How to maintain Model Effectiveness after deployed

When we ready to deploy a good predictive model which given a good accuracy score on training and testing dataset, there is one more problem to solve.  How long will this model effectively solve the problem with the same high accuracy and what is the strategy to maintain the model accuracy. We also need to…

Machine learning models grow more powerful every week, but the earliest models and the most recent state-of-the-art models share the exact same dependency: data quality. The maxim “garbage in – garbage out” coined decades ago, continues to apply today. Recent examples of data verification shortcomings abound, including JP Morgan/Chase’s 2013 fiasco and this lovely…