Information Maximization Using Data Stream Partitioning
This paper proposes an information maximization method to extract useful informations for streaming signals. The information maximization method is to extract signals using stream partitioning function and standard normalization function. The stream partitioning function divides streaming signals into intervals. Signal coefficient and interval coefficient are used at generating a interval by using stream partitioning function. Signal coefficient determines the quantity of data included in a signal interval. Interval coefficient determines the starting point of a signal interval. The standard normalization function determines a signal range by using the probability density function for streaming signals. The information maximization method becomes resistant to outliers and missing values. A normal signal can be extracted effectively for streaming signals by the Information maximization method. The effectiveness of proposed method are also presented by several experimental results.