LQM Registry / decision
Evidence-Backed Adaptive Model Router
Selects a champion or adaptive ensemble from candidate performance histories using recency weighting, optional regime conditioning, minimum-evidence guards and switching margins.
Model key
lq.decision.adaptive-routerVersion1.0.0
Runtimepython-process
Model contract
Inputs are explicit.
{
"type": "object",
"required": [
"candidates"
],
"properties": {
"candidates": {
"type": "object"
},
"current_regime": [],
"incumbent": {
"type": [
"string",
"null"
]
}
}
}Output contract
Outputs are explicit.
{
"type": "object",
"required": [
"status",
"selected_candidate",
"switch",
"ensemble_weights",
"candidate_scores",
"decision_basis"
]
}Default parameters
{
"direction": "lower",
"decay": 0.9699999999999999733546474089962430298328399658203125,
"regime_weight": 0.34999999999999997779553950749686919152736663818359375,
"switch_margin": 0.0200000000000000004163336342344337026588618755340576171875,
"min_observations": 3,
"eta": 2
}Example input
{
"current_regime": "high-vol",
"incumbent": "ar1",
"candidates": {
"ar1": [
{
"value": 0.6999999999999999555910790149937383830547332763671875,
"regime": "low-vol"
},
{
"value": 0.9499999999999999555910790149937383830547332763671875,
"regime": "high-vol"
},
{
"value": 0.90000000000000002220446049250313080847263336181640625,
"regime": "high-vol"
}
],
"trend": [
{
"value": 0.75,
"regime": "low-vol"
},
{
"value": 0.59999999999999997779553950749686919152736663818359375,
"regime": "high-vol"
},
{
"value": 0.57999999999999996003197111349436454474925994873046875,
"regime": "high-vol"
}
]
}
}