Forecasting in Python: from Linear Regressions to Time Series

Objectives

​

• Invite the student to use and understand the statistical and computational qualities of the main algorithms associated with linear regressions and time series.

• To familiarize the student with some aspects of the mathematical language used in Data Science and Machine Learning.

• Train the student in the programming techniques necessary to implement the models studied.

​

Syllabus

​

1. From the implementation point of view the student will learn:

• Programming: Basic Python

• Most used library for data science: Pandas

• Learning good programming practices for data science

• Familiarization with common and necessary processes in data science

• Application of some models without pre-made libraries

• State-of-the-art libraries application for all the models seen in the course

• Sufficient knowledge to learn and implement widely used data science models outside of the course

​

2. From a mathematical point of view we will study the following topics:

• Linear regressions

• Statistical qualities behind the least squares estimator

• Some algorithms to calculate it

• Time Series: general concepts

• Buys-Ballot and least squares estimation

• Moving-Average

• Associated Vectors

• ARIMA

• Invitation to stochastic processes