A Generic Ranking Service on Scientific Datasets

Different ranking algorithms have been proposed to fulfil the need of ranking. The problem is that most of the existing algorithms and models are just applicable on a specific data. When the data is imbalanced and heterogeneous, finding the records belonging to the minority class is significant especially in failure cases. So considering ranking as a classification problem of predicting the specific relevance score for any category, we are going to propose a generic ranking service. In this model, a metric learning based ranking model is proposed which can be used on wide range of scientific data sets. A real world imbalanced and heterogeneous data set is used to prove the efficiency of model.