TOP2VEC: New way of topic modellingFew years back, it was very difficult to extract Subjects/Topics/Concepts of thousands of unannotated free text documents. Best and simple way was to make some human sit, go thru each articles, understand and annotate Topics. Indeed it was time consuming and prone to subjectivity of perception we humans have. Although many attempts were made in past with simple algorithms like pLSA to treat this a
