The Application Of Data Mining In Development Of Drug Directions

The development of Medication is a thing that need spend lots of money and containsgreat venture. The term of development of Medication all most need ten years in foreigncountry and the expense of that is 3 to 5 hundred million, some can attain to billions. Theexpense of development makes up 15% to 20% of the total sale. From above, we can seethat the development of Medication is a difficult job. On the contrary, people have lots ofdrugs to cure kinds of diseases through accumulation and innovation. There are lots of drugresources in "Chinese Pharmacopoeia", "Japanese Pharmacopoeia, "U.S. Pharmacopoeia"and so on. From those resources we can come to a conclusion that a drug may cure manykinds of diseases(same drug different disease) and a disease can be cure by many kinds ofdrug(same disease different drug). That is common in Traditional Chinese Medicine. Thatproves the relationship of composition of drug and disease is complex. If we prove up therelationship and know effect of each composition upon the disease, we may get a neweffective drug by recombining compositions. Data Mining Technologies find interesting knowledge from large datum stored in databaseor data warehouse. Data Mining Technologies make use of statistics, artificial intelligence,machine learning and database to find hidden patterns, relationship, complete predictive modeling andvaluable knowledge. Principal Component Analysis (PCA) tries to reduce dimensions to improve efficiencyby principle of losing least information. We should attract attention to the result of PCA thatdoesn't show which variable is most important, interesting and doesn't show how to compareone variable's weightiness with another. This paper will discuss the conception that equals theproduct of multiplying eigenvector by eigenvalue. The formula is follow.Ui* λ i i∑ =p1λiThis paper bases on the Medicine Manage System that belongs to the ChangChung Maternity HospitalInformation System. The project gets data from system. This paper takes example for cold andtests system. System gets 22 item, sets weight and come out the anticipated result. Because theChangChung Maternity Hospital is just a small hospital, the number of drugs are limited. If weget large data resources, the result may be better.