Social Internet of Things (SIoT) underpins numerous novel applications and systems administration administrations for the IoT in an all the more incredible and gainful way. In this paper, we have presented a progressive structure for highlight extraction in SIoT enormous information utilizing map-diminished system alongside a directed classifier display. Additionally, a Gabor channel is utilized to lessen clamor and undesirable information from the database, and Hadoop Map Reduce has been utilized for mapping and decreasing enormous databases, to enhance the productivity of the proposed work.
Besides, the component choice has been performed on a sifted informational index by utilizing Elephant Herd Optimization. The proposed framework engineering has been actualized utilizing Linear Kernel Support Vector Machine-based classifier to group the information and for foreseeing the proficiency of the proposed work. From the outcomes, the most extreme exactness, specificity, and affectability of our work is 98.2%, 85.88%, and 80%, besides investigated time and memory, and these outcomes have been contrasted and the current writing.