make their lives easier. This article presents a system to determine the patient’s imagined words in the brain without him/her physically expressing by EEG signal and machine learning. The imagined words in the mind to text mapping is converted into a classification problem among a predesigned set of words and classified by using machine learning algorithms. The decoding/classification of EEG signals to identify imagined words is carried out KNN classifier, and Random forest classifier. The classification accuracy shows that the random forest classifier achieved better classification accuracy in comparison with KNN.
EEG Signals, Machine Learning, Imagined Words, KNN Classifier, Random Forest Classifier.