|
Name | Evaluation method | Advantage | Disadvantage |
|
Subjective Logic Model | Express the influence of trust parameter factor on trust value in fact space and idea space | More in line with common sense | Difficulty dealing with massive collaborative cheating and defamation |
Information Entropy Model | Constructing a trust evaluation algorithm based on entropy increment | Vulnerability was assessed and characterized | Using entropy to represent trust is not comprehensive enough |
Weighted Average Model | Fit the weighted average formula of each parameter factor and directly calculate the trust value | The method is simple to execute and the algorithm is efficient | The weight determination process is complicated |
Bayesian Model | Use Bayes’ theorem to combine prior probabilities with new evidence to get new probabilities | The method is simple and the algorithm is efficient. | Poor subjectivity |
Fuzzy Theory | Updating trust vectors using fuzzy logic | Takes into account the “fuzzy” properties of trust itself | Difficulty making rules |
Game Theory | Establish a game model for the acquisition of trust information of nodes | Comprehensive evaluation and high accuracy | The game model building process is complicated |
Machine Learning | Build a machine learning model as an evaluation algorithm | Strong intuition, high evaluation accuracy, closer to “human” evaluation | Large amount of calculation and low evaluation efficiency |
|