Supervised Machine Learning

Train a model with known input and output data, to predict(label) known output from new data.
Supervised Machine Learning algorithms must be used when have past data with outcomes (labels in machine learning terminology) and you want to predict the outcomes for the future(New and unlabeled data). Supervised Machine Learning Models often use a loss function to compare the output of the model to the ground truth.

💡 Understanding Supervised Machine Learning:
We have have input data X and output data y, model finds a map from X to y by training via updating it’s model parameters instead of programmed logic.

💡 Components of Supervised Machine Learning:
We have m observations(As rows of data) and n features(As defined columns)
X is an m × n matrix
y is a vector of size m used for mapping from X to y
optimizer, I.e. gradient descent or BFGS, is used to find the optimum parameters


  • Classification
    • Used with Categorical Target Variables.
    • Classification Problems: The goal is to predict discrete labels(classify outcomes into different classes).
  • Regression
    • Used with Continuous Target Variables.
    • Regression Problem: The goal is to predict continuous values.