Applying Machine Learning In the Educational Sector to Support Weak Performing Students

Description

Machine learning techniques are widely used to support applications in different sectors such as banking, finance, education, engineering, and medicine. In the educational sector, a significant amount of data is being generated for managing, educating, supporting, and maintaining the sustainability of institutions. The application of machine learning in the educational domain is also called the Educational Data Mining (EDM). This project aims to apply machine learning techniques to the student performance record to identify weak-performing students. Next, a course material recommendation system has been introduced for offering suitable learning material to the student. The proposed model also considers the text analysis techniques to analyze the course material complexity that best fits according to the student’s learning ability.

In this context, the dataset has been obtained from Kaggle and you can download the dataset using the download link. First, the dataset has been analyzed and preprocessed to make it clean. Next, the feature selection technique has been implemented to reduce the dimensions of the dataset. Further, a modified Fuzzy C Means (FCM) algorithm is implemented to categorize the students into three grades i.e. High, Mid, and Low. Next, a text-based dataset is prepared to demonstrate the course material. The text dataset is also preprocessed to eliminate the special character removal and stop word removal. Additionally, the Term Frequency and Inverse Document Frequency (TF-IDF) for selecting the text features from the course material. The extracted features are next used with a Natural Language Processing (NLP) based approach to measure the complexity score of the course material. Next, the Cosine similarity matching technique is implemented to recommend suitable course material to the student.

The implementation of the project is performed using Python technology and the performance has been measured. The accuracy and training time is used to demonstrate the performance of the given system.