Affiliations: Hunan Post and Telecommunication College, Nanhu Road, Tianxin District, Changsha, Hunan 410015, China | E-mail: lnpt51r@126.com
Correspondence:
[*]
Corresponding author: Hunan Post and Telecommunication College, Nanhu Road, Tianxin District, Changsha, Hunan 410015, China. E-mail: lnpt51r@126.com.
Abstract: The employment situation of fresh college graduates is affected by many factors. In this paper, on the basis of decision tree, the C4.5 method was used to analyze the employment factors of fresh college graduates. An improved C4.5 model was designed by simplifying the calculation formula of the C4.5 method and combining the error tolerance. Experiments were performed on the actual data of fresh college graduates. The results found that the practice level had a great impact on the employment of fresh college graduates, so the training of the practice level should be focused on before graduation. The results of the prediction models showed that the improved C4.5 method had a smaller training error than ID3 and C4.5 methods, a significantly higher prediction accuracy (88.39%), higher precision, recall rate, and F1 value, and a shorter running time (1.642 s); the improved model remained a high accuracy even when the data volume increased. The experimental results verify the reliability of the improved C4.5 model in predicting the employment situation of fresh college graduates. The model can be applied in actual employment guidance.
Keywords: Fresh graduates, employment, decision tree, error tolerance, practice level
