A Mutual Conditional Probability based document ranking model using Map-Reduce Framework

Abstract

As the volume of biomedical text increases exponentially, automatic indexing becomes increasingly important. However, existing approaches do not distinguish central(or core) concepts from concepts that were mentioned in passing.Since, the size of biomedical literature creates a new challenge for researchers to find and extract gene-based disease ranking documents in historical databases. Also, the multilevel gene documents suffer from noisy and duplicate features, it is difficult to rank and summarize the relevant phrases within the multiple document sets. Most of the text ranking and prediction algorithms are sensitive to noise, outliers, high dimensionality and uncertainty. In this proposed approach, a novel biomedical document ranking and prediction model was implemented to find interesting patterns on high dimensional biomedical dataset. Experimental results proved that the proposed probabilistic ranking model has high accuracy compared to traditional text ranking and prediction models.