Research on the Application of Artificial Intelligence and Big Data Visualization Technology in Human Resource Planning

This study constructs a dynamic human resource prediction model (ADHRP) based on the attention mechanism, integrates multimodal data to achieve accurate prediction, and uses three-dimensional dynamic visualization technology to intuitively present the results. The experimental data set covers structured data of internal employees of a multinational company, unstructured text such as resumes, and external data on industry talent demand trends. After data preprocessing, the ADHRP model is used for prediction and compared with traditional regression models (linear regression, logistic regression), common deep learning models (LSTM, GRU) and more advanced models (Transformer-XL and graph neural network, GNN). The results show that ADHRP has an accuracy of 87%, a recall rate of 85%, and an F1-score of 0.86 in predicting employee turnover rate, which are 7%, and 6.5% higher than LSTM, 6.2%, and 5.9% higher than Transformer-XL, and 4.6%, and 4.3% higher than GNN; in job demand prediction, the error rate is as low as 4.5%, which is much better than traditional regression models, LSTM, GRU, Transformer-XL and GNN. Three-dimensional dynamic visualization effectively displays job vacancies, employee skill distribution, and changes in labor costs at different times. This study provides enterprises with efficient human resource planning solutions to help them improve their competitiveness.