산림식물자원 기능성의 학술적 근거에 기반한 가치평가 방법론 및 위계분기 알고리즘 개발

초록

Background: Forest plant resources have been valued through economic indicators, while a framework for quantifying the strength of evidence (SoE) for their bioactivities has remained underdeveloped. Existing approaches typically provide resource-effect mapping without quantitative evidence grading or are restricted to a single study design and risk-of-bias (RoB) assessment framework. Objectives: We developed a semi-automated, human-in-the-loop evidence assessment methodology and hierarchy-branching algorithm for extracting and grading evidence on the functional value of forest plant resources, demonstrated through a proof-of-concept case. Methods: A six-step procedure (S1~S6) was designed: (S1) a forest plant resource ontology with composite identifiers combining scientific name, synonyms, taxonomy, parts, origin, and extraction conditions, harmonized with a South Korean plant ontology and agro-food thesaurus; (S2) a hierarchical ontology of seven functionality themes (i.e., functional categories at the ontology level) and seven study-design tiers, each linked to a predefined risk-of-bias (RoB) assessment instrument (Cochrane RoB 2, ROBINS-I, Newcastle- Ottawa Scale, JBI, SYRCLE, or ToxRTool); (S3) automated search-query kit generation for PubMed, Scopus, ScienceON, and KCI; (S4) literature collection with deduplication; (S5) three-tier rule-based matrix extraction; and (S6) hierarchy-branching SoE grading on a seven-grade scale. The methodology was applied to the anti- inflammatory functionality of Lonicera caerulea fruit. Results: Twenty-three articles were retrieved across four databases. After re-classification of one study (Bell 2019) from case-control to quasi-clinical, the quasi-clinical tier formed the highest tier satisfying the threshold of ≥2 low-or-medium RoB articles. After auxiliary-factor adjustment for consistency, effect size, precision, directness, and dose–response relationship, the final SoE grade was Possible-convincing. Tier-specific evidence counts and the grading trace were stored as a reproducible audit trail. Conclusions: The methodology converts fragmented literature into ontology-aligned, hierarchy-aware SoE grades through a reproducible, traceable workflow, complementing existing evidence-synthesis tools and supporting forest-resource R&D, functional-food evaluation, and environmental-health risk-benefit assessment and evidence integration.

키워드