Estimation of understanding of words by lifelog analysis for vocabulary visualizing

Abstract

In this study, we made an estimate understanding of words for visualization of individual vocabulary. We have hypothesized that each of words which a person understand and does not are concentrating on certain word concept. Therefore, we proposed a method for estimating understanding of words with hierarchical neural network and word2vec, and succeeded in estimating the understanding of words by accuracy of over about 96%. From these results, it has been shown the validity of the hypothesis and that it is possible to estimate the understanding of word by life log analysis.