The short answer is No. Machine learning is where the traditional Statistical modeling of data meets the algorithmic and computational field of data science. Such statistical modeling has been used for a long time probably since mid 60’s or even before to model some real world process. Actuarial modeling in the insurance industry is a good example, where a lot of data about general health, longevity, personal habits are used to model and determine insurance premiums. Statisticians and Actuaries have been doing this unsexy work for modeling for decades with none of the pomp and attention that Machine learning has been getting the last few years.
So if Machine learning is mostly about model building then why is there all the recent hoopla? Well it is not just a model building based on the data, there is also some learning involved by turning the model parameters to achieve better prediction accuracy. This learning part could be somewhat thought of as Artificial Intelligence since the algorithm learns on its own from the data without human intervention.
In the recent years, new algorithms are being invented that create complex and computationally intensive models that are very good at detecting and parsing subtle patterns in the data. This coincided with the exponential increase in computational resources at very low cost through cloud computing. There is also an exponential increase in the data generated in fields like Biostatistics, Bioinformatics, FinTech etc that started exceeding existing means of analysis. This troika of events precipitated in bringing Machine Learning to the mainstream. Researchers and Data Scientist from these diverse disciplines started looking at Machine learning to help them in their data analysis and classification tasks. The advent of sites like Kaggle brought open competitions into the world of data modeling and it really helped machine learning take off.
We at Optisol, are building a team of data analysis and machine learning experts to satisfy our customer’s growing need for analyzing and classifying their data and build data-centric decision support practices. It has been an interesting journey for us in the last one year and we are excited about the prospects of this interdisciplinary blend of Statistics, Computation, and Data to answer our customer’s questions.