In order to reduce the waiting rate for calls from customer service centers and improve service quality, a method for call volume prediction based on dual-modal Transformer model is proposed. This is in response to the issue that existing algorithms cannot achieve medium and long-term call volume prediction. Firstly, the real call volume data from a certain ISP is collected and preprocessed; then, beneficial features are constructed through bimodal feature fusion; finally, multiple models are used for call volume prediction. Additionally, a variety of measurement indicators are used to analyze the prediction results. The results show that compared with other algorithms, the Transformer model performs better, providing significant guidance for the rational allocation of ISP resources. It also makes it easier to achieve higher customer satisfaction and loyalty.
万方数据
维普
