Topic modeling and feature engineering
Could I say that topic modeling is one of the feature engineering techniques in NLP?
1 answers ( 0 marked as helpful)
Hi Abdulrahman!
You could say that but it's not exactly the same.
The extracted topics could indeed be represented as numerical values, which can then be used as features in tasks like clasification or clustering. In that sense, topic modeling can be used as a feature engineering technique since it transforms raw text into a more structured representation.
However, feature engineering usually involves manually or semi-automatically crafting features. Topic modeling is more like an unsupervised learning technique or dimensionality reduction method. So, more accurately would be to describe it as a form of feature extraction. But what it extracts could indeed be used as engineered features for ML models.
Hope that helps.
Best,
Ivan
You could say that but it's not exactly the same.
The extracted topics could indeed be represented as numerical values, which can then be used as features in tasks like clasification or clustering. In that sense, topic modeling can be used as a feature engineering technique since it transforms raw text into a more structured representation.
However, feature engineering usually involves manually or semi-automatically crafting features. Topic modeling is more like an unsupervised learning technique or dimensionality reduction method. So, more accurately would be to describe it as a form of feature extraction. But what it extracts could indeed be used as engineered features for ML models.
Hope that helps.
Best,
Ivan