Machine learning

- Machine Learning or ML is one of the major tools of the trade for data scientists. ML is a branch of Artificial Intelligence or AI. AI is a field that studies ways for computers to mimic human intelligence. In fact, ML is one of the mechanisms that implements the idea of AI. In essence, ML is an algorithm manifested in the form of a computer program that can learn from experience to improve its performance.

ML has a significance in data science because its performance is largely driven by and dependent on the quality and quantity of data available. By the way, if you haven't already noticed, ML gets its experience by the data fed into its software. Take fraud detection. You have so many data points to consider such as customer gender, purchase and payment history, location, time, among other things.

There are also tens of millions of customers you need to keep track of at once. And the time is of the essence here. The sooner you detect the fraud, the less damage a criminal can inflict. Processing that much data at such a high speed is well beyond human capacity and ability and ML is a great solution to this problem. It can very easily monitor numerous targets at the same time and is extremely good at number crunching.

The precision of detection gets better and better as ML can look back on previous decisions and fine tunes its parameters to produce a more accurate detection result. This process of reflection and correction is what we call learning in ML. However, ML alone cannot accomplish this impressive feat. It needs all the necessary infrastructures to be able to handle the volume, velocity, and voracity requirements of adequately processing input data.

And this is only made possible by the use of tools such as cloud computing, distributed systems like Hadoop and distributed processing solutions such as Spark in addition to ML.