PART 01 Getting Started with Prophet
__01.
History of time series forecasting
____01-1 Time Series Forecasting Overview
__1.1.1 Dependent Data Problem
____01-2 Moving average and exponential smoothing
____01-3 Autoregressive Cumulative Moving Average (ARIMA)
____01-4 ARCH/GARCH
____01-5 Neural Network
____01-6 Prophet
____01-7 Recent Developments
__1.7.1 NeuralProphet
__1.7.2 Google's "Robust Large-Scale Time Series Forecasting"
__1.7.3 LinkedIn's SilverKite/GreyKite
__1.7.4 Uber's Orbit
__02.
Prophet Start
____02-1 Download the Colab notebook file
____02-2 Building a Simple Prophet Model
____02-3 Interpreting the Prediction Data Frame
____02-4 Understanding the Component Plot
__03.
How Prophet Works
____03-1 Why Facebook Built Prophet
____03-2 Analyst-in-the-loop Forecast
____03-3 Prophet Formula
__3.4.1 Linear growth
__3.4.2 Logistic Growth
__3.4.3 Seasonality
__3.4.4 Public holidays
PART 02 Seasonality, Tuning, and Advanced Features
__04.
Data processing on a non-daily basis
____04-1 Monthly data usage
____04-2 Use data with a cycle shorter than one day
____04-3 Using data with regular missing intervals
__05.
Seasonality handling
____05-1 Additive vs. Multiplicative Seasonality
____05-2 Seasonality control by Fourier order
____05-3 Added custom seasonality
____05-4 Added conditional seasonality
____05-5 Seasonal Regulation
__5.5.1 Global seasonal regulation
__5.5.2 Local seasonal regulation
__06.
Predicting the effect of public holidays
____06-1 Added basic national holidays
____06-2 Additional local government (state/province) holidays
____06-3 Create a custom holiday
____06-4 Create holiday
____06-5 Public Holiday Regulations
__6.5.1 Global Holiday Regulations
__6.5.2 Individual Holiday Regulations
__07.
Growth mode adjustment
____07-1 Linear growth application
____07-2 Logistic function
____07-3 Applying logistic growth to predict saturation
__7.3.1 Increasing Logistic Growth
__7.3.2 Fluctuating upper limit cap
__7.3.3 Decreasing Logistic Growth
____07-4 Flat growth application
____07-5 Create a custom trend
__08.
Adjusting trend change points
____08-1 Automatic detection of trend change points
__8.1.1 Default Change Detection
____08-2 Changes Regulations
____08-3 Set custom change point location
__09.
Add explanatory variables
____09-1 Add binary value variable
____09-2 Add continuous variables
Interpretation of coefficients ____09-3
__10.
Outliers and Special Events
____10-1 Correcting outliers that cause seasonal variation
____10-2 Correcting outliers that cause wide uncertainty intervals
____10-3 Automatic outlier detection
__10.3.1 Windsor
__10.3.2 Standard Deviation
__10.3.3 Moving average
__10.3.4 Standard deviation of error
Modeling ____10-4 outliers as special events
____10-5 Modeling the Impact of COVID-19 Lockdowns
__11.
Uncertainty interval handling
____11-1 Trend Uncertainty Modeling
____11-2 Seasonal Uncertainty Modeling
PART 03 DIAGNOSIS AND ASSESSMENT
__12.
Run cross validation
____12-1 k-fold cross-validation
____12-2 Forward chained cross-validation
____12-3 Creating a Prophet Cross-Validation DataFrame
____12-4 Parallel Cross Validation
__13.
Performance Indicator Evaluation
____13-1 Understanding the Prophet Indicator
__13.1.1 MSE
__13.1.2 RMSE
__13.1.3 MAE
__13.1.4 MAPE
__13.1.5 MdAPE
__13.1.6 SMAPE
__13.1.7 Coverage
__13.1.8 Selecting the optimal indicator
____13-2 Creating a Prophet Performance Metrics Dataframe
____13-3 Irregular cutoff processing
____13-4 Hyperparameter Tuning with Grid Search
__14.
Prophet commercialization
____14-1 Save the model
____14-2 Fit Model Update

This book will guide you through installing and setting up Prophet on your computer, and building your initial model with just a few lines of code.
Next, we'll explore more advanced features, such as visualizing forecast results, adding holidays, seasonality, trend inflection points, and handling outliers.
You will also learn why and how to modify each basic parameter.
You will also learn to understand Fourier series and how they model seasonality, and how to apply them.
You will further learn how to optimize more complex models by tuning hyperparameters and including additional regressors in your model.
You'll run diagnostics to evaluate model performance and learn useful features for deploying Prophet in real-world production environments.
This will enable you to develop sophisticated and accurate forecasting models using concise, understandable, and reproducible code, leveraging raw time series datasets.

★Main topics covered in this book★

Chapter 1: The Evolution of Time Series Forecasting: From Early Attempts to Understand Time Series Data to the Present, Major Algorithm Developments
Chapter 2: Getting Started with Prophet: The Basic Process of How Prophet Works and How to Build Your First Model
Chapter 3: How Prophet Works: Why Meta Developed Prophet and the Analyst-in-the-Loop, Predictive Algorithm Formula
Chapter 4: Processing Non-Daily Data: How to Process and Use Various Periodic Data Other Than Daily Data
Chapter 5: Seasonality Handling: The Concept and Control of Seasonality, a Key Element of the Prophet Model
Chapter 6: Predicting the Effect of Holidays: Setting Default, Regional, and Custom Holidays and Reflecting the Effect of Holidays
Chapter 7: Adjusting Growth Modes: Linear, Logistic, and Flat Trend Modes and Their Applications
Chapter 8: Controlling Trend Changes: Adjusting Trend Line Rigidity and Flexibility to Build Optimal Models for Different Situations
Chapter 9 Adding Explanatory Variables: How to Add Multivariate Input Data Using Explanatory Variables
Chapter 10 Outliers and Special Events: Outlier Problems and Automatic Detection and Prophet-Based Processing Techniques
Chapter 11 Handling Uncertainty Intervals: Quantifying Model Uncertainty and Visualizing Techniques
Chapter 12: Implementing Cross-Validation: Cross-Validation Methods and Prophet for Time Series Data Characteristics
Chapter 13: Evaluating Performance Metrics: Optimizing Models and Improving Accuracy Using Performance Metrics and Cross-Validation
Chapter 14: Prophet Productization: Model Saving and Update Methods and Dashboard Sharing Techniques Using Plotly

★This book's target audience★

* Business managers and data analysts who want to learn time series forecasting using Python or R.
* Data scientists and analysts who want to understand the characteristics of time series data and the differences from other data types.
* Machine learning engineers who want to apply machine learning and predictive modeling to real-world projects.
* Software engineers who want to use predictive techniques while handling time series data in the software development process.