Big Data Real Time Ingestion and Machine Learning

Data arrives in all shapes and sizes. Many time data are acquired sequentially - as an infinite ever growing stream. This real time stream data needs to be processed sequentially by taking the data source and splitting it up along temporal boundaries into finite chunks or windows. Take examples from stock market, sensors or Twitter feed data. Rather waiting for data to be collected as a whole at a long periodic interval, streaming analysis let us identify patterns - and make decisions based on them - as data start arriving. When data are nonstationary, and patterns change over time, streaming analyses adapt. At scales, where storing raw data becomes impractical, streaming analysis let us persist only smaller, more targeted representations. This work describes machine learning approaches to analyze streams of data with an intuitive parameterization. Linear regression and K-means clustering concepts are redefined to the context of streaming.