Course Outline
Introduction
This module offers an overview of the appropriate scenarios for employing 'machine learning', key considerations, and its implications, including advantages and limitations. It covers data types (structured, unstructured, static, streamed), data quality and volume, the distinction between data-driven and user-driven analytics, comparisons between statistical and machine learning models, the challenges of unsupervised learning, the bias-variance trade-off, iterative evaluation, cross-validation strategies, and the paradigms of supervised, unsupervised, and reinforcement learning.
CORE TOPICS
1. Grasping Naive Bayes
- Foundations of Bayesian methods
- Probability concepts
- Joint probability
- Conditional probability and Bayes' theorem
- The Naive Bayes algorithm
- Classification using Naive Bayes
- The Laplace estimator
- Handling numerical features with Naive Bayes
2. Grasping Decision Trees
- The divide-and-conquer approach
- The C5.0 decision tree algorithm
- Selecting optimal splits
- Pruning decision trees
3. Grasping Neural Networks
- Transition from biological to artificial neurons
- Activation functions
- Network structure
- Determining the number of layers
- Direction of data flow
- Determining nodes per layer
- Training networks via backpropagation
- Deep Learning
4. Grasping Support Vector Machines
- Classification via hyperplanes
- Maximizing the margin
- Handling linearly separable data
- Handling non-linearly separable data
- Applying kernels for non-linear spaces
5. Grasping Clustering
- Clustering as a machine learning objective
- The k-means clustering algorithm
- Utilizing distance for cluster assignment and updates
- Determining the optimal number of clusters
6. Evaluating Classification Performance
- Interpreting classification prediction data
- Deep dive into confusion matrices
- Assessing performance using confusion matrices
- Metrics beyond accuracy
- The kappa statistic
- Sensitivity and specificity
- Precision and recall
- The F-measure
- Visualizing performance trade-offs
- ROC curves
- Predicting future performance
- The holdout method
- Cross-validation
- Bootstrap sampling
7. Optimizing Standard Models for Enhanced Performance
- Leveraging caret for automated parameter tuning
- Developing a basic tuned model
- Customizing the tuning workflow
- Enhancing model output via meta-learning
- Concepts of ensembles
- Bagging
- Boosting
- Random forests
- Training random forests
- Assessing random forest performance
ADDITIONAL TOPICS
8. Classification via Nearest Neighbors
- The kNN algorithm
- Distance calculation
- Selecting the appropriate k value
- Data preparation for kNN
- The lazy nature of the kNN algorithm
9. Classification Rule-Based Methods
- The separate-and-conquer strategy
- The One Rule algorithm
- The RIPPER algorithm
- Deriving rules from decision trees
10. Fundamentals of Regression
- Simple linear regression
- Ordinary least squares estimation
- Correlations
- Multiple linear regression
11. Regression and Model Trees
- Incorporating regression into tree structures
12. Association Rule Learning
- The Apriori algorithm for association rules
- Evaluating rule significance – support and confidence
- Generating rule sets using the Apriori principle
Supplementary Materials
- Spark, PySpark, MLlib, and Multi-armed bandits
Requirements
Proficiency in Python
Testimonials (7)
I thoroughly enjoyed the training and appreciated the deeper dive into the subject of Machine Learning. I appreciated the balance between theory and practical applications, especially the hands-on coding sessions. The trainer provided engaging examples and well-designed exercises that enhanced the learning experience. The course covered a wide range of topics, and Abhi demonstrated excellent expertise by answering all questions with clarity and ease.
Valentina
Course - Machine Learning
I appriciated the exercise that help me to undersand the theory and apply it step by step . as well the way the trainer explained everything in a simple and clear manner. It was easy to follow even though I'm not very experienced with Python, still, I didn't want to miss the opportunity to learn something that relly interests me. I also appreciated the variety of information provided and the trainer’s availability to explain and support us in understanding the concepts. After this course, machine learning concepts are much clear to me, and now I feel like I have a direction and a better undersantind of the topic.
Cristina
Course - Machine Learning
At the end of the training, I could see the real-life use-case of the subjects presented.
Daniel
Course - Machine Learning
I liked the pace, I liked the balance between theory and practice, the main topics covered and the way the trainer was able to put everything into balance. I also really like your training infrastructure, very practical to work with VMs
Andrei
Course - Machine Learning
Keeping it short and simple. Creating intuition and visual models around the concepts (decision tree graph, linear equations, calculating y_pred manually to prove how the model works).
Nicolae - DB Global Technology
Course - Machine Learning
It helped me achieve my goal of understanding ML. Much respect for Pablo for giving a proper introduction in this topic, since it becomes obvious after 3 days of training how vast this topic is. I have also enjoyed A LOT the idea of virtual machines you have provided, which had very good latency! It allowed every coursant to do experiments at their own pace.
Silviu - DB Global Technology
Course - Machine Learning
The way practical part, seeing the theory materializing into something practical is great.
