Machine Learning Based Admission Data Processing for Early Forecasting Students' Learning Outcomes

Introduction

Industry 4.0 is shaped inseparably with data and data analysis, posing challenges for organizations (including universities) that How could they be able to improve their operational capacity, effective management, and minimize the risk of failure through efficiently handling data?

In the era of big data technology, the demand for universities to innovate their governance models and improve governance efficiency has become an urgent issue for managers. In training management, universities need to digitalize (digital transformation) management information, create large database systems to organize training management and support decision making instantly and accurately. The problem is how to use, analyze and exploit this data source effectively to adapt the management in education and improve education efficiency.

A new field that emerged to solve this problem, is data mining in education (Romero & Ventura, 2010). The field uses data mining models, machine learning techniques to extract potential knowledge in educational data. Since then, this field has been growing and obtaining many significant achievements.

Machine learning and big data mining is a rapidly developing field, which is the intersection of many related fields such as databases, statistics, machine learning, algorithms, and other related ones to extract useful knowledge from large data sets. Other names can also be used for data mining and knowledge exploration such as Knowledge discovery in databases (KDD), Knowledge extraction (KE), Data analysis or samples (Data/pattern analysis - DA/PA), or Business intelligence (BI), see Lu et. al. (2022), Schild et. al. (2022).

Over the past two decades, significant progress has been made in the field of machine learning. The field has arisen as a method of choice for developing practical software for computer vision, speech recognition, natural language processing, robotic control, and other applications. With the positive impact of the increase in the amount of educational data through digitalization, there are quite a few areas in which machine learning can positively impact education. It can be affirmed that this is an inevitable trend that demonstrates the development of education and training associated with technology.

The machine learning model of selection of the factors affecting students’ output, the Naive Bayes classification model are recommended by Harvey & Kumar (2019). This model can be used to implement early intervention for students. Others model from Đambić et. al. (2016) got also positive appreciation. Another study using two data mining algorithms Naïve Bayes and Logistic Regression also gave some positive results in predicting learning outcomes and predicting forced drop-out from school (Uyen & Tam, 2019)

Entrance admission is one of the important activities of any higher education institution. Every year all higher education institutions in Vietnam develop admissions schemes for their schools. The purpose of the admissions scheme is to recruit an adequate number of students according to the allocated criteria and, more importantly, to recruit the right students who wish to study the appropriate subject.

Admissions schemes of universities are usually assigned to the training management department and training organizers to plan and implement. In some universities, there is also a dedicated center/department/unit that organizes the implementation of the university's admissions tasks. The actual process carried out in most educational institutions is based on the experience of the officer in charge and, if any, on the calculation of admission figures in previous years. This leads to undesirable admissions results.

By approaching university admissions problems from the perspective of data mining and the application of machine learning techniques, the problem of making admission plans and related problems are extraordinarily complex issues. In Vietnam and around the world, this problem is less common in education research.

In this article, we apply machine learning techniques to solve several problems in university admissions, in which data is sampled from Hanoi Metropolitan University (in Vietnam) as a case study. In the framework of this paper, based on machine learning technology, our research questions are: