Machine learning challenges with Big Data

The developments in the information technology in recent years have led to the collection of massive amounts of data. Big data attracts attention from industry, academia and government as big data technologies allows us to process large volumes of data to get deeper insights for decision making through data analytics. Machine Learning(ML) forms an important component of data analytics because it gets the computers to learn from the trained data and act, and it improves from past experience without being explicitly programmed every time. ML algorithms are used to uncover more fine-grained patterns and make accurate predictions and also they have the ability to improve the performance of other parts of the knowledge discovery. The characteristics of Big data such as volume, variety, velocity, veracity form obstacles to traditional machine learning algorithms for performing data analytics. In this paper we present major challenges that are presented by big data to machine learning such as scalability of the algorithm for large data sets and the ability to be computed in a distributed and real time processing environment. We also review some of the Machine learning methods, frameworks and tools. The identified challenges faced by the ML algorithms pave the way for new opportunities to develop novel algorithms which are adapted for Big Data.