Machine Learning and Analytics Update / Visual Analysis of Data Relationships - Roger Dev, Sr Architect, LexisNexis Risk SolutionsIn this talk I will discuss our latest developments in the Machine Learning and Analytics Library. Specifically:
Gaussian Process Regression – A non-linear regression bundle, accelerated using Random Fourier Features. This is a flexible kernel-based regression method, with enhanced scalability.
HPCC Systems Causality Toolkit – Our official release of the Causality toolkit. Provides synthetic data generation and powerful probability analysis, as well as causal model analysis, and causal inference.
Then, I will move into visual analysis of data relationship. The Joint Probability Space of a multi-variate dataset is a remarkably complicated object, encompassing everything that is knowable from the data alone. We use the powerful analytic and visualization capabilities of the HPCC Systems Causality Toolkit to examine the nature of probabilistic and causal relationships within datasets. We start by examining synthetic data with known relationships in order to recognize essential patterns. Then we move on to natural datasets to see how those patterns can be recognized, and how they differ from those observed in synthetic data.
Copyright © 2022 LexisNexis Risk Solutions Group
Музыка
Фильмы
ТВ Шоу
Категории видео
