| Election type | Approval |
|---|---|
| Culture | Identity (ID) |
| Candidates | {50} |
| Voters | {100} |
| If applicable, committee size is 10 | |
| Instances | None |
| We have: 250 elections from the Disjoint Model; 225 elections from the Noise Model with Hamming distance; 225 elections from the Truncated Urn Model; 200 elections from Euclidean Model. We use (p, φ)-resampling elections as a background dataset, which consists of 241 elections 250 + 225 + 225 + 200 + 241 = 1141 | |
| Parameters | None |
| the Disjoint Model (50 for each $g \in \{2, 3, 4, 5, 6\}$ with $\phi \in (0.05, 1/g)$); Noise Model with Hamming distance (25 for each $p \in \{0.1, 0.2, . . . , 0.9\}$ with $\phi \in (0, 1)$); Truncated Urn Model (25 for each $p \in \{0.1, 0.2, . . . , 0.9\}$ with $\alpha \in (0, 1)$); Euclidean Model (100 for 1D-Uniform, with radius in (0.0025, 0.25), and 100 for 2D-Square, with radius in (0.005, 0.5)); $(p, \phi)$-resampling elections, with $p$ and $\phi$ parameters as follows: 1. $p$ is chosen from $\{0, 0.1, 0.2, . . . , 0.9, 1\}$ and $\phi$ is chosen from the interval $(0, 1)$, $1$ or 2. $\phi$ is chosen from $\{0, 0.25, 0.5, 0.75, 1\}$ and $p$ is chosen from the interval $(0, 1)$. | |
| Notes | Sections 4, 5, 6. Experiment involved 1) computing max approval score, 2) computing max cohesiveness level, 3) computing the number of voters belonging to at least one cohesive group, 4) measuring PAV runtime. Disjoint model is based on resampling model |