Classification in Machine Learning

The output variables in machine learning can also be referred to as independent variables.

Linear regression is a model that assumes a non-linear relationship between input and output.

In simple linear regression, the model parameters are denoted by w = (0', 0$).

Learning linear regression involves finding the optimal parameter w ∗ using unsupervised learning techniques.

To learn linear regression, one needs a cost function and a gradient descent algorithm.

The training set for linear regression provided involves the relationship between housing prices and the size of the house in square feet.

In the context of linear regression, the cost function quantifies how well our model predicts the data.

The goal of gradient descent is to maximize the cost function.

The logistic function is typically used in the context of classification problems rather than regression problems.

The parameter 'w' is typically estimated using closed-form solutions rather than iterative optimization algorithms like gradient descent.

The cost function is also known as the loss function because it quantifies the errors or discrepancies in our model's predictions.

Gradient descent aims to find the parameters that minimize the cost function by iteratively updating them in the direction of steepest descent.

In logistic regression, the output variable y is a continuous variable.

The logistic function, also known as the sigmoid function, maps any real-valued number to a value between 0 and 1.

The decision boundary in logistic regression is always a linear boundary.

In logistic regression, the goal is to minimize the mean squared error between predicted probabilities and true labels.

The logistic function is used to predict the probability of the negative class.

The threshold value of 0.5 is used to convert the predicted probability into a class label.