In this fourth episode of the Deep Learning Fundamentals series, we discuss non linear classification problems.
Real world problems that can be solved with a simple linear binary classification are very limited.
Non linear classification problems can be either binary (i.e. prediction is either 0 or 1) or multi class (i.e. prediction is more than two outcomes).
We will show how regression analysis causes the prediction formula to increase in complexity for data sets with high input features. This will lead to an intractable number of parameters in the optimisation algorithm.
Episode 1 link: https://www.youtube.com/watch?v=7FtQ-Ski0gE&t=5s
Episode 2 link: https://www.youtube.com/watch?v=DapdsyWNQK8&t=3s
Episode 3 link: https://www.youtube.com/watch?v=zrLKOYOIloQ
Intro music: https://www.bensound.com
#classification #binaryclassification #regression #deeplearning #neuralnetworks #deepneuralnetworks #logisticregression #machinelearning #datascience