Evaluating Corporate Employee Performance Using Big Data-Enabled KPCA-LSTM Models

The long-term growth of the business and the enhancement of the management system are closely tied to enterprise employee performance assessment, a crucial tool for the development of contemporary private firms. The problem of evaluating the performance of employees in an enterprise setting is fundamentally a predictive regression problem; a deterministic optimization algorithm cannot determine the optimal network parameters, and the current approach to evaluating employee performance in an enterprise setting is prone to local optimality. This research presents an employee performance evaluation for enterprises using the Kernel Principal Component Analysis – Spotted Hyena Optimization – Long Short-Term Memory (KPCA-SHO-LSTM) model. This paper describes the method of building a performance evaluation system for enterprise employees. It begins by analyzing the steps involved in constructing the system and selecting the evaluation indices. Next, the Kernel Principal Component Analysis (KPCA) and Spotted Hyena Optimization (SHO) algorithms are employed to search for optimal parameters for the Long Short-Term Memory (LSTM) network and improve the performance evaluation model. Finally, a design is proposed to evaluate the algorithm's performance using optimized covariates and an adaptive segment function. The approach is evaluated against other algorithms in the context of enterprise employee performance evaluation. The findings demonstrate that the proposed method exhibits the lowest error accuracy, as well as the ideal convergence speed and iteration number.