Knowledge Building via Optimally Clustered Word Embedding with Hierarchical Clustering

Given a particular domain, drawing out information from a vast amount of data is not an easy task. Data may be adequate and abundant but analysis of data requires a great deal of work. Therefore, manual construction of a knowledge base on any particular domain is more time consuming and it requires much human intervention. The process can be semi-automated by effectively using word embedding to identify the semantics. The words can then be clustered using hierarchical clustering. This study proposes a novel approach on how knowledge discovery can be semiautomated for the restaurant domain by effectively identifying the optimal number of clusters from hierarchical clustering of words extracted from restaurant reviews.