Download models/model_metadata.json from AL-mozahem/HRP-Stage07-Occupancy-Forecast: direct link, hf CLI and curl.
- Browser
- Download file 3.35 kB
-
https://huggingface.co/AL-mozahem/HRP-Stage07-Occupancy-Forecast/resolve/main/models/model_metadata.json
- Command line
-
hf download hf://AL-mozahem/HRP-Stage07-Occupancy-Forecast/models/model_metadata.json
-
curl -L -o model_metadata.json https://huggingface.co/AL-mozahem/HRP-Stage07-Occupancy-Forecast/resolve/main/models/model_metadata.json
3.35 kB
| { | |
| "stage": "Stage 07 — Tree-Based + Ensemble Regression", | |
| "task": "Forecast realized occupancy rate for the next 7 days by room type", | |
| "target": "target_next_7d_occupancy_rate", | |
| "feature_columns": [ | |
| "city_id", | |
| "property_type", | |
| "star_rating", | |
| "hotel_avg_rating_prior", | |
| "hotel_reviews_count_prior", | |
| "hotel_amenity_count", | |
| "hotel_prior_confirmation_rate", | |
| "hotel_recent_rejection_rate", | |
| "favorite_count_prior", | |
| "bed_type", | |
| "bed_count", | |
| "room_size_sqm", | |
| "view_type", | |
| "breakfast_included", | |
| "is_refundable", | |
| "smoking_allowed", | |
| "max_occupancy", | |
| "room_amenity_count", | |
| "total_quantity", | |
| "base_price_per_night", | |
| "prior_7d_occupancy_rate", | |
| "prior_30d_occupancy_rate", | |
| "prior_90d_occupancy_rate", | |
| "prior_7d_booking_count", | |
| "prior_30d_booking_count", | |
| "prior_30d_cancel_rate", | |
| "prior_30d_no_show_rate", | |
| "prior_30d_avg_lead_days", | |
| "prior_30d_avg_nights", | |
| "prior_30d_avg_guests", | |
| "prior_30d_avg_sold_price", | |
| "prior_30d_revenue", | |
| "next_7d_avg_planned_rate", | |
| "rate_to_base_ratio", | |
| "minimum_stay_nights", | |
| "stop_sell_days_next_7d", | |
| "on_books_room_nights", | |
| "on_books_occupancy_next_7d", | |
| "available_room_nights_at_snapshot", | |
| "availability_ratio_next_7d", | |
| "forecast_month", | |
| "forecast_dayofweek", | |
| "forecast_weekend_days", | |
| "holiday_share_next_7d", | |
| "annual_sin", | |
| "annual_cos", | |
| "days_since_last_booking" | |
| ], | |
| "categorical_columns": [ | |
| "city_id", | |
| "property_type", | |
| "bed_type", | |
| "view_type", | |
| "forecast_month", | |
| "forecast_dayofweek" | |
| ], | |
| "numeric_columns": [ | |
| "star_rating", | |
| "hotel_avg_rating_prior", | |
| "hotel_reviews_count_prior", | |
| "hotel_amenity_count", | |
| "hotel_prior_confirmation_rate", | |
| "hotel_recent_rejection_rate", | |
| "favorite_count_prior", | |
| "bed_count", | |
| "room_size_sqm", | |
| "breakfast_included", | |
| "is_refundable", | |
| "smoking_allowed", | |
| "max_occupancy", | |
| "room_amenity_count", | |
| "total_quantity", | |
| "base_price_per_night", | |
| "prior_7d_occupancy_rate", | |
| "prior_30d_occupancy_rate", | |
| "prior_90d_occupancy_rate", | |
| "prior_7d_booking_count", | |
| "prior_30d_booking_count", | |
| "prior_30d_cancel_rate", | |
| "prior_30d_no_show_rate", | |
| "prior_30d_avg_lead_days", | |
| "prior_30d_avg_nights", | |
| "prior_30d_avg_guests", | |
| "prior_30d_avg_sold_price", | |
| "prior_30d_revenue", | |
| "next_7d_avg_planned_rate", | |
| "rate_to_base_ratio", | |
| "minimum_stay_nights", | |
| "stop_sell_days_next_7d", | |
| "on_books_room_nights", | |
| "on_books_occupancy_next_7d", | |
| "available_room_nights_at_snapshot", | |
| "availability_ratio_next_7d", | |
| "forecast_weekend_days", | |
| "holiday_share_next_7d", | |
| "annual_sin", | |
| "annual_cos", | |
| "days_since_last_booking" | |
| ], | |
| "selected_algorithm": "catboost", | |
| "selected_params": { | |
| "depth": 6, | |
| "iterations": 120, | |
| "l2_leaf_reg": 5, | |
| "learning_rate": 0.055 | |
| }, | |
| "model_kind": "catboost_native", | |
| "synthetic_training_data": true, | |
| "target_window": "forecast_start through forecast_end (7 days)", | |
| "leakage_rule": "Every feature must be known at snapshot_at. Bookings created after snapshot_at belong only to the target outcome.", | |
| "production_warning": "Retrain and temporally validate on real HRP snapshots before production demand forecasting." | |
| } |