PredictMod BCOs

← Go Back to the PredictMod Home Page.

PredictMod utilizes BioCompute Objects (BCOs) to ensure standardized documentation of machine learning pipelines.

BioCompute Objects were created as an IEEE standard for standardizing and sharing computations and analyses. In PredictMod, BCOs are used to fully describe all details of a model, including the data set used for training and associated provenance, model type and associated hyperparameters used to create the model, software requirements, and any other information required to fully replicate the model. Within the platform, model authors are given expansive freedom in creating the BCO for a description of submitted models. By adhering to BCO standards, researchers and practitioners can confidently reproduce results, validate findings, and collaborate effectively, thereby advancing scientific rigor and trustworthiness in computational biology and related disciplines. Our pipelines leverage standardized preprocessing, rigorous model training, and thorough evaluation to create an intervention outcome prediction model, supporting reproducibility and transparency in its approach.

You can search for and download BCOs for each PredictMod model at https://www.biocomputeobject.org/bcodbs.