Çekicin AI · M2 — cost model 1.2 (replace / repair + price range)
A transparent, rule-based model: damage type + size → action (polish / paintless dent repair / repaint / plastic repair / repair+paint / replace) → labour hours × country rates + paint (TR: ATO tariff with market floor) + part price quantiles from 1,100 real listings.
| file | |
|---|---|
cost_model.json |
all numbers with sources (TR / DE / GB / US) |
cost_estimate.js |
reference implementation (browser / Worker / app) |
cost_estimate.py |
Python port, tested identical to the JS on 6 cases |
import json; from cost_estimate import estimate_cost
M = json.load(open("cost_model.json"))
estimate_cost(M, [{"cls": "dent", "area": 0.05, "part": "door"}], "TR")
Evaluation: new-part range covers 57.9% of real listings out of sample (split-half backtest), midpoint error ×1.36 — dataset https://huggingface.co/datasets/sosa123454321/cekicin-m2-cost-eval. Limitations: asking prices not invoices; one reference car per country; decision thresholds not validated against assessors.
Part of the Çekicin AI demo (6 roadmap modules). Not in the Çekicin mobile app yet. AI Lab: https://sosa123454321-cekicin-site.static.hf.space/ai-lab/status.html