Title of the course: NEURAL NETWORKS FOR TIME SERIES ANALYSIS
Abstract
In this mini-course we will study artificial neural networks as a tool for time series classification and forecasting. We will review theoretical concepts of feedforward, convolutional and recurrent neural networks, their modern architectures and implement them in Python. The emphasis of the mini-course is on practical applications, so we will use before mentioned algorithms to forecast stock prices movements and build a real asset trading strategy and backtest it. After attending this course, students will understand basic pipeline of machine learning based time series analysis: data preprocessing, fitting the model, evaluation of results and will be able to use their own models for building algorithmic trading strategies.
Details for each of the four lessons
Lecture 1: Introduction to machine learning and time series analysis
Lecture 2: Data preparation and feedforward neural networks
Lecture 3: Convolutional and recurrent neural networks
Lecture 4: Building a trading strategy and further applications
Dates, hours and Rooms
21/3 - 1530-1730 - Meeting Room 2nd floor
23/3 - 1530-1730 - Meeting Room 2nd floor
28/3 - 1530-1730 - Meeting Room 2nd floor
29/3 - 1530-1730 - Meeting Room 2nd floor
