arXiv subjects

Reginald Neil C. Recario

Publications and source records attributed to Reginald Neil C. Recario.

3 recordsLinked to original sources

arXiv2026

Determinants of Starting Salaries for Filipino Graduates: An Explainable Machine Learning Approach

Filipino graduates face a persistent disconnect between educational preparation and labor market outcomes, where starting salary is a key signal of entry-level valuation. Current Philippine research is dominated by descriptive tracer studies that document employment rates but do not explain the determinants of pay. We address this gap using a crowd-sourced survey dataset of graduate responses whose noisy, self-reported nature makes it a challenging prediction target. Applying machine learning to this problem, we identify job role and industry as the dominant determinants of starting salary, significantly outweighing institutional prestige. The strength of this finding is its central contribution: it is corroborated by three independent lines of evidence, namely SHAP attributions, the heavy reliance of the best ensemble on occupational text, and a Natural Language Inference reformulation. These results suggest that career guidance and policy should prioritize sector-specific skills over institutional brand.

cs.CY↗

arXiv2026

Is the Statistical Advantage Worth the Cost? An Empirical Comparison of KANs and MLPs for Structured Data Classification

This study presents an empirical benchmarking comparison between Kolmogorov-Arnold Networks (KANs) and Multi-Layer Perceptrons (MLPs) on structured tabular classification tasks. Motivated by the growing interest in KANs as an alternative function-approximating architecture, we evaluate their out-of-the-box performance on twelve publicly available datasets spanning binary, multiclass, multilabel, and ordinal problems. Both models were trained under standardized preprocessing, architecture, and fixed hyperparameter settings, with performance assessed using test accuracy and F1-Score, paired hypothesis testing, and effect size analysis. Results show that KANs statistically outperform MLPs in binary and multiclass domains and achieve a significant aggregate advantage across all datasets. However, the observed medium effect size (d = -0.46) raises an important cost-benefit consideration: while KANs offer superior generalization through adaptive spline-based mappings, this advantage comes with substantially higher parameter and computational complexity relative to the MLP baseline. These findings suggest KANs are the preferred choice for high-precision applications, while MLPs remain a robust and efficient option for resource-constrained environments. Future work should extend this analysis to additional data modalities to further refine these architectural selection criteria.

cs.LG↗

arXiv2015

Social Loafing Among Members of Undergraduate Software Engineering Groups: Persistence of Perception Seven Years After

We surveyed 169 undergraduate students who are enrolled in various courses. They were members of software engineering groups formed to solve various real-world computational problems by implementing software projects as part of the requirements of the course. This time, our analysis show that task visibility is negatively associated with social loafing while contributions, dominance, aggression and sucker effect are positively correlated. We further found out that perception of social loafing exists and still persists among members of computer programming groups. Compared to our 2008 analysis, we provide in this paper detailed analysis based on demographic parameters such as gender, course taken, age group, type of residence (urban or rural), and region of residence. The implication of this result is that aside from the usual problems that an instructor faces in teaching software engineering-related courses, the presence of social loafing also adds to the impediment of teaching effectiveness. Thus, it is imperative that instructors and course designers consider the implications associated with social loafing when designing group projects.

cs.CY↗