Machine Learning
Data Wrangled values elegantly brought together into a score to detect new typology while guarding against confirmed typology past.
![Machine Learning](/images/machine-learning/machine-learning-performance.png)
Jube takes a novel approach to artificial intelligence, ultimately Supervised Learning, yet blending anomaly detection with confirmed class data to ensure datasets of sufficient amounts of class data. Using data archived from its processing, Jube searches for optimal input variables, hidden layers and processing elements. The result is small, optimal, generalised and computationally inexpensive models for efficient real-time recall. The approach taken by Jube allows artificial intelligence’s benefits to be available very early in an implementation’s lifecycle. It avoids over-fitting models to typology long since passed.
![Machine Learning configuration.](/images/machine-learning/machine-learning-configuration.png)
![Machine Learning statistics.](/images/machine-learning/machine-learning-statistics.png)
![Machine Learning configuration.](/images/machine-learning/machine-learning-selected-model-with-variables.png)
![Machine Learning statistics.](/images/machine-learning/machine-learning-promoted-model-testing.png)
![Machine Learning variable relationships.](/images/machine-learning/machine-learning-variable-relationships.png)
![Machine Learning FaaS example.](/images/machine-learning/machine-learning-faas.png)