IR systems follow the process of matching the user's query with the index terms representing the documents in the data collection. There are different similarity models (Boolean, Vector, and Probabilistic similarities). These similarity models vary from being geared toward supporting the exact result set matches to supporting approximate matches (ranking).

Which one of these models you think most practical to match users query to documents (primarily unstructured data) from the internet of things (IoT) using general-purpose search engines? Discuss the rationale behind your choice of the selected model.