{
  "stage": "Stage 01 — Supervised Classification — Linear/Probabilistic",
  "task": "Booking cancellation risk",
  "target": "target_cancelled",
  "synthetic_training_data": true,
  "production_warning": "The artifacts are technically deployable, but predictive quality is not production-validated until retrained and evaluated on real HRP history.",
  "feature_columns": [
    "city_id",
    "property_type",
    "star_rating",
    "hotel_avg_rating_prior",
    "hotel_reviews_count_prior",
    "hotel_amenity_count",
    "bed_type",
    "bed_count",
    "room_size_sqm",
    "view_type",
    "smoking_allowed",
    "breakfast_included",
    "base_price_per_night",
    "effective_price_per_night",
    "price_to_base_ratio",
    "max_occupancy",
    "total_quantity",
    "available_quantity_at_booking",
    "availability_ratio_at_booking",
    "minimum_stay_nights",
    "stop_sell",
    "is_refundable",
    "guests_count",
    "rooms_count",
    "number_of_nights",
    "total_price",
    "days_until_checkin",
    "checkin_month",
    "checkin_dayofweek",
    "checkin_weekend",
    "customer_prior_bookings",
    "customer_prior_cancellations",
    "customer_prior_cancel_rate",
    "customer_prior_completed",
    "customer_prior_no_show",
    "customer_avg_prior_spend",
    "customer_favorites_count",
    "customer_has_favorited_hotel",
    "customer_account_age_days",
    "customer_last_login_recency_days",
    "email_verified",
    "phone_verified"
  ],
  "categorical_columns": [
    "city_id",
    "property_type",
    "bed_type",
    "view_type",
    "checkin_month",
    "checkin_dayofweek"
  ],
  "numeric_columns": [
    "star_rating",
    "hotel_avg_rating_prior",
    "hotel_reviews_count_prior",
    "hotel_amenity_count",
    "bed_count",
    "room_size_sqm",
    "smoking_allowed",
    "breakfast_included",
    "base_price_per_night",
    "effective_price_per_night",
    "price_to_base_ratio",
    "max_occupancy",
    "total_quantity",
    "available_quantity_at_booking",
    "availability_ratio_at_booking",
    "minimum_stay_nights",
    "stop_sell",
    "is_refundable",
    "guests_count",
    "rooms_count",
    "number_of_nights",
    "total_price",
    "days_until_checkin",
    "checkin_weekend",
    "customer_prior_bookings",
    "customer_prior_cancellations",
    "customer_prior_cancel_rate",
    "customer_prior_completed",
    "customer_prior_no_show",
    "customer_avg_prior_spend",
    "customer_favorites_count",
    "customer_has_favorited_hotel",
    "customer_account_age_days",
    "customer_last_login_recency_days",
    "email_verified",
    "phone_verified"
  ],
  "non_feature_columns": [
    "booking_id",
    "created_at",
    "check_in_date",
    "check_out_date",
    "user_id",
    "hotel_id",
    "room_type_id",
    "target_cancelled"
  ],
  "leakage_columns_never_use": [
    "status",
    "confirmed_at",
    "rejected_at",
    "cancelled_at",
    "completed_at",
    "hotel_response_note",
    "booking_status_history.to_status after created_at"
  ],
  "best_classifier": "logistic_regression",
  "risk_probability_model": "lda",
  "risk_threshold": 0.265,
  "risk_levels": {
    "low": [
      0.0,
      0.35
    ],
    "medium": [
      0.35,
      0.65
    ],
    "high": [
      0.65,
      1.0
    ]
  }
}