ML Market Map — Korea Equity Clusters for 2026-07-16

A daily unsupervised machine-learning read of the Korea Equity market: 1,166 stocks grouped into 7 consensus clusters (KMeans + Gaussian-mixture + hierarchical, over robust-scaled PCA features) for 2026-07-16. Descriptive, not predictive — there is no buy or sell signal. Research, not investment advice.

Markets: US Equity · China A-Share · Japan Equity · Hong Kong Equity · Taiwan Equity · Korea Equity · UK Equity · India Equity

Today's clusters

• high cash yield · cheap (high E/P) · high margin — 456 names; mostly Consumer Cyclical; drivers: val cfop z (+0.87), val ep z (+0.68), profit margin (+0.54)
• falling earnings · expensive (low E/P) · high leverage — 271 names; mostly Industrials; drivers: ni growth (−1.10), val ep z (−0.89), cfo growth (−0.88)
• high turnover · high margin · high beta — 236 names; mostly Industrials; drivers: turnover to mcap (+0.80), profit margin (+0.74), beta 252 d (+0.72)
• low margin · expensive (low E/P) · rising earnings — 87 names; mostly Healthcare; drivers: profit margin (−4.58), val ep z (−1.93), ni growth (+1.05)
• strong 1y momentum · high turnover · rising (3m) — 60 names; mostly Technology; drivers: ret 252d (+3.88), ret 126d (+3.09), turnover to mcap (+2.63)
• rising (3m) · high volatility (3m) · high turnover — 30 names; mostly Consumer Cyclical; drivers: vol of vol 63 (+3.45), ret 63d (+3.29), rv 63 (+2.96)
• high downside vol · expensive (low E/P) · high volatility (3m) — 26 names; mostly Technology; drivers: vol of vol 63 (+3.69), downside vol 63 (+3.27), val ep z (−3.17)

Machine-readable data (free, read-only JSON)

The full map, per-ticker cluster assignments with confidence and anomaly scores, and PCA structure are published as open JSON for automated and AI-analyst consumption:

• /api/ml/latest?market=kr — clusters, names, sector mix, representative tickers, top features
• /api/ml/tickers?market=kr — per-ticker cluster, confidence, anomaly score, day-over-day migration, tenure
• /api/ml/pca?market=kr — PCA components, explained variance, feature loadings

Descriptive market-structure research only. Unsupervised clustering finds structure, not direction; a tight cluster or an anomaly is a starting point for research, never a trade signal.