In　recent　years,　a　rapid　rise　in　the　incidence　of　Large　for　gestational　age　(LGA)　neonate　is　reported,　and　health　care　professionals　employed　themselves　to　discover　the　cause.　Utmost　of　the　previous　studies　were　cohort　or　observational　studies　that　employed　simple　linear　or　multivariate　regression　models,　and　very　few　of　them　employed　machine　learning　(ML)　schemes.　Therefore,　this　research　proposes　to　use　1　expert-driven　and　7　automated　feature　selection　schemes　with　well-known　ML　classifiers　using　10　and　30　folds　cross-validation.　The　induced　results　were　compared　with　existing　baselines,　and　Wilcoxon　signed-rank　test　and　the　Friedman　test　were　also　introduced　to　verify　the　results.　The　ranked　20　features　of　the　proposed　expert-driven　feature　selection　scheme　outperformed　amongst　7　automated　feature　selection　schemes　with　a　prediction　precision,　accuracy,　and　AUC　scores　of　0.94606,　0.84529,　and　0.86492,　respectively.　Out　of　twenty　features,　eleven　features　were　found　similar　to　twenty　ranked　features　of　the　automated　feature　selection　schemes　subsets.　The　classification　results　of　the　extracted　features　were　utmost　identical　to　the　results　of　twenty　features　subset　proposed　by　the　expert-driven　feature　selection　scheme.　Therefore,　we　suggest　pediatricians　to　refresh　LGA　diagnosis　process　with　the　proposed　scheme　because　of　its　practical　usage　and　maximum　expert　involvement.