Forecasting of GPU Prices Using Transformer Method

GPU or VGA (graphic processing unit) is a vital component of computers and laptops, used for tasks such as rendering videos, creating game environments, compiling large amounts code. The price GPU/VGA has fluctuated significantly since the start COVID-19 pandemic in 2020, due part to increased demand GPUs remote work online activities. Furthermore, accurate forecasting can have broader implications beyond computer hardware industry, with potential applications investment decision-making, production planning, pricing strategies manufacturers. This research aims forecast future prices using deep learning-based time series Transformer model. We use daily NVIDIA RTX 3090 Founder Edition test case. historical 8, 16, 30 days. Moreover, we compare results model two other models, RNN LSTM. found that days; gets higher coefficient correlation (CC) 0.8743, lower root mean squared error (RMSE) value 34.68, absolute percentage (MAPE) 0.82 compared LSTM These suggest an effective efficient method predicting prices.