Machine Learning with MATLAB

Choose from a range of classification, clustering, and regression algorithms, including “shallow” neural nets (up to three layers), among other machine learning models. Use classification and regression apps to interactively train, compare, tune, and export models for further analysis, integration, and deployment. If you prefer to write code, feature selection and parameter tuning can help you improve models even more.

Use known interpretability methods (Shapley values, generalized additive models, LIME, and partial dependency graphs) to overcome the problematic black-box nature of a machine learning system. Verify that the model is making predictions with the right evidence, and look for model biases that were not obvious during training.

Using hyperparameter tuning approaches such as Bayesian optimization, it automatically generates features from training data and optimizes models. For signal or picture data and feature selection techniques (neighborhood component analysis (NCA), minimum redundancy maximum relevance (MRMR), and sequential feature selection), use specific feature extraction techniques (wavelet scattering).

Machine learning models through MATLAB function blocks and native Simulink blocks will help you verify and validate your high-fidelity simulations faster. With the help of statistics and machine learning models, you can generate C or C++ code for the whole machine learning algorithm, including pre- and post-processing steps.

With minimal code changes, tall arrays that train machine learning models can help you fit large data sets into memory. Parallel computing on your desktop, in clusters, or in the cloud can help you speed up statistical computations and model training.