Train Ensemble Members node
This node takes your training data and trains a user defined amount of candidate networks. The goal is to later utilize methods model ensembling methods to yield a bootstrap aggregated model that best generatizes to unseen data
This node allows users to load well data for training, testing, or blind validation. The data must have been processed using QApp’s internal QLog preprocessing module.
This node performs an unsupervised lithology clustering process. Models are logged into the model management database and retrievable for use in real-time or ROP optimizer testing.
Model Interpretation node
This node generates a variety of analytics that allow the user to develop a deeper understanding of the model’s ability to learn features and how these learnings impact the final predictions.
Users can connect various trained models and perform an interactive analysis of the predictions at each well location. Users can then connect the corresponding model simulations to an ensemble node and generate the desired realization.
This node leverages a ROP targeted QLog model and a lithology clustering model to perform a validation of an ROP optimization process.
Feature Engineering node
This node takes your training data and performs feature scaling and applies Quantico’s proprietary feature engineering methodology to add spatial and physical context.
Bulk Well Ingestion Node
A version of well ingestion that allows for multiple wells to be loaded at one time.
This node allows users to load well data for training, testing, or blind validation. The data must have been processed using QApp’s internal QLog preprocessing module
Feature Engineering Node
Creates and attaches features generated from seismic and data to training table.
Parallel Seismic Simulation Node
This node leverages parallel multi-processing capabilities to significantly boost the simulation speed accross models simultaneously.
Well Simulation node
This node gives users the ability to use a previously trained model to predict properties at the wellbore. Users can load models from the model orchestration catalog and create single model realizations or bootstrap aggregated results