IBM Introduction to Machine Learning

This module introduces learners to the importance of good, quality data in machine learning. Participants will learn common techniques to retrieve, clean, and prepare data for preliminary analysis and hypothesis testing. By the end of the module, learners will be equipped to retrieve data from multiple sources, apply feature engineering, handle categorical and ordinal features, and use techniques for detecting and dealing with outliers.

• Retrieve and clean data for analysis
• Apply feature engineering and handle missing values
• Describe and use common feature selection techniques
• Understand the importance of feature scaling and use scaling techniques

Learners will be introduced to supervised machine learning for regression, focusing on training regression models to predict continuous outcomes. Participants will gain an understanding of error metrics, best practices for model comparison, train and test splits, and regularization techniques. By the end of the module, learners will be able to differentiate uses and applications of classification and regression, describe and use linear regression models, and articulate the benefits of regularization.

• Train regression models to predict continuous outcomes
• Use error metrics for model comparison
• Apply best practices for model comparison and regularization

This module delves into supervised machine learning for classification, guiding learners in training predictive models to classify categorical outcomes. Participants will engage in hands-on exercises focusing on best practices for classification, including train and test splits, and handling data sets with unbalanced classes. By the end of the module, learners will be proficient in using logistic regression models, decision tree and tree-ensemble models, and various ensemble methods for classification.

• Train predictive models to classify categorical outcomes
• Use error metrics for model comparison and handling unbalanced classes
• Apply best practices for classification

Learners will explore unsupervised machine learning, discovering insights from data sets without a target or labeled variable. The module covers clustering and dimension reduction algorithms for unsupervised learning, providing participants with the knowledge to select the most suitable algorithm for their data. By the end of the module, learners will understand problems suitable for unsupervised learning approaches, the curse of dimensionality, and common clustering and dimensionality-reduction algorithms.

• Explore insights from data sets without a target variable
• Understand clustering and dimension reduction algorithms
• Select the most suitable algorithm for their data