BA-PHERD

“The project aims to design and develop an effective machine learning approach for the identification of disease biomarkers using circulating miRNA as predictors of disease onset, progression or therapeutic efficacy monitoring. The BA-PHERD project will design a complete methodology supporting the development of analytics pipelines, from data acquisition and ingestion to model learning, evaluation and sharing, aiming to address machine learning predictive tasks involving one or several of the following complexity aspects: (a) structured complex data as input, (b) massive datasets, with many examples or many input features and views (c) heterogeneous and non-independent input data for the analysis process, also representing different biological entities (d) different non-independent sensitive data sources, possibly subject to different data processing and strict data governance policies (e) the need to balance task quality and data protection.”