AI-Driven Career Guidance System: A Predictive Model for Student Subject Recommendations Based on Academic Performance and Aspirations

Abstract

This study presents an AI-based predictive model designed to forecast potential career paths for students based on academic performance, extracurricular activities, and personal aspirations. Traditional career guidance methods, such as standardized assessments and counselling, often lack the personalization necessary to cater to individual student needs, particularly in a rapidly changing educational landscape. To overcome these limitations, this research employs advanced neural network architectures, specifically Encoder-Decoder Long Short-Term Memory (LSTM) models, to analyse data on academic performance, demographic factors, and student aspirations. The dataset used in this study includes academic scores in math, reading, and writing, along with demographic information such as gender, parental education level, and test preparation status. The data was pre-processed to handle missing values, standardize scores, and encode categorical variables, ensuring the model's inputs were clean and consistent. Feature engineering was performed to extract meaningful insights, such as average performance across subjects, deviations from class averages, and alignment of performance with potential career paths. The core of the model is an Encoder-Decoder LSTM architecture designed to capture temporal dependencies and sequence patterns within the student data. The encoder processes sequences of academic scores and demographic features, generating a context vector that encapsulates the student's academic profile. The decoder then uses this vector to produce recommendations for potential career paths and subject choices, offering personalized guidance tailored to each student's strengths and aspirations. Key findings indicate that the AI-driven recommendations closely align with traditional guidance while offering enhanced personalization. The model achieved a high accuracy rate, demonstrating its ability to understand and predict optimal subject choices for students. Comparisons with traditional methods showed that the AI model provided more individualized guidance, reflecting each student’s unique academic journey and personal goals. This research highlights the transformative potential of AI in educational decision-making, providing actionable insights that can help students make more informed choices about their academic and career paths. Ethical considerations, such as data privacy and bias mitigation, were addressed throughout the study to ensure responsible AI deployment. Measures were taken to anonymize student data and regularly audit the model for biases related to gender and socio-economic status. These steps are critical in maintaining the integrity and fairness of AI-driven recommendations in educational contexts. This research underscores the importance of integrating ethical guidelines into AI systems, ensuring they serve as trusted tools that enhance, rather than undermine, educational equity and personalization.