• Import and organize data
• Preprocess time-based signals and extract key features in the time and frequency domains
• Build classification and regression model using Statistic and Machine Learning toolbox
• Ensemble data and train model to predict remaining useful life (RUL) with Predictive Maintenance Toolbox

DAY 1 of 3: Signal Preprocessing and Feature Extraction for Data Analytics with MATLAB

Explore and Analyze Signals (Time Series) in MATLAB

Objectives: Learn to easily import and visualize multiple signals or time series data sets to gain insights into the features and trends in the data.

• Import, visualize, and browse signals to gain insights
• Make measurements on signals
• Compare multiple signals in the time and frequency domain
• Perform interactive spectral analysis
• Extract regions of interest for focused analysis
• Recreate analysis with auto-generated MATLAB scripts

Preprocess Signals to Improve Data Set Quality

Objectives: Learn techniques to clean signal sets with operations such as resampling, removing outliers, and filling gaps.

• Perform resampling to ensure common time base across signals
• Work with non-uniformly sampled data
• Find gaps in data and remove or fill gaps
• Remove noise and unwanted frequency content
• Perform wavelet denoising
• Use the envelope spectrum to perform fault analysis
• Locate outlier values in data and replace them with acceptable data
• Locate signal changepoints and use boundaries to automatically create signal segments

Extract Features from Signals

Objectives: Apply different techniques in time and frequency domains to extract features. Become familiar with the spectral analysis tools in MATLAB and explore ways to bring out features for multiple signals.

• Locate peaks
• Locate desired signals from patterns in the time and spectral domains
• Use spectral analysis to extract features from signals
• Perform classification using supervised learning
• Use the Classification Learner app to interactively train and evaluate neural networks

Day 2 of 3: Machine Learning and Predictive Maintenance with MATLAB

Importing and Organizing Data

Objectives: Bring data into MATLAB and organize it for analysis, including normalizing data and removing observations with missing values.

• Data types
• Tables
• Categorical data
• Data preparation

Finding Natural Patterns in Data

Objectives: Use unsupervised learning techniques to group observations based on a set of explanatory variables and discover natural patterns in a data set.

• Unsupervised learning
• Clustering methods
• Cluster evaluation and interpretation

Building Classification Models

Objectives: Use supervised learning techniques to perform predictive modelling for classification problems. Evaluate the accuracy of a predictive model.

• Supervised learning
• Training and validation
• Classification methods

DAY 3 of 3: Machine Learning and Predictive Maintenance with MATLAB

Improving Predictive Models

Objectives: Reduce the dimensionality of a data set. Improve and simplify machine learning models.

• Cross validation
• Hyperparameter optimization
• Feature transformation
• Feature selection
• Ensemble learning

Building Regression Models

Objectives: Use supervised learning techniques to perform predictive modelling for continuous response variables.

• Parametric regression methods
• Nonparametric regression methods
• Evaluation of regression models

Estimating Time to Failure

Objectives: Explore data to identify features and train decision models to predict remaining useful life (RUL).

• Data Organization & Labeling using Data Ensembles
• Condition Indicator Design
• Remaining Useful Life (RUL) Estimator Models
• RUL Estimation Using Dynamic Models