Neutral Network and Fuzzy System are simulation system of intelligence of the people, but the former is from bottom to top, and the latter is from top to bottom.Both of them can establish a non-linear model when they are given a system' s input/output data.But Fuzzy System is different from Neutral Network in some aspects. Neutral Network can learn from examples and has a strong ability of studying. However, it is a set of no models. To make Neutral Network' s results satisfied, it needs a lot of trained data. On the other hand, the relationship of input/output obtained by Neutral Network is hardly to express in an easy accepted way.On the contrary, Fuzzy System is based on an easily expressed way - 'if-then' , which is easily accepted by people. Its constringency is great, because Fuzzy System is based on the knowledge of experts. However, it is a challenge on how to create and adjust the membership functions and fuzzy rules automatically.This paper studies the advantages and disadvantages of Fuzzy System and Neutral Network, and then we combine them organically. At the same time these theories are applied to The Automatically Making Test Paper-System and complete a span from theory to practice. |