Stream Based Machine Learning Models - Why Not?

90% of the businesses that run machine learning models are bath ar batch based. Changing that would mean having a different machilonig model and completely different assumptions. There is a reason why, for example, moders that are used in health careand f and finta can make damage if they perform wrong predictions. And so imagine anoline learn system that keeps learning from the data that flows in Imagine, if that data has poor quality or statistical properties, start diverging from time zero to time ten. What happens is that your model, youknow, that difference, that divergence, is reflected from your model. And so your mothers start behaving differently as well. So these are the things