Loading...
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 44745 (19)
- University: Sharif University of Technology
- Department: Computer Engineering
- Advisor(s): Habibi, Jafar; Abtahi, Ebrahim
- Abstract:
- Nowadays, prediction markets and decision markets are among the common methods for predicting future events. There are many studies on these markets and some models are developed for analyzing them. Score rule is a classic method that is used for many years for information elicitation in weather forecasting, economical predictions and predicting sport games’ results. In this thesis, a study on score rules, prediction and decision markets was conducted. These studies resulted in a new model for predicting the value of a variable in future. This model was designed as a proper score rule in its primary form. Then features which were discussed in this new literature were defined, and it was discussed how to use defined score rules in a market scoring rule. These studies were also conducted in markets for deciding and a score rule was introduced for this context. In addition, a case study for predicting results of 11th Iran presidential election in 2013 was performed. This case study predicted the order of nominees in election correctly and it predicted the percentage of different nominees’ and spoilt votes with average absolute error of 6.53%. This prediction had a better performance compared with a notable number of other predictions
- Keywords:
- Prediction Market ; Decision Making Market ; Election Prediction ; Variable Value Prediction ; Truthful Mechanism ; Score Rule ; Game Theory
Digital Object List
-
محتواي کتاب
- view
Bookmark