| Summary: | It is indispensable to understand where the development of the Environmental-Social-Governance (further- ESG) concept is moving. People need to know where to develop in the future, and also understand whether the development of the ESG concept is going the right way. The primary objectives of the paper are to analyze ESG terminology from various sources, constract the ESG concept development model, and discover the relationship between tokens and the papers’ ESG presentation sequence. Four compression methods were considered in the paper: Doc2Vec from the Gensim library, MDS with cosine distance measure, tSNE with cosine distance measure, pre-compression of texts using Doc2Vec into a ten-dimensional space, followed by compression using the tSNE method with a cosine distance measure. The research analysis result shows 65 tokens and 14 bigrams. The experiment showed that dividing paragraphs into 70 clusters achieves the best clustering quality. The results show that the most vital link is between the words “ESG” and “firms”, as well as “stock”, “value”, “social”. “environmental”, and “risk”. It also found a strong connection with tokens “investment decisions” and “performance”, “significant company” and “effect”. With the increasing interconnections links between the “ESG-value-environmental-significant- influence”, the authors suppose that investors are becoming more environmentally conscious when making investment decisions. Although some challenges persist, including inconsistency in terminology, a considerable amount of data to analyze and heterogenous rating standards, ESG investing is likely to play an essential role in new environmentally-oriented values.
|
|---|