Machine learning based Time series prediction using Holt-Winters Exponential Smoothing with Multiplicative Seasonality

Forecasting precision is essential tool in revenue management systems. For instance, predicting the demand of capacity available in future for airplanes became essential in order to increase the revenue by setting up suitable infrastructure standards. Future planning of airplane capacity expansion depends upon accurate prediction of passengers who boards the flight. This prediction is beneficial for Airline agencies for effective planning of future capacity thereby in order to increase the financial growth. With the recent development of Oracle Machine learning time series prediction algorithms, it is possible to accurately predict passengers. In this paper, we analyze the data of the time series using Holt-Winters's precision methods. The use of the Holt-Winters seasonal model in predicting passenger level is analyzed and the predicted results were compared with the seven seasonal trends. Finally, the results show that the Holt-Winters exponential model with a multiplicative tendency and multiplicative seasonality is effective in predicting passenger level with high accuracy compared to remaining methods. Customers around the world take advantage of Oracle's machine learning capabilities in the database to solve complex and important data-driven problems.