Year: 2023
Type: article
Abstract: The importance of Educational Data Mining (EDM), a new interdisciplinary study field that builds on various other disciplines, is rising. It is directly connected to data mining (DM), which is crucial to finding knowledge in databases (KDD). This data is expanding exponentially and may include valuable hidden information for users (both teachers and students). Such knowledge can easily be recognized as models, patterns, or any other type of representational scheme that enables improved system exploitation. It is discovered that data mining can be used to make similar discoveries, giving rise to EDM. To get the best outcomes in this complicated setting, many approaches and learning algorithms are typically applied. In recent times, educational systems have witnessed a surge in using artificial intelligence (AI) systems, especially for extracting relevant information. One such AI system is EDM, which combines various techniques to support the capture, processing, and analysis of these record sets. The primary method used in EDM is machine learning, which has been applied more frequently since the emergence of big data to extract useful information from a vast amount of data. Machine learning has been used for decades in data processing in various contexts. Educational data mining tools and algorithms can be used to assess student academic achievement. This study offers a fresh approach to forecasting student success in middle school Portuguese and mathematics subjects. Hyperparameter tuning of classifiers is essential to overcome the misclassification of conventional classifiers. In order to predict student performance in the UCI dataset, this work proposes a Bayesian-optimized KNN and a random forest classifier. For random forest and KNN, the attained accuracy is 87% and 73%, respectively.
Authors: Safira Begum, Sunita S Padmannavar
Language: English
Subfield: Computer Science Applications
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: bronze
OpenAlex ID: W4379524247