Time Series Analysis, Forecasting, and Machine Learning

Description

In this course, you will learn :

• ETS and Exponential Smoothing Models
• Holt's Linear Trend Model and Holt-Winters
• Autoregressive and Moving Average Models (ARIMA)
• Seasonal ARIMA (SARIMA), and SARIMAX
• Auto ARIMA
• The statsmodels Python library
• The pmdarima Python library
• Machine learning for time series forecasting
• Deep learning (ANNs, CNNs, RNNs, and LSTMs) for time series forecasting
• Tensorflow 2 for predicting stock prices and returns
• Vector autoregression (VAR) and vector moving average (VMA) models (VARMA)
• AWS Forecast (Amazon's time series forecasting service)
• FB Prophet (Facebook's time series library)
• Modeling and forecasting financial time series
• GARCH (volatility modeling)

Syllabus :

1. Time Series Basics

• Time Series Basics Section Introduction
• What is a Time Series?
• Modeling vs. Predicting
• Why Do We Care About Shapes?
• Types of Tasks
• Power, Log, and Box-Cox Transformations
• Power, Log, and Box-Cox Transformations in Code
• Forecasting Metrics
• Financial Time Series Primer
• Price Simulations in Code
• Random Walks and the Random Walk Hypothesis
• The Naive Forecast and the Importance of Baselines
• Naive Forecast and Forecasting Metrics in Code

2. Exponential Smoothing and ETS Methods

• Exponential Smoothing Section Introduction
• Exponential Smoothing Intuition for Beginners
• SMA Theory
• SMA Code
• EWMA Theory
• EWMA Code
• SES Theory
• SES Code
• Holt's Linear Trend Model (Theory)
• Holt's Linear Trend Model (Code)
• Holt-Winters (Theory)
• Holt-Winters (Code)
• Walk-Forward Validation
• Walk-Forward Validation in Code
• Application: Sales Data
• Application: Stock Predictions
• SMA Application: COVID-19 Counting
• SMA Application: Algorithmic Trading
• Exponential Smoothing Section Summary

3. ARIMA

• Autoregressive Models - AR(p)
• Moving Average Models - MA(q)
• ARIMA
• ARIMA in Code
• Stationarity
• Stationarity in Code
• ACF (Autocorrelation Function)
• PACF (Partial Autocorrelation Funtion)
• ACF and PACF in Code
• Auto ARIMA and SARIMAX
• Model Selection, AIC and BIC
• Auto ARIMA in Code
• Auto ARIMA in Code (Stocks)
• ACF and PACF for Stock Returns
• Auto ARIMA in Code (Sales Data)
• How to Forecast with ARIMA

4. Vector Autoregression (VAR, VMA, VARMA)

• VAR and VARMA Theory
• VARMA Code
• VARMA Econometrics Code 
• Granger Causality
• Granger Causality Code

5. Machine Learning Methods

• Supervised Machine Learning: Classification and Regression
• Autoregressive Machine Learning Models
• Machine Learning Algorithms: Linear Regression
• Machine Learning Algorithms: Logistic Regression
• Machine Learning Algorithms: Support Vector Machines
• Machine Learning Algorithms: Random Forest
• Extrapolation and Stock Prices
• Machine Learning for Time Series Forecasting in Code
• Forecasting with Differencing
• Application: Sales Data
• Application: Predicting Stock Prices and Returns
• Application: Predicting Stock Movements
• Machine Learning Section Summary

6. Deep Learning: Artificial Neural Networks (ANN)

• Artificial Neural Networks: Section Introduction
• The Neuron
• Forward Propagation
• The Geometrical Picture
• Activation Functions
• Multiclass Classification
• ANN Code Preparation
• Feedforward ANN for Time Series Forecasting Code
• Feedforward ANN for Stock Return and Price Predictions Code
• Human Activity Recognition Dataset
• Human Activity Recognition: Code Preparation
• Human Activity Recognition: Data Exploration
• Human Activity Recognition: Multi-Input ANN
• Human Activity Recognition: Feature-Based Model
• Human Activity Recognition: Combined Model
• How Does a Neural Network "Learn"?
• Artificial Neural Networks: Section Summary

7. Deep Learning: Convolutional Neural Networks (CNN)

• What is Convolution?
• What is Convolution? (Pattern-Matching)
• What is Convolution? (Weight Sharing)
• Convolution on Color Images
• Convolution for Time Series and ARIMA
• CNN Architecture
• CNN Code Preparation
• CNN for Time Series Forecasting in Code
• CNN for Human Activity Recognition

8. Deep Learning: Recurrent Neural Networks (RNN)

• Simple RNN / Elman Unit
• Aside: State Space Models vs. RNNs
• RNN Code Preparation
• RNNs: Understanding by Implementing (Paying Attention to Shapes)
• GRU and LSTM 
• LSTMs for Time Series Forecasting in Code
• LSTMs for Time Series Classification in Code
• The Unreasonable Ineffectiveness of Recurrent Neural Networks

9. VIP: GARCH

• ARCH Theory 
• GARCH Theory
• GARCH Code Preparation 
• GARCH Code 
• A Deep Learning Approach to GARCH

10. VIP: AWS Forecast

• Data Model
• Creating an IAM Role
• Code pt 1 (Getting and Transforming the Data)
• Code pt 2 (Uploading the data to S3)
• Code pt 3 (Building your Model)
• Code pt 4 (Generating and Evaluating the Forecast)
• AWS Forecast Exercise

11. VIP: Facebook Prophet

• How does Prophet work?
• Prophet: Code Preparation
• Prophet in Code: Data Preparation
• Prophet in Code: Fit, Forecast, Plot
• Prophet in Code: Holidays and Exogenous Regressors
• Prophet in Code: Cross-Validation
• Prophet in Code: Changepoint Detection
• Prophet: Multiplicative Seasonality, Outliers, Non-Daily Data
• (The Dangers of) Prophet for Stock Price Prediction

Show more