A NOVEL APPROACH OF WEAKLY SUPERVISED MODEL FOR COMPARABLE ENTITY MINING

To compare things which are having same function but differ in their properties is part of human decision making process.Differentiating products having same job but differing in their capabilities or properties is difficult task.Consider someone want to purchase online mobile phone then he or she should know which mobile phone to compare with how many and which alternatives.This is common activity but requires high knowledge skills.So to deal with this difficulty we developed a novel way to automatically get comparable entities from comparative questionsThe proposed system will consider comparators which are alises of each other .Given users input entity, we find comparable entities for that entity. This method achieves high precision and high recall.It is also going to rank results according to relevance to users requirements. This method achieves more than eighty percent F1-measure in both comparative question identification and comparable entity extraction.