The environment could differ owing to past and present market demands in situations of
rapid change. Detailed examination of whether large volume data and traditional forecast
models are suitable for the high-tech industry is important. Because the product life cycle
of the high-tech industry product is so short, forecasting results are influenced by shifts in
market demand cycles, and traditional forecasting methods cannot explain the real
situation. These forecasting models, listed in Table 1, each have various strengths and
weaknesses. Therefore, as noted above, these methods always suffer from insufficiencies and
ambiguities when applied to real-world studies. This work attempts to solve the above
problems and develop a new forecasting method that only requires short-term, current and
limited data.