Introduction
Fact Retrieving It requires the retrieval system not only from data (fact) collection data or facts, but also from existing basic data or facts Deliberate, deduct new data or facts. For example, there is a fact that there is a fact that Li Ming is a student at school. Students in 2A schools learn foreign languages. If the system is a fact retrieval system, it should be able to answer "Li Mingxue Foreign Language?" Asked for a user. Truth search is the most complex in intelligence retrieval. It requires data and facts in the system to store in a natural language or close to natural language. Not only must be deposited in various data or factual units, but also deposits the semantic relationship, syntact relationships, and various relevant background knowledge. Allow users to ask for natural language and use natural language answers. More importantly, the system must have a certain logical reasoning ability and natural language understanding.
Practice
The fact search is a fairly complex process. It is often done by manuality. The specific approach is: First use the retrieval tool, reference tool book, database or other channels to find relevant original data, facts, or literature, then analyze comparisons, to cut fine, go to fake, and finally provide "facts" to users. Practical computerization fact retrieval system is being developed, and some experimental systems have been established. For example, Wa Woods and others have designed a natural language intelligence search system called Lunar in 1972. The purpose is to help geologists compare and evaluate Apollo-11 moon rockets on moon rock and soil composition ingredients. Chemical analysis data. It can syntactic analysis, semantic interpretation, and practical answers to geologist commonly used questions, with certain practicality.
