Skip to content

Latest commit

 

History

History
16 lines (16 loc) · 1.97 KB

README.md

File metadata and controls

16 lines (16 loc) · 1.97 KB
Raw

Collective intelligence research

The repository contains data and scripts for the study From Prediction Markets to Interpretable Collective Intelligence by Alexey V. Osipov and Nikolay N. Osipov (arXiv:2204.13424 [cs.GT]).

  • CI.RData is an R workspace that contains some data preloaded from our database into tables data12 and data13 (we use R x64 4.1.2 for Windows). This workspace also contains all our scripts from the files described below.

Data

In spite of the fact that our goal is the creation of a play-money prediction system for scientific research, we rely on the detailed historical data of the largest real-money prediction market. We believe that their great volume and level of detail are necessary for our studies.

  • Data/pmdata.backup is a backup of our database (we use PostgreSQL x64 9.3 for Windows).
  • Data/InCSV.7z contains the same database in CSV format.
  • Data/DataHandling.r contains scripts for uploading the data into R and for further working with them. Before uploading the PostgreSQL data, they must be restored from Data/pmdata.backup.

Model

We implement and verify our ideas in R.

  • Model/MuSgmEstimation.r contains scripts for eliciting estimates of μ and σ from market states.
  • Model/GetGraph.r contains scripts for obtaining graphs of μ and σ estimates where "time" is aggregated V + and V -.
  • Model/VotingModel.r contains scripts for simulation of our model.
  • Model/ModelExample.r contains scripts for reproducing Table 3 from the article. This file should be loaded from CI.RData workspace.

Demonstration application

In order to see some of our analysis in real time via a shiny app, download all the files, open CI.RData and use source to read app.R.