DEVELOPMENT OF A DIGITAL INFORMATION SYSTEM BASED ON HYBRID SEARCH FOR ASSESSING THE CONSTRUCTION CONDITION OF BUILDINGS
Abstract. The paper proposes a digital information system that automates the assessment of whether the construction condition of multi-storey buildings complies with current regulatory documents. The system is based on a formal model that links the stages from the building under inspection to the expert’s decision. Relevant code provisions are retrieved by a hybrid algorithm that combines lexical and vector search, while the levels of compliance and risk are calculated by dedicated operators. In a pilot test, the hybrid search achieved Precision@k, Recall@k and MRR values of 0.67, and the time required for a full inspection was reduced by approximately 82%. Keywords: digital information system, hybrid search, normative database, compliance operator, risk assessment, expert control. Annotatsiya. Maqolada ko‘p qavatli binolarning qurilish holatini amaldagi me’yoriy hujjatlar talablariga muvofiqligini baholash jarayonini avtomatlashtiruvchi raqamli axborot tizimi taklif etiladi. Tizim tekshirilayotgan obyektdan ekspert qaroriga qadar bo‘lgan bosqichlarni bog‘lovchi formal modelga tayanadi. Kerakli me’yoriy bandlarni topish uchun leksik va vektorli qidiruvni birlashtirgan gibrid algoritm qo‘llanilgan, muvofiqlik va xavf darajasi esa alohida operatorlar yordamida hisoblanadi. Pilot sinovda gibrid qidiruvning Precision@k, Recall@k va MRR ko‘rsatkichlari 0,67 ni tashkil etdi, to‘liq tekshiruvga sarflanadigan vaqt esa taxminan 82 % ga qisqardi. Kalit so‘zlar: raqamli axborot tizimi, gibrid qidiruv, me’yoriy baza, muvofiqlik operatori, xavfni baholash, ekspert nazorati.
Authors
- Qo'rg'onova Zulayxo Rustamjon qizi
- Yunusaliyev Elmurad Mahammatyakubovich
Institutions
- Namangan davlat texnika universiteti (UZ)
- Fergana State Technical University (UZ)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-10-31
- DOI
- https://doi.org/10.5281/zenodo.23837052
- Primary Topic
- Engineering and Information Technology
- Type
- article
- Field-Weighted Citation Impact
- 0.00