Privacy Preservation using Deep Learning Methods

Privacy Preservation using Deep Learning Methods

Amreen Ayisha M1, Ankit Kumar2, Gasti Harshini3, Keerthana S4, Mrs. Samatha R Swamy5

12345 ayishaamreen78@gmail.com, ankit.astaim@gmail.com, harshinisweety61@gmail.com, twinklekee026@gmail.com, samatha-ise@dsatm.edu.in

1234Student, Department of Information Science and Engineering, DSATM, Bangalore-88, Karnataka

5Faculty, Department of Information Science and Engineering, DSATM, Bangalore-88, Karnataka

Abstract- Heart disease is a major concern worldwide , and the analysis of heart disease - related data can provide valuable insights for healthcare professionals and researchers. However , sharing such data poses a significant risk to individuals’ privacy , as personal information can be used to re-identify individuals. To address this issue, techniques as k-anonymization, randomization is employed in our project protecting individuals privacy and deriving meaningful analysis of data. We have created the random data using Faker library. Providing the specific attributes it randomly generates the values. We have generated 10000 rows for four columns. Privacy preserving techniques are applied to this data and which is shown in the background, We have also included the prediction of heart disease which is a website. This predicts the presence or absence of the disease based on realistic data. Our website also provides the information about heart disease. When the required data is entered by the user, that data can be viewed in encrypted form except the age all the other factors are encrypted for privacy concerns. This is to show how the data is encrypted. Our results show that k-anonymization and randomization techniques can be used together to achieve privacy-preserving analysis of heart related data , making it possible to derive valuable insights from such data without compromising individuals’ privacy.

Keywords- K-anonymization, Randomization, Prediction, Encrypted data, Security, Faker library.