LQM Registry / research
Walk-Forward AR(1) Research Study
Out-of-sample one-step AR(1) evaluation against a naive last-value baseline with explicit look-ahead guard and LQBench skill metrics.
Model key
lq.research.walk-forward-ar1Version1.0.0
Runtimepython-process
Model contract
Inputs are explicit.
{
"type": "object",
"required": [
"values"
],
"properties": {
"values": {
"type": "array",
"items": {
"type": "number"
},
"minItems": 20
}
}
}Output contract
Outputs are explicit.
{
"type": "object",
"required": [
"evaluation_protocol",
"rmse",
"mae",
"naive_rmse",
"skill_score_rmse_vs_naive",
"directional_accuracy",
"lookahead_guard"
]
}Default parameters
{
"initial_window": 12,
"expanding_window": true,
"rolling_window": 12
}Example input
{
"values": [
100,
100.5,
101.2000000000000028421709430404007434844970703125,
101,
101.7999999999999971578290569595992565155029296875,
102.400000000000005684341886080801486968994140625,
102.099999999999994315658113919198513031005859375,
103,
103.400000000000005684341886080801486968994140625,
103.900000000000005684341886080801486968994140625,
104.099999999999994315658113919198513031005859375,
104.7999999999999971578290569595992565155029296875,
105.2000000000000028421709430404007434844970703125,
105,
105.7000000000000028421709430404007434844970703125,
106.2999999999999971578290569595992565155029296875,
106.7999999999999971578290569595992565155029296875,
107,
107.599999999999994315658113919198513031005859375,
108.099999999999994315658113919198513031005859375,
108,
108.7999999999999971578290569595992565155029296875,
109.2999999999999971578290569595992565155029296875,
109.7000000000000028421709430404007434844970703125
]
}