Machine Learning with R
Dive into data science for finance with this focused course, centered on practical R applications and machine learning for portfolio management. Kick off with foundational R packages, then quickly move to model construction with H2O and interpretation with LIME. Utilize caret and xgboost for predictor analysis, and leverage randomForest for feature importance evaluation.
Each session is designed to reinforce data handling, from manipulation with dplyr and tidyr to hypothesis testing with broom and multcomp. Learn to craft recommendation algorithms with recipes and fine-tune financial models using optimization and sensitivity analysis techniques.
Conclude by mastering Shiny for interactive UIs, and gain proficiency in deploying robust apps with rsconnect and shinytest. This course equips you with the R toolkit necessary to build and deploy sophisticated portfolio optimization models
Curriculum
- Predictions with LIME (6:02)
- Creating an explainer with LIME (14:15)
- Explainer with LIME Part2 (8:50)
- Illustrating the significance of predictors with caret (16:18)
- Depicting the weight of variables with xgboost (6:52)
- Displaying the relevance of features with LIME (6:05)
- Mapping out key predictors' influence (1:04)
- Showcasing predictor impact (13:27)
- randomForest (6:03)
- Variable Importance Plots, focused on tidy, customizable plots (8:17)
- Identify the Research Problem & Formulate the Hypotheses (11:55)
- Determine the Required Data & Collect Data with httr & rvest (9:50)
- Clean & Organize the Data with dplyr, tidyr, &, data.table (1:58)
- Statistical Tests & Method Selection with stats & MASS (7:26)
- Variable Selection & Model Parameters with caret & glmnet (9:04)
- Perform the Analysis with nnet, lme4, vegan, & nlme (16:18)
- Interpret Findings & Check Assumptions with broom (6:38)
- Confirm or Reject Hypotheses with multcomp & Car (12:25)
- Reporting with rmarkdown (13:27)
- Reporting with ggplot2 plotly (13:35)
- Optimization at Various Thresholds (9:04)
- Explaining The Optimization Results (14:29)
- Sensitivity Analysis (10:20)
- Adjusting Parameters to Retest the Assumptions (18:37)
- Iterating with pmap() (6:13)
- Overview (6:13)
- BSPF & SetUp (10:20)
- Recipes for Discretization (8:39)
- Creating A Recipe (1:58)
- Binning with step_discretize() (13:27)
- Dummy Variables & One Hot Encoding (9:04)
- Bake() function (11:55)
- Accessing the Binning Strategy Using tidy() (12:25)
- Correlation & Visualization (16:18)
- Data Manipulation Part1 (6:38)
- Data Manipulation Part2 (9:50)
- Data Manipulation Part3 (13:35)
- Visualize Discretized Correlation with ggplot2 and Plotly (16:18)
- Creating a Worksheet (5:31)
- Filling Out the Worksheet Part1 (8:46)
- Filling Out the Worksheet Part2 (1:21)
- Filling Out the Worksheet Part3 (13:08)
- Conclusion of the code and framework (4:28)
- Intro (9:33)
- SetUp fluidPage, sidebarLayout, and mainPanel (6:05)
- Server Logic with Shiny (4:28)
- Debugging (8:50)
- User Testing Part1 (1:21)
- User Testing Part2 (8:54)
- Choose a Platform (8:54)
- Quantmod & Tidyquant (14:51)
- PortfolioAnalytics (8:54)
- ShinyWidgets (6:52)
- Shinyjs (8:13)
- DT Package (6:38)
- Deployment and Production rsconnect (9:50)
- Deployment and Production shinytest (7:26)
- Deployment and Production packrat (1:58)
- Shinymanager (6:38)
- Shinymanager Part2 (8:17)
- Database connection with pool (12:25)
- Final Comments (11:55)