High Performance Time Series – Matt Dancho- Business Science

Salepage link: At HERE. Archive: https://archive.is/wip/IBDQOTotal sizes: 715 MB – include: Buy now $169 $897, High Performance Time Series – Matt Dancho- Business Science Course.Become the Time Series Expertfor your organizationThe High-Performance Time Series Forecasting Course is an amazing course designed to teach Business Analysts and Data Scientists how to reduce forecast error using state-of-the-art forecasting techniques that have won competitions. You’ll undergo a complete transformation learning the most in-demand skills that organizations need right now. Time to accelerate your career.Undergo a Complete TransformationBy learning forecasting techniques that get resultsWith High-Performance Forecasting, you will undergo a complete transformation by learning the most in-demand skills for creating high-accuracy forecasts.Through this course, you will learn and apply:Machine Learning & Deep LearningFeature EngineeringVisualization & Data WranglingTransformationsHyper Parameter TuningForecasting at Scale (Time Series Groups)How it worksYour path to becoming an Expert Forecaster is simplified into 3 streamlined steps.1 Time Series Feature Engineering2 Machine Learning for Time Series3 Deep Learning for Time SeriesPart 1Time Series Feature EngineeringFirst, we build your time series feature engineering skills. You learn:Visualization: Identifying features visually using the most effective plotting techniquesData Wrangling: Aggregating, padding, cleaning, and extending time series dataTransformations: Rolling, Lagging, Differencing, Creating Fourier Series, and moreFeature Engineering: Over 3-hours of content on introductory and advanced feature engineeringPart 2Machine Learning for Time SeriesNext, we build your time series machine learning skills. You learn:17 Algorithms: 8 hours of content on 17 TOP Algorithms. Divided into 5 groups:ARIMAProphetExponential Smoothing – ETS, TBATS, Seasonal DecompositionMachine Learning – Elastic Net, MARS, SVM, KNN, Random Forest, XGBOOST, Cubist, NNET & NNETARBoosted Algorithms – Prophet Boost & ARIMA BoostHyper Parameter Tuning: Strategies to reduce overfitting & increase model performanceTime Series Groups: Scale your analysis from one time series to hundredsParallel Processing: Needed to speed up hyper parameter tuning and forecasting at scaleEnsembling: Combining many algorithms into a single super learnerPart 3Deep Learning for Time SeriesNext, we build your time series deep learning skills. You learn:GluonTS: A state-of-the-art forecasting package that’s built on top of mxnet (made by Amazon)Algorithms: Learn DeepAR, DeepVAR, NBEATS, and more!Challenges & Cheat SheetsNext, we build your time series machine learning skills. You learn:Cheat Sheets: Developed to make your forecasting workflow reproducible on any problemChallenges: Designed to test your abilities & solidify your knowledgeSummary of what you getA methodical training plan that goes from concept to production ($10,000 value)Part 1 – Feature Engineering with TimetkPart 2 – Machine Learning with ModeltimePart 3 – Deep Learning with GluonTSChallenges & Cheat SheetsCourse CurriculumWelcome to High Performance Time Series!High-Performance Time Series – Become the Time Series Expert for Your Organization (2:34)Private Slack Channel – How to JoinVideo Subtitles (Captions)What is a High-Performance Forecasting System?[IMPORTANT] System Requirements – R + Python Requirements & Common IssuesWould You Like To Become An Affiliate (And Earn 20% On Your Sales)?PrerequisitesPrerequisite – Data Science for Business Part 1Getting HelpGetting Help (IMPORTANT!!!)Module 0 – Introduction to High-Performance ForecastingHigh-Performance Forecasting – What You’re Learning, Why You’re Learning It (0:43)0.1 Forecast Competition ReviewThe Forecasting Competition Review & Course Progression (3:34) 2014 Kaggle Walmart Recruiting Challenge (5:11) 2018 M4 Competition (3:37) 2018 Kaggle Wikipedia Website Traffic Forecasting Competition (4:30) 2020 M5 Competition (5:59) 5 Key Takeaways from the Forecast Competition Review (5:41)0.2 Course Projects – Google Analytics, Email Subscribers, & Sales Forecasting The Business Case – Developing a Best-in-Class Forecasting System (3:03)0.3 What Tools are in Your Toolbox? Timetk: Time Series Data Preparation, Visualization, & Preprocessing (5:54) Modeltime: Time Series Machine Learning (5:25) GluonTS: Time Series Deep Learning (2:01) [Cheat Sheet] Forecasting WorkflowModule 01 – Time Series JumpstartTime Series Jumpstart (0:54)1.1 Time Series Project Setup Project Setup (2:28) Course Data (File Download) (1:02) R Package Installation – Part 1 (File Download) (5:26) R Package Installation – Part 2 (5:14) Jumpstart Setup (File Download) (0:44)1.2 Business Understanding & Dataset Terminology Establish Relationships, Part 1 – Google Analytics Summary Dataset (4:11) Establish Relationships, Part 2 – Google Analytics Top 20 Pages (5:23) Build Relationships – Mailchimp & Learning Lab Events (4:49) Generate Course Revenue – Transaction Revenue & Product Events (3:03)Code Checkpoint (File Download) (0:54)1.3 TS Jumpstart: Dive into Forecasting Email Subscribers! Read This! – Time Series Jumpstart Intent Time Series Jumpstart – Setup (File Download) (3:20) Libraries & Data (3:13)1.3.1 Exploratory Data Analysis for Time Series EDA for Time Series (1:08) Summarize By Time (5:46) Time Series Summary Diagnostics (4:47) Pad by Time (4:08) Visualize the Time Series (3:12)1.3.2 Evaluation & Train/Test Windows Evaluation Window – Filter By Time (4:43) Time Series Train/Test Split (4:53)1.3.3 Forecasting with Prophet Training a Prophet Model with Modeltime (4:21) Modeltime Forecasting Workflow – Round 1 (7:43)1.3.4 Forecasting with Feature Engineering  Visualizing Seasonality (4:34) Feature Engineering – Part 1 (5:45) Feature Engineering – Part 2 (5:51) Machine Learning with Workflows (3:35) Modeltime Forecasting Workflow – Round 2 (5:59)1.3.5 Recap & Code Checkpoint – Module 01 – TS Jumpstart Here’s where you are going. (3:11) Code Checkpoint (File Download)✨[Part 1] Time Series with Timetk Welcome to Part 1 – Time Series with Timetk! (2:17)Module 02 – Time Series Visualization Setup (File Download) & Overview – Visualization (2:11) Data Preparation – Part 1 (4:29) Data Preparation – Part 2 (3:23)2.1 Time Series Plots [MUST KNOW FUNCTION]  [MUST KNOW] Plotting Time Series (5:31) Plotting with Transformations (4:37) Adjusting the Smoother (6:11) Smoother for Groups (1:54) Interactive & Static Plots (2:00)2.2 Autocorrelation Plots ACF & PACF Concepts – Autocorrelation & Partial Autocorrelation ACF & PACF Plotting (7:49) Lag Adjustment (1:24) CCF Plotting – Cross Correlations (7:58)2.3 Seasonality Plots Seasonality Box Plot (5:52) Seasonality Violin Plot (0:53)2.4 Anomaly Plots Anomaly Plot Basics (4:50) Getting the Anomaly Data (2:00) Working with Grouped Data (1:43)2.5 STL Decomposition & Regression Plots STL Decomposition Plot (4:44) STL Decomposition – Grouped Time Series (2:11)2.6 Regression Plots [SECRET WEAPON FOR FEATURE ENGINEERING] [SECRET WEAPON] Time Series Regression Plot (7:08) Time Series Regression Plot – Grouped Time Series (4:05)2.7 Code Checkpoint – Module 02 – Visualization Code Checkpoint (File Download)Module 03 – Time Series Data Wrangling Setup (File Download) & Overview – Data Wrangling (2:34)3.1 Summarise By Time [MUST KNOW]  Single & Grouped Time Series Summarizations (4:37) Using Across (to Summarize Wide-Format Tibbles by Time) (5:11) Weekly/Monthly/Quarterly/Yearly Aggregations (3:33) Floor, Ceiling, Round (5:15)3.2 Pad by Time Filling in Gaps (2:54) From Low-Frequency to High-Frequency (3:36)3.3 Filter By Time Zooming & Slicing (5:14) Offsetting by Time (2:01)3.4 Mutate By Time Extrapolate the Mean, Median, Max, Min By Time (7:57)3.5 Joining By Time Combining Subscribers & Web Traffic (3:48) Inspecting the Join (3:00) Formatting the Join for Feature Relationships (5:49) Join Cross Correlations (3:22)3.6 Time Series Index Operations Making a Time Series (4:39) Making a Holiday Sequence (3:14) Time Offsets (3:01) Making a Future Time Series (3:12)3.7 Forecasting with Future Frames  The Future Frame (2:47) [FORECAST SPOTLIGHT] Forecasting with the Future Frame (6:53)3.8 Code Checkpoint – Module 03 – Data Wrangling Code Checkpoint (File Download)Module 04 – Transformations for Time Series Setup (File Download) & Overview – Transformations (2:15) Libraries & Data (2:12)4.1 Variance Reduction Transformations – Log & Box Cox [MUST KNOW]  Why is Variance Reduction Important? (4:43) Log – Log (and Log1P) Transformation (4:17) Log – Assessing the Benefit of Log1P Transformation (2:51) Log – Groups & Inversion (3:43) Box Cox – What is the Box Cox Transformation? (2:34) Box Cox – Assessing the Benefit (4:04) Box Cox – Inversion (2:05) Box Cox – Managing Grouped Transformations & Inversion (8:36)4.2 Rolling & Smoothing Transformations Introduction to Rolling & Smoothing (1:49) Rolling Windows – What is a Moving Average? (File Download) (3:53) Rolling Windows – Moving Average & Median Applied (8:53) Loess Smoother (7:02) Rolling Correlation – Slidify, Part 1 (4:16) Rolling Correlation – Slidify, Part 2 (7:40) [BUSINESS SPOTLIGHT] The Problem with Forecasting using a Moving Average (6:43)4.3 Range Reduction Transformations Introduction to Normalization & Standardization (0:59) What is Normalization? [Min = 0, Max = 1] (4:50) What is Standardization? [Mean = 0, Standard Deviation = 1] (2:31)4.4 Imputation & Outlier Cleaning Introduction to Imputation & Outlier Cleaning (0:44) Imputation – Time Series NA Repair (6:40) Anomalies – Time Series Outlier Cleaning (7:22) Anomalies – When to Remove Outliers (5:21)4.5 Lags & Differencing Transformations [MUST KNOW]  Introduction to Lags & Differencing (1:08) Lags – What is a Lag? (1:49) Lags – Lag Detection with ACF/PACF (3:54) Lags – Regression with Lags (5:06) Differencing – Growth vs Change (4:00) Differencing – Acceleration (6:22) Differencing – Comparing Multiple Time Series (4:44) Differencing – Inversion (0:57)4.6 Fourier Series [MUST KNOW]  Introduction to the Fourier Series (7:23) Fourier Regression (4:24)4.7 Constrained Interval Forecasting [FORECAST SPOTLIGHT]  What is the Log Interval Transformation? (5:47) Visualizing the Transformation (4:12) Transformations & Preprocessing (5:09) Modeling (6:29) Preparing Future Data (3:36) Making Predictions (1:05) Combining the Forecast Data (4:08) Estimating Confidence Intervals (8:24) Visualizing Confidence Intervals (2:10) Inverting the Log Interval Transformation (4:08)4.8 Code Checkpoint – Module 04 – Transformations Code Checkpoint (File Download)⛰️ Challenge #1 – Exploring Transactions & Web Page Traffic Challenge #1 Discussion (File Download) (4:21) Solution – Part 1 (File Download) (7:18) Solution – Part 2: Begins at “Identify Relationships” (7:51)Module 05 – Introduction to Feature Engineering (for Time Series) Setup (File Download) & Overview – Intro to Feature Engineering (2:30) Data Prep, Part 1 – Log Standardize (5:27) Data Prep, Part 2 – Getting Ready to Clean (5:01) Data Prep, Part 3 – Targeted Cleaning with Between Time (4:18)5.1 Time-Based Features (Trend & Seasonal/Calendar) [MUST KNOW]  The Time Series Signature (7:55) Feature Removal (3:28) Linear Trend (2:10) Non-Linear Trend – Basis Splines (4:41) Non-Linear Trend – Natural Splines (Stiffer than Basis Splines) (4:29) Seasonal Features – Weekday & Month (3:21) Seasonal Features – Combining with Trend (5:23)5.2 Interactions Interaction Features – Spikes Every Other Wednesday (7:35)5.3 Fourier Features Selecting & Adding Fourier Frequency Features (4:21) Modeling & Visualizing the Fourier Effects (2:07)5.4 Autocorrelated Lag Features Selecting & Adding Lag Features (6:59) Modeling & Visualizing the Lag Effects (5:20)5.5 Special Event Features Preparing Event Data for Analysis (6:34) Visualizing Events (2:57) Modeling & Visualizing Event Effects (2:08) Fixing the Spline (2:07)5.6 External Regressors (Xregs) Transforming Xregs (5:05) Joining Xregs (1:49) Examining Cross Correlations (1:53) Modeling with Xregs (3:28) Visualizing PageViews vs Optins & Modeling Lags (6:58)5.7 Recommended Model Features Collecting the Recommended Model (3:44) Saving the Model Artifact (2:28)5.8 Code Checkpoint – Module 05 – Introduction to Feature EngineeringCode Checkpoint (File Download)Module 06 – Advanced Feature Engineering WorkflowForecasting Workflow [CHEAT SHEET] ️ (3:40) Setup (File Download) & Overview – Advanced Feature Engineering (1:43) Data Preparation (4:42)6.1 Creating the “Full” Dataset – Extending & Adding Lagged Features & Events [IMPORTANT]  The “Full” Dataset (2:50)Extending – Future Frame (3:21) Adding Lag Features (4:02) Add Lagged Rolling Features (5:03) Add Events (External Regressors) (2:57) Format Column Names (3:09)6.2 Separate into Modeling Data & Forecast Data Data Prepared / Future Data Split (2:48)6.3 Separate into Training Data & Testing DataTrain / Test Split (3:55)6.4 Recipes – Feature Engineering Pipeline StepsRecipes Intro (2:41)Step – Time Series Signature Features (5:48)Step – Feature Removal (3:10)Step – Standardization (2:11)Step – One-Hot Encoding (1:55)Step – Interaction Features (2:28)Step – Fourier Series Features (2:03)6.5 Building the Spline Model Model Spec: LM Model (1:02) Recipe Spec: Spline Features (5:59) Workflow: Spline Recipe + LM Model (2:49)6.6 Introduction to Modeltime Workflow Modeltime Table & Calibration (2:08) Forecasting the Test Data (2:40) Measuring the Test Accuracy (1:19) Comparing the Training & Testing Accuracy (1:32)6.7 Building the Lag Model Recipe Spec: Lag Features (3:00) Workflow: Lag Recipe+ LM Model (2:40) Modeltime: Comparing Spline & Lag Models (4:23)6.8 Forecasting the Future Refitting the Models (4:37) Transformation Inversion (5:23) Visualizing the Forecast in the Original Scale (1:59) Overfitting (An Optional Fix)6.9 Saving the Artifacts Creating an Artifact List, Part 1 (4:34) Creating an Artifact List, Part 2 (3:11) Organizing the Artifacts List (1:57) Saving the Artifacts (1:28)6.10 Code Checkpoint – Module 06 – Advanced Feature Engineering Code Checkpoint (File Download)⛰️ Challenge #2 – Feature Engineering & Modeltime Workflow [YOU’VE GOT THIS!] Challenge Discussion, Part 1 (File Download) – Feature Preparation (5:11) Challenge Discussion, Part 2 – Feature Engineering & Modeling (4:56)Challenge #2 – SolutionSolution, Part 1 (File Download) – Collect & Prepare Data (3:49) Solution, Part 2 – Visualizations (3:19) Solution, Part 3A – Create Full Dataset (5:46) Solution, Part 3B – Visualize the Full Dataset (3:47) Solution, Part 4 – Model/Forecast Data Split (1:05) Solution, Part 5 – Train/Test Data Split (0:56) Solution, Part 6 – Feature Engineering (4:18) Solution, Part 7 – Modeling: Spline Model (6:08)Solution, Part 8 – Modeling: Lag Model (2:25)Solution, Part 9 – Modeltime (4:03)Solution, Part 10 – Forecast (6:49)Challenge #2 Bonus – RegularizationRegularization, Part 1 (File Download) – Model: GLMnet (4:01) Regularization, Part 2 – Improving the Lag Model with GLMNet (5:28) Regularization, Part 3 – Forecasting the Future Data with GLMNet + Lag Recipe (3:02)Part 1 Complete – You rock!  WOOO HOOO – You crushed it!✨[Part 2] Machine Learning for Time Series with ModeltimPicking Up From Part 1 (Project Download)Module 07 – Modeltime Workflow [DEEP DIVE]  Setup – Modeltime Workflow [In-Depth] (1:25) Overview – Modeltime Workflow [In-Depth] (1:16) Libraries & Artifacts Preparation (2:33)7.1 Making Models – Object Types & Requirements Model Requirements for Modeltime (1:34) Parsnip Object Models – Univariate (3:37)Workflow Objects – Multivariate, Date-Based Features (7:14)Workflow Object – Multivariate, External Features (4:53)7.2 Modeltime Table Modeltime Table – Key Requirements (4:27)7.3 Calibration Table Calibration Table – How It Works (3:29)7.4 Measuring Model Accuracy [IMPORTANT!!!] Primary Accuracy Metrics & Uses [SUPER IMPORTANT] (7:40) Custom Metric Sets using Yardstick (3:54) Customizing the Accuracy Table Output (3:28)7.5 Forecasting the Test Data Modeltime Forecast – How It Works (6:22) Customizing the Forecast Visualization (5:00)7.6 Model Refitting & Forecasting Refitting – How It Works (3:02) Making the Forecast (5:20)7.7 Code Checkpoint – Module 07A – Modeltime Workflow [In-Depth] Code Checkpoint (File Download)7.8 New Features of Modeltime 0.1.0 – Module 07B  Setup (File Download) – Modeltime New Features (1:53) Expedited Forecasting – Modeltime Table (5:20) Expedited Forecasting – Skip Straight to Forecasting (2:20) Visualizing a Fitted Model (2:57) Calibration – In-Sample vs Out-of-Sample Accuracy (5:25) Residual Diagnostics – Getting Residuals (2:16) Residuals – Time Plot (2:39) Residuals – Plot Customization (2:29) Residuals – ACF Plot (4:06) Residuals – Seasonality Plot (3:50)7.9 Code Checkpoint – Module 07B – Modeltime New Features! Code Checkpoint (File Download)Module 08 – ARIMA Setup (File Download) (0:40) ARIMA Training Overview (1:29) Libraries & Artifacts Setup (1:49)8.1 ARIMA Concepts  Auto-Regressive Functions: ar() & arima() (5:15) Auto-Regressive (AR) Modeling with Linear Regression (3:11) Single-Step Forecast for AR Models (4:43) Multi-Step Recursive Forecasting for AR Models (4:44) Integration (Differencing) (5:42) Moving Average (MA) Process (Error Modeling) (7:36) Seasonal ARIMA (SARIMA) (4:29) Adding XREGS (SARIMAX) (4:44)8.2 ARIMA in Modeltime Setting Up Basic ARIMA in Modeltime (4:45) Trying Different ARIMA Parameters (5:11) About AIC (Akaike Information Criterion) (3:42)8.3 Modeltime Auto ARIMA Implementing Auto ARIMA in Modeltime (1:49) How Auto ARIMA Works – Lazy Grid Search (1:27) Comparing ARIMA & Auto ARIMA (3:15)Adding Fourier Features to Pick Up More than 1 Seasonality (3:49)Adding Event Features to Improve R-Squared (Variance Explained) (1:33)Refitting & Reviewing the Forecast (2:57) Adding Month Features to Account for February Increase – BEST MAE 0.564 (3:35)8.4 Recap – ARIMA ARIMA Strengths & Weaknesses (and Strategies that Worked) (3:56) Saving Artifacts – Best ARIMA Model (3:28)8.5 Code Checkpoint – Module 08 – ARIMA Code Checkpoint (File Download)Module 09 – Prophet Setup (File Download) (0:27) Prophet Training Overview (0:51)Libraries & Artifacts (2:02)9.1 Prophet with Modeltime Prophet Regression: prophet_reg() (3:23) Modeltime Workflow (2:02)Adjusting the Key Prophet Parameters (5:13)9.2 Prophet Concepts  Extracting the Prophet Model from Modeltime (3:11) Visualizing the Effect of Key Parameters on the Prophet Model (5:48) Understanding Prophet Components & Additive Model (2:37)9.3 Back to Modeling with Prophet – XREGS! Fitting Prophet w/ Events (2:19) Comparing No Events vs Events – BEST MAE 0.488 (w/ Events) (3:05) Making the Forecast (2:10)9.4 Recap – Prophet Logging (Saving) Your Progress (2:40)Recap – Prophet Strengths & Weaknesses (3:02)9.5 Checkpoint – Module 09 – Prophet Code Checkpoint (File Download)Module 10 – Exponential Smoothing, TBATS, & Seasonal Decomposition Setup (File Download) (0:18) Overview – Exponential Smoothing (0:35) Libraries & Artifacts (1:37)10.1 Exponential Smoothing The Exponential Weighting Function (4:50) Applying the Exponential Weighting Function to Make a Forecast (2:41) ETS Model: exp_smoothing() (3:52) Visualizing the ETS Model (4:48)10.2 TBATS TBATS Model: seasonal_reg() (3:36) Visualizing the TBATS Model (2:48)10.3 Seasonal Decomposition Models Seasonal Decomposition & Multiple Seasonality Time Series (MSTS) Objects (2:28) STLM ETS Model (2:33) STL Plot & Relationship to STLM ETS Model (2:49) STLM ARIMA Model (1:55) STLM ARIMA – Adding XREGS (1:08)10.4 Evaluation Preparing the Test Forecast Visualization (3:30) Comparing Multiple Models – ETS, TBATS, STLM ARIMA & ETS – BEST MAE 0.523 (TBATS) (3:45) Refitting – Examining the Future Forecasts (3:34)10.5 Recap – ETS, TBATS, Seasonal Decomp Saving Artifacts (2:22) Strengths & Weaknesses – ETS, TBATS, Seasonal Decomp (2:05)10.6 Code Checkpoint – Module 10 – ETS, TBATS, & Seasonal DecompositionCode Checkpoint (File Download)⛰️ Challenge #3 – ARIMA + Prophet + ETS + TBATSChallenge #3 Discussion, Part 1 (File Download) – Start through ARIMA (5:32)Challenge #3 Discussion, Part 2 – Prophet to End of Challenge (2:33)Challenge #3 – SolutionSolution, Part 1 – Train/Test Setup (Solution File Download) (1:55) Solution, Part 2 – ARIMA (Model 1): Basic Auto ARIMA (3:03) Solution, Part 3 – ARIMA (Model 2): Auto ARIMA + Adding Product Events (2:14) Solution, Part 4 – ARIMA (Model 3): Auto ARIMA + Events + Seasonality (2:08)Solution, Part 5 – ARIMA (Model 4): Forcing Seasonality with Manual ARIMA (1:17)Solution, Part 6 – ARIMA (Model 5): Auto ARIMA + Events + Fourier Series (0:57)Solution, Part 7 – ARIMA – Modeltime Workflow (2:26)Solution, Part 8 – ARIMA – Forecast Review (3:18)Solution, Part 9 – Prophet Models: Basic (6), Yearly Seasonality (7), Events (8), Events + Fourier (9) (2:52)Solution, Part 10 – Prophet – Modeltime Workflow (1:38)Solution, Part 11 – Prophet – Forecast Review (3:13)Solution, Part 12 – Exponential Smoothing Models: ETS (10), TBATS (11) (3:24)Solution, Part 13 – Exponential Smoothing – Modeltime Workflow (1:45)Solution, Part 14 – Exponential Smoothing – Forecast Review (1:30)Solution, Part 15 – Forecasting the Future Data – ARIMA, Prophet & ETS/TBATS (3:40)Solution, Part 16 – Final Review – ARIMA, Prophet, & ETS/TBATS (2:47)Challenge #3 BONUS – ARIMA & Prophet vs Linear Model Bonus, Part 1 (File Download) – Adding the LM from Challenge #2 (4:43) Bonus, Part 2 – Why is the LM forecast high in March? (4:41)11.0 Machine Learning Algorithms [IMPORTANT]Welcome to Machine Learning for Time Series (File Download) (5:22)11.1 Elastic Net Algorithm (GLMNet) – Linear GLMNet – Model Spec (3:43) GLMNet – Spline & Lag Workflows (2:40) GLMNet – Calibration, Accuracy, & Plot (4:06) GLMNet – Tweaking Parameters – BEST MAE 0.519 (Lag Model) (2:33)*** Plotting Utility *** – Let’s make a helper function to speed evaluation up! calibrate_and_plot() (5:50) Visualizing the Effect of Parameter Adjustments (3:19)11.2 Multiple Adaptive Regression Splines (MARS) – LinearWe come from MARS (3:30)MARS – A Simple Example (6:55)MARS – Spline & Lag Models – BEST MAE 0.518 (Spline Model) (4:28)11.3 Support Vector Machine (SVM) – PolynomialSVM Polynomial – Model Specification (2:54)SVM Poly – Tweaking Parameters – BEST MAE 0.615 (Spline Model) – BOOO (5:09)11.4 Support Vector Machine (SMV) – Radial Basis Function 16% Improvement – SVM RBF vs SVM Poly (2:29) SVM RBF – Parameter Tweaking (3:11)SVM RBF – Lag Model – BEST MAE 0.520 (Spline Model) – Niiiice! (1:55)11.5 [Important Concept] KNN & Tree-Based Algorithms – The Problem with Predicting Time Series TrendsStrengths/Weakness – KNN & Tree-Based Algorithms Can’t Predict Beyond the Min/Max (1:24) KNN vs GLMNET – Making Sample Data with Trend (2:08) KNN vs GLMNET – Making Simple Trend Models (4:12) KNN vs GLMNET – Visualize the Trend Predictions w/ Modeltime – Yikes, GLMNET just schooled KNN (4:14)11.5 K-Nearest Neighbors (KNN) – Similarity (Distance) Based KNN – Spline Model (3:30)KNN – Tweaking Key Parameters (5:52)KNN – Lag Model – BEST MAE 0.558 (Spline Model) (2:05)You’re kicking butt… But, don’t forget to take breaks [COFFEE BREAK] With Bill Murray11.6 Random Forest (Tree-Based) RF – Spline Model (4:27) RF – Lag Model – 32% Better vs Spline Model (3:11) RF – Tweaking Parameters – BEST MAE 0.516 (Lag Model) (4:02)11.7 XGBoost (Gradient Boosting Machine) – Tree-Based XGBoost – Spline & Lag Models (5:00) XGBoost – Tweaking Parameters – 0.484 MAE (Lag Model) (6:35) XGBoost – Tweaking Parameters 2 – BEST MAE 0.484 (Lag Model) (3:32)11.8 Cubist – Combo of Trees (Rules) + Linear Models at Nodes Cubist – Spline & Lag Models – 0.514 MAE out of the gate! (4:53) Cubist – Tweaking Parameters – OPTIMAL MAE / R-SQUARED (0.524 / 0.316) (5:48)11.9 Neural Net (NNET) – Like a Linear Regression but Better NNET – Spline & Lag Models (4:57) NNET – Tweaking Parameters – BEST MAE 0.553 (Spline Model) (5:39)11.10 NNETAR – Combining AR Terms with a NNET! What the heck is NNETAR? (NNET + ARIMA – IMA = NNETAR) (2:22) NNETAR – Model, Recipe, & Workflow (4:11) NNETAR – Tweaking AR Parameters (2:24)NNETAR – Tweaking NNET Parameters – BEST MAE 0.512 (4:13)11.11 Modeltime Experimentation ReviewOrganizing in a Modeltime Table (4:22)Updating the s Programmatically (4:02Model Selection – Process & Tips (using Accuracy Table) (3:39) Model Inspection – Process & Tips (using Test Forecast Visualization) (3:03)Model Inspection – Visualizing the Future Forecast (5:42)11.12 Saving Your Work – Artifacts! Saving Models (2:34) Saving your calibrate_and_plot() function (1:29)11.13 Checkpoint – Module 11 – Machine Learning Algorithms Code Checkpoint (File Download)12.0 Boosted Algorithms – Prophet Boost & ARIMA Boost Boosted Algorithms – A Powerful Technique for Improving Performance (3:37)12.1A Prophet Baseline Model Baseline: Best Prophet Model (2:38) [Pro Tip] How to Fix a Broken Model (2:50) Prophet Baseline – Best Model MAE 0.488 (0:54)12.1B Prophet Boost Recipe for Prophet Boost (3:33) Model Strategy – Using XGBOOST for Seasonality/XREG Modeling (4:39) Workflow – No Parameter Tweaking (3:41) [KEY CONCEPT] Prophet Boost – Modeling Trend with Prophet, Residuals with XGBoost (3:00) Prophet Boost – Tweaking Parameters – BEST MAE 0.457 (6:33)12.2 ARIMA Boost Modeling Strategy – ARIMA for trend, XGBOOST for XREGS (3:50) ARIMA Boost – Model Specification (5:57) ARIMA Boost – Tweaking Parameters – BEST MAE 0.523 (4:34)12.3 Boosted Models – Modeltime Workflow Modeltime – Accuracy Evaluation & Identifying Broken Models (2:43) Modeltime – Forecast Test Data (2:10) Modeltime – Refitting & Forecasting Future (3:08) Save Your Work (1:26)12.4 Code Checkpoint – Boosted AlgorithmsCode Checkpoint (File Download)13.0 Hyper Parameter Tuning & Cross Validation – For Time Series Hyperparameter Tuning for Time Series (File Downloads) (3:56)[CHEAT SHEET] Hyperparameter Tuning Workflow (4:47) Getting Started – Setup & Workflow (3:09)13.1 Reviewing 28 Models (It’s Easy with Modeltime) Combining Our Artifacts – 28 Models! (3:06) Accuracy Review & Hyperparameter Tuning Candidate Selection (This Used to Take Me Weeks To Do) (4:36)13.2 [SEQUENTIAL MODELS] NNETAR – Hyperparameter Tuning Process What are Sequential Models? (& Why do we need to tune them differently?) (2:55) Extracting the Workflow from a Modeltime Table: pluck_modeltime_model() (1:40)Time Series Cross Validation (TSCV) Specification, Part 1: time_series_cv() (4:34) Time Series Cross Validation (TSCV), Part 2: plot_time_series_cv_plan() (4:14) Identify Tuning Parameters – Recipe Spec (3:07)Identify Tuning Parameters – Model Spec (5:14)Make a Grid for Parameters – Grid Spec (5:55)13.2.1 – NNETAR Tuning, Round 1 – Default ParamsGrid Latin Hypercube Specification: grid_latin_hypercube() (3:19)Tuning Workflow Preparation (3:30)Tune Grid & Show Reults (7:24)Visualize the Parameter Results (3:24)13.2.2 NNETAR Tuning, Round 2 – Finding the Sweet Spot!Update Grid Parameter Ranges (8:13)Parallel Processing – Speed-Up Tuning (5:13)Speed Comparison (Parallel vs Series) – 3.4X Speed Boost (44 sec vs 151 sec)Review Parameters vs Performance Metrics (1:09)NNETAR – Train the Final Model – Best RMSE 0.507 (4:15)13.3 [NON-SEQUENTIAL MODELS] Prophet Boost – Hyperparameter Tuning ProcessWhat are Non-Sequential Models? (2:44)Model Extraction: pluck_modeltime_model() (1:04)K-Fold Cross Validation (Use with Non-Sequential Models ONLY) (4:23)Prophet Boost – Recipe (1:10)Prophet Boost – Model Spec (Identify Parameters for Tuning) (3:57)13.3.1 Prophet Boost Tuning, Round 1 – Default ParametersGrid Specification – Grid Latin Hypercube w/ Default Parameters (4:52)Tuning the Grid (in Parallel) (6:18)Visualize Results – Learning Rate Dominates ⚡ (2:58)13.3.2 Prophet Boost Tuning, Round 2 – Controlling Learning Rate Grid Specification – Controlling Learning Rate (4:45)Hyperparameter Tuning – Round 2 – We can see parameter trends! (3:17)13.3.3 Prophet Boost Tuning, Round 3 – Honing InGrid Specification & Tuning – Honing the parameter ranges in (5:49)Best RMSE Model (Central Tendency) – MAE 0.466, RMSE 0.630, RSQ 0.450 (6:13)Best R-Squared Model (Variance Explained) – MAE 0.464, RMSE 0.643, RSQ 0.459 (2:42)13.4 Saving Our ProgressRecap & Saving the Models (6:53)13.5 Code Checkpoint – Model 13 – Hyperparameter TuningCode Checkpoint (File Download)14.0 Ensemble Time Series Models (Stacking)Competition Ensembling Review (5:57)What is an Ensemble Model? (7:21)Modeltime Ensemble: Documentation (2:01) Forecasting Cheat Sheet Upgrade ️ [Download Here] (1:00)14.1 Model Performance ReviewCode Setup [File Download] (6:49)Reviewing Models – Combining Tables & Organizing Results (4:24)Reviewing Models – Making Sub-Model Selections (7:46)14.2 Average EnsembleMean Ensemble – RMSE 0.640 vs 0.630 (Best Submodel) (5:00)Median Ensemble – RMSE 0.648 vs 0.630 (Best Submodel) (2:23)14.3 Weighted Average EnsemblesIntroduction to Weighted Ensembles (1:02)Loading Selection (4:29)Accuracy Assessment – RMSE 0.628 vs RMSE 0.630 (Baseline) (2:37)14.4.A Stacked Ensembles – Stacking ProcessIntroduction to Meta-Learner Ensembling with Modeltime Ensemble (3:57)Resampling: Time Series Cross Validation (TSCV) Strategy (5:17)Making Sub-Model CV Predictions – modeltime_fit_resamples() (4:27)Resampling & Sub-Model Prediction: K-Fold Strategy (6:28)Linear Regression Stack – TSCV – RMSE 1.00 (Ouch!) (7:16) Linear Regression Stack – K-Fold – RMSE 0.651 (Much Better, but We Can Do Better) (3:25)14.4.B Stacked Ensembles – Stacking with Tunable AlgorithmsGLMNET Stack – RMSE 0.641 (On the right track) (6:38)Modeltime Ensemble – In-Sample Prediction Error – Bug Squashed (1:10)Random Forest Stack – RMSE 0.587!!! (7% improvement) (4:33)Neural Net Stack – RMSE 0.643 (4:05)XGBoost Stack – RMSE 0.585!!! (4:29)Cubist Stack – RMSE 0.649 (3:11)SVM Stack – RMSE 0.608!! (3:26)14.5 Multi-Level StackingLevel 2 – Model Evaluation & Selection (4:27)Level 3 – Weighted Ensemble Creation, Evaluation, & Selection – RMSE 0.595 (Level 2 RF is New Baseline RMSE 0.585) (3:34)14.6 Modeltime Workflow for EnsemblesEnsemble Calibration (4:45)Ensemble Refitting, Method 1: Retraining Submodels Only (5:43)Ensemble Refitting, Method 2: Retraining both Sub-Models & Super-Learners (5:33)14.7 Saving Your WorkSave the Multi-Level Ensemble (1:27)Object Size: 50MB! Here’s why. (3:15)14.8 Code Checkpoint – Module 14 – Ensemble Methods Code Checkpoint [File Download]15.0 Forecasting at Scale – Time Series Groups [Panel Data]Welcome to Module 15 – Forecasting at Scale using Panel Data (Non-Recursive) Strategies (2:30)Setup [File Download] (4:30)15.1 Data Understanding & PreparationData Understanding (4:33)Data Prep, Part 1: Padding by Group | Ungrouped Log Transformation (3:53)Data Prep, Part 2: Extend by Group (2:44)Data Prep, Part 3: Fourier Features & Lag Features by Group (6:03)Data Prep, Part 4: Rolling Features by Group | Adding a Row ID (4:59)Future & Prepared Data – Preparation (7:34)15.2 Time Splitting – Train/TestTime Series Split (Train/Test) (3:50)15.3 Preprocessing & RecipesCleaning Outliers by Group (5:18)Recipe, Part 1: Time Series Calendar Features (3:24)Recipe, Part 2: Normalization (Standardization) & Categorical Encoding (5:36)15.4 Modeling: Make 7 Panel ModelsPanel Model 1: Prophet with Regressors (2:11)UPDATE: HARDHAT 1.0.0 FIXPanel Model 2: XGBoost (2:41)Panel Model 3: Prophet Boost (1:57)Panel Model 4: SVM (Radial) (2:02)Panel Model 5: Random Forest (1:31)Panel Model 6: Neural Net (1:27)Panel Model 7: MARS (1:27)Accuracy Check – This will help us