Accurate　Human　Activity　Recognition　(HAR)　is　a　critical　challenge　with　wide-ranging　applications　in　healthcare,　assistive　technologies,　and　human-computer　interaction.　Traditional　feature　extraction　methods　often　struggle　to　capture　the　complex　spatial　and　temporal　dynamics　of　human　movements,　leading　to　suboptimal　classification　performance.　To　address　this　limitation,　this　study　introduces　a　novel　encoding　approach　using　Locality-Constrained　Linear　Coding　(LLC)　to　enhance　the　discriminative　power　of　hand-crafted　features　extracted　from　low-cost　wearable　sensors—an　accelerometer　and　a　gyroscope.　The　proposed　LLC-based　encoding　scheme　enables　robust　feature　representation,　improving　the　accuracy　of　HAR　models.　The　encoded　features　are　classified　using　a　diverse　set　of　Machine　Learning　(ML)　and　Deep　Learning　(DL)　algorithms,　including　Support　Vector　Machine　(SVM),　Logistic　Regression　(LR),　Naive　Bayes　(NB),　K-Nearest　Neighbours　(KNN),　AdaBoost,　Gradient　Boosting　Machine　(GBM),　and　Deep　Belief　Network　(DBN).　Extensive　quantitative　evaluations　demonstrate　that　LLC　significantly　outperforms　conventional　feature　encoding　techniques,　leading　to　improved　classification　accuracy.　Among　the　tested　models,　DBN　achieves　a　state-of-the-art　accuracy　of　99%,　highlighting　its　superiority　for　HAR　tasks.　The　contributions　of　this　research　are　threefold:　(1)　it　establishes　the　necessity　of　an　advanced　encoding　scheme　(LLC)　for　feature　enhancement　in　HAR,　(2)　it　provides　a　rigorous　comparative　analysis　of　multiple　ML　and　DL　classifiers,　and　(3)　it　introduces　a　scalable　and　cost-effective　HAR　framework　suitable　for　real-world　applications.　Performance　is　comprehensively　assessed　using　robust　evaluation　metrics　such　as　Area　Under　the　Curve　(AUC),　Classification　Accuracy　(CA),　F1　Score,　Precision,　Recall,　and　Matthews　Correlation　Coefficient　(MCC).　The　findings　of　this　study　offer　new　insights　into　feature　encoding　for　HAR,　setting　a　foundation　for　future　advancements　in　sensor-based　activity　recognition.　©　2025　The　Author(s).　Healthcare　Technology　Letters　published　by　John　Wiley　&　Sons　Ltd　on　behalf　of　The　Institution　of　Engineering　and　Technology.