There has been a significant rise in internet usage in recent years, which has led to the presence of data theft and the diversity of counterfeit materials. This has resulted the proliferation of cybercrimes and the theft of personal data via social media, e-mail, and phishing websites that are similar to the websites commonly used to grab user data details like that of a credit card or login ID. Phishing, a prevalent form of cybercrime, poses a danger to online security through the theft of personal information, and with the emergence of the COVID-19 virus, which has led to people and organizations being drawn towards the Internet and many people and companies being forced to work remotely, it has led to an increase in the existing phishing threats. Previously, hackers took advantage of the situation to infiltrate the devices of people and companies in numerous ways, which caused huge financial losses and damage to organizations. Based on previous results and research, Machine Learning (ML) is selected by researchers as an efficient method for identifying malicious software web pages from original web pages. This paper presents 30 characteristics of websites, which are analyzed using a correlation matrix to determine the relationship between variables. Feature selection is performed through a wrapper method and Extra Tree Classifiers (ETC) to identify the top-ranked characteristics (Features) for website classification. To evaluate web pages, various machine learning techniques such as Random Forest Tree (RF), Multilayer Perceptron (MLP), Decision Tree (DT), and Support Vector Machine (SVM) are used. The results of monitoring indicate that MLP, a deep neural network, outperforms all other techniques in terms of performance.

Data security has become a significant issue nowadays with the increase of information capacity and its transmission rate. The most common and widely used techniques in the data security fields is cryptography. Cryptography is the process of concealing and transmitting data in an appropriate format, so that only authorized people can access and process it. The main goal of the cryptographic process is protecting data from being hijacked and altered. This paper proposes an algorithm for encrypting data through the use of Deoxyribo Nucleic Acid (DNA) sequence and four-dimensional hyper chaotic system. Whereby, the hyper chaotic system is applied to generate a binary sequence which is later passed to a permutation function for the key generation of the first level encryption. The proposed encryption algorithm includes several intermediate steps, which are binary-coded form and the generation of arbitrary keys. Experimental results were analyzed by calculating encryption time, key generation time, histogram and correlation coefficient entropy. Furthermore, the proposed text encryption algorithm is implemented on two different students’ datasets to improve the security of educational systems.  Finally, experimental and comparative studies have shown that, the proposed encryption algorithm reported a uniform encrypted text distribution and correlation coefficient values  nearer to ‘0’, which are close to the theoretical optimal value.

With the integration of cloud computing approaches in the healthcare systems, medical images are now processed and stored remotely on third-party servers. For such digital medical image data, privacy, protection, and security must be maintained by using image encryption methods. The aim of this paper is to design and apply a robust medical encryption framework to enhance the protection of medical image transformation and the patient information confidentiality. The proposed Framework encrypt the digital medical images using DNA computation and hyperchaotic RKF-45 random sequence approach. For which, the DNA computation is enhanced by applying hyperchaoticRKF-45 random key to the different Framework phases. The simulation results on different medical images were measured with various security analyses to prove the proposed framework randomness and coherent. Simulation results showed the ability of the hyperchaotic DNA encryption framework to withstand multiple electronic attacks with high performance compared to its counterparts of encryption algorithms. Finally, simulation and comparative studies have shown that, the proposed cryptography framework reported UACI and NPCR values 33.327 and 99.603 respectively, which are near to the theoretical optimal value.