Companies from various industries are struggling with predictive models. It takes a lot of time, iterations and experience for skilled data scientists to build and maintain such models since they are not static over time. Many try to find a shortcut, applying a brute force computing approach (AutoML) that automatically creates tons of models with different settings to find the one that works the best. It’s very computing intensive, increasing exponentially with prediction volumes.
Introducing TIM, Tangent Information Modeler. TIM builds geometric structures based on your data and automatically detects the relevant data features and feature combinations. TIM’s model building process is organic and generates accurate models in one pass. This is unique and inherently different to widely used model selection techniques, where many models are generated and then scored on their performance in a cross validation stage. Unlike results of most other popular modeling techniques, TIM’s models are not black boxes. TIM generates a formula you can interpret, providing insights into the dynamics hidden in the data.
TIM is adaptive and fast. It can update your models daily, hourly, or even in higher frequency, for a growing number of individual predictive objects. TIM is so lean on computing resources that you can do real-time predictions with your data.
TIM exists as a plug-in, web-service and web application. It has an API and it can be embedded in your appliance.
Are you curious?
Ask for an online demo now.