Students’ academic success prediction

Keywords: Student' academic, success

Background

In this study, we investigate the use of Artificial Neural Networks (ANN) for predicting student outcomes in terms of graduation or dropout or enrolled. The dataset used for analysis contains various student-related attributes, such as marital status, inflation, parents’ qualifications, and socioeconomic factors. The goal is to develop an ANN model capable of accurately classifying students as either graduates or dropouts based on their individual characteristics. The author preprocesses the dataset by cleaning and transforming the raw data into a suitable format for ANN training. The input variables are all numerical, so they have been standardized to rescale them, and as the target variable is categorical, it was label encoded. The dataset is then split into training and testing sets, ensuring an appropriate ratio of 75:25 to avoid overfitting. After splitting the data, the author again converts the target variable into binary matrix of 0’s and 1’s using Keras library of np.utils model. A sequential model of ANN has been built using 5 layers, out of which 3 are hidden layers, and 1 is input layer, and the last is output layer. The activation function used involve ReLu in the input layers, and SoftMax in the output layer. The accuracy of the whole model came to be 73 percent.

Aim & Objectives

Aim
To Predict the Students’ academic success Using DL Techniques
Objective
Employ a deep learning ANN to predict students' academic success, enhancing early intervention strategies.
Develop an ANN model to analyze diverse student data for accurate academic success prognosis.
Investigate and refine the ANN's features to optimize students' academic success prediction and support.

Research Methodology

ANN Technique used

Software & Tools

Jupyter, Google Collab

— Preview truncated. The complete document (full chapters, references, diagrams, and appendices) is shared with clients as part of project delivery. —