This document provides recommendations for applying machine learning:
1) Understand the problem you are trying to solve by defining what the problem is exactly and what is the required solution.
2) Understand the available data by analyzing the data size and performing some descriptive statistics.
3) Define algorithms to test based on problem description and available data such as regression, classification, clustering, etc.
4) Design a test by splitting data into training, validation, and testing sets. Evaluate models on test data and select the best performing one.