Please enable JavaScript.

Coggle requires JavaScript to display documents.

Machine Learning (Definition (Well-posed learning problem is defined as…

- - - - Classification
        
        We're trying to predict a discrete value output zero or one, malignant or benign.
      - Regression
        
        We're trying to predict a continuous valued output.
    - - Neat mathematical trick that will allow a computer to deal with an infinite number of features
- - - - x^i
        to denote the “input” variables
        input features
      - y^i
        to denote the “output”
        target variable
- - - - 1/(2m) * ∑ (h(x) - y)^2
      - measure the accuracy of our hypothesis function
      - Thus as a goal, we should try to minimize the cost function. Find our global minimum.
        
        Graph J(θ0, θ1)
        
        Graph J(θ1):
        
        (θ0 = 0)
      - Gradient Descent
        
        θJ : = θJ − α * (∂J(θ0 , θ1) /θJ∂)
        
        j = 0,1 represents the feature index number
        α = learning rate. It controls how big a step we take when updating my parameter theory j
        (∂/θJ∂)J(θ0 , θ1) = derivative term (частная производная)
        
        batch gradient descent