Portal de Periódicos da CAPES

An Integrated Approach for Big Data Classification and Security Using Optimized Random Forest and DSSE Algorithm

2023; Springer Science+Business Media; Linguagem: Inglês

10.1007/978-981-99-1051-9_1

1876-1119

S. Castro, R. Pushpalakshmi,

Machine Learning and Data Classification

The present digital era handles a massive amount of data every day from various sources. These enormous volumes are termed as big data which is heterogeneous, is dynamic, and includes numerous valuable insights. Handling massive raw data is quite complex, and there might be a chance to miss the important aspects. Data processing methods like clustering and classification models reduce the burden and handle the big data effectively. Recently, numerous clustering and classification models are evolved; however, attaining the maximum classification accuracy for better performance is the main objective of every research work. Similarly, big data security gains more attention equal to the classification process. Encryption procedures will enhance data security so that classified data can be stored securely in the cloud environment. In this research work, a big data classification approach is presented using a random forest algorithm and secured the classified results in the cloud using the DSS encryption technique to attain maximum accuracy and security. The features for the classification process are obtained through a whale optimization algorithm which selects the optimal features and enhances the classification accuracy. The proposed model attains enhanced performance in terms of 98.47% accuracy, 96.48% precision, 96.58% recall, and 96.53% F1-score. Also, the proposed DSS encryption attains better encryption and decryption performances in terms of throughput, encryption, and decryption time compared to existing Encrypting File System standard algorithm (EFSSA).