Cloud services for children early development via wearable devices
Date of Issue2016
School of Computer Engineering
In the early child development, what the parents talk to infants would affect different aspects of their mental development. In order to help parents to trace the mental development and understand the natural language environment around infants, we decided to train machines to identify what the attitude of parents when they are talking to infants and what the influences on baby of their words. Five categories, emotion, intelligence, character, language and art was classified and used to identify the influences on infants. According to the comparison among four different algorithms of Part-of-Speech tagging, Trigrams’n’Tags (TnT) was chosen as the core method of semantic analysis. In the implementation of sentiment analysis, a data-driven algorithm was designed. Because the training materials is the most pivotal element in the machine learning algorithms, the most time-consuming part of this project, tagging training materials, took a very significant role to increase the accuracy of sentiment and semantic identification. The design and implementation of sentiment and semantic analysis enhance the availability and usability of the application to help users in analyzing what they said to infants and improving what they would say to infants.
Final Year Project (FYP)
Nanyang Technological University