Predict the Shipment Forecast using Time-Series Data in Machine Learning

Abstract

In the point of large big data and massive increase the rate of time series data flow in the upcoming market program business, mining of related data and real time data[1] are been briefly explained. This paper proposes predicted the value of the trader marketing when we reach the expected the value and increase the rate of accuracy. For this purpose we can use the time series algorithm in machine learning and gets regular item sets by using the corresponding of Map reduce [2], which consumes less space and will not increase the time overhead. The usage of CPU is improved by using the thread calling algorithm and batch algorithm, it meets deep business opportunities and requirement processing or feature model based on the requirements of traders. Thus our results indicates that the model not only explained the time series data stream[4],it also helps traders to get to a confirmation that they can achieve data quickly and achieve accuracy trade off's. This paper proposes a new demand forecasting model which is an extension of the traditional exponential diffusion models [5]. We examined the forecasting performance of the models just after the release of the item when the small number of model calibration data is available. This paper shows that the model which we proposed has the thing of enabling early decision making and best performance