Bias in machine learning

Abstract

Data in today's world is the new black gold which is growing exponentially. This growth can be attributed to the growth of existing data, and new data in a structured and unstructured format from multiple sources such as social media, Internet, documents and the Internet of Things. The flow of data must be collected, processed, analyzed, and finally presented in real time to ensure that the consumers of the data are able to take informed decisions in today’s fast-changing environment. Machine learning techniques are applied to the data using the context of the problem to be solved to ensure that fast arriving and complex data can be analyzed in a scientific manner using statistical techniques. Using machine learning algorithms that iteratively learn from data, hidden patterns can be discovered. The iterative aspect of machine learning is important because as models are exposed to new data, they are able to independently adapt and learn to produce reliable decisions from new data sets.