Description:1. INTRODUCTION

Drug Formula are being used to create drugs. As we know, drug formulas are the key important factor in drug discovery, these drug formulas must be kept secured without being leaked with unauthorized parties. Due to the significant investments and high commercial values involved in drug discovery, privacy is an important factor which is not there in existing system. The main aim of our project is to Secure and detect whether the Drug Formula is active or not by Data Mining Algorithms. The overall process is as follows. Initially the Drug Owner as well as Drug Tester will register to the system with their credentials and logins to the system. The Drug owner uploads the Drug Formula composition in form of CSV file which acts as the Dataset for our Project. The file is stored in our system and we are training our dataset by two algorithms- SVM and NB.

The Support Vector Machine (SVM) and Naive Bayes (NB) algorithm are used to detect whether the Drug Formula is active or not by comparing the existing Data set in the system. Before Testing the Drug, the tester will send a request to owner for getting permissions for Testing. In final we will get trained data and accuracy for the uploaded data which will be tested by Drug Tester. Admin will approve the drug component finally and let us know the file was active or not. The actual outcome of project is accurate compared to other classifier models and we minimize the risk of unauthorized disclosure of Drug Formula during the SVM and NB training so that avoiding Privacy issue.

2. EXISTING WORKS

In the existing system real time data sets (or) existing datasets of known drug formulas were used to train the SVM classifier, and the trained SVM classifier can be used for new drug compound visual scanning. Due to the significant investments and high commercial values involved in drug discovery, privacy is an important factor. When a researcher sends some chemical compounds for SVM classification, it is important to ensure that the potential new drug compounds will not be leaked to a third-party, such as a competing pharmaceutical corporation. As we use real time datasets in existing system there is no privacy preserving for the drug component. which leads to access of Drug Formula to Unauthorized parties.

Also, Tester comes to know about the entire drug formula before releasing the drug in existing system. Also, one of the drawbacks of existing system is that the new drug composition cannot be found. Thus, in existing system there is a Privacy issue, not accurate and Leads to Drug Formula Leakage which has been overcome by Support Vector Machine (SVM) and Naive Bayes (NB) Algorithms in the proposed system.

3. DESCRIPTION ABOUT PROPOSED WORK

The main aim of our project is to detect whether the given Drug Formula is active or not by using Data Mining Algorithms such as Support Vector Machine (SVM) and Naïve Bayes (NB) in a privacy preserving way. The Project is to facilitate drug manufacturers to securely outsource their formulas for storage done by SVM and NB training. The given Drug Formula are tested and Trained by SVM and NB Algorithms in a privacy-preserving way without any Drug Formula Leakage. The outcome of this project is accurate compared to other classifier models and used to detect the given drug formula is active or not.
The Proposed work is mainly used to
• To allow the system to securely use multiple drug Formulas in a privacy preserving way without sharing to unauthorized parties.
• To check whether the Drug component is active or not with accuracy.
• To minimize the risk of unauthorized disclosure during Support Vector machine and Naive Bayes Training.

In this proposed work, we propose a framework for securing Drug components by Data mining algorithms such as Support Vector Machine (SVM) and Naïve Bayes (NB). Specifically, it is designed to allow the system to securely use multiple drug formula providers’ drug formulas to train SVM and NB provided by the analytical model provider. In our approach, we design secure computation protocols to allow the system to perform commonly used integer and fraction computations. To securely train the SVM, we design a secure SVM parameter selection protocol to select two SVM parameters and construct a secure sequential minimal optimization protocol to privately refresh both selected SVM parameters.

The trained SVM classifier can be used to determine the existing drug chemical compound. On the other hand, NB classifier can be used to detect the active Drug component. These two algorithms are used to train the uploaded drug dataset and in final we will get trained data and accuracy for that. Lastly, we proven that the proposed system minimizes the risk of unauthorized disclosure of Drug Formula during the SVM and NB training so that multiple pharmaceutical corporations won’t reveal the drug components in detail avoiding Privacy issue.

In this project basically there are three users involved: Drug Owner, Drug Tester and the Admin. Initially the Drug owner and Drug tester will register their personal details to the system and Logins to system with their credentials. Those details will be stored into the database. The Drug owner then upload the Drug composition in the form of .CSV file. That data set contains the formula and we have to mention the type of class (Class A, Class B). While uploading the file the Drug composition are read and it is trained with existing data set by Support Vector Machine (SVM) and Naive Bayes (NB) Algorithms to train the system. The trained data and accuracy will be sent to the owner from python server. Based on the composition the user is prompted whether the Drug Composition uploaded was a new file or existing file. On the other hand, Drug Tester can view the Owner details and can request the owner for test. The Drug owner can view the Tester profile and based on preference he can allow or Deny the Tester for Drug Testing. After confirmation the Drug tester test the Drug and will send results to Admin for Approve. The admin can login and view accuracy for Drug composition and can approve or reject. On Approving the Drug owner can upload the Drug Details to Clinic.

4. CLAIMS

1. The proposed system gives more accuracy and preserves the privacy in Drug Component than the existing system. The existing system have the issue in the field of privacy is that sometimes the Drug Formula can be leaked to Unauthorised parties.
2. But in proposed method, since the system is trained with existing Dataset, the trained model has enough knowledge to preserve privacy and can even detect the active Drug component by usage of Naïve Bayes.
3. So that there is no way drug Formula cannot be leaked and also these system gives almost right accuracy values from Drug composition

ABSTRACT

Data mining enables healthcare insurers to recognize fraud and abuse because of its higher performance in image classification. Basically, Drug Formula are being used to create drugs. When a researcher sends some chemical compounds for SVM classification, it is important to ensure that the potential new drug compounds will not be leaked to a third-party, such as a competing pharmaceutical corporation. Thus, in existing system there is a Privacy issue, not accurate and Leads to Drug Formula Leakage which has been overcome in our proposed system. This Project is used to check whether the drug formulas being used to create a drug is currently active or not in privacy preserving way. Drug Formula are being given in form of trained dataset (binary data set) to the tester to check the particular drug is active or not. Support Vector machine (SVM) and Naive Bayes (NB) algorithm are used to give the drug Formula s in the form of trained data set (binary data set) to the tester. As the formula are given in the form of trained data set to the tester there is no possibility of leakage of drug formula to the unauthorized parties. Finally, the Tester tells whether the drug component is active or not.
, Claims:1. The proposed system gives more accuracy and preserves the privacy in Drug Component than the existing system. The existing system have the issue in the field of privacy is that sometimes the Drug Formula can be leaked to Unauthorised parties.
2. But in proposed method, since the system is trained with existing Dataset, the trained model has enough knowledge to preserve privacy and can even detect the active Drug component by usage of Naïve Bayes.
3. So that there is no way drug Formula cannot be leaked and also these system gives almost right accuracy values from Drug composition