A Learning to Rank Approach for Pharmacist Assignment

With people focus much more on their health, the need of Chinese medicine is increasing heavily. There are thousands of kinds of Chinese medicine prescriptions. Different pharmacists are familiar with different prescriptions and a single pharmacist is not likely to deal with all the prescriptions well. Therefore, there is a need to find the most proper pharmacist for each prescription so that the quality and efficiency for pharmacists dealing with prescriptions can be improved. To solve the problem, we propose a novel approach by leveraging learning to rank algorithm. The model built by our approach can be used to automatically recommend which pharmacist is the most proper for an unknown labeled prescription. With experiments on a Chinese medicine dataset, we demonstrate that our approach can better achieve pharmacist assignment. In particular, when compared with the baseline, our approach can achieve an improvement of over 300% in terms of MAP. With the learning to rank approach, we can achieve automated pharmacist assignment for different kinds of Chinese medicine prescriptions and improve the quality and efficiency for pharmacists dealing with prescriptions.