D4.7 Machine Learning Algorithms over Semi Honest Operation Modes – Final Version

This deliverable (D4.7 Machine Learning Algorithms over Semi Honest Operation Modes – Final Version) comprises the MUSKETEER Machine Learning Library (MMLL) needed to execute the distributed learning under POMs 4, 5 and 6, as well as some demonstration scripts to check the correct execution of the code. The list of available algorithms contains the models committed in this deliverable (Linear models (Linear Regression, Logistic Classifier, Ridge Regression), Kernel methods (Kernel Regression, Support Vector Machines) and Clustering (Kmeans)). The design and usage of the new models incorporated in this version of the library is analogous to the already available ones in the previous version, so the integration and use of the new methods do not present any special difficulty with respect to what is already integrated in the platform. MMLL has been successfully integrated in the MUSKETEER Client connector developed in WP7 and it is fully operative over the “pycloudmessenger” communication service implemented in WP3.