ABSTRACT
This study presents a single, mobile-based Senior High School Strands Recommendation System that harnesses prescriptive analytics—specifically a Restricted Boltzmann Machine (RBM)—to convert Grade 10 students’ interests, aptitudes, and key performance indicators into transparent, prioritized strand suggestions (e.g., STEM, ABM, HUMSS, TVL), addressing the persistent challenge in the Philippines—underscored by Pascual (2014)—of students struggling to select tracks that genuinely align with their strengths; by pairing concise, student-friendly data gathering with an RBM that learns latent patterns across profiles, the system not only predicts affinity but also prescribes next steps, delivering ranked options with short rationales that explain why a strand rises to the top, how specific competencies map to typical course demands, and which targeted actions could mitigate any gaps; the mobile delivery ensures access at home, in school, and during guidance sessions, while a clear interface, plain-language tooltips, and example pathways help students connect current choices to plausible futures, building confidence rather than pressure; for counselors and administrators, cohort-level dashboards summarize common interest–aptitude mismatches, reveal areas where orientation or remediation may help, and support more focused, empathetic conversations without replacing professional judgment; methodologically, the RBM’s compact architecture enables efficient learning from modest datasets, and the prescriptive layer translates modeled affinities into actionable recommendations that students can immediately use, emphasizing timeliness, clarity, and agency; ethically, the design follows minimal and relevant data collection, explicit consent, opt-out controls, and continuous bias checks that compare recommendations across demographic slices and adjust feature weights when disparities emerge, while model performance is monitored through validation against subsequent student satisfaction, persistence, and early outcomes within chosen strands; practically, the application guides users through a brief questionnaire, returns results in seconds, and links each recommended strand to concise descriptions of core subjects, skill emphasis, workload expectations, and illustrative careers, encouraging reflection and informed dialogue with families and mentors; strategically, the project aims to lower decision anxiety, strengthen fit between learners and learning environments, and increase the efficiency of guidance resources by fusing evidence, personalization, and explainability in a format students can trust; ultimately, by narrowing information gaps, elevating self-knowledge, and giving schools a scalable, data-informed ally, the system seeks to improve match quality in strand selection, enhance engagement and self-efficacy during Senior High School, and contribute over time to more coherent educational trajectories and career development for Filipino youth.
DOI 10.5281/zenodo.17452873
