Summary

In this chapter, we used the mlr and OpenML packages from R to build an entire machine learning workflow for solving a multilabel semantic scene classification problem. The mlr package offered a rich collection of machine learning algorithms and evaluation measures that helped us in quick implementation and facilitated a faster experimentation process to get the best model for the problem. The package also offered many wrapper functions to handle the multilabel problem. Building real-world machine learning models using a robust framework such as the one in mlr helps in speeding the implementation and provides a structure to the complete project. Further, using OpenML, we could reproduce a research work using the already available dataset and code, and then modify it according to our need. Such a platform offers the ability to collaborate at scale with researchers all over the world. At the end, we could also upload our own machine learning flows with others for them to pick it up...