Choosing a machine learning model without consulting a voodoo witch

After a full day at a data conference, our group found ourselves engaged in a deep discussion during our train ride home. Among us was a gentleman in his fifties, who revealed himself as the Chief Technology Officer (CTO) of a company with a strong focus on data-driven products. As we chatted, he posed a question that lingered in our minds: how do we decide which model to use for classification, prediction, or regression? Some of us responded casually, suggesting we simply choose whatever works best. However, this response didn’t satisfy him; he emphasized the need for a more systematic and rational approach to model selection. While I offered a brief explanation at the time, the conversation was cut short by the end of our journey and our exhaustion. In this post, I aim to delve deeper into this question and provide a comprehensive answer. <a href="https://medium.com/@uriitai/choosing-a-machine-learning-model-without-consulting-a-voodoo-witch-dac2264b883a">Click Here</a>