16 September 2016, Bangalore – MathWorks today introduced Release 2016b (R2016b) with new capabilities that simplify working with big data in MATLAB. Engineers and scientists can now more easily work with data too big to fit in memory. R2016b also includes additional features in Simulink; a new product, Risk Management Toolbox; and updates and bug fixes to 83 other products.
Tall arrays now provide a way to work naturally with out-of-memory data using familiar MATLAB functions and syntax, removing the need to learn big data programming. Engineers and scientists can use tall arrays with hundreds of math, statistics, and machine learning algorithms. Code can run on Hadoop clusters or be integrated directly into Spark applications.
R2016b also includes a timetable data container for indexing and synchronizing time-stamped tabular data; string arrays to help manipulate, compare, and store text data efficiently; and new functions for preprocessing data.
“Companies are awash in data, but struggle to take advantage of it to build better predictive models and gain deeper insights,” says David Rich, MATLAB marketing director, MathWorks. “With R2016b, we’ve lowered the bar to allow domain experts to work with more data, more easily. This leads to improved system design, performance, and reliability.”
MATLAB Product Family Updates Include:
o Tall arrays for manipulating data too big to fit in memory
o Timetable data container for indexing and synchronizing time-stamped tabular data
o Ability to define local functions in scripts for improved code reuse and readability
o Capabilities for running MATLAB code from Java programs with the MATLAB Engine API for Java
· MATLAB Mobile: Data logging from iPhone and Android sensors on the MathWorks Cloud
· Database Toolbox: Graph database interface for retrieving Neo4j data
· MATLAB Compiler: Support for deploying MATLAB applications, including tall arrays, on a Spark cluster
· Parallel Computing Toolbox: Ability to process big data with tall arrays in parallel on your desktop and on servers and Spark clusters with MATLAB Distributed Computing Server
· Statistics and Machine Learning Toolbox: Big data algorithms for processing out-of-memory data including dimension reduction, descriptive statistics, k-means clustering, linear regression, logistic regression, and discriminant analysis
· Statistics and Machine Learning Toolbox: Bayesian optimization for automatically tuning machine learning algorithm parameters, and neighborhood component analysis (NCA) for choosing machine learning model features
· Statistics and Machine Learning Toolbox: Automatic C/C++ code generation support for SVM and logistic regression models with MATLAB Coder