Clustering Methods For Drug Information Discovery

Abstract

Clustering methods for drug discovery is simply said as classification ways of objects, particularly the drugs information, into clusters. It is built based on some methods in order to get the similarity or dissimilarity, in another word, it tends to compute the distance between the objects. With the similarity information in the large set of data collected from bio-medical, we will able to learn the relationships between entities or extract unique object to decide to get further into it. There are several methods of clustering, each type has a specific convention to adapt with different types of data. For drug information, especially main text data, the used method belongs to density-based clustering. For the use case, after getting the result from clusters, we can use it to predict the drug trends or simply provide the decision of providing treatments for patients. The DrugBank database is a detailed database on small molecule and biotech drugs, it is really helpful in providing drugs information to apply the clustering method.