January 2019
Intermediate to advanced
294 pages
6h 43m
English
We have come across many budding data scientists who would build a model and, in the name of evaluation, are just content with the overall accuracy. However, that's not the correct way to go about evaluating a model. For example, let's say there's a dataset that has got a response variable that has two categories: customers willing to buy the product and customers not willing to buy the product. Let's say that the dataset has 95% of customers not willing to buy the product and 5% of customers willing to buy it. Let's say that the classifier is able to correctly predict the majority class and not the minority class. So, if there are 100 observations, TP=0, TN= 95, and the rest misclassified, this will ...
Read now
Unlock full access