Marketing Strategy Support System based on Temporal Fusion Transformer

Nowadays, the marketing strategy support system is designed to aid marketing professionals and businesses in developing the effective marketing strategies. However, the existing Customer Relationship Management (CRM) system limited by data quality issues due to the vast amounts of available data coupled with the intricacies of market dynamics. Hence, this research proposes the marketing strategy support system namely Temporal Fusion Transformer (TFT) based on stock price data. Initially, the input data is collected from Nifty500 stocks dataset which is available at Kaggle platform. Then, the data is preprocessed by using min-max normalization to rescale the data into a specific range. After that, the proposed TFT is introduced to forecast the stock prices by using self-attention mechanism and decoder. Finally, the proposed TFT predicted the technical indicators including Money Flow Index (MFI), Relative Strength Index (RSI), Resistance Level (RL) and Fibonacci Retracement (FR) effectively for efficient stock price forecasting. From the results, the proposed TFT provided outstanding results in terms of Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), Mean Square Error (MSE), and Root Mean Square Error (RMSE) with 16.82, 5.19, 9.31, and 2.73 respectively when compared with existing Long Short-Term Memory and Transformer with Particle Swarm Optimization (LTPNet) model.