The controls, also called the constraints (inward connection to the top edge of the box), of the process include software and hardware limitations, privacy issues, and the difficulties related to processing of the text that is presented in the from of natural language.
The mechanisms (inward connection to the bottom edge of the box) of the process include proper techniques, software tools, and domain expertise.
The primary purpose of text mining (within the context of knowledge discovery) is to process unstructured (textual) data (along with structured data, if relevant to the problem being addressed and available) to extract meaningful and actionable patterns for better decision making.