<![CDATA[Micro, small, and medium enterprises (MSMEs) continue to face low digital-technology adoption, with conventional incubation programs still relying on generic training and manual assessment rather than automated, evidence-based evaluation. This study develops and validates an AI-based startup incubation model (AI-SIM), an integrated ICT platform functioning as an intelligent decision-support system for MSME digital-readiness classification, adaptive training recommendation, geospatial analysis, and student-MSME experiential learning within one operational environment. Data were collected from 4,276 MSMEs across Lampung Province, Indonesia; after strict two-level deduplication to prevent data leakage, 961 unique records were retained for modeling. A random forest classifier grouped MSMEs into three readiness categories (Traditional, Stable, Digital Ready) using two feature configurations, achieving 33.68% (10 features; 5 fold CV 36.75%) and 30.57% accuracy (7 manuscript-aligned features; 5 fold CV 37.23%), both near the 33.3% random-chance baseline. Feature importance analysis showed predictive weight spread across all seven features, led by digital infrastructure and monthly revenue, with no dominant predictor. The system was validated through field deployment, training, and focus group discussions with 84 student and MSME participants, confirming its practical value as a real-time incubation dashboard, while the classification component requires richer behavioral and transactional data to improve accuracy in future iterations.]]>
Copyrights © 2026
