Category: articles
-
0.6.0 – Basic Neural Network Feature Space Topologies
One caveat with the binary and multi-class logistic regression is that it only able to classify the data if the data is linearly separable: If the data points are non-linearly separable, binary/multi-class logistic regression will only be able to find the best linear line or hyperplane to separate the data…
-
0.5.0 – Regularized Linear Regression
In order to mitigate overfitting our linear model we can add one or more regularization terms to our loss function. The regularization terms incentivize lower the weights in the coefficient/weight matrices. Why does one want to avoid large coefficients? When you have large weights in you model this lead to…
-
0.4.0 – Multi-class Logistic Regression
Often we want to predict the label/class of a data point where there are more than two labels. Taking a look at binary linear regression, we see it is only created for when there are only two labels to choose from. From binary logistic regression we have: When there only…
-
0.3.0 – Binary Logistic Regression
Sometimes it useful to predict a class given input data, unlike in linear and polynomial regression where you predict a number given input data. To predict binary class labels (either 1 or 0) one can use logistic regression. Its worth noting that while polynomial regression is kind of a extension…
-
0.2.0 – Polynomial Regression
Different datasets have different underlying structures, while some have linear underlying structures other can have nonlinear structures: The differences are notable. When applying linear regression to both cases, it is apparent that linear regression is not a good way to model the nonlinear dataset. To capture the non-linearity better we…
-
0.1.0 – Linear Regression
Linear regression is one of the oldest and most fundamental forms of linear modelling, developed/discovered by none other than Carl Fredriech Gauss. Much of modern machine learning is just a beefed up version of linear regression so its worth understanding it well. Linear regression is used to approximate the best…
kbclock.work 