نموذج تجريبي فقط — لا تدخل بيانات حقيقية أو رقمًا مدنيًا حقيقيًا. لا دفع ولا إصدار وثائق ولا رسائل فعلية.
DEMO ONLY — use fictional data. No real payments, policies, or SMS.
Random Forest · AUC 0.9433 · n = 1,999

Pre-Purchase Insurance Risk Scoring — Omani Motor Market
02555100
—
Risk Score
—
—
—
1
Risk Score = Σ (wi × Si)

= wage·Sage + wexp·Sexp + wregion·Sregion + wcar_type·Scar_type + wcar_age·Scar_age + wclaims·Sclaims

wi = Random Forest feature-importance weight · Si = raw risk score of the feature value (0–100)
2
Σ wi = 0.23 + 0.22 + 0.20 + 0.13 + 0.12 + 0.10 = 1.00
3
4
Score < 25
Claim rate 1.3% · n = 678
25 ≤ Score < 55
Claim rate 22.8% · n = 631
Score ≥ 55
Claim rate 86.5% · n = 690
5
6
Model: Random Forest Classifier (scikit-learn)
Dataset: n = 1,999 (synthetic, Omani market patterns)
Split: 80/20 stratified
Performance: AUC = 0.9433, Recall = 0.8600, F1 = 0.8037
Cross-validation: 5-fold CV AUC = 0.9332 ± 0.0062
Confusion matrix: TN = 208, FP = 42, FN = 21, TP = 129
Musab Albadi · UWE Bristol · UFCFMJ-15-M
Random Forest Classifier · 6 Pre-Purchase Features