| [Visitor (223.146.*.*)]answers [Chinese ] | Time :2020-03-04 | Core: Dempster synthesis rules-Fusion of multiple subjects (which can be predictions of different people, data from different sensors, output from different classifiers, etc.)
Advantages: 1. The required prior data is more intuitive and easier to obtain than in probabilistic reasoning theory (I don't know why); it meets a weaker condition than Bayes' probability theory, that is, "the probability does not have to be met.
2. Can fuse a variety of data and knowledge (as described in the core)
3. Has the ability to directly express "uncertain" and "don't know". This information is expressed in the mass function and is retained during the synthesis of evidence (not yet felt).
Disadvantages: 1. The evidence must be independent
2. There is no very strong theoretical support for the rules of evidence synthesis, and there is still considerable controversy over its rationality and validity.
3. "Exponential explosion problem" in calculation
4. The results obtained in some cases are contrary to common sense, such as "Zadeh's Paradox" |
|
