Introduction to Trading, Machine Learning & GCP 2주차

Supervised Learning with BigQuery ML

Introduction to Supervised Learning

What is forecasting?

ML terminology

• Label : the correct answer typically from historical data (can be number, string, etc.)
• Feature : other columns of data for the model to learn from
• Model : a computer-determined recipe to get from features to label
• Model Types : (will cover soon)
• Training : showing the model lots of examples for it to learn the relationship
• Weight : Adjustable parameter of a model
• Evaluation : how the model performs on a set of known labels
• Prediction : using a trained model to predict on unknown labels

Choosing the right model and BQML

Decision tree to help guide us (there are many different types of models out there that you may not see on this tree)

Additional Practice: BigQuery ML model types

Forecasting Stock Prices using Regression in BQML