A number of quizes to get familiar with the Numpy library.
Implementation of basic matrix operations, such as addition and multiplication without the standard methods included in the Numpy library.
Implementation of Linear Regression with Gradient Descent Algorithm. Methods implemented in this module calculate the Cost of a linear function and apply the Gradient Descent Algorithm in order to find the line that best describes a Data Set, which can be expressed through a linear function.
Implementation of Multivariate Regression with Gradient Descent Algorithm. The methods that are implemented in this module allow linear regression to be applied on datasets that contain more than one feature. To achieve this feature scaling is applied on the data set.
With the aid of the Gradient Descent Algorithm, one can input initial theta values on the data set and iterate a number of times in order to get theta values that would help find a line that can fit better on the data set.
Implementation of Logistic Regression.
The purpose of logistic regression is to help classify the values of the data set in a specific class for instance black or white. The methods implemented in this module, achieve this by using the sigmoid function as the base of the cost function.
In order to train our model Gradient Descent Algorithm is applied on the data set. After taking as input initial theta values, which are chosen for our data set, the gradient descent algorithm calculates the partial derivatives of the cost function including the linear hypothesis. The gradients scaled by a scalar are subtracted from the given theta values.
After gradient descent is applied on the dataset and better thetas have been generated, the decision boundary and the accuracy of the model can be calculated with the aid of the following methods implemented on the module:
plot_decision_boundary(thetas, data_set)
print_accuracy(thetas, data_set)