In the current age of technology, various diseases in the body are also on the rise. Tumours that cause more discomfort in the body are set to increase the discomfort of most patients. Patients experience different effects depending on the tumour size and type. Future developments in the medical field are moving towards the development of tools based on IoT devices. These advances will in the future follow special features designed based on multiple machine learning developed by artificial intelligence. In that order, an improved algorithm named Internet of Things-based enhanced machine learning is proposed in this paper. What makes it special is that it involves separate functions to diagnose each type of tumour. It analyzes and calculates things like the size, shape, and location of the tumour. Cure from cancer is determined by the stage at which we find cancer. Early detection of cancer has the potential to cure quickly. At a saturation point, the proposed Internet of Things-based enhanced machine learning model achieved 94.56% of accuracy, 94.12% of precision, 94.98% of recall, 95.12% of F1-score, and 1856 ms of execution time. The simulation is conducted to test the efficacy of the model, and the results of the simulation show that the proposed Internet of Things-based enhanced machine learning obtains a higher rate of intelligence than other methods.

No abstract

Through Advanced Persistent Threats (APTs), which can reveal data alteration, destruction, or Denial of Service attacks through the examples of exposed hardware and software, the information technology model advances. Moving Target (MTD) is a promising risk-reduction strategy that primarily relies on APTs by utilizing dynamic and randomization techniques on properties that are collaborated. Although there are various MTD approaches to implement the blind random mutation, it still produces better performance overhead as well as poor defense utility. Additionally, APT is a unique assault strategy that was typically developed by hacking groups to steal data or deactivate systems for enormous originalities and uniform countries. APT is a multi-stage, long-term representative, and it is difficult to identify attacks effectively using an outmoded approach. In this paper, Conditional Dingo Optimization Algorithm Deep Residual Network (CDOA-based DRN) is devised for APT detection. Moreover, correlation Tversky index-based similarity is designed for performing feature fusion. The hybrid optimization algorithm effectively increases the performance and reduces various real-world issues. Testing accuracy, True Positive Rate, and False Positive Rate of the newly developed CDOA-based DRN are 95.43%, 96.34%, and 91.43%, respectively, for better performance.

No abstract