MACHINE LEARNING IN R
In-person events in Wiesbaden or online seminar on 1 day: €1,090 per person (net)
The seminar is aimed at R users who want to use machine learning algorithms in R.
Dates for Open Training - Crash Course: 07.03.2024, 17.10.2024
LEARNING OBJECTIVES AND AGENDA
Goals:
Overview of the different machine learning methods
Supervised vs. unsupervised learning
Possibilities for overfitting control
Non-black-box methods (ALM, GLM, recursive partitioning, etc.)
Black-box methods (neural networks, boosting and bagging, random forests, SVMs, etc.)
Deep Learning in R: Notes on Tensorflow and the use of mxnet
Specification of hyperparameters (e.g. with Grid Search, Random Grid Search)
Getting to know the mlr3 framework (caret/tidymodels on request)
Case study
IN-HOUSE SEMINAR
Seminars held at the customer's location
€1,390.00
per day up to 4 participants plus statutory VAT
All content of the in-house seminars is individually tailored and taught to specific target groups .
Intensive follow-up support enables participants to implement their knowledge in the shortest possible time.
Recommended seminar duration: 1 day
Rental fees for training notebook (on request): 60,- Euro (per day, per training computer)
WORKSHOP
You tell us your topics!
Price on request
plus statutory VAT and travel expenses if applicable
All workshop content is individually tailored and taught to specific target groups .
We are happy to conduct the workshop at your location, in Wiesbaden or online.
Rental fees for training notebook (on request): 60,- Euro (per day, per training computer)
Day 1
Overview of algorithms in machine learning
Supervised Learning: Overview
Non-black-box methods I: OLS methods + extensions (ALM, GLM etc.), recursive partitioning (trees), regularized methods (lasso, ridge regression etc.).
Black-box methods: neural networks, random forests, boosting and bagging, support vector machines
Deep Learning
Day 2
Overfitting control
Hyperparameter tuning (grid search, random grid search)
Model evaluation: parameters for model evaluation, confusion matrix, ROC curve, etc.
Hyperparameter Tuning: Grid Search and Random Search
Pipelines: Creating pipelines consisting of various data preparation and analysis steps
Machine Learning Frameworks in R: The mlr3 framework (on request: caret/tidymodels
On demand : Cox regression: Models for modeling customer churn (churn prediction)
CONTENTS
This course is aimed at anyone who wants to use machine learning to identify and use optimized models for a problem.
This course deals specifically with the modeling phase in the CRISP-DM process , in which the prepared data is then actually modeled.
Due to the considerable range of different modeling methods , most of which can still be specified via hyperparameters, this step is by no means trivial! The focus of the modeling phase is therefore on selecting suitable modeling methods and finding the best hyperparameter values.
This search phase is complicated by the varying modeling times required depending on the method and the problem of overfitting. Strategies are taught to control overfitting .
The Machine Learning Algorithms in R seminar makes getting started easier, as it's often unclear in advance which methods are suitable and how to adjust their hyperparameters. To enable the use of many different modeling methods and the variation of hyperparameters, you'll learn techniques like Grid Search and Random Grid Search , as well as the development of your own methods. Here, you'll use Cross-Validation (CV) methods to control overfitting.
Practical implementation is now supported by two data science frameworks in R: the mlr3 framework and the caret/tidymodels framework. Both frameworks allow for easy control of numerous algorithms and their hyperparameter tuning. The results can then be easily compared, and the best model can be selected based on typical fit metrics (accuracy, AUC, etc.). In addition to these two frameworks, we also demonstrate how model tuning can be performed using custom methods. The mlr3 framework also allows the use of pipelines : various data preparation and analysis steps can be automated.
Implementation : In addition to an overview of the relevant modeling techniques and typical application scenarios, the practical implementation is tested using a case study (supervised learning) in R.