Machine Learning Intro

Machine learning is not only about writing algorithms, actually if an algorithm tells the machine what to do exactly; then it is explicitly programmed not learned how to interact.

Key Terminology:

1. Model: a set of patterns learned from data.
2. Algorithm - a specific ML process used to train a model.
3. Training data - the dataset from which the algorithm learns the model.
4. Test data - a new dataset for reliably evaluating model performance.
5. Features - Variables (columns) in the dataset used to train the model.
6. Target variable - A specific variable you’re trying to predict.
7. Observations - Data points (rows) in the dataset.

3 elements for great machine learning:

1. A skilled designer
2. Cleaning and analyzing data
3. Avoid overfitting

Machine learning basic steps:

1. Exploratory Analysis: study your dataset carefully
2. Data Cleaning: good data is the key for good models
3. Feature Engineering: generate new features from the old features
4. Algorithm Selection
5. Model Training

*Resource: elitedatascience