The hype around AI/ML that is gripping the present-day IT industry is very real, but sometimes it’s only a marketing tactic leveraged by vendors to grab the attention of the buyers. It doesn’t always require a complex ML model to get things done. That is not to say that ML models are not useful, they are incredibly useful, and therefore should be reserved for complex things that cannot be accomplished with simpler models.
Instead, companies often apply advanced Machine Learning models to do things that could very easily be eliminated. That’s just an overkill, or worse, a waste. Alternately, a lot can be achieved easily with strategic use of linear regression methods, telling apart the essential from the non-essential for example. In some cases, linear regression can achieve even better results than complex Machine Learning.
Justin Warren, a long time Field Day delegate, has a very interesting read on this. In his article- “Linear Regression Is Better Than Machine Learning”, he looks at the AI/ML bubble in IT through his witty and humorous lens. He writes,
I remain convinced that 90% of what vendors are calling machine learning or AI is three linear regressions in a trench coat.
Give his article “Linear Regression Is Better Than Machine Learning” a read for his point of view on this.