QQP Model

Manage finishing-level parameters, weights, and model versions.

Active model — v102
Trained on 0 dossiers · 01/06/2026 · MAE: · R²:

Active QQPs

36 active parameters used in finishing level extraction.

NameCategoryTypeWeightDossiersSourceActive

Solar Panel Count

solar_panel_count

numeric
+0.329
AI discovered

Hvac System Type

hvac_system_type

numeric
+0.000
AI discovered

Building Use Type

building_use_type

numeric
+0.000
AI discovered

Is Heritage Protected

is_heritage_protected

numeric
+0.220
AI discovered

Total Livable Area

total_livable_sqm

Size & Layout
numeric · m2
+0.203
Seeded

Entrance Hall Size

entrance_hall_sqm

Size & Layout
numeric · m2
+0.210
Seeded

Living Room Size

living_room_sqm

Size & Layout
numeric · m2
+0.311
Seeded

Kitchen Size

kitchen_sqm

Size & Layout
numeric · m2
+0.536
Seeded

Master Bedroom Size

master_bedroom_sqm

Size & Layout
numeric · m2
+0.457
Seeded

Average Bedroom Size

avg_bedroom_sqm

Size & Layout
numeric · m2
+0.465
Seeded

Largest Bathroom Size

largest_bathroom_sqm

Size & Layout
numeric · m2
+0.181
Seeded

Garage Size

garage_sqm

Size & Layout
numeric · m2
+0.206
Seeded

Terrace/Balcony Size

terrace_balcony_sqm

Size & Layout
numeric · m2
+0.551
Seeded

Circulation Ratio

circulation_ratio

Size & Layout
numeric · ratio
+0.190
Seeded

Floor Count

floor_count

Size & Layout
numeric · count
+0.510
Seeded

Bedroom Count

bedroom_count

Room Count & Composition
numeric · count
+0.147
Seeded

Bathroom Count

bathroom_count

Room Count & Composition
numeric · count
+0.120
Seeded

Separate Toilet Count

toilet_count

Room Count & Composition
numeric · count
+0.132
Seeded

Bathroom/Bedroom Ratio

bathroom_per_bedroom_ratio

Room Count & Composition
numeric · ratio
+0.509
Seeded

Has Separate Dining

has_separate_dining

Room Count & Composition
boolean
-0.121
Seeded

Has Office/Study

has_office

Room Count & Composition
boolean
+0.476
Seeded

Has Dressing Room

has_dressing

Room Count & Composition
boolean
+0.111
Seeded

Has Laundry Room

has_laundry_room

Room Count & Composition
boolean
+0.313
Seeded

Has Wellness Area

has_wellness

Room Count & Composition
boolean
+0.099
Seeded

Has Basement

has_basement

Room Count & Composition
boolean
+0.259
Seeded

Has Garage

has_garage

Room Count & Composition
boolean
+0.155
Seeded

Kitchen Appliance Count

kitchen_appliance_count

Equipment & Features
numeric · count
+0.574
Seeded

Has Kitchen Island

has_kitchen_island

Equipment & Features
boolean
-0.106
Seeded

Bathroom Luxury Score

bathroom_luxury_score

Equipment & Features
score · score
+0.476
Seeded

Has Fireplace

has_fireplace

Equipment & Features
boolean
+0.193
Seeded

Has Open Kitchen

has_open_kitchen

Equipment & Features
boolean
+0.233
Seeded

Built-in Storage Count

built_in_storage_count

Equipment & Features
numeric · count
+0.288
Seeded

Living/Total Ratio

living_to_total_ratio

Proportionality
numeric · ratio
+0.140
Seeded

Wet Room/Total Ratio

wet_room_to_total_ratio

Proportionality
numeric · ratio
+0.582
Seeded

Outdoor/Indoor Ratio

outdoor_to_indoor_ratio

Proportionality
numeric · ratio
+0.294
Seeded

Average Room Size

avg_room_size

Proportionality
numeric · m2
-0.486
Seeded

Proposed QQPs

AI-suggested parameters not yet activated. Accept to create a new QQP definition; reject to dismiss.

No pending proposals. Process more dossiers to generate suggestions.

Model versions

History of weight calibration runs. The latest version is used for re-evaluating dossiers.

v102
Active
0 dossiers · 01/06/2026
MAE RMSE

connect-v1 de-biased: intercept 1.2824→1.0274 (−0.255) to center the F-median on the CED-median 1.04 (€2150), removing the systematic +0.255 / +19% overshoot. Weights unchanged from v101. Rollback: set v101 is_active=true.

v1010 dossiers · 30/05/2026
MAE RMSE

Connect-Value gewichten (mei 2026) + APPARTEMENT-CENTRERING. Intercept 1.2824 = doel-F 0.96 voor gemiddeld appartement (compenseert systematische negatieve QQP-bias t.o.v. huis-geijkte score-0). Gewichten uit Connect-Excel; intercept gecentreerd op 15-dossier appartement-gemiddelden. Σw·gem=-0.3224.

v225 dossiers · 21/05/2026
MAE 0.059RMSE 0.0660.9345 predictions

Ridge regression (λ=1) from 5 dossiers. MAE=0.059, R²=0.934.

v2124 dossiers · 20/05/2026
MAE 1.014RMSE 1.188-0.90214 predictions

Auto-calibrated from 24 dossiers. 34 QQPs calibrated.

v2024 dossiers · 20/05/2026
MAE 1.014RMSE 1.188-0.90214 predictions

Auto-calibrated from 24 dossiers. 34 QQPs calibrated.

v1923 dossiers · 20/05/2026
MAE 1.014RMSE 1.188-0.90214 predictions

Auto-calibrated from 23 dossiers. 34 QQPs calibrated.

v1824 dossiers · 20/05/2026
MAE 1.014RMSE 1.188-0.90214 predictions

Auto-calibrated from 24 dossiers. 34 QQPs calibrated.

v1724 dossiers · 20/05/2026
MAE 1.014RMSE 1.188-0.90214 predictions

Auto-calibrated from 24 dossiers. 34 QQPs calibrated.

v1624 dossiers · 20/05/2026
MAE 1.014RMSE 1.188-0.90214 predictions

Auto-calibrated from 24 dossiers. 34 QQPs calibrated.

v1522 dossiers · 20/05/2026
MAE 1.058RMSE 1.227-1.25013 predictions

Auto-calibrated from 22 dossiers. 34 QQPs calibrated.