Enlighten and empower your audience with visual narratives that will bring your insights to life.
With Logical Glue’s white-box machine learning platform, improving the communication and visualisation of your valuable insights is easier than ever before.
With Logical Glue, you will also gain access to new exciting techniques with Fuzzy Logic and Artificial Neural Networks.
We use the best techniques in statistics and computational intelligence to create the most accurate predictive models, while making them easy to apply for data scientists and domain experts alike.
We will show you exactly how useful your data is with a ranked list of features to understand what data you should care about the most, and which data you should spend less time and money on in future.
The models you build are hosted in the cloud and ready to use straight away, without any complex deployment process. The Logical Glue API allows easy integration into tools and workflows, such as an underwriting process.
We are the only platform to use Type-II Fuzzy Logic, which generates human-readable rules that give a clear understanding of transactions and customers, and the reasoning behind predictions.
The platform employs Fuzzy Logic which aims to replicate the process of human thinking and natural language, as well as capture the approximate, inexact nature of the real world.
Fuzzy Logic systems are universal approximators, but one of their main advantages is the ability to represent the prediction in a human readable form, unlike Neural Networks and Support Vector Machines. Importantly, Fuzzy Logic is not a probabilistic technique, but focuses on the degree of truth of an event or condition.
Fuzzy Logic is a particularly effective method for modelling uncertainty. Since its appearance in the 1960s, Fuzzy Logic has continuously grown in terms of popularity and number of applications.
The platform employs Artificial Neural Networks (ANNs), computational models that mimic central nervous systems to estimate or approximate unknown functions.
ANNs are described by a system of interconnected neurons which compute outcomes from a set of inputs. They are capable of learning, as well as pattern recognition, thanks to their adaptive nature, and are universal approximators, meaning they can approximate any function with a given accuracy.