பிரிட்டிஷ் ஜர்னல் ஆஃப் ரிசர்ச் திறந்த அணுகல்

சுருக்கம்

Formulating Data Science problems to maximize Business impact – Case studies

Mirza Rahim Baig

Artificial Intelligence and Data Science seem to be ubiquitous now with every company investing in Data Science
capabilities and data scientists. While it is tempting to throw the latest in AI at every business problem, not many
business situations and applications require the use of sophisticated methods and state of the art techniques. In many
situations, using a Deep Learning solution may just not be feasible. Indeed, designing a Data science based solution is
among the most important aspects of data based problem solving. Formulating the problem the right way is the critical
first step and has an extremely high impact on the solution design and eventually the efficacy of it and the benefit to
business. Should the formulation be supervised ML or unsupervised? Should you replicate a state of the art model
from scratch for your own domain or are better off using transfer learning? Do you even need Machine Learning for the
problem, or is a good rule based solution? These are critical questions to ask while beginning to solve a problem using
Data Science. .