Bot for Diagnosing Dyscalculia

Abstract

Learning disability has been gaining an increasing amount of attention in society as professionals begin to contribute student's unsatisfactory academic performance to scientific reasons other than student's ethics. While dyslexia, a learning disability for reading and writing, is well-known to the public, Dyscalculia, a learning disability that affects a person's ability to understand numbers and mathematical concept, is still not under the radar of many researchers. This project aims to propose a new method for diagnosing Dyscalculia by utilizing bot to increase public awareness and provide an alternative solution to clinical examination. With the accessible diagnostic test deployed on a website in the form of web chat, this project aims to increase the chance of victims getting the correct diagnosis. Incorporating the Algebra Game, the bot will be launched on a website with web chat enabled via Microsoft Bot Framework to serve as a pre-test for Dyscalculia. The question set is taken from the CS Challenge 2017 in which the data collected was used as a model for analysis. Upon user's request or at the end of the test, the bot will evaluate user's performance on this test by utilizing the data previously collected. Since computerized test lacks normal human interaction with a fixed message flow, the usage of a conversational bot overcomes this restraint by implementing the Language Understanding Intelligent Service (LUIS) to determine the meaning of user's input. The bot allows interruptions to be handle upon user's request and smoothen the delivery. Although this project successfully provides an alternative for the diagnostic test, it has room for further improvement such as including more variety of question format and having different question sets for different ages. Due to the limitation of data collected, this project concludes that more research on Dyscalculia should be performed to gain a better understanding of the symptoms and how it can be diagnosed.