628 article

 628

🎤 Texto que debes decir — Slide 1 (Title)

 

Good morning everyone.

My name is Augusto Paolo Bernal Parraga, and today I will present our research titled:

“Explainable Rule-Based Computational Framework for Early Detection of Presuicidal Behavioral States.”

This study proposes an interpretable computational framework designed to support the early identification of presuicidal behavioral manifestations in adolescent educational contexts.

The research was conducted in secondary education institutions in Manabí, Ecuador, using behavioral observations, psychosocial surveys, interviews, and institutional records from 200 adolescents.

Unlike opaque predictive systems, our approach emphasizes explainability, transparency, and rule-based interpretation through finite behavioral states and deterministic transitions.

The framework models behavioral variability, psychosocial vulnerability, and emotional instability as structured computational states capable of supporting preventive educational decision-making.

Importantly, the objective of the framework is not clinical diagnosis, but rather the support of early preventive intervention processes in educational environments.

 

🎤 Texto que debes decir — Slide 2 (Background and Related Work)

Adolescent presuicidal behavior represents an important public health and educational challenge worldwide.

One of the main difficulties in educational contexts is that warning signs often emerge gradually through subtle behavioral changes rather than explicit verbal expressions.

Indicators such as emotional instability, school disengagement, irritability, social withdrawal, and behavioral variability may appear before severe crisis episodes.

Because of this, schools become critical environments for early preventive identification and psychosocial monitoring.

However, many current preventive practices still rely heavily on subjective observation and non-standardized interpretation methods.

At the same time, recent advances in explainable artificial intelligence and educational decision-support systems have created new possibilities for interpretable behavioral monitoring.

In sensitive domains such as adolescent mental health, explainable and rule-based systems are generally preferred over opaque machine learning approaches because educators and professionals need to understand how classifications are generated.

Despite these advances, most existing computational approaches focus either on clinical populations or on black-box predictive models that prioritize prediction accuracy over interpretability.

Therefore, this study addresses an important research gap by proposing an interpretable rule-based computational framework based on observable psychoeducational indicators and behavioral transitions in school environments.

🎤 Texto que debes decir — Slide 3 (Methodology)

This study followed an exploratory case-control design focused on the computational representation of presuicidal behavioral manifestations in adolescents.

The research was conducted in 19 public secondary education institutions in Manabí, Ecuador, involving a total of 200 adolescents.

To obtain a broader contextual perspective, information was collected not only from students, but also from teachers, legal guardians, and institutional records.

The study combined multiple data collection methods, including participant observation, psychosocial surveys, structured interviews, and institutional behavioral reports.

After data collection, the observed manifestations were organized into psychoeducational indicators associated with behavioral variability and psychosocial vulnerability conditions.

The framework relies on four major dimensions:

school-related behavioral changes,
emotional and verbal manifestations,
family and social interaction patterns,
and psychosocial vulnerability indicators.

Rather than treating behavioral manifestations as isolated observations, the framework categorizes them according to persistence, intensity, and frequency.

This allowed behavioral variability to be transformed into structured computational input suitable for rule-based classification and state-transition analysis.

The main objective was to develop an interpretable computational framework capable of modeling behavioral states and transitions to support preventive educational decision-making.

🎤 Texto que debes decir — Slide 4 (Results — Behavioral Variability)

The first stage of the analysis focused on detecting observable behavioral variability among adolescents.

The study involved 200 adolescents from 19 public secondary schools in Manabí, Ecuador.

The results showed that 46 percent of adolescents exhibited observable behavioral changes during the monitoring period.

In addition, 30.5 percent presented psychosocial vulnerability indicators associated with presuicidal behavioral risk.

Among the most frequent associated factors were emotional instability, psychosocial school stress, disengagement from academic activities, and social withdrawal.

These findings suggest that behavioral variability is relatively common in adolescent educational contexts and may provide valuable early-warning information for preventive intervention.

The framework then classified adolescents into four typological behavioral states represented by the finite-state structure.

These states combined two dimensions:

emotional regulation
and observable behavioral variability.

From a computational perspective, this allowed behavioral manifestations to be represented as interpretable and structured preventive states suitable for educational monitoring systems.

Overall, the findings support the relevance of explainable rule-based approaches for early preventive identification in educational environments.

 

🎤 Texto que debes decir — Slide 5 (Proposed Computational Framework)

This slide presents the proposed explainable rule-based computational framework developed in this study.

The framework receives two main categories of input:

psychoeducational indicators
and behavioral variability indicators.

These indicators include emotional instability, social withdrawal, school disengagement, psychosocial stressors, and observable behavioral changes over time.

The information is processed through a Rule-Based Transition System, or RBT, which classifies adolescents into interpretable behavioral states.

One of the main advantages of the framework is its explainability.

Unlike black-box predictive systems, the classification process is based on explicit logical rules and observable conditions that educators and professionals can understand and interpret.

The framework is also structured around a finite-state behavioral model composed of four typological states.

State A represents calm behavior without observable changes.

State B corresponds to restless behavior without behavioral changes.

State C represents calm behavior with behavioral changes.

And State D corresponds to restless behavior accompanied by behavioral changes and emotional instability.

Behavioral variability is modeled through deterministic transitions between these states according to persistence, intensity, and psychosocial vulnerability conditions.

Overall, the framework transforms complex behavioral manifestations into structured computational states capable of supporting preventive educational monitoring and decision-support processes.

 

🎤 Texto que debes decir — Slide 6 (Results — Framework Evaluation)

This slide summarizes the evaluation of the proposed rule-based computational framework.

The results demonstrated that the framework was capable of consistently classifying behavioral typological states using interpretable rule-based transitions.

One important observation is the strong alignment between observed behavioral conditions and the framework’s computational classifications across the four behavioral states.

This indicates that the finite-state representation successfully captured meaningful behavioral variability patterns in educational contexts.

Another key contribution of the framework is its interpretability.

Unlike opaque predictive systems, the classification logic is transparent and understandable because decisions are generated through explicit behavioral rules and observable psychoeducational indicators.

The analysis also showed that emotional and verbal indicators generated the strongest contribution to behavioral classification, followed by school-related indicators and family-social conditions.

Psychosocial vulnerability indicators additionally contributed to identifying higher-risk behavioral configurations.

From a preventive perspective, the framework demonstrated practical applicability for educational monitoring and early intervention support.

The results suggest that explainable rule-based approaches may provide a feasible alternative for sensitive educational contexts where transparency, ethical alignment, and professional supervision are essential.

Overall, the framework combines interpretability, structured behavioral modeling, and preventive applicability within a computationally explainable architecture.

 

🎤 Texto que debes decir — Slide 7 (Framework Application in Schools)

This slide illustrates how the proposed framework can be applied within real educational environments.

The process begins with the collection of information from multiple educational sources.

These sources include student behavioral data, psychosocial surveys, structured interviews, teacher observations, and institutional records.

The collected information is then processed through the Rule-Based Transition System, or RBT framework.

The framework analyzes behavioral indicators and psychosocial conditions using interpretable transition rules between behavioral states.

Instead of generating opaque predictions, the system classifies adolescents into understandable preventive behavioral conditions.

This allows educational professionals to identify behavioral variability, psychosocial vulnerability, and transitions toward higher-risk states.

Once risk conditions are identified, the framework supports several preventive educational actions.

These include early intervention processes, psychosocial support, continuous monitoring, and institutional follow-up strategies.

One important advantage is that the framework is designed to support professional decision-making rather than replace clinical or educational judgment.

Overall, the system demonstrates how explainable computational approaches can contribute to safer educational environments, preventive support systems, and improved student well-being.

🎤 Texto que debes decir — Slide 8 (Discussion)

The discussion of the findings highlights the importance of explainability in sensitive educational contexts related to adolescent behavioral monitoring.

One important consideration is that presuicidal behavioral manifestations are highly dynamic, multifactorial, and context-dependent.

Because of this complexity, educational professionals require systems that are transparent, interpretable, and ethically aligned with preventive educational practices.

In this context, the proposed rule-based framework offers several important advantages.

First, the behavioral classifications are generated through explicit rules and controlled transitions rather than opaque predictive mechanisms.

This improves transparency and allows educators, counselors, and institutional professionals to understand the logic behind each behavioral state.

Second, the framework prioritizes preventive educational support instead of deterministic clinical prediction.

The objective is not to label students, but rather to identify behavioral variability patterns that may require additional monitoring, support, or psychosocial intervention.

Another important contribution is the integration of computational modeling with educational practice.

The framework bridges psychoeducational observation, institutional monitoring, and explainable computational representation within a unified preventive architecture.

Finally, the results suggest that explainable rule-based approaches may represent a more appropriate alternative for educational environments where ethical responsibility, human supervision, and contextual interpretation are essential.

🎤 Texto que debes decir — Slide 9 (Limitations and Future Work)

Although the proposed framework demonstrated promising results, several limitations must be considered.

First, the study was conducted within a specific regional context involving 200 adolescents from public secondary schools in Manabí, Ecuador.

Therefore, broader validation across different educational and cultural contexts is still necessary.

Another limitation is that part of the information relied on observational and psychoeducational reports, which may introduce variability in interpretation and behavioral assessment.

In addition, although the rule-based structure improves interpretability, some rare or highly individualized behavioral manifestations may not be fully represented within the current transition system.

It is also important to emphasize that the framework is not intended to function as a clinical diagnostic system.

Instead, its role is to support preventive educational monitoring and early intervention processes under professional supervision.

Regarding future work, several important directions emerge from this study.

Future research could incorporate larger and more diverse datasets from multiple educational regions and institutional contexts.

Additional psychoeducational indicators, digital behavior patterns, and academic engagement variables could also strengthen the framework.

Another promising direction is the integration of explainable artificial intelligence methods with rule-based architectures to create hybrid interpretable systems.

Finally, future longitudinal studies and real-world institutional deployments could help evaluate long-term preventive impact and practical applicability in educational monitoring systems.

 

🎤 Texto que debes decir — Slide 10 (Conclusion)

In conclusion, this study demonstrates that explainable rule-based computational approaches can provide valuable support for early preventive identification in adolescent educational contexts.

The proposed framework successfully transformed psychoeducational indicators and behavioral variability into interpretable computational states and transitions.

One of the main contributions of the framework is its transparency.

Because classifications are generated through explicit rules and observable conditions, educational professionals can understand and interpret the behavioral logic behind each state.

The framework also demonstrated practical applicability for preventive educational monitoring and early intervention support.

Importantly, the system was designed to complement professional judgment rather than replace human decision-making.

Another important contribution is the integration of computational modeling, psychoeducational observation, and preventive educational practice within a unified explainable architecture.

Overall, the findings suggest that interpretable computational frameworks may represent an important direction for safer educational environments, earlier preventive support, and improved adolescent well-being.

Thank you very much for your attention.