Majority Judgment for Typescript

This typescript package helps to resolve polls using Majority Judgment.

Features

• Efficient Majority Judgment algorithm, scales well to billions of participants
• Configure whether to favor adhesion or contestation (default)
• Balance proposal tallies using a static default grade or the median grade
• Room for Central Judgment and Usual Judgment
• Unit-tested (run npm run converage)

Install

npm install scalable-majority-judgment

Get started

Majority judgment from ballots

import {
    TallyCollector,
    MajorityJudgmentDeliberator,
    IDeliberator,
    IResult,
} from "majority-judgment";

const proposalAmount: number = 5;
const mentionAmount: number = 7;
const voterAmount: number = 1000;
const tally: TallyCollector = new TallyCollector(proposalAmount, mentionAmount);
fillTallyCollectorWithRandomVote(tally, voterAmount);

const deliberator: IDeliberator = new MajorityJudgmentDeliberator();
const result: IResult = deliberator.deliberate(tally);

for (let i: number = 0; i < proposalAmount; i++)
    console.log(
        `Proposal at index ${i} obtains the rank ${result.proposalResults[i].rank} with the majority mention at index ${result.proposalResults[i].analysis.medianMentionIndex}`
    );

function fillTallyCollectorWithRandomVote(tally: TallyCollector, voterAmount: number): void {
    const applyRandomVoteForEachProposal = () => {
        let randomProposalIndex: number;
        let randomMentionIndex: number;

        for (let i: number = 0; i < proposalAmount; i++) {
            randomProposalIndex = Math.floor(Math.random() * proposalAmount);
            randomMentionIndex = Math.floor(Math.random() * mentionAmount);
            tally.collect(randomProposalIndex, randomMentionIndex);
        }
    };

    for (let i: number = 0; i < voterAmount; i++) applyRandomVoteForEachProposal();
}

Known datas

import {
    MajorityJudgmentDeliberator,
    IDeliberator,
    IResult,
    ITally,
    Tally,
    Proposal,
} from "majority-judgment";

const meritProfileSample: bigint[] = [4n, 0n, 2n, 1n];
// index 0 = worst mention, 4 vote
// max index (3) = best mention, 1 vote

const tally: ITally = new Tally([
    new Proposal(meritProfileSample),
    new Proposal([2n, 1n, 0n, 4n]),
    new Proposal([3n, 2n, 1n, 1n]),
]);

const deliberator: IDeliberator = new MajorityJudgmentDeliberator();
const result: IResult = deliberator.deliberate(tally);
const proposalAmount: number = tally.proposalAmount;

for (let i: number = 0; i < proposalAmount; i++)
    console.log(
        `Proposal at index ${i} obtains the rank ${result.proposalResults[i].rank} with the majority mention at index ${result.proposalResults[i].analysis.medianMentionIndex}`
    );

// Note: the sum vote in proposal must be equal. If not, the deliberator will throw an error.

Majority judgment where voters did not vote for every proposal

import { ITally, Proposal, NormalizedTally } from "majority-judgment";

// If the sum vote in proposal are not equal. You can use NormalizedTally.
const tally: ITally = new NormalizedTally([
    new Proposal([4n, 0n, 2n, 10n]),
    new Proposal([20n, 10n, 0n, 4n]),
    new Proposal([3n, 200n, 100n, 1n]),
]);

// (...) Same logic

Contribute

Usual git flow: clone, tinker, request a merge.

Authors

This package was made by MieuxVoter, a French nonprofit.