Machine learning (ML) is a branch of artificial intelligence (AI) which did not just appear when James Cameron made the movie ‘The Terminator’ in 1984. In fact, it has existed for many decades. Early statisticians would recognize this as statistical learning from historical data. It allows systems to learn and improve without directly being programmed. Lately there has been a resurgence on this field due to higher memory and computational power of today’s computer systems.
ML has the ability to collect and analyse huge quantities of data in a very short time span. Harnessing the power of ML is becoming a critical component in the ever demanding and evolving business culture of today, as businesses seek continual efficiency gains. ML has huge potentials in the field of classification, regression and predictive analytics, including forecasting such diverse areas as the share markets, weather, and house prices. Face and image recognition and medical are also applications.
It is a multidisciplinary field with foundations in Mathematics and Computer Science. Statistics, probability, linear algebra, calculus and matrix all form the basis of ML. Unlike traditional computing where algorithms are programmed to resolve problems, ML algorithms allow computers to train themselves based on input data, and build models to predict or classify the data into specific categories.
ML is generally classified into broad categories. The most widely used categories are supervised and unsupervised learning. In supervised learning, algorithms are trained using labelled input and output data. The algorithm learns by comparing the calculated results with actual output and modifies the model accordingly. The raw data for supervised learning is divided into two parts, the first part is for training the algorithm, and the other is used to test the trained model.
Unsupervised learning is the second type of ML, where unlabeled data is used to train the model. The algorithm explorers and finds some structure in the data and tends to make clusters with new labels. Have you ever wondered how Netflix recommends movies based on your tastes, or Ebay and Amazon knows what you will buy before you do? Turns out they use unsupervised ML algorithms to predict your tastes.
This is just a beginner’s guide of what ML is all about, hopefully it helps build your curiosity in this field. Remember, most big organizations such as banks and government with large amounts of data know the true value of ML technology. They are already harnessing it’s power to make predictions, or find data anomalies that don’t fit the usual pattern, to spot risks early.
About the Author
Vinay Singh is a Senior Solutions Consultant at Intergy Consulting, who is experienced in helping clients by providing solutions in the area of data analytics, reporting and much more!
If you wish to talk to Vinay or one of our other consultants to explore how Machine Learning can help your business, please contact us today!
