November 2019
Intermediate to advanced
296 pages
7h 52m
English
In this experiment, the dataset is a two-dimensional binary of clusters. The points in each cluster have been sampled from the Gaussian distribution. Although the same standard deviation is shared by all the base distributions that generate each cluster, the center of the distribution is shifted to make them separated clusters. For logistic regression to work well, the samples of each cluster need to be generated from the same shape distribution, as we explained previously. The following screenshot shows the samples from each cluster:

The rectangle points belong to class 1, while the triangle points belong to class 2. ...
Read now
Unlock full access