UMEED

People with pre-existing health conditions or economic backgrounds find it difficult to have access to health and jobs. So, we made an affordable and accessible virtual conversation therapy for people with speech disorders such as autism spectrum disorder or aphasia post-stroke. UMEED: A VR Game Using NLP and Latent Semantic Analysis for Conversation Therapy for Patients use a solid pipeline to achieve the result. There are multiple interactive environments with avatars/chatbots to converse with to achieve a set of tasks to achieve basic proficiency in oral expressions, cognition, naming and identification, arithmetic and auditory comprehension. As the interactive environment is chosen, the user's words are recorded and converted to text in real-time. Through topic modeling, coherence, latent semantic analysis and singular value decomposition, we initiate real-time conversations. After working with an undergraduate researcher from Johns Hopkins, we arranged the data in document word matrix format and graphical analysis for easy analysis and improvement. Heavy return on investments through our efficient economic model based on short terms and long terms profits.