TF-IDF is a widely used text vectorization technique for feature extraction in machine learning. It converts text data, such as entity names, into numerical feature vectors. The technique calculates a weight for each term by multiplying its Term Frequency (how often it appears in the text) by its Inverse Document Frequency (downweighting common terms across the entire dataset). This process emphasizes the importance of unique, semantically meaningful terms over common words, making it crucial for analyzing entity relevance and creating precise relationship graphs.