matter
Built from the latest training run on 2026-09-30 17:45 UTC, not from a released model set. The released documentation is at https://matter.indigo.engineering/docs/.
Matter builds 23 machine learning models predicting outcomes across four domains: affordability and family, benefits enrollment, employment, and health.
Each model page includes key metrics, decile lift charts, feature importance, and prediction distributions.
Every classifier score also comes with a likelihood segment: Most likely, Somewhat likely, or Least likely. The segments come in their own segments table, under the same column names as the scores. Most likely holds the top X of records by score, where X is the model’s prevalence (its mean score) up to a cap of 0.33. Somewhat likely holds the next X, and Least likely holds the rest.
A machine-readable catalog of these models, with the same metrics, is at models.json.
models
affordability & family excessive costs: child care auc: 0.90
affordability & family excessive costs: groceries auc: 0.67
affordability & family monthly housing cost r²: 0.30
affordability & family utility shut-off warning auc: 0.76
affordability & family has children at home auc: 0.85
affordability & family owns home auc: 0.87
benefits enrolled in aca auc: 0.71
benefits enrolled in chip auc: 0.76
benefits enrolled in medicaid auc: 0.71
benefits enrolled in medicare auc: 0.90
benefits enrolled in snap auc: 0.75
benefits receives eitc auc: 0.88
employment is employed auc: 0.83
employment w2 hourly status auc: 0.73
employment manual occupation auc: 0.76
employment professional occupation auc: 0.78
employment service occupation auc: 0.71
health does not visit pcp auc: 0.68
health insurance: employer sponsored auc: 0.78
health insurance: government sponsored auc: 0.84