Part 2/14:
Traditional search engines began with simple keyword-based retrieval methods like TF-IDF (Term Frequency-Inverse Document Frequency). These methods ranked documents based on how often certain keywords appeared, providing a basic level of relevance. However, they lacked an understanding of the actual intent behind user queries, often leading to mismatched results.
By the early 2000s, advancements introduced NLP techniques, enabling computers to process human language more effectively. Google’s autocomplete feature launched around 2004 exemplifies how understanding at the word level evolved. Later, ontologies and knowledge graphs further enriched search by capturing relationships between entities, such as linking a person to their organization.