WITHDRAWN: College student employment data platform based on FPGA and machine learning

Xu Lin

DOI: https://doi.org/10.1016/j.micpro.2020.103471

IF: 3.503

2020-11-01

Microprocessors and Microsystems

Abstract:<p>Student employment, the primary concern of higher education and early prediction method to provide institutions for students of employment, is always the need for timely action. In this paper, Bayesian methods, multilayer perception and Sequential Minimal Optimization (SMO), ensemble methods, and decision trees predict students' employment. The algorithm different classification techniques in data mining as finding the most suitable for this problem. For this purpose, the data set is socio-economic status, will be developed in the traditional parameters such as academic performance and some additional emotional skills of parameters. Comparative analysis of the conclusion J48 is, what is described is best suited to predict employment in a comfortable and 70.19% accuracy. Random forest time (0.02 sec) is less model building, slightly better prediction accuracy (71.30 percent) is that has, is difficult to explain the high construction time (0.11). J48 algorithm classifies various applications and will be used to perform an accurate result of the classification. To clarify, the algorithm of J48 is always one of the best machine learning algorithms to check the data. If it is being used for illustrative purposes, it occupies more storage space and consumes the performance and accuracy in data classification. Besides, compassion, health problems, economic problems, communication problems, and stress management functions are significant emotional parameters that affect employment.</p>

computer science, theory & methods,engineering, electrical & electronic, hardware & architecture

What problem does this paper attempt to address?

The problem this paper attempts to address is predicting the employment situation of university students. It aims to predict students' employability using different classification techniques such as Bayesian methods, multilayer perceptron, Sequential Minimal Optimization (SMO), ensemble methods, and decision trees. The authors seek to identify the most suitable data mining technique for this problem and analyze various factors affecting students' employment, including traditional academic performance parameters and socio-economic status, as well as some additional emotional skill parameters such as empathy, health issues, economic problems, communication issues, and stress management functions. Through these analyses, the study hopes to provide universities with an effective method to identify students who may face employment difficulties in advance, thereby enabling timely intervention measures.