Large-Scale Weekly Evaluation

Up to 100 stocks × 4 models (GROQ · OpenAI · Meta · LSTM), 7-day horizon. Full prediction rows are stored in PostgreSQL. When each target date arrives, rows auto-match to actual prices and per-model accuracy recalculates.

Loading…

Run

—

Status

—

Total Predictions

—

Matched

—

Pending (awaiting date)

—

Per-model accuracy (matched rows only)

Model	Samples	Matched	Pending	Dir %	MAE	MAPE	RMSE	R²	≤5%	≤10%

Full predictions table

All stock × model rows from PostgreSQL

Model Status

#	Symbol	Model	Pred Date	Target	Current	Predicted	Direction	Actual	Abs Err	% Err	Status

—