Machine Learning Essentials Boot Camp / Part 1: Preparing Your Data (TTML5510)

Course Objectives

Objectives

Throughout the course you will explore:

• Data Encoding: Dive into data encoding to seamlessly translate diverse information into a machine-friendly format.
• Data Manipulation Mastery: You'll get comfortable with encoding, scaling, and normalizing data. By the end of the course, the curse of dimensionality will no longer be a challenge.
• Quality Analysis Confidence: Learn how to identify and remove duplicates, handle null values, manage outliers, and work with dates in your data. You'll be a pro at maintaining clean datasets.
• Feature Analysis Wizardry: Discover how to identify unused columns, detect low variance ones, and understand multicollinearity. By the end of the workshop, feature selection will feel like second nature.
• Pipeline Proficiency: Gain a deep understanding of the critical role of pipelines in machine learning and develop the skills to create and implement your own data preprocessing pipelines.
• Machine Learning Basics: Get introduced to the fundamentals of machine learning, understand k-fold cross-validation, master the art of partitioning data, and learn how to prevent data leakage. You'll be set to step confidently into the world of machine learning.

Course Prerequisites

Audience

This course is geared for data scientists and business professionals seeking to leverage data insights in decision-making. It's also ideal for software developers wanting to diversify their skills into the exciting field of machine learning. Whether you're a student eager to jumpstart your career or an experienced professional looking to enhance your data-driven strategies, our hands-on workshop offers a valuable learning experience to transform you into a confident data handler and problem-solver.

Pre-Requisites

This is an intermediate-level program, designed to prepare attendees for a deeper dive into next-level, heavy hands-on machine learning courses and workshops. Attendees should have practical, hands-on experience working with Python for Data Science, pandas and numpy.

Take Before: Students should have incoming practical skills aligned with those in the course(s) below, or should have attended the following course(s) as a pre-requisite:

• TTPS4873      Fast Track to Python for Data Science
• TTPS4874      Applied Python for Data Science

Take After/ Follow On: The following course(s) might serve as good next steps in your learning journey.

• TTML5511    Machine Learning Essentials Boot Camp (Part 2): Machine Learning Deep Dive (3 days)
• TTML5503    Introduction to AI, AI Programming and Machine Learning (3 days)

Next Steps / Follow-on Courses: We offer a wide variety of related or follow-on courses and learning paths for Machine Learning, Data Science, Advanced Python, Artificial Intelligence, AI for Business, Generative AI, GPT, AutoGPT, AGI, Applied AI, Azure OpenAI, Google BARD, AI for developers, testers, data analytics, deep learning, programming, intelligent automation and many other related topics.  Please see our catalog for the current AI & Machine Learning Courses, Learning Journeys & Skills Roadmaps, list courses and programs.

Course Agenda

1. Getting Started with Data

• Explore the role and importance of data in machine learning.
• Encoding data: Transform raw data into a format suitable for analytics.
• Dealing with the curse of dimensionality: Navigate high-dimensional spaces effectively.
• Scaling and normalizing data: Standardize data for consistent analysis.
• Hands-on Activity / Lab

2. Structural Analysis

• Delve into the intricate patterns that define data.
• Importing libraries: Equip yourself with the right tools for data manipulation.
• Importing data: Initiate the first steps of data-driven exploration.
• Conducting basic data investigation: Peek into the essence of your dataset.
• Utilizing relevant tools for data structure analysis: Get acquainted with state-of-the-art tools to dissect data structure.
• Hands-on Activity / Lab

3. Quality Analysis

• Refine data sets by spotting and fixing errors.
• Identifying and removing duplicates: Ensure uniqueness in your dataset.
• Handling null values and missing data: Fill the gaps in your data with precision.
• Detecting and managing outliers: Understand and manage extreme data points.
• Working with dates in data: Harness the power of time-series data.
• Hands-on Activity / Lab

4. Exploratory Data Analysis

• Dive deep into data to extract meaningful insights.
• Conducting univariate analysis: Analyze one variable at a time.
• Conducting bivariate analysis: Discover relationships between two variables.
• Conducting multivariate analysis: Understand complex data interactions.
• Using pivot tables for data analysis: Summarize data visually and numerically.
• Understanding correlation: Measure linear relationships between variables.
• Understanding mutual information: Gauge dependency between variables.
• Hands-on Activity / Lab

5. Data Features

• Pinpoint the most impactful data components.
• Identifying and dropping unused columns: Streamline data for efficiency.
• Detecting and handling low variance or no variance columns: Maintain data variability.
• Understanding multicollinearity (VIF): Ensure independent predictor variables.

6. Feature Selection

• Prioritize the most relevant data features for robust models.
• Using wrappers (RFE, Forward, Backward selection): Implement dynamic feature selection.
• Using filters (Statistical tests): Opt for features based on statistical relevance.
• Using embedded methods: Integrate feature selection into algorithm functionality.
• Understanding unsupervised feature selection methods: Navigate feature selection without target variables.
• Hands-on Activity / Lab

7. Feature Importance

• Gauge the significance of different data features in prediction.
• Understanding dimensionality reduction: Simplify data without losing information.
• Using Principal Component Analysis (PCA): Transform data to highlight variance.
• Using Linear Discriminant Analysis (LDA): Optimize class separability.
• Hands-on Activity / Lab

8. Encoding, Scaling, and Skewness

• Tailor data formats for better compatibility with machine learning algorithms.
• Encoding categorical variables: Convert categories into numerical values.
• Scaling numerical variables: Maintain consistency in data magnitude.
• Detecting and correcting skewness in data: Normalize data distributions.
• Hands-on Activity / Lab

9. Pipelines

• Streamline machine learning workflows with seamless data transitions.
• Understanding the role of pipelines in machine learning: Appreciate the significance of efficient workflows.
• Creating and implementing data preprocessing pipelines: Process data in a structured manner.
• Using pipelines for efficient cross-validation and hyperparameter tuning: Optimize model parameters with ease.
• Hands-on Activity / Lab

10. Introduction to Machine Learning

• Lay the groundwork for next-level machine learning practices.
• Understanding k-fold cross-validation: Assess model performance effectively.
• Using resampling techniques: Balance dataset disparities.
• Dividing data into training and test sets: Create a structured environment for model training and evaluation.
• Identifying and preventing data leakage: Maintain the integrity of your datasets.
• Understanding the basic types and applications of machine learning models
• Capstone Project: Develop an end-to-end machine learning model: Apply the course skills to develop a complete data-driven projects.

Course Materials

Setup Made Simple! Learning Experience Platform (LXP) 

All applicable course software, digital courseware files or course notes, labs, data sets and solutions, live coaching support channels and rich extended learning and post training resources are provided for you in our “easy access, no install required” online Learning Experience Platform (LXP), remote lab and content environment.