Building an accurate predictive model could mean combing through a near infinite combination of data transformations, features, algorithms and tuning parameters. DataRobot narrows down the search universe based on the characteristics of the training dataset and prediction target. It then executes only the most relevant end-to-end procedures for fitting a model (called Modeling Blueprints), to deliver the best predictive model in the fastest time possible. DataRobot uses cloud computing to cost-effectively evaluate thousands of Modeling Blueprints in parallel. It then systematically applies a cross-validation framework to accurately compare the performance of even the most diverse modeling techniques. This will be added to the tools section of Research Resources Subject Tracer™ Information Bot. This will be added to Prediction Markets Subject Tracer™. This will be added to Artificial Intelligence Resources Subject Tracer™. This will be added to Bot Research Subject Tracer™. This will be added to Statistics Resources and Big Data Subject Tracer™.
posted by Marcus Zillman |
2:58 AM
