What does TUBR do?
Making predictive analytics accessible for everyone
Our Stats
TUBR Accuracy
80-85% accuracy from first processing of data (will improve over time). In comparison, traditional ML accuracy is only 75%.
Minimum Amount of Data
Traditional prediction systems require full and complete data. We need only 20% of that data to generate predictions and business insights.
TUBR vs traditional time-series ML
Traditional ML
Statistical Methodologies
Statistical ML uses large amounts of data to identify trends that have occurred in the past to determine if the same trend is going to occur again.
Model depends on data
Assumes available data to be the only variables to represent what is happening in an area.
Can’t handle chaos
Does not account for chaos in an area.
Requires lots of data
Requires the full data to make predictions.
TUBR ML
Physics Methodologies
Physics-based methodologies use information about the dynamics of the area in which it’s predicting to inform the model as to what’s happening, enabling the ML to identify a changing trend before it’s been seen with few data points.
Model built dynamically
Understands all the elements that occur in an area and brings understanding from the available data.
Accepts chaos
Accounts for chaos in an area (and in your data!).
Requires whatever data you have
Accepts sparse or partial inputs so the data you have is enough.
Three simple steps to success
Provide your data
No manual data crunching with complete or incomplete data
We analyse data
Our advanced algorithms are accurate with small and large data sets
Get your predictions
In just 6 days you can get your accuracy to over 85% predictions and insights