Scoring Methodology
Complete documentation of EvidenX metrics, algorithms, and evidence classification. Designed for reproducibility and auditability.
EvidenX Core Metrics
These metrics describe the evidence landscape — they do not recommend or score quality.
Measures the overall robustness of the evidence base. Combines volume, quality, diversity, access, and recency.
Indicates how consolidated the science is. High EMI = systematic reviews, meta-analyses, and RCTs exist.
Opportunity for new original research. High PRF = open field. Low PRF = saturated topic.
Percentage of articles with verified legal open access via Unpaywall, PMC, CORE.
Evidence Stages
Based on EMI score, we classify the research field into one of five stages:
Evidence Hierarchy Pyramid
EvidenX classifies studies into 5 layers based on methodological design.
Free Science Signal (FSS)
FSS measures verified legal open access. Unlike simple OA labels, FSS validates actual PDF availability.
Example Distribution
Validation Sources
🔍 Live Scoring Audit
Test the ARS (Anchor Relevance Score) algorithm. See exactly how articles are scored.
ARS — Anchor Relevance Score
function computeARS(article, anchors) {
let score = 0;
// Title matches (+1 each)
anchors.forEach(a => { if (title.includes(a)) score += 1; });
// Abstract matches (+2 each)
anchors.forEach(a => { if (abstract.includes(a)) score += 2; });
// Co-occurrence bonus (+2-3)
if (matchedInAbstract >= 2) score += 3;
// Recency bonus (+1 if ≤5 years)
if (currentYear - article.year <= 5) score += 1;
// Layer multiplier
score *= { A: 1.5, B: 1.5, C: 1.2, D: 1.0, E: 1.0 }[article.layer];
return score;
}Sorting Algorithm
// EvidenX sorts results by:
// 1. Open Access first (verified legal access)
// 2. Then by ARS (relevance score)
// 3. Then by Layer (A > B > C > D > E)
// 4. Then by Year (newest first)