Research On Potential Knowledge Discovery Of Drug That Based On Deep Learning

Drugs are a special substance that is closely related to public health and safety.With the continuous improvement of public health awareness,people’s attention to medication is also increasing.People enjoy the drug brought about by the treatment of diseases,enhance the physical benefits of the same time,the additional effect of drugs on the human body is still people uneasy.Disappointingly,despite today’s science and technology to an unprecedented level in human history,it is still difficult to fully grasp the nature of all drugs.The extra malignancies effects of drugs on the human body are known as adverse drug reactions,which is one of the public’s top concerns about drug safety.According to the US Food and Drug Administration announced that adverse drug reactions have become the fifth leading cause of human death.How to early detection and solve the potential adverse drug reactions,has become a drug manufacturers,regulators,doctors and patients,as well as the whole society in front of the hot issues.Drugs before the listing of clinical trials is the most important means of controlling the occurrence of adverse drug reactions,but because of its limitations,still can not reveal all the adverse events.With the advent of the Web2.0 era and the rapid growth of medical literature resources,which contains a lot of valuable resources.The use of computer means to automatically mining useful information contained in these data has become an important way to tap the adverse drug reactions.At the same time,drug repositioning,that is,the potential for the extraction of existing drugs,has also become a hot topic of concern.The development of new drugs is a cost and a huge project,if you can use the existing drugs to treat other diseases,will bring countless economic benefits at the same time,to bring more people health,which is extremely practical significance.The use of computer methods to tap the potential drug treatment goals,and then delivered to the medical staff for clinical validation,can greatly reduce the medical staff’s work pressure,improve overall efficiency,therefore,has become an important means of research on drug repositioning.It is encouraging to note that computer science,represented by in-depth learning,has made significant progress in recent years and has achieved remarkable results in areas such as computer vision,speech recognition,and natural language processing.The main work of thispaper is to introduce the deep learning into the research of drug implied nature,and use the convolution neural network,Word2 Vec,LDA topic model and so on.We can use multiple drug-related databases and tools to dig up the adverse drug reactions and potential of drug treatment goals.At the end of this paper,the performance of the methods used in this paper is validated by comparing experiments and verifying the results in the relevant database.