Study on Business Strategy Quantification Using Topic Modeling and Word Embedding: Focusing on 'Virtual Reality' and 'Augmented Reality'

摘要

This study analyzed articles related to "Virtual Reality" and "Augmented Reality," the main elements of the Fourth Industrial Revolution, and reviewed the views that the media has by using text mining techniques. For this, 14,443 news articles on "Virtual Reality" and "Augment Reality" from 2011 to 2020 were collected from the BigKinds of Korea Press Foundation. For the analysis of collected datasets, the LDA algorithm-based topical modeling techniques implemented in Python3 languages were used. The number of topics suitable for the collected dataset using coherence was determined to be 14. The calculated topic was given the title of the topic through expert survey, and word vector was created using Word2vec, one of the word-embedding techniques. This word vector was then used to quantify the analyzed topic. Through this, quantitative expressions of strategic perspectives were made and changes in topics around the year 2016 were visualized when the issue of the "fourth industrial revolution" began to be discussed.