An improved CNN architecture for exploring the influence of quality-oriented college education to work promotion probability: a case study in Guangdong province, China

Work promotion is determinant from enterprise evaluation on job abilities of the employees. The persons who work in efficiency and suitable to the company is usually originated from being well trained when they were undergraduate in the quality-oriented college cultivation. This paper studies the influence of quality-oriented college education to the work promotion in enterprises, based on the case in Guangdong province of China. An empirical study is performed using data collected from thousands of fresh employees in various types of companies. The available data is pertinent to several training items covered in college and is relevant to the promotion in the companies. For big data analysis, a novel network architecture is built up to extract the data information. Hyperparameters are well trained involving the convolution filters and the activation functions. A dual-metric scalable information entropy optimization strategy is proposed for model optimization. The aim of model prediction is to meet the standard of the company’s natures, scales and academic divisions. It is revealed that the promotion chance for the fresh employees is discriminated with over 90% prediction accuracy in relation to their undergraduate trained items. We conclude that the work promotion probability is highly influenced by the practical items of college cultivation. The proposed modeling framework has been successful in big data analysis for forecasting work promotion opportunities, and it is regarded as a potential tool to assist companies and enterprises in making various evaluations of the employees.