<![CDATA[This study addresses a critical limitation in ascariasis surveillance by introducing a mathematical modeling framework based on non-laboratory diagnostic indicators for early risk prediction in resource-constrained settings. Conventional detection methods rely heavily on laboratory diagnostics, which are often inaccessible in rural healthcare systems, leading to delayed intervention and persistent transmission. To overcome this challenge, the study develops a K-nearest neighbors (KNN) model interpreted as a distance-based classification approach within an epidemiological risk space, where similarity reflects shared exposure patterns among individuals. The model utilizes secondary data from 315 cases with laboratory-confirmed outcomes and incorporates key variables including sex, body mass index, livestock exposure, waste disposal practices, and environmental risk factors. Model performance was evaluated using 10-fold cross-validation across multiple configurations (k = 1, 3, 5, 7, 9) and assessed through AUC, accuracy, precision, recall, F1-score, and Matthews correlation coefficient. The results demonstrate that the k = 9 configuration achieves the most balanced performance, with an AUC of 0.924 and F1-score of 0.852, indicating strong discriminative capability and classification stability. The study contributes to epidemiological risk modeling by formalizing KNN within a structured risk-space interpretation and demonstrates that non-laboratory indicators can support reliable prediction. This approach offers a computationally efficient and interpretable tool for rural health surveillance and data-driven decision-making.]]>
